4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * The -crypt option to the passwd command line tool has been removed.
30 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
32 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
33 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
34 The inverse ciphers use AES decryption for wrapping, and
35 AES encryption for unwrapping.
39 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
40 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
41 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
42 OpenSSL 3.0 these are replaced by the more generic functions
43 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
44 The old versions have been converted to deprecated macros that just call the
49 * The security callback, which can be customised by application code, supports
50 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
51 in the "other" parameter. In most places this is what is passed. All these
52 places occur server side. However there was one client side call of this
53 security operation and it passed a DH object instead. This is incorrect
54 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
55 of the other locations. Therefore this client side call has been changed to
56 pass an EVP_PKEY instead.
60 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
61 interface. Their functionality remains unchanged.
65 * Added new option for 'openssl list', '-providers', which will display the
66 list of loaded providers, their names, version and status. It optionally
67 displays their gettable parameters.
71 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
72 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
73 type is internally recognised so the workaround is no longer needed.
75 Functionality is still retained as it is, but will only work with
76 EVP_PKEYs with a legacy internal key.
80 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
81 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
85 * Changed all "STACK" functions to be macros instead of inline functions. Macro
86 parameters are still checked for type safety at compile time via helper
91 * Remove the RAND_DRBG API
93 The RAND_DRBG API did not fit well into the new provider concept as
94 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
95 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
96 and some of its API calls are rather low-level. This holds in particular
97 for the callback mechanism (RAND_DRBG_set_callbacks()).
99 Adding a compatibility layer to continue supporting the RAND_DRBG API as
100 a legacy API for a regular deprecation period turned out to come at the
101 price of complicating the new provider API unnecessarily. Since the
102 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
105 *Paul Dale and Matthias St. Pierre*
107 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
108 as well as actual hostnames.
112 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
113 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
114 conversely, silently ignore DTLS protocol version bounds when configuring
115 TLS-based contexts. The commands can be repeated to set bounds of both
116 types. The same applies with the corresponding "min_protocol" and
117 "max_protocol" command-line switches, in case some application uses both TLS
120 SSL_CTX instances that are created for a fixed protocol version (e.g.
121 TLSv1_server_method()) also silently ignore version bounds. Previously
122 attempts to apply bounds to these protocol versions would result in an
123 error. Now only the "version-flexible" SSL_CTX instances are subject to
124 limits in configuration files in command-line options.
128 * Deprecated the `ENGINE` API. Engines should be replaced with providers
133 * Reworked the recorded ERR codes to make better space for system errors.
134 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
135 given code is a system error (true) or an OpenSSL error (false).
139 * Reworked the test perl framework to better allow parallel testing.
141 *Nicola Tuveri and David von Oheimb*
143 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
144 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
148 * 'Configure' has been changed to figure out the configuration target if
149 none is given on the command line. Consequently, the 'config' script is
150 now only a mere wrapper. All documentation is changed to only mention
153 *Rich Salz and Richard Levitte*
155 * Added a library context `OSSL_LIB_CTX` that applications as well as
156 other libraries can use to form a separate context within which
157 libcrypto operations are performed.
159 There are two ways this can be used:
161 - Directly, by passing a library context to functions that take
162 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
164 - Indirectly, by creating a new library context and then assigning
165 it as the new default, with `OSSL_LIB_CTX_set0_default`.
167 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
168 apart from the functions directly related to `OSSL_LIB_CTX`, accept
169 NULL to indicate that the default library context should be used.
171 Library code that changes the default library context using
172 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
173 second call before returning to the caller.
175 _(Note: the library context was initially called `OPENSSL_CTX` and
176 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
180 * Handshake now fails if Extended Master Secret extension is dropped
185 * Dropped interactive mode from the 'openssl' program. From now on,
186 the `openssl` command without arguments is equivalent to `openssl
191 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
192 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
193 While the old function names have been retained for backward compatibility
194 they should not be used in new developments
195 because their return values are confusing: Unlike other `_cmp()` functions
196 they do not return 0 in case their arguments are equal.
200 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
201 EC_GROUP_get_field_type().
205 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
206 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
207 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
208 Applications should rely on the library automatically assigning a suitable
209 EC_METHOD internally upon EC_GROUP construction.
213 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
214 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
215 assigned internally without application intervention.
216 Users of EC_GROUP_new() should switch to a different suitable constructor.
220 * Add CAdES-BES signature verification support, mostly derived
221 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
223 *Filipe Raimundo da Silva*
225 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
229 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
230 functions are not widely used and now OpenSSL automatically perform this
231 conversion when needed.
235 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
236 EC_KEY_precompute_mult(). These functions are not widely used and
237 applications should instead switch to named curves which OpenSSL has
238 hardcoded lookup tables for.
242 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
243 should instead use the L<EC_POINT_mul(3)> function.
247 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
248 that are not applicable to the new provider model. Applications should
249 instead use EVP_default_properties_is_fips_enabled() and
250 EVP_default_properties_enable_fips().
254 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
255 is set, an unexpected EOF is ignored, it pretends a close notify was received
256 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
260 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
261 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
262 used and applications should instead use the
263 L<EC_POINT_set_affine_coordinates(3)> and
264 L<EC_POINT_get_affine_coordinates(3)> functions.
268 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
269 arrays to be more easily constructed via a series of utility functions.
270 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
271 the various push functions and finally convert to a passable OSSL_PARAM
272 array using OSSL_PARAM_BLD_to_param().
276 * The security strength of SHA1 and MD5 based signatures in TLS has been
277 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
278 working at the default security level of 1 and instead requires security
279 level 0. The security level can be changed either using the cipher string
280 with @SECLEVEL, or calling SSL_CTX_set_security_level().
284 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
285 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
286 internal keys, if they correspond to one of those built in types.
290 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
291 contain a provider side internal key.
295 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
296 They are old functions that we don't use, and that you could disable with
297 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
301 * Project text documents not yet having a proper file name extension
302 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
303 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
307 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
308 have been converted to Markdown with the goal to produce documents
309 which not only look pretty when viewed online in the browser, but
310 remain well readable inside a plain text editor.
312 To achieve this goal, a 'minimalistic' Markdown style has been applied
313 which avoids formatting elements that interfere too much with the
314 reading flow in the text file. For example, it
316 * avoids [ATX headings][] and uses [setext headings][] instead
317 (which works for `<h1>` and `<h2>` headings only).
318 * avoids [inline links][] and uses [reference links][] instead.
319 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
321 [ATX headings]: https://github.github.com/gfm/#atx-headings
322 [setext headings]: https://github.github.com/gfm/#setext-headings
323 [inline links]: https://github.github.com/gfm/#inline-link
324 [reference links]: https://github.github.com/gfm/#reference-link
325 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
326 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
328 *Matthias St. Pierre*
330 * The test suite is changed to preserve results of each test recipe.
331 A new directory test-runs/ with subdirectories named like the
332 test recipes are created in the build tree for this purpose.
336 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
337 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
338 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
340 *David von Oheimb, Martin Peylo*
342 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
343 The legacy OCSP-focused and only partly documented API is retained.
344 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
348 * Added `util/check-format.pl`, a tool for checking adherence to the
349 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
350 The checks performed are incomplete and yield some false positives.
351 Still the tool should be useful for detecting most typical glitches.
355 * BIO_do_connect and BIO_do_handshake have been extended:
356 If domain name resolution yields multiple IP addresses all of them are tried
357 after connect() failures.
361 * All of the low level RSA functions have been deprecated including:
363 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
364 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
365 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
366 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
367 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
368 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
369 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
370 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
371 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
372 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
373 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
374 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
375 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
376 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
377 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
378 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
379 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
380 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
381 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
382 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
383 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
384 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
385 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
386 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
387 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
388 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
389 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
390 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
392 Use of these low level functions has been informally discouraged for a long
393 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
394 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
395 L<EVP_PKEY_decrypt(3)>.
399 * X509 certificates signed using SHA1 are no longer allowed at security
401 In TLS/SSL the default security level is 1. It can be set either
402 using the cipher string with @SECLEVEL, or calling
403 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
404 a call to SSL_CTX_use_certificate() will fail if the security level is not
406 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
407 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
412 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
413 modified to use PKEY APIs. These commands are now in maintenance mode
414 and no new features will be added to them.
418 * The command line utility rsautl has been deprecated.
419 Instead use the pkeyutl program.
423 * The command line utilities genrsa and rsa have been modified to use PKEY
424 APIs. They now write PKCS#8 keys by default. These commands are now in
425 maintenance mode and no new features will be added to them.
429 * All of the low level DH functions have been deprecated including:
431 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
432 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
433 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
434 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
435 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
436 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
437 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
438 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
439 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
440 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
441 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
442 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
443 DH_meth_get_generate_params and DH_meth_set_generate_params.
445 Use of these low level functions has been informally discouraged for a long
446 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
447 and L<EVP_PKEY_derive(3)>.
451 * All of the low level DSA functions have been deprecated including:
453 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
454 DSA_get_default_method, DSA_set_method, DSA_get_method,
455 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
456 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
457 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
458 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
459 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
460 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
461 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
462 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
463 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
464 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
465 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
468 Use of these low level functions has been informally discouraged for a long
469 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
470 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
474 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
475 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
476 This means that applications don't have to look at the curve NID and
477 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
478 However, they still can, that EVP_PKEY_set_alias_type() call acts as
479 a no-op when the EVP_PKEY is already of the given type.
481 Parameter and key generation is also reworked to make it possible
482 to generate EVP_PKEY_SM2 parameters and keys without having to go
483 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
484 However, code that does the latter will still work as before.
488 * Deprecated low level ECDH and ECDSA functions. These include:
490 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
491 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
494 Use of these low level functions has been informally discouraged for a long
495 time. Instead applications should use the EVP_PKEY_derive(3),
496 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
500 * Deprecated the EC_KEY_METHOD functions. These include:
502 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
503 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
504 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
505 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
506 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
507 EC_KEY_METHOD_get_verify.
509 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
513 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
514 and EVP_PKEY_decrypt() instead.
515 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
516 and EVP_PKEY_encrypt() instead.
520 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
521 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
522 a new formulation to include all the things it can be used for,
523 as well as words of caution.
527 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
528 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
532 * All of the low level HMAC functions have been deprecated including:
534 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
535 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
538 Use of these low level functions has been informally discouraged for a long
539 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
540 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
541 and L<EVP_MAC_final(3)>.
545 * Over two thousand fixes were made to the documentation, including:
546 - Common options (such as -rand/-writerand, TLS version control, etc)
547 were refactored and point to newly-enhanced descriptions in openssl.pod.
548 - Added style conformance for all options (with help from Richard Levitte),
549 documented all reported missing options, added a CI build to check
550 that all options are documented and that no unimplemented options
552 - Documented some internals, such as all use of environment variables.
553 - Addressed all internal broken L<> references.
557 * All of the low level CMAC functions have been deprecated including:
559 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
560 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
562 Use of these low level functions has been informally discouraged for a long
563 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
564 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
565 and L<EVP_MAC_final(3)>.
569 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
570 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
573 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
574 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
575 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
576 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
577 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
578 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
579 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
580 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
581 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
582 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
584 Use of these low level functions has been informally discouraged
585 for a long time. Applications should use the EVP_DigestInit_ex(3),
586 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
590 * Corrected the documentation of the return values from the `EVP_DigestSign*`
591 set of functions. The documentation mentioned negative values for some
592 errors, but this was never the case, so the mention of negative values
595 Code that followed the documentation and thereby check with something
596 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
600 * All of the low level cipher functions have been deprecated including:
602 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
603 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
604 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
605 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
606 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
607 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
608 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
609 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
610 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
611 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
612 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
613 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
614 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
615 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
616 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
617 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
618 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
619 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
620 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
621 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
622 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
623 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
624 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
625 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
626 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
627 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
628 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
629 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
630 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
632 Use of these low level functions has been informally discouraged for
633 a long time. Applications should use the high level EVP APIs, e.g.
634 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
635 equivalently named decrypt functions instead.
637 *Matt Caswell and Paul Dale*
639 * Removed include/openssl/opensslconf.h.in and replaced it with
640 include/openssl/configuration.h.in, which differs in not including
641 <openssl/macros.h>. A short header include/openssl/opensslconf.h
642 was added to include both.
644 This allows internal hacks where one might need to modify the set
645 of configured macros, for example this if deprecated symbols are
646 still supposed to be available internally:
648 #include <openssl/configuration.h>
650 #undef OPENSSL_NO_DEPRECATED
651 #define OPENSSL_SUPPRESS_DEPRECATED
653 #include <openssl/macros.h>
655 This should not be used by applications that use the exported
656 symbols, as that will lead to linking errors.
660 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
661 used in exponentiation with 512-bit moduli. No EC algorithms are
662 affected. Analysis suggests that attacks against 2-prime RSA1024,
663 3-prime RSA1536, and DSA1024 as a result of this defect would be very
664 difficult to perform and are not believed likely. Attacks against DH512
665 are considered just feasible. However, for an attack the target would
666 have to re-use the DH512 private key, which is not recommended anyway.
667 Also applications directly using the low level API BN_mod_exp may be
668 affected if they use BN_FLG_CONSTTIME.
673 * Most memory-debug features have been deprecated, and the functionality
674 replaced with no-ops.
678 * Added documentation for the STACK API.
682 * Introduced a new method type and API, OSSL_ENCODER, to
683 represent generic encoders. An implementation is expected to
684 be able to encode an object associated with a given name (such
685 as an algorithm name for an asymmetric key) into forms given by
686 implementation properties.
688 Encoders are primarily used from inside libcrypto, through
689 calls to functions like EVP_PKEY_print_private(),
690 PEM_write_bio_PrivateKey() and similar.
692 Encoders are specified in such a way that they can be made to
693 directly handle the provider side portion of an object, if this
694 provider side part comes from the same provider as the encoder
695 itself, but can also be made to handle objects in parametrized
696 form (as an OSSL_PARAM array of data). This allows a provider to
697 offer generic encoders as a service for any other provider.
701 * Added a .pragma directive to the syntax of configuration files, to
702 allow varying behavior in a supported and predictable manner.
703 Currently added pragma:
707 This allows dollar signs to be a keyword character unless it's
708 followed by a opening brace or parenthesis. This is useful for
709 platforms where dollar signs are commonly used in names, such as
710 volume names and system directory names on VMS.
714 * Added functionality to create an EVP_PKEY from user data. This
715 is effectively the same as creating a RSA, DH or DSA object and
716 then assigning them to an EVP_PKEY, but directly using algorithm
717 agnostic EVP functions. A benefit is that this should be future
718 proof for public key algorithms to come.
722 * Change the interpretation of the '--api' configuration option to
723 mean that this is a desired API compatibility level with no
724 further meaning. The previous interpretation, that this would
725 also mean to remove all deprecated symbols up to and including
726 the given version, no requires that 'no-deprecated' is also used
727 in the configuration.
729 When building applications, the desired API compatibility level
730 can be set with the OPENSSL_API_COMPAT macro like before. For
731 API compatibility version below 3.0, the old style numerical
732 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
733 For version 3.0 and on, the value is expected to be the decimal
734 value calculated from the major and minor version like this:
736 MAJOR * 10000 + MINOR * 100
740 -DOPENSSL_API_COMPAT=30000 For 3.0
741 -DOPENSSL_API_COMPAT=30200 For 3.2
743 To hide declarations that are deprecated up to and including the
744 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
745 given when building the application as well.
749 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
750 access to certificate and CRL stores via URIs and OSSL_STORE
753 This adds the following functions:
755 - X509_LOOKUP_store()
756 - X509_STORE_load_file()
757 - X509_STORE_load_path()
758 - X509_STORE_load_store()
759 - SSL_add_store_cert_subjects_to_stack()
760 - SSL_CTX_set_default_verify_store()
761 - SSL_CTX_load_verify_file()
762 - SSL_CTX_load_verify_dir()
763 - SSL_CTX_load_verify_store()
767 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
768 The presence of this system service is determined at run-time.
772 * Added functionality to create an EVP_PKEY context based on data
773 for methods from providers. This takes an algorithm name and a
774 property query string and simply stores them, with the intent
775 that any operation that uses this context will use those strings
776 to fetch the needed methods implicitly, thereby making the port
777 of application written for pre-3.0 OpenSSL easier.
781 * The undocumented function NCONF_WIN32() has been deprecated; for
782 conversion details see the HISTORY section of doc/man5/config.pod
786 * Introduced the new functions EVP_DigestSignInit_ex() and
787 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
788 EVP_DigestVerifyUpdate() have been converted to functions. See the man
789 pages for further details.
793 * Over two thousand fixes were made to the documentation, including:
794 adding missing command flags, better style conformance, documentation
797 *Rich Salz, Richard Levitte*
799 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
800 X25519, X448, Ed25519 and Ed448.
804 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
809 * Deprecated the public definition of ERR_STATE as well as the function
810 ERR_get_state(). This is done in preparation of making ERR_STATE an
815 * Added ERR functionality to give callers access to the stored function
816 names that have replaced the older function code based functions.
818 New functions are ERR_get_error_func(), ERR_peek_error_func(),
819 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
820 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
821 and ERR_peek_last_error_all().
823 These functions have become deprecated: ERR_get_error_line_data(),
824 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
825 ERR_func_error_string().
829 * Extended testing to be verbose for failing tests only. The make variables
830 VERBOSE_FAILURE or VF can be used to enable this:
832 $ make VF=1 test # Unix
833 $ mms /macro=(VF=1) test ! OpenVMS
834 $ nmake VF=1 test # Windows
838 * Added several checks to X509_verify_cert() according to requirements in
839 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
840 (which may be done by using the CLI option `-x509_strict`):
841 * The basicConstraints of CA certificates must be marked critical.
842 * CA certificates must explicitly include the keyUsage extension.
843 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
844 * The issuer name of any certificate must not be empty.
845 * The subject name of CA certs, certs with keyUsage crlSign,
846 and certs without subjectAlternativeName must not be empty.
847 * If a subjectAlternativeName extension is given it must not be empty.
848 * The signatureAlgorithm field and the cert signature must be consistent.
849 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
850 must not be marked critical.
851 * The authorityKeyIdentifier must be given for X.509v3 certs
852 unless they are self-signed.
853 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
857 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
858 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
862 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
863 used even when parsing explicit parameters, when loading a encoded key
864 or calling `EC_GROUP_new_from_ecpkparameters()`/
865 `EC_GROUP_new_from_ecparameters()`.
866 This prevents bypass of security hardening and performance gains,
867 especially for curves with specialized EC_METHODs.
868 By default, if a key encoded with explicit parameters is loaded and later
869 encoded, the output is still encoded with explicit parameters, even if
870 internally a "named" EC_GROUP is used for computation.
874 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
875 this change, EC_GROUP_set_generator would accept order and/or cofactor as
876 NULL. After this change, only the cofactor parameter can be NULL. It also
877 does some minimal sanity checks on the passed order.
882 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
883 An attack is simple, if the first CMS_recipientInfo is valid but the
884 second CMS_recipientInfo is chosen ciphertext. If the second
885 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
886 encryption key will be replaced by garbage, and the message cannot be
887 decoded, but if the RSA decryption fails, the correct encryption key is
888 used and the recipient will not notice the attack.
889 As a work around for this potential attack the length of the decrypted
890 key must be equal to the cipher default key length, in case the
891 certifiate is not given and all recipientInfo are tried out.
892 The old behaviour can be re-enabled in the CMS code by setting the
893 CMS_DEBUG_DECRYPT flag.
897 * Early start up entropy quality from the DEVRANDOM seed source has been
898 improved for older Linux systems. The RAND subsystem will wait for
899 /dev/random to be producing output before seeding from /dev/urandom.
900 The seeded state is stored for future library initialisations using
901 a system global shared memory segment. The shared memory identifier
902 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
903 the desired value. The default identifier is 114.
907 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
908 when primes for RSA keys are computed.
909 Since we previously always generated primes == 2 (mod 3) for RSA keys,
910 the 2-prime and 3-prime RSA modules were easy to distinguish, since
911 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
912 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
913 This avoids possible fingerprinting of newly generated RSA modules.
917 * Correct the extended master secret constant on EBCDIC systems. Without this
918 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
919 negotiate EMS will fail. Unfortunately this also means that TLS connections
920 between EBCDIC systems with this fix, and EBCDIC systems without this
921 fix will fail if they negotiate EMS.
925 * Changed the library initialisation so that the config file is now loaded
926 by default. This was already the case for libssl. It now occurs for both
927 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
928 OPENSSL_init_crypto() to suppress automatic loading of a config file.
932 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
933 where the former acts as a replacement for ERR_put_error(), and the
934 latter replaces the combination ERR_put_error()+ERR_add_error_data().
935 ERR_raise_data() adds more flexibility by taking a format string and
936 an arbitrary number of arguments following it, to be processed with
941 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
942 to check if a named provider is loaded and available. When called, it
943 will also activate all fallback providers if such are still present.
947 * Enforce a minimum DH modulus size of 512 bits.
951 * Changed DH parameters to generate the order q subgroup instead of 2q.
952 Previously generated DH parameters are still accepted by DH_check
953 but DH_generate_key works around that by clearing bit 0 of the
954 private key for those. This avoids leaking bit 0 of the private key.
958 * Significantly reduce secure memory usage by the randomness pools.
962 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
967 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
968 algorithms. An implementation of a key exchange algorithm can be obtained
969 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
970 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
971 the older EVP_PKEY_derive_init() function. See the man pages for the new
972 functions for further details.
976 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
980 * Removed the function names from error messages and deprecated the
983 * Removed NextStep support and the macro OPENSSL_UNISTD
987 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
988 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
989 Also removed "export var as function" capability; we do not export
990 variables, only functions.
994 * RC5_32_set_key has been changed to return an int type, with 0 indicating
995 an error and 1 indicating success. In previous versions of OpenSSL this
996 was a void type. If a key was set longer than the maximum possible this
1001 * Support SM2 signing and verification schemes with X509 certificate.
1005 * Use SHA256 as the default digest for TS query in the ts app.
1009 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1010 This checks that the salt length is at least 128 bits, the derived key
1011 length is at least 112 bits, and that the iteration count is at least 1000.
1012 For backwards compatibility these checks are disabled by default in the
1013 default provider, but are enabled by default in the fips provider.
1014 To enable or disable these checks use the control
1015 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1019 * Default cipher lists/suites are now available via a function, the
1020 #defines are deprecated.
1024 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1025 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1026 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1030 * Join the directories crypto/x509 and crypto/x509v3
1034 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1035 This changes the size when using the genpkey app when no size is given. It
1036 fixes an omission in earlier changes that changed all RSA, DSA and DH
1037 generation apps to use 2048 bits by default.
1041 * Added command 'openssl kdf' that uses the EVP_KDF API.
1045 * Added command 'openssl mac' that uses the EVP_MAC API.
1049 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1050 as default directories. Also added the command 'openssl info'
1051 for scripting purposes.
1055 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1056 deprecated. These undocumented functions were never integrated into the EVP
1057 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1058 Bi-directional IGE mode. These modes were never formally standardised and
1059 usage of these functions is believed to be very small. In particular
1060 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1061 is ever used. The security implications are believed to be minimal, but
1062 this issue was never fixed for backwards compatibility reasons. New code
1063 should not use these modes.
1067 * Add prediction resistance to the DRBG reseeding process.
1071 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1072 mandated by IEEE Std 1619-2018.
1076 * Added newline escaping functionality to a filename when using openssl dgst.
1077 This output format is to replicate the output format found in the `*sum`
1078 checksum programs. This aims to preserve backward compatibility.
1080 *Matt Eaton, Richard Levitte, and Paul Dale*
1082 * Removed the heartbeat message in DTLS feature, as it has very
1083 little usage and doesn't seem to fulfill a valuable purpose.
1084 The configuration option is now deprecated.
1088 * Changed the output of 'openssl {digestname} < file' to display the
1089 digest name in its output.
1093 * Added a new generic trace API which provides support for enabling
1094 instrumentation through trace output. This feature is mainly intended
1095 as an aid for developers and is disabled by default. To utilize it,
1096 OpenSSL needs to be configured with the `enable-trace` option.
1098 If the tracing API is enabled, the application can activate trace output
1099 by registering BIOs as trace channels for a number of tracing and debugging
1102 The 'openssl' application has been expanded to enable any of the types
1103 available via environment variables defined by the user, and serves as
1104 one possible example on how to use this functionality.
1106 *Richard Levitte & Matthias St. Pierre*
1108 * Added build tests for C++. These are generated files that only do one
1109 thing, to include one public OpenSSL head file each. This tests that
1110 the public header files can be usefully included in a C++ application.
1112 This test isn't enabled by default. It can be enabled with the option
1113 'enable-buildtest-c++'.
1117 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1121 * Add KMAC to EVP_MAC.
1125 * Added property based algorithm implementation selection framework to
1130 * Added SCA hardening for modular field inversion in EC_GROUP through
1131 a new dedicated field_inv() pointer in EC_METHOD.
1132 This also addresses a leakage affecting conversions from projective
1133 to affine coordinates.
1135 *Billy Bob Brumley, Nicola Tuveri*
1137 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1138 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1139 those algorithms that were already supported through the EVP_PKEY API
1140 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1141 and scrypt are now wrappers that call EVP_KDF.
1145 * Build devcrypto engine as a dynamic engine.
1147 *Eneas U de Queiroz*
1149 * Add keyed BLAKE2 to EVP_MAC.
1153 * Fix a bug in the computation of the endpoint-pair shared secret used
1154 by DTLS over SCTP. This breaks interoperability with older versions
1155 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1156 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1157 interoperability with such broken implementations. However, enabling
1158 this switch breaks interoperability with correct implementations.
1160 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1161 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1165 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1169 * Change the license to the Apache License v2.0.
1173 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1175 - Major releases (indicated by incrementing the MAJOR release number)
1176 may introduce incompatible API/ABI changes.
1177 - Minor releases (indicated by incrementing the MINOR release number)
1178 may introduce new features but retain API/ABI compatibility.
1179 - Patch releases (indicated by incrementing the PATCH number)
1180 are intended for bug fixes and other improvements of existing
1181 features only (like improving performance or adding documentation)
1182 and retain API/ABI compatibility.
1186 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1190 * Remove the 'dist' target and add a tarball building script. The
1191 'dist' target has fallen out of use, and it shouldn't be
1192 necessary to configure just to create a source distribution.
1196 * Recreate the OS390-Unix config target. It no longer relies on a
1197 special script like it did for OpenSSL pre-1.1.0.
1201 * Instead of having the source directories listed in Configure, add
1202 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1207 * Add GMAC to EVP_MAC.
1211 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1215 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1216 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1217 to facilitate the continued use of MACs through raw private keys in
1218 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1222 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1223 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1227 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1228 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1229 are retained for backwards compatibility.
1233 * AES-XTS mode now enforces that its two keys are different to mitigate
1234 the attacked described in "Efficient Instantiations of Tweakable
1235 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1236 Details of this attack can be obtained from:
1237 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1241 * Rename the object files, i.e. give them other names than in previous
1242 versions. Their names now include the name of the final product, as
1243 well as its type mnemonic (bin, lib, shlib).
1247 * Added new option for 'openssl list', '-objects', which will display the
1248 list of built in objects, i.e. OIDs with names.
1252 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1253 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1258 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1259 improves application performance by removing data copies and providing
1260 applications with zero-copy system calls such as sendfile and splice.
1264 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1265 option is set, openssl cleanses (zeroize) plaintext bytes from
1266 internal buffers after delivering them to the application. Note,
1267 the application is still responsible for cleansing other copies
1268 (e.g.: data received by SSL_read(3)).
1272 * `PKCS12_parse` now maintains the order of the parsed certificates
1273 when outputting them via `*ca` (rather than reversing it).
1277 * Deprecated pthread fork support methods. These were unused so no
1278 replacement is required.
1280 - OPENSSL_fork_prepare()
1281 - OPENSSL_fork_parent()
1282 - OPENSSL_fork_child()
1289 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1293 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1295 * Certificates with explicit curve parameters are now disallowed in
1296 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1300 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1301 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1302 conversely, silently ignore DTLS protocol version bounds when configuring
1303 TLS-based contexts. The commands can be repeated to set bounds of both
1304 types. The same applies with the corresponding "min_protocol" and
1305 "max_protocol" command-line switches, in case some application uses both TLS
1308 SSL_CTX instances that are created for a fixed protocol version (e.g.
1309 TLSv1_server_method()) also silently ignore version bounds. Previously
1310 attempts to apply bounds to these protocol versions would result in an
1311 error. Now only the "version-flexible" SSL_CTX instances are subject to
1312 limits in configuration files in command-line options.
1316 * Handshake now fails if Extended Master Secret extension is dropped
1321 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1323 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1325 * Fixed segmentation fault in SSL_check_chain()
1326 Server or client applications that call the SSL_check_chain() function
1327 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1328 dereference as a result of incorrect handling of the
1329 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1330 or unrecognised signature algorithm is received from the peer. This could
1331 be exploited by a malicious peer in a Denial of Service attack.
1336 * Added AES consttime code for no-asm configurations
1337 an optional constant time support for AES was added
1338 when building openssl for no-asm.
1339 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1340 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1341 At this time this feature is by default disabled.
1342 It will be enabled by default in 3.0.
1346 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1348 * Revert the change of EOF detection while reading in libssl to avoid
1349 regressions in applications depending on the current way of reporting
1350 the EOF. As the existing method is not fully accurate the change to
1351 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1352 branch and will be present in the 3.0 release.
1356 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1357 when primes for RSA keys are computed.
1358 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1359 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1360 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1361 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1362 This avoids possible fingerprinting of newly generated RSA modules.
1366 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1368 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1369 while reading in libssl then we would report an error back to the
1370 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1371 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1372 therefore give a hint as to what went wrong.
1376 * Check that ed25519 and ed448 are allowed by the security level. Previously
1377 signature algorithms not using an MD were not being checked that they were
1378 allowed by the security level.
1382 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1383 was not quite right. The behaviour was not consistent between resumption
1384 and normal handshakes, and also not quite consistent with historical
1385 behaviour. The behaviour in various scenarios has been clarified and
1386 it has been updated to make it match historical behaviour as closely as
1391 * *[VMS only]* The header files that the VMS compilers include automatically,
1392 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1393 that the C++ compiler doesn't understand. This is a shortcoming in the
1394 compiler, but can be worked around with `__cplusplus` guards.
1396 C++ applications that use OpenSSL libraries must be compiled using the
1397 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1398 functions. Otherwise, only functions with symbols of less than 31
1399 characters can be used, as the linker will not be able to successfully
1400 resolve symbols with longer names.
1404 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1405 The presence of this system service is determined at run-time.
1409 * Added newline escaping functionality to a filename when using openssl dgst.
1410 This output format is to replicate the output format found in the `*sum`
1411 checksum programs. This aims to preserve backward compatibility.
1413 *Matt Eaton, Richard Levitte, and Paul Dale*
1415 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1420 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1422 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1423 number generator (RNG). This was intended to include protection in the
1424 event of a fork() system call in order to ensure that the parent and child
1425 processes did not share the same RNG state. However this protection was not
1426 being used in the default case.
1428 A partial mitigation for this issue is that the output from a high
1429 precision timer is mixed into the RNG state so the likelihood of a parent
1430 and child process sharing state is significantly reduced.
1432 If an application already calls OPENSSL_init_crypto() explicitly using
1433 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1436 *Matthias St. Pierre*
1438 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1439 used even when parsing explicit parameters, when loading a encoded key
1440 or calling `EC_GROUP_new_from_ecpkparameters()`/
1441 `EC_GROUP_new_from_ecparameters()`.
1442 This prevents bypass of security hardening and performance gains,
1443 especially for curves with specialized EC_METHODs.
1444 By default, if a key encoded with explicit parameters is loaded and later
1445 encoded, the output is still encoded with explicit parameters, even if
1446 internally a "named" EC_GROUP is used for computation.
1450 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1451 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1452 NULL. After this change, only the cofactor parameter can be NULL. It also
1453 does some minimal sanity checks on the passed order.
1458 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1459 An attack is simple, if the first CMS_recipientInfo is valid but the
1460 second CMS_recipientInfo is chosen ciphertext. If the second
1461 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1462 encryption key will be replaced by garbage, and the message cannot be
1463 decoded, but if the RSA decryption fails, the correct encryption key is
1464 used and the recipient will not notice the attack.
1465 As a work around for this potential attack the length of the decrypted
1466 key must be equal to the cipher default key length, in case the
1467 certifiate is not given and all recipientInfo are tried out.
1468 The old behaviour can be re-enabled in the CMS code by setting the
1469 CMS_DEBUG_DECRYPT flag.
1474 * Early start up entropy quality from the DEVRANDOM seed source has been
1475 improved for older Linux systems. The RAND subsystem will wait for
1476 /dev/random to be producing output before seeding from /dev/urandom.
1477 The seeded state is stored for future library initialisations using
1478 a system global shared memory segment. The shared memory identifier
1479 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1480 the desired value. The default identifier is 114.
1484 * Correct the extended master secret constant on EBCDIC systems. Without this
1485 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1486 negotiate EMS will fail. Unfortunately this also means that TLS connections
1487 between EBCDIC systems with this fix, and EBCDIC systems without this
1488 fix will fail if they negotiate EMS.
1492 * Use Windows installation paths in the mingw builds
1494 Mingw isn't a POSIX environment per se, which means that Windows
1495 paths should be used for installation.
1500 * Changed DH_check to accept parameters with order q and 2q subgroups.
1501 With order 2q subgroups the bit 0 of the private key is not secret
1502 but DH_generate_key works around that by clearing bit 0 of the
1503 private key for those. This avoids leaking bit 0 of the private key.
1507 * Significantly reduce secure memory usage by the randomness pools.
1511 * Revert the DEVRANDOM_WAIT feature for Linux systems
1513 The DEVRANDOM_WAIT feature added a select() call to wait for the
1514 /dev/random device to become readable before reading from the
1515 /dev/urandom device.
1517 It turned out that this change had negative side effects on
1518 performance which were not acceptable. After some discussion it
1519 was decided to revert this feature and leave it up to the OS
1520 resp. the platform maintainer to ensure a proper initialization
1521 during early boot time.
1523 *Matthias St. Pierre*
1525 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1527 * Add build tests for C++. These are generated files that only do one
1528 thing, to include one public OpenSSL head file each. This tests that
1529 the public header files can be usefully included in a C++ application.
1531 This test isn't enabled by default. It can be enabled with the option
1532 'enable-buildtest-c++'.
1536 * Enable SHA3 pre-hashing for ECDSA and DSA.
1540 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1541 This changes the size when using the genpkey app when no size is given. It
1542 fixes an omission in earlier changes that changed all RSA, DSA and DH
1543 generation apps to use 2048 bits by default.
1547 * Reorganize the manual pages to consistently have RETURN VALUES,
1548 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1549 util/fix-doc-nits accordingly.
1551 *Paul Yang, Joshua Lock*
1553 * Add the missing accessor EVP_PKEY_get0_engine()
1557 * Have apps like 's_client' and 's_server' output the signature scheme
1558 along with other cipher suite parameters when debugging.
1562 * Make OPENSSL_config() error agnostic again.
1566 * Do the error handling in RSA decryption constant time.
1570 * Prevent over long nonces in ChaCha20-Poly1305.
1572 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1573 for every encryption operation. RFC 7539 specifies that the nonce value
1574 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1575 and front pads the nonce with 0 bytes if it is less than 12
1576 bytes. However it also incorrectly allows a nonce to be set of up to 16
1577 bytes. In this case only the last 12 bytes are significant and any
1578 additional leading bytes are ignored.
1580 It is a requirement of using this cipher that nonce values are
1581 unique. Messages encrypted using a reused nonce value are susceptible to
1582 serious confidentiality and integrity attacks. If an application changes
1583 the default nonce length to be longer than 12 bytes and then makes a
1584 change to the leading bytes of the nonce expecting the new value to be a
1585 new unique nonce then such an application could inadvertently encrypt
1586 messages with a reused nonce.
1588 Additionally the ignored bytes in a long nonce are not covered by the
1589 integrity guarantee of this cipher. Any application that relies on the
1590 integrity of these ignored leading bytes of a long nonce may be further
1591 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1592 is safe because no such use sets such a long nonce value. However user
1593 applications that use this cipher directly and set a non-default nonce
1594 length to be longer than 12 bytes may be vulnerable.
1596 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1602 * Add DEVRANDOM_WAIT feature for Linux systems
1604 On older Linux systems where the getrandom() system call is not available,
1605 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1606 Contrary to getrandom(), the /dev/urandom device will not block during
1607 early boot when the kernel CSPRNG has not been seeded yet.
1609 To mitigate this known weakness, use select() to wait for /dev/random to
1610 become readable before reading from /dev/urandom.
1612 * Ensure that SM2 only uses SM3 as digest algorithm
1616 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1618 * Change the info callback signals for the start and end of a post-handshake
1619 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1620 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1621 confused by this and assume that a TLSv1.2 renegotiation has started. This
1622 can break KeyUpdate handling. Instead we no longer signal the start and end
1623 of a post handshake message exchange (although the messages themselves are
1624 still signalled). This could break some applications that were expecting
1625 the old signals. However without this KeyUpdate is not usable for many
1630 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1632 * Timing vulnerability in DSA signature generation
1634 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1635 timing side channel attack. An attacker could use variations in the signing
1636 algorithm to recover the private key.
1638 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1643 * Timing vulnerability in ECDSA signature generation
1645 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1646 timing side channel attack. An attacker could use variations in the signing
1647 algorithm to recover the private key.
1649 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1654 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1655 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1656 of two gigabytes and the error handling improved.
1658 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1659 categorized as a normal bug, not a security issue, because the DRBG reseeds
1660 automatically and is fully functional even without additional randomness
1661 provided by the application.
1663 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1665 * Add a new ClientHello callback. Provides a callback interface that gives
1666 the application the ability to adjust the nascent SSL object at the
1667 earliest stage of ClientHello processing, immediately after extensions have
1668 been collected but before they have been processed. In particular, this
1669 callback can adjust the supported TLS versions in response to the contents
1674 * Add SM2 base algorithm support.
1678 * s390x assembly pack: add (improved) hardware-support for the following
1679 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1680 aes-cfb/cfb8, aes-ecb.
1684 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1685 parameter is no longer accepted, as it leads to a corrupt table. NULL
1686 pem_str is reserved for alias entries only.
1690 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1691 step for prime curves. The new implementation is based on formulae from
1692 differential addition-and-doubling in homogeneous projective coordinates
1693 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1694 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1695 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1696 to work in projective coordinates.
1698 *Billy Bob Brumley, Nicola Tuveri*
1700 * Change generating and checking of primes so that the error rate of not
1701 being prime depends on the intended use based on the size of the input.
1702 For larger primes this will result in more rounds of Miller-Rabin.
1703 The maximal error rate for primes with more than 1080 bits is lowered
1706 *Kurt Roeckx, Annie Yousar*
1708 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1712 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1713 moving between systems, and to avoid confusion when a Windows build is
1714 done with mingw vs with MSVC. For POSIX installs, there's still a
1715 symlink or copy named 'tsget' to avoid that confusion as well.
1719 * Revert blinding in ECDSA sign and instead make problematic addition
1720 length-invariant. Switch even to fixed-length Montgomery multiplication.
1724 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1725 step for binary curves. The new implementation is based on formulae from
1726 differential addition-and-doubling in mixed Lopez-Dahab projective
1727 coordinates, modified to independently blind the operands.
1729 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1731 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1732 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1733 EC_METHODs to implement their own specialized "ladder step", to take
1734 advantage of more favorable coordinate systems or more efficient
1735 differential addition-and-doubling algorithms.
1737 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1739 * Modified the random device based seed sources to keep the relevant
1740 file descriptors open rather than reopening them on each access.
1741 This allows such sources to operate in a chroot() jail without
1742 the associated device nodes being available. This behaviour can be
1743 controlled using RAND_keep_random_devices_open().
1747 * Numerous side-channel attack mitigations have been applied. This may have
1748 performance impacts for some algorithms for the benefit of improved
1749 security. Specific changes are noted in this change log by their respective
1754 * AIX shared library support overhaul. Switch to AIX "natural" way of
1755 handling shared libraries, which means collecting shared objects of
1756 different versions and bitnesses in one common archive. This allows to
1757 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1758 doesn't affect the way 3rd party applications are linked, only how
1759 multi-version installation is managed.
1763 * Make ec_group_do_inverse_ord() more robust and available to other
1764 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1765 mitigations are applied to the fallback BN_mod_inverse().
1766 When using this function rather than BN_mod_inverse() directly, new
1767 EC cryptosystem implementations are then safer-by-default.
1771 * Add coordinate blinding for EC_POINT and implement projective
1772 coordinate blinding for generic prime curves as a countermeasure to
1773 chosen point SCA attacks.
1775 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1777 * Add blinding to ECDSA and DSA signatures to protect against side channel
1778 attacks discovered by Keegan Ryan (NCC Group).
1782 * Enforce checking in the pkeyutl command line app to ensure that the input
1783 length does not exceed the maximum supported digest length when performing
1784 a sign, verify or verifyrecover operation.
1788 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1789 I/O in combination with something like select() or poll() will hang. This
1790 can be turned off again using SSL_CTX_clear_mode().
1791 Many applications do not properly handle non-application data records, and
1792 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1793 around the problems in those applications, but can also break some.
1794 It's recommended to read the manpages about SSL_read(), SSL_write(),
1795 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1796 SSL_CTX_set_read_ahead() again.
1800 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1801 now allow empty (zero character) pass phrases.
1805 * Apply blinding to binary field modular inversion and remove patent
1806 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1810 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1811 binary and prime elliptic curves.
1815 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1816 constant time fixed point multiplication.
1820 * Revise elliptic curve scalar multiplication with timing attack
1821 defenses: ec_wNAF_mul redirects to a constant time implementation
1822 when computing fixed point and variable point multiplication (which
1823 in OpenSSL are mostly used with secret scalars in keygen, sign,
1824 ECDH derive operations).
1825 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1828 * Updated CONTRIBUTING
1832 * Updated DRBG / RAND to request nonce and additional low entropy
1833 randomness from the system.
1835 *Matthias St. Pierre*
1837 * Updated 'openssl rehash' to use OpenSSL consistent default.
1841 * Moved the load of the ssl_conf module to libcrypto, which helps
1842 loading engines that libssl uses before libssl is initialised.
1846 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1850 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1852 *Ingo Schwarze, Rich Salz*
1854 * Added output of accepting IP address and port for 'openssl s_server'
1858 * Added a new API for TLSv1.3 ciphersuites:
1859 SSL_CTX_set_ciphersuites()
1860 SSL_set_ciphersuites()
1864 * Memory allocation failures consistently add an error to the error
1869 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1870 in libcrypto when run as setuid/setgid.
1874 * Load any config file by default when libssl is used.
1878 * Added new public header file <openssl/rand_drbg.h> and documentation
1879 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1881 *Matthias St. Pierre*
1883 * QNX support removed (cannot find contributors to get their approval
1884 for the license change).
1888 * TLSv1.3 replay protection for early data has been implemented. See the
1889 SSL_read_early_data() man page for further details.
1893 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1894 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1895 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1896 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1897 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1898 configuration has been separated out. See the ciphers man page or the
1899 SSL_CTX_set_ciphersuites() man page for more information.
1903 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1904 in responder mode now supports the new "-multi" option, which
1905 spawns the specified number of child processes to handle OCSP
1906 requests. The "-timeout" option now also limits the OCSP
1907 responder's patience to wait to receive the full client request
1908 on a newly accepted connection. Child processes are respawned
1909 as needed, and the CA index file is automatically reloaded
1910 when changed. This makes it possible to run the "ocsp" responder
1911 as a long-running service, making the OpenSSL CA somewhat more
1912 feature-complete. In this mode, most diagnostic messages logged
1913 after entering the event loop are logged via syslog(3) rather than
1918 * Added support for X448 and Ed448. Heavily based on original work by
1923 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1924 objects loaded. This adds the functions OSSL_STORE_expect() and
1925 OSSL_STORE_find() as well as needed tools to construct searches and
1926 get the search data out of them.
1930 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1931 version of OpenSSL should review their configuration settings to ensure
1932 that they are still appropriate for TLSv1.3. For further information see:
1933 <https://wiki.openssl.org/index.php/TLS1.3>
1937 * Grand redesign of the OpenSSL random generator
1939 The default RAND method now utilizes an AES-CTR DRBG according to
1940 NIST standard SP 800-90Ar1. The new random generator is essentially
1941 a port of the default random generator from the OpenSSL FIPS 2.0
1942 object module. It is a hybrid deterministic random bit generator
1943 using an AES-CTR bit stream and which seeds and reseeds itself
1944 automatically using trusted system entropy sources.
1946 Some of its new features are:
1947 - Support for multiple DRBG instances with seed chaining.
1948 - The default RAND method makes use of a DRBG.
1949 - There is a public and private DRBG instance.
1950 - The DRBG instances are fork-safe.
1951 - Keep all global DRBG instances on the secure heap if it is enabled.
1952 - The public and private DRBG instance are per thread for lock free
1955 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1957 * Changed Configure so it only says what it does and doesn't dump
1958 so much data. Instead, ./configdata.pm should be used as a script
1959 to display all sorts of configuration data.
1963 * Added processing of "make variables" to Configure.
1967 * Added SHA512/224 and SHA512/256 algorithm support.
1971 * The last traces of Netware support, first removed in 1.1.0, have
1976 * Get rid of Makefile.shared, and in the process, make the processing
1977 of certain files (rc.obj, or the .def/.map/.opt files produced from
1978 the ordinal files) more visible and hopefully easier to trace and
1979 debug (or make silent).
1983 * Make it possible to have environment variable assignments as
1984 arguments to config / Configure.
1988 * Add multi-prime RSA (RFC 8017) support.
1992 * Add SM3 implemented according to GB/T 32905-2016
1993 *Jack Lloyd <jack.lloyd@ribose.com>,*
1994 *Ronald Tse <ronald.tse@ribose.com>,*
1995 *Erick Borsboom <erick.borsboom@ribose.com>*
1997 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1998 as documented in RFC6066.
1999 Based on a patch from Tomasz Moń
2001 *Filipe Raimundo da Silva*
2003 * Add SM4 implemented according to GB/T 32907-2016.
2004 *Jack Lloyd <jack.lloyd@ribose.com>,*
2005 *Ronald Tse <ronald.tse@ribose.com>,*
2006 *Erick Borsboom <erick.borsboom@ribose.com>*
2008 * Reimplement -newreq-nodes and ERR_error_string_n; the
2009 original author does not agree with the license change.
2013 * Add ARIA AEAD TLS support.
2017 * Some macro definitions to support VS6 have been removed. Visual
2018 Studio 6 has not worked since 1.1.0
2022 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2023 without clearing the errors.
2027 * Add "atfork" functions. If building on a system that without
2028 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2029 requirements. The RAND facility now uses/requires this.
2037 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2038 not possible to disable entirely. However, it's still possible to
2039 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2042 To disable, configure with 'no-ui-console'. 'no-ui' is still
2043 possible to use as an alias. Check at compile time with the
2044 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2045 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2049 * Add a STORE module, which implements a uniform and URI based reader of
2050 stores that can contain keys, certificates, CRLs and numerous other
2051 objects. The main API is loosely based on a few stdio functions,
2052 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2053 OSSL_STORE_error and OSSL_STORE_close.
2054 The implementation uses backends called "loaders" to implement arbitrary
2055 URI schemes. There is one built in "loader" for the 'file' scheme.
2059 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2060 then adjusted to work on FreeBSD 8.4 as well.
2061 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2062 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2066 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2067 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2068 error code calls like this:
2070 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2072 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2073 that can be encoded in C. For the foreseeable future, this will only
2076 *Richard Levitte and Tim Hudson*
2078 * Removed BSD cryptodev engine.
2082 * Add a build target 'build_all_generated', to build all generated files
2083 and only that. This can be used to prepare everything that requires
2084 things like perl for a system that lacks perl and then move everything
2085 to that system and do the rest of the build there.
2089 * In the UI interface, make it possible to duplicate the user data. This
2090 can be used by engines that need to retain the data for a longer time
2091 than just the call where this user data is passed.
2095 * Ignore the '-named_curve auto' value for compatibility of applications
2098 *Tomas Mraz <tmraz@fedoraproject.org>*
2100 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2101 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2102 alerts across multiple records (some of which could be empty). In practice
2103 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2104 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2105 support this at all. Supporting it adds significant complexity to the
2106 record layer, and its removal is unlikely to cause interoperability
2111 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2112 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2113 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2118 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2119 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2121 *Richard Levitte, Andy Polyakov*
2123 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2128 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2129 platform rather than 'mingw'.
2133 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2134 success if they are asked to add an object which already exists
2135 in the store. This change cascades to other functions which load
2136 certificates and CRLs.
2140 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2141 facilitate stack unwinding even from assembly subroutines.
2145 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2146 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2150 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2151 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2152 which is the minimum version we support.
2156 * Certificate time validation (X509_cmp_time) enforces stricter
2157 compliance with RFC 5280. Fractional seconds and timezone offsets
2158 are no longer allowed.
2162 * Add support for ARIA
2166 * s_client will now send the Server Name Indication (SNI) extension by
2167 default unless the new "-noservername" option is used. The server name is
2168 based on the host provided to the "-connect" option unless overridden by
2169 using "-servername".
2173 * Add support for SipHash
2177 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2178 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2179 prevent issues where no progress is being made and the peer continually
2180 sends unrecognised record types, using up resources processing them.
2184 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2185 using the algorithm defined in
2186 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2190 * Heartbeat support has been removed; the ABI is changed for now.
2192 *Richard Levitte, Rich Salz*
2194 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2198 * The RSA "null" method, which was partially supported to avoid patent
2199 issues, has been replaced to always returns NULL.
2206 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2208 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2209 used even when parsing explicit parameters, when loading a encoded key
2210 or calling `EC_GROUP_new_from_ecpkparameters()`/
2211 `EC_GROUP_new_from_ecparameters()`.
2212 This prevents bypass of security hardening and performance gains,
2213 especially for curves with specialized EC_METHODs.
2214 By default, if a key encoded with explicit parameters is loaded and later
2215 encoded, the output is still encoded with explicit parameters, even if
2216 internally a "named" EC_GROUP is used for computation.
2220 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2221 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2222 NULL. After this change, only the cofactor parameter can be NULL. It also
2223 does some minimal sanity checks on the passed order.
2228 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2229 An attack is simple, if the first CMS_recipientInfo is valid but the
2230 second CMS_recipientInfo is chosen ciphertext. If the second
2231 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2232 encryption key will be replaced by garbage, and the message cannot be
2233 decoded, but if the RSA decryption fails, the correct encryption key is
2234 used and the recipient will not notice the attack.
2235 As a work around for this potential attack the length of the decrypted
2236 key must be equal to the cipher default key length, in case the
2237 certifiate is not given and all recipientInfo are tried out.
2238 The old behaviour can be re-enabled in the CMS code by setting the
2239 CMS_DEBUG_DECRYPT flag.
2244 * Use Windows installation paths in the mingw builds
2246 Mingw isn't a POSIX environment per se, which means that Windows
2247 paths should be used for installation.
2252 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2254 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2255 This changes the size when using the genpkey app when no size is given. It
2256 fixes an omission in earlier changes that changed all RSA, DSA and DH
2257 generation apps to use 2048 bits by default.
2261 * Prevent over long nonces in ChaCha20-Poly1305.
2263 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2264 for every encryption operation. RFC 7539 specifies that the nonce value
2265 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2266 and front pads the nonce with 0 bytes if it is less than 12
2267 bytes. However it also incorrectly allows a nonce to be set of up to 16
2268 bytes. In this case only the last 12 bytes are significant and any
2269 additional leading bytes are ignored.
2271 It is a requirement of using this cipher that nonce values are
2272 unique. Messages encrypted using a reused nonce value are susceptible to
2273 serious confidentiality and integrity attacks. If an application changes
2274 the default nonce length to be longer than 12 bytes and then makes a
2275 change to the leading bytes of the nonce expecting the new value to be a
2276 new unique nonce then such an application could inadvertently encrypt
2277 messages with a reused nonce.
2279 Additionally the ignored bytes in a long nonce are not covered by the
2280 integrity guarantee of this cipher. Any application that relies on the
2281 integrity of these ignored leading bytes of a long nonce may be further
2282 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2283 is safe because no such use sets such a long nonce value. However user
2284 applications that use this cipher directly and set a non-default nonce
2285 length to be longer than 12 bytes may be vulnerable.
2287 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2293 * Added SCA hardening for modular field inversion in EC_GROUP through
2294 a new dedicated field_inv() pointer in EC_METHOD.
2295 This also addresses a leakage affecting conversions from projective
2296 to affine coordinates.
2298 *Billy Bob Brumley, Nicola Tuveri*
2300 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2301 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2305 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2309 * Remove the 'dist' target and add a tarball building script. The
2310 'dist' target has fallen out of use, and it shouldn't be
2311 necessary to configure just to create a source distribution.
2315 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2317 * Timing vulnerability in DSA signature generation
2319 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2320 timing side channel attack. An attacker could use variations in the signing
2321 algorithm to recover the private key.
2323 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2328 * Timing vulnerability in ECDSA signature generation
2330 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2331 timing side channel attack. An attacker could use variations in the signing
2332 algorithm to recover the private key.
2334 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2339 * Add coordinate blinding for EC_POINT and implement projective
2340 coordinate blinding for generic prime curves as a countermeasure to
2341 chosen point SCA attacks.
2343 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2345 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2347 * Client DoS due to large DH parameter
2349 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2350 malicious server can send a very large prime value to the client. This will
2351 cause the client to spend an unreasonably long period of time generating a
2352 key for this prime resulting in a hang until the client has finished. This
2353 could be exploited in a Denial Of Service attack.
2355 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2360 * Cache timing vulnerability in RSA Key Generation
2362 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2363 a cache timing side channel attack. An attacker with sufficient access to
2364 mount cache timing attacks during the RSA key generation process could
2365 recover the private key.
2367 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2368 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2373 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2374 parameter is no longer accepted, as it leads to a corrupt table. NULL
2375 pem_str is reserved for alias entries only.
2379 * Revert blinding in ECDSA sign and instead make problematic addition
2380 length-invariant. Switch even to fixed-length Montgomery multiplication.
2384 * Change generating and checking of primes so that the error rate of not
2385 being prime depends on the intended use based on the size of the input.
2386 For larger primes this will result in more rounds of Miller-Rabin.
2387 The maximal error rate for primes with more than 1080 bits is lowered
2390 *Kurt Roeckx, Annie Yousar*
2392 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2396 * Add blinding to ECDSA and DSA signatures to protect against side channel
2397 attacks discovered by Keegan Ryan (NCC Group).
2401 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2402 now allow empty (zero character) pass phrases.
2406 * Certificate time validation (X509_cmp_time) enforces stricter
2407 compliance with RFC 5280. Fractional seconds and timezone offsets
2408 are no longer allowed.
2412 * Fixed a text canonicalisation bug in CMS
2414 Where a CMS detached signature is used with text content the text goes
2415 through a canonicalisation process first prior to signing or verifying a
2416 signature. This process strips trailing space at the end of lines, converts
2417 line terminators to CRLF and removes additional trailing line terminators
2418 at the end of a file. A bug in the canonicalisation process meant that
2419 some characters, such as form-feed, were incorrectly treated as whitespace
2420 and removed. This is contrary to the specification (RFC5485). This fix
2421 could mean that detached text data signed with an earlier version of
2422 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2423 signed with a fixed OpenSSL may fail to verify with an earlier version of
2424 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2425 and use the "-binary" flag (for the "cms" command line application) or set
2426 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2430 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2432 * Constructed ASN.1 types with a recursive definition could exceed the stack
2434 Constructed ASN.1 types with a recursive definition (such as can be found
2435 in PKCS7) could eventually exceed the stack given malicious input with
2436 excessive recursion. This could result in a Denial Of Service attack. There
2437 are no such structures used within SSL/TLS that come from untrusted sources
2438 so this is considered safe.
2440 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2446 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2448 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2449 effectively reduced to only comparing the least significant bit of each
2450 byte. This allows an attacker to forge messages that would be considered as
2451 authenticated in an amount of tries lower than that guaranteed by the
2452 security claims of the scheme. The module can only be compiled by the
2453 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2455 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2461 * Add a build target 'build_all_generated', to build all generated files
2462 and only that. This can be used to prepare everything that requires
2463 things like perl for a system that lacks perl and then move everything
2464 to that system and do the rest of the build there.
2468 * Backport SSL_OP_NO_RENGOTIATION
2470 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2471 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2472 changes this is no longer possible in 1.1.0. Therefore the new
2473 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2474 1.1.0 to provide equivalent functionality.
2476 Note that if an application built against 1.1.0h headers (or above) is run
2477 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2478 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2482 * Removed the OS390-Unix config target. It relied on a script that doesn't
2487 * rsaz_1024_mul_avx2 overflow bug on x86_64
2489 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2490 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2491 Analysis suggests that attacks against RSA and DSA as a result of this
2492 defect would be very difficult to perform and are not believed likely.
2493 Attacks against DH1024 are considered just feasible, because most of the
2494 work necessary to deduce information about a private key may be performed
2495 offline. The amount of resources required for such an attack would be
2496 significant. However, for an attack on TLS to be meaningful, the server
2497 would have to share the DH1024 private key among multiple clients, which is
2498 no longer an option since CVE-2016-0701.
2500 This only affects processors that support the AVX2 but not ADX extensions
2501 like Intel Haswell (4th generation).
2503 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2504 was originally found via the OSS-Fuzz project.
2509 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2511 * bn_sqrx8x_internal carry bug on x86_64
2513 There is a carry propagating bug in the x86_64 Montgomery squaring
2514 procedure. No EC algorithms are affected. Analysis suggests that attacks
2515 against RSA and DSA as a result of this defect would be very difficult to
2516 perform and are not believed likely. Attacks against DH are considered just
2517 feasible (although very difficult) because most of the work necessary to
2518 deduce information about a private key may be performed offline. The amount
2519 of resources required for such an attack would be very significant and
2520 likely only accessible to a limited number of attackers. An attacker would
2521 additionally need online access to an unpatched system using the target
2522 private key in a scenario with persistent DH parameters and a private
2523 key that is shared between multiple clients.
2525 This only affects processors that support the BMI1, BMI2 and ADX extensions
2526 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2528 This issue was reported to OpenSSL by the OSS-Fuzz project.
2533 * Malformed X.509 IPAddressFamily could cause OOB read
2535 If an X.509 certificate has a malformed IPAddressFamily extension,
2536 OpenSSL could do a one-byte buffer overread. The most likely result
2537 would be an erroneous display of the certificate in text format.
2539 This issue was reported to OpenSSL by the OSS-Fuzz project.
2544 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2546 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2547 platform rather than 'mingw'.
2551 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2552 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2553 which is the minimum version we support.
2557 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2559 * Encrypt-Then-Mac renegotiation crash
2561 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2562 negotiated where it was not in the original handshake (or vice-versa) then
2563 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2564 and servers are affected.
2566 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2571 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2573 * Truncated packet could crash via OOB read
2575 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2576 cipher is being used, then a truncated packet can cause that host to
2577 perform an out-of-bounds read, usually resulting in a crash.
2579 This issue was reported to OpenSSL by Robert Święcki of Google.
2584 * Bad (EC)DHE parameters cause a client crash
2586 If a malicious server supplies bad parameters for a DHE or ECDHE key
2587 exchange then this can result in the client attempting to dereference a
2588 NULL pointer leading to a client crash. This could be exploited in a Denial
2591 This issue was reported to OpenSSL by Guido Vranken.
2596 * BN_mod_exp may produce incorrect results on x86_64
2598 There is a carry propagating bug in the x86_64 Montgomery squaring
2599 procedure. No EC algorithms are affected. Analysis suggests that attacks
2600 against RSA and DSA as a result of this defect would be very difficult to
2601 perform and are not believed likely. Attacks against DH are considered just
2602 feasible (although very difficult) because most of the work necessary to
2603 deduce information about a private key may be performed offline. The amount
2604 of resources required for such an attack would be very significant and
2605 likely only accessible to a limited number of attackers. An attacker would
2606 additionally need online access to an unpatched system using the target
2607 private key in a scenario with persistent DH parameters and a private
2608 key that is shared between multiple clients. For example this can occur by
2609 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2610 similar to CVE-2015-3193 but must be treated as a separate problem.
2612 This issue was reported to OpenSSL by the OSS-Fuzz project.
2617 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2619 * ChaCha20/Poly1305 heap-buffer-overflow
2621 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2622 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2623 crash. This issue is not considered to be exploitable beyond a DoS.
2625 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2630 * CMS Null dereference
2632 Applications parsing invalid CMS structures can crash with a NULL pointer
2633 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2634 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2635 structure callback if an attempt is made to free certain invalid encodings.
2636 Only CHOICE structures using a callback which do not handle NULL value are
2639 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2644 * Montgomery multiplication may produce incorrect results
2646 There is a carry propagating bug in the Broadwell-specific Montgomery
2647 multiplication procedure that handles input lengths divisible by, but
2648 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2649 and DH private keys are impossible. This is because the subroutine in
2650 question is not used in operations with the private key itself and an input
2651 of the attacker's direct choice. Otherwise the bug can manifest itself as
2652 transient authentication and key negotiation failures or reproducible
2653 erroneous outcome of public-key operations with specially crafted input.
2654 Among EC algorithms only Brainpool P-512 curves are affected and one
2655 presumably can attack ECDH key negotiation. Impact was not analyzed in
2656 detail, because pre-requisites for attack are considered unlikely. Namely
2657 multiple clients have to choose the curve in question and the server has to
2658 share the private key among them, neither of which is default behaviour.
2659 Even then only clients that chose the curve will be affected.
2661 This issue was publicly reported as transient failures and was not
2662 initially recognized as a security issue. Thanks to Richard Morgan for
2663 providing reproducible case.
2668 * Removed automatic addition of RPATH in shared libraries and executables,
2669 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2673 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2675 * Fix Use After Free for large message sizes
2677 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2678 message larger than approx 16k is received then the underlying buffer to
2679 store the incoming message is reallocated and moved. Unfortunately a
2680 dangling pointer to the old location is left which results in an attempt to
2681 write to the previously freed location. This is likely to result in a
2682 crash, however it could potentially lead to execution of arbitrary code.
2684 This issue only affects OpenSSL 1.1.0a.
2686 This issue was reported to OpenSSL by Robert Święcki.
2691 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2693 * OCSP Status Request extension unbounded memory growth
2695 A malicious client can send an excessively large OCSP Status Request
2696 extension. If that client continually requests renegotiation, sending a
2697 large OCSP Status Request extension each time, then there will be unbounded
2698 memory growth on the server. This will eventually lead to a Denial Of
2699 Service attack through memory exhaustion. Servers with a default
2700 configuration are vulnerable even if they do not support OCSP. Builds using
2701 the "no-ocsp" build time option are not affected.
2703 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2708 * SSL_peek() hang on empty record
2710 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2711 sends an empty record. This could be exploited by a malicious peer in a
2712 Denial Of Service attack.
2714 This issue was reported to OpenSSL by Alex Gaynor.
2719 * Excessive allocation of memory in tls_get_message_header() and
2720 dtls1_preprocess_fragment()
2722 A (D)TLS message includes 3 bytes for its length in the header for the
2723 message. This would allow for messages up to 16Mb in length. Messages of
2724 this length are excessive and OpenSSL includes a check to ensure that a
2725 peer is sending reasonably sized messages in order to avoid too much memory
2726 being consumed to service a connection. A flaw in the logic of version
2727 1.1.0 means that memory for the message is allocated too early, prior to
2728 the excessive message length check. Due to way memory is allocated in
2729 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2730 to service a connection. This could lead to a Denial of Service through
2731 memory exhaustion. However, the excessive message length check still takes
2732 place, and this would cause the connection to immediately fail. Assuming
2733 that the application calls SSL_free() on the failed connection in a timely
2734 manner then the 21Mb of allocated memory will then be immediately freed
2735 again. Therefore the excessive memory allocation will be transitory in
2736 nature. This then means that there is only a security impact if:
2738 1) The application does not call SSL_free() in a timely manner in the event
2739 that the connection fails
2741 2) The application is working in a constrained environment where there is
2742 very little free memory
2744 3) The attacker initiates multiple connection attempts such that there are
2745 multiple connections in a state where memory has been allocated for the
2746 connection; SSL_free() has not yet been called; and there is insufficient
2747 memory to service the multiple requests.
2749 Except in the instance of (1) above any Denial Of Service is likely to be
2750 transitory because as soon as the connection fails the memory is
2751 subsequently freed again in the SSL_free() call. However there is an
2752 increased risk during this period of application crashes due to the lack of
2753 memory - which would then mean a more serious Denial of Service.
2755 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2756 (CVE-2016-6307 and CVE-2016-6308)
2760 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2761 had to be removed. Primary reason is that vendor assembler can't
2762 assemble our modules with -KPIC flag. As result it, assembly
2763 support, was not even available as option. But its lack means
2764 lack of side-channel resistant code, which is incompatible with
2765 security by todays standards. Fortunately gcc is readily available
2766 prepackaged option, which we firmly point at...
2770 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2772 * Windows command-line tool supports UTF-8 opt-in option for arguments
2773 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2774 (to any value) allows Windows user to access PKCS#12 file generated
2775 with Windows CryptoAPI and protected with non-ASCII password, as well
2776 as files generated under UTF-8 locale on Linux also protected with
2781 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2782 have been disabled by default and removed from DEFAULT, just like RC4.
2783 See the RC4 item below to re-enable both.
2787 * The method for finding the storage location for the Windows RAND seed file
2788 has changed. First we check %RANDFILE%. If that is not set then we check
2789 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2790 all else fails we fall back to C:\.
2794 * The EVP_EncryptUpdate() function has had its return type changed from void
2795 to int. A return of 0 indicates and error while a return of 1 indicates
2800 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2801 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2802 off the constant time implementation for RSA, DSA and DH have been made
2803 no-ops and deprecated.
2807 * Windows RAND implementation was simplified to only get entropy by
2808 calling CryptGenRandom(). Various other RAND-related tickets
2811 *Joseph Wylie Yandle, Rich Salz*
2813 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2814 and `OPENSSL_LH_`, respectively. The old names are available
2815 with API compatibility. They new names are now completely documented.
2819 * Unify TYPE_up_ref(obj) methods signature.
2820 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2821 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2822 int (instead of void) like all others TYPE_up_ref() methods.
2823 So now these methods also check the return value of CRYPTO_atomic_add(),
2824 and the validity of object reference counter.
2826 *fdasilvayy@gmail.com*
2828 * With Windows Visual Studio builds, the .pdb files are installed
2829 alongside the installed libraries and executables. For a static
2830 library installation, ossl_static.pdb is the associate compiler
2831 generated .pdb file to be used when linking programs.
2835 * Remove openssl.spec. Packaging files belong with the packagers.
2839 * Automatic Darwin/OSX configuration has had a refresh, it will now
2840 recognise x86_64 architectures automatically. You can still decide
2841 to build for a different bitness with the environment variable
2842 KERNEL_BITS (can be 32 or 64), for example:
2844 KERNEL_BITS=32 ./config
2848 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2849 256 bit AES and HMAC with SHA256.
2853 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2857 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2861 * To enable users to have their own config files and build file templates,
2862 Configure looks in the directory indicated by the environment variable
2863 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2864 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2865 name and is used as is.
2869 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2870 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2871 X509_CERT_FILE_CTX was removed.
2875 * "shared" builds are now the default. To create only static libraries use
2876 the "no-shared" Configure option.
2880 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2881 All of these option have not worked for some while and are fundamental
2886 * Make various cleanup routines no-ops and mark them as deprecated. Most
2887 global cleanup functions are no longer required because they are handled
2888 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2889 Explicitly de-initing can cause problems (e.g. where a library that uses
2890 OpenSSL de-inits, but an application is still using it). The affected
2891 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2892 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2893 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2894 COMP_zlib_cleanup().
2898 * --strict-warnings no longer enables runtime debugging options
2899 such as REF_DEBUG. Instead, debug options are automatically
2900 enabled with '--debug' builds.
2902 *Andy Polyakov, Emilia Käsper*
2904 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2905 have been moved out of the public header files. New functions for managing
2906 these have been added.
2910 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2911 objects have been moved out of the public header files. New
2912 functions for managing these have been added.
2916 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2917 have been moved out of the public header files. New functions for managing
2918 these have been added.
2922 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2923 moved out of the public header files. New functions for managing these
2928 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2932 * Removed the mk1mf build scripts.
2936 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2937 it is always safe to #include a header now.
2941 * Removed the aged BC-32 config and all its supporting scripts
2945 * Removed support for Ultrix, Netware, and OS/2.
2949 * Add support for HKDF.
2951 *Alessandro Ghedini*
2953 * Add support for blake2b and blake2s
2957 * Added support for "pipelining". Ciphers that have the
2958 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2959 encryptions/decryptions simultaneously. There are currently no built-in
2960 ciphers with this property but the expectation is that engines will be able
2961 to offer it to significantly improve throughput. Support has been extended
2962 into libssl so that multiple records for a single connection can be
2963 processed in one go (for >=TLS 1.1).
2967 * Added the AFALG engine. This is an async capable engine which is able to
2968 offload work to the Linux kernel. In this initial version it only supports
2969 AES128-CBC. The kernel must be version 4.1.0 or greater.
2973 * OpenSSL now uses a new threading API. It is no longer necessary to
2974 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2975 are two supported threading models: pthreads and windows threads. It is
2976 also possible to configure OpenSSL at compile time for "no-threads". The
2977 old threading API should no longer be used. The functions have been
2978 replaced with "no-op" compatibility macros.
2980 *Alessandro Ghedini, Matt Caswell*
2982 * Modify behavior of ALPN to invoke callback after SNI/servername
2983 callback, such that updates to the SSL_CTX affect ALPN.
2987 * Add SSL_CIPHER queries for authentication and key-exchange.
2991 * Changes to the DEFAULT cipherlist:
2992 - Prefer (EC)DHE handshakes over plain RSA.
2993 - Prefer AEAD ciphers over legacy ciphers.
2994 - Prefer ECDSA over RSA when both certificates are available.
2995 - Prefer TLSv1.2 ciphers/PRF.
2996 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3001 * Change the ECC default curve list to be this, in order: x25519,
3002 secp256r1, secp521r1, secp384r1.
3006 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3007 disabled by default. They can be re-enabled using the
3008 enable-weak-ssl-ciphers option to Configure.
3012 * If the server has ALPN configured, but supports no protocols that the
3013 client advertises, send a fatal "no_application_protocol" alert.
3014 This behaviour is SHALL in RFC 7301, though it isn't universally
3015 implemented by other servers.
3019 * Add X25519 support.
3020 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3021 for public and private key encoding using the format documented in
3022 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3023 key generation and key derivation.
3025 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3030 * Deprecate SRP_VBASE_get_by_user.
3031 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3032 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3033 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3034 seed, even if the seed is configured.
3036 Users should use SRP_VBASE_get1_by_user instead. Note that in
3037 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3038 also that even though configuring the SRP seed attempts to hide
3039 invalid usernames by continuing the handshake with fake
3040 credentials, this behaviour is not constant time and no strong
3041 guarantees are made that the handshake is indistinguishable from
3042 that of a valid user.
3046 * Configuration change; it's now possible to build dynamic engines
3047 without having to build shared libraries and vice versa. This
3048 only applies to the engines in engines/, those in crypto/engine/
3049 will always be built into libcrypto (i.e. "static").
3051 Building dynamic engines is enabled by default; to disable, use
3052 the configuration option "disable-dynamic-engine".
3054 The only requirements for building dynamic engines are the
3055 presence of the DSO module and building with position independent
3056 code, so they will also automatically be disabled if configuring
3057 with "disable-dso" or "disable-pic".
3059 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3060 are also taken away from openssl/opensslconf.h, as they are
3065 * Configuration change; if there is a known flag to compile
3066 position independent code, it will always be applied on the
3067 libcrypto and libssl object files, and never on the application
3068 object files. This means other libraries that use routines from
3069 libcrypto / libssl can be made into shared libraries regardless
3070 of how OpenSSL was configured.
3072 If this isn't desirable, the configuration options "disable-pic"
3073 or "no-pic" can be used to disable the use of PIC. This will
3074 also disable building shared libraries and dynamic engines.
3078 * Removed JPAKE code. It was experimental and has no wide use.
3082 * The INSTALL_PREFIX Makefile variable has been renamed to
3083 DESTDIR. That makes for less confusion on what this variable
3084 is for. Also, the configuration option --install_prefix is
3089 * Heartbeat for TLS has been removed and is disabled by default
3090 for DTLS; configure with enable-heartbeats. Code that uses the
3091 old #define's might need to be updated.
3093 *Emilia Käsper, Rich Salz*
3095 * Rename REF_CHECK to REF_DEBUG.
3099 * New "unified" build system
3101 The "unified" build system is aimed to be a common system for all
3102 platforms we support. With it comes new support for VMS.
3104 This system builds supports building in a different directory tree
3105 than the source tree. It produces one Makefile (for unix family
3106 or lookalikes), or one descrip.mms (for VMS).
3108 The source of information to make the Makefile / descrip.mms is
3109 small files called 'build.info', holding the necessary
3110 information for each directory with source to compile, and a
3111 template in Configurations, like unix-Makefile.tmpl or
3114 With this change, the library names were also renamed on Windows
3115 and on VMS. They now have names that are closer to the standard
3116 on Unix, and include the major version number, and in certain
3117 cases, the architecture they are built for. See "Notes on shared
3118 libraries" in INSTALL.
3120 We rely heavily on the perl module Text::Template.
3124 * Added support for auto-initialisation and de-initialisation of the library.
3125 OpenSSL no longer requires explicit init or deinit routines to be called,
3126 except in certain circumstances. See the OPENSSL_init_crypto() and
3127 OPENSSL_init_ssl() man pages for further information.
3131 * The arguments to the DTLSv1_listen function have changed. Specifically the
3132 "peer" argument is now expected to be a BIO_ADDR object.
3134 * Rewrite of BIO networking library. The BIO library lacked consistent
3135 support of IPv6, and adding it required some more extensive
3136 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3137 which hold all types of addresses and chains of address information.
3138 It also introduces a new API, with functions like BIO_socket,
3139 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3140 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3141 have been adapted accordingly.
3145 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3150 * CRIME protection: disable compression by default, even if OpenSSL is
3151 compiled with zlib enabled. Applications can still enable compression
3152 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3153 using the SSL_CONF library to configure compression.
3157 * The signature of the session callback configured with
3158 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3159 was explicitly marked as `const unsigned char*` instead of
3164 * Always DPURIFY. Remove the use of uninitialized memory in the
3165 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3169 * Removed many obsolete configuration items, including
3170 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3171 MD2_CHAR, MD2_INT, MD2_LONG
3173 IDEA_SHORT, IDEA_LONG
3174 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3176 *Rich Salz, with advice from Andy Polyakov*
3178 * Many BN internals have been moved to an internal header file.
3180 *Rich Salz with help from Andy Polyakov*
3182 * Configuration and writing out the results from it has changed.
3183 Files such as Makefile include/openssl/opensslconf.h and are now
3184 produced through general templates, such as Makefile.in and
3185 crypto/opensslconf.h.in and some help from the perl module
3188 Also, the center of configuration information is no longer
3189 Makefile. Instead, Configure produces a perl module in
3190 configdata.pm which holds most of the config data (in the hash
3191 table %config), the target data that comes from the target
3192 configuration in one of the `Configurations/*.conf` files (in
3197 * To clarify their intended purposes, the Configure options
3198 --prefix and --openssldir change their semantics, and become more
3199 straightforward and less interdependent.
3201 --prefix shall be used exclusively to give the location INSTALLTOP
3202 where programs, scripts, libraries, include files and manuals are
3203 going to be installed. The default is now /usr/local.
3205 --openssldir shall be used exclusively to give the default
3206 location OPENSSLDIR where certificates, private keys, CRLs are
3207 managed. This is also where the default openssl.cnf gets
3209 If the directory given with this option is a relative path, the
3210 values of both the --prefix value and the --openssldir value will
3211 be combined to become OPENSSLDIR.
3212 The default for --openssldir is INSTALLTOP/ssl.
3214 Anyone who uses --openssldir to specify where OpenSSL is to be
3215 installed MUST change to use --prefix instead.
3219 * The GOST engine was out of date and therefore it has been removed. An up
3220 to date GOST engine is now being maintained in an external repository.
3221 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3222 support for GOST ciphersuites (these are only activated if a GOST engine
3227 * EGD is no longer supported by default; use enable-egd when
3230 *Ben Kaduk and Rich Salz*
3232 * The distribution now has Makefile.in files, which are used to
3233 create Makefile's when Configure is run. *Configure must be run
3234 before trying to build now.*
3238 * The return value for SSL_CIPHER_description() for error conditions
3243 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3245 Obtaining and performing DNSSEC validation of TLSA records is
3246 the application's responsibility. The application provides
3247 the TLSA records of its choice to OpenSSL, and these are then
3248 used to authenticate the peer.
3250 The TLSA records need not even come from DNS. They can, for
3251 example, be used to implement local end-entity certificate or
3252 trust-anchor "pinning", where the "pin" data takes the form
3253 of TLSA records, which can augment or replace verification
3254 based on the usual WebPKI public certification authorities.
3258 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3259 continues to support deprecated interfaces in default builds.
3260 However, applications are strongly advised to compile their
3261 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3262 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3263 or the 1.1.0 releases.
3265 In environments in which all applications have been ported to
3266 not use any deprecated interfaces OpenSSL's Configure script
3267 should be used with the --api=1.1.0 option to entirely remove
3268 support for the deprecated features from the library and
3269 unconditionally disable them in the installed headers.
3270 Essentially the same effect can be achieved with the "no-deprecated"
3271 argument to Configure, except that this will always restrict
3272 the build to just the latest API, rather than a fixed API
3275 As applications are ported to future revisions of the API,
3276 they should update their compile-time OPENSSL_API_COMPAT define
3277 accordingly, but in most cases should be able to continue to
3278 compile with later releases.
3280 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3281 0x10000000L and 0x00908000L, respectively. However those
3282 versions did not support the OPENSSL_API_COMPAT feature, and
3283 so applications are not typically tested for explicit support
3284 of just the undeprecated features of either release.
3288 * Add support for setting the minimum and maximum supported protocol.
3289 It can bet set via the SSL_set_min_proto_version() and
3290 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3291 MaxProtocol. It's recommended to use the new APIs to disable
3292 protocols instead of disabling individual protocols using
3293 SSL_set_options() or SSL_CONF's Protocol. This change also
3294 removes support for disabling TLS 1.2 in the OpenSSL TLS
3295 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3299 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3303 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3304 and integrates ECDSA and ECDH functionality into EC. Implementations can
3305 now redirect key generation and no longer need to convert to or from
3308 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3309 include the ec.h header file instead.
3313 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3314 ciphers who are no longer supported and drops support the ephemeral RSA key
3315 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3319 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3320 opaque. For HMAC_CTX, the following constructors and destructors
3323 HMAC_CTX *HMAC_CTX_new(void);
3324 void HMAC_CTX_free(HMAC_CTX *ctx);
3326 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3327 destroy such methods has been added. See EVP_MD_meth_new(3) and
3328 EVP_CIPHER_meth_new(3) for documentation.
3331 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3332 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3333 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3334 an already created structure.
3335 2) For consistency with the majority of our object creators and
3336 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3337 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3338 for deprecated builds.
3342 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3343 cryptographic operations to be performed asynchronously as long as an
3344 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3345 further details. Libssl has also had this capability integrated with the
3346 introduction of the new mode SSL_MODE_ASYNC and associated error
3347 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3348 pages. This work was developed in partnership with Intel Corp.
3352 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3353 always enabled now. If you want to disable the support you should
3354 exclude it using the list of supported ciphers. This also means that the
3355 "-no_ecdhe" option has been removed from s_server.
3359 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3360 SSL_{CTX_}set1_curves() which can set a list.
3364 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3365 curve you want to support using SSL_{CTX_}set1_curves().
3369 * State machine rewrite. The state machine code has been significantly
3370 refactored in order to remove much duplication of code and solve issues
3371 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3372 further details). This change does have some associated API changes.
3373 Notably the SSL_state() function has been removed and replaced by
3374 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3375 SSL_set_state() has been removed altogether. The previous handshake states
3376 defined in ssl.h and ssl3.h have also been removed.
3380 * All instances of the string "ssleay" in the public API were replaced
3381 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3382 Some error codes related to internal RSA_eay API's were renamed.
3386 * The demo files in crypto/threads were moved to demo/threads.
3390 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3393 *Matt Caswell, Rich Salz*
3395 * New ASN.1 embed macro.
3397 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3398 structure is not allocated: it is part of the parent. That is instead of
3406 This reduces memory fragmentation and make it impossible to accidentally
3407 set a mandatory field to NULL.
3409 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3410 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3411 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3416 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3420 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3421 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3422 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3423 DES and RC4 ciphersuites.
3427 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3428 This changes the decoding behaviour for some invalid messages,
3429 though the change is mostly in the more lenient direction, and
3430 legacy behaviour is preserved as much as possible.
3434 * Fix no-stdio build.
3435 *David Woodhouse <David.Woodhouse@intel.com> and also*
3436 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3438 * New testing framework
3439 The testing framework has been largely rewritten and is now using
3440 perl and the perl modules Test::Harness and an extended variant of
3441 Test::More called OpenSSL::Test to do its work. All test scripts in
3442 test/ have been rewritten into test recipes, and all direct calls to
3443 executables in test/Makefile have become individual recipes using the
3444 simplified testing OpenSSL::Test::Simple.
3446 For documentation on our testing modules, do:
3448 perldoc test/testlib/OpenSSL/Test/Simple.pm
3449 perldoc test/testlib/OpenSSL/Test.pm
3453 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3454 are used; the latter aborts on memory leaks (usually checked on exit).
3455 Some undocumented "set malloc, etc., hooks" functions were removed
3456 and others were changed. All are now documented.
3460 * In DSA_generate_parameters_ex, if the provided seed is too short,
3463 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3465 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3466 from RFC4279, RFC4785, RFC5487, RFC5489.
3468 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3469 original RSA_PSK patch.
3473 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3474 era flag was never set throughout the codebase (only read). Also removed
3475 SSL3_FLAGS_POP_BUFFER which was only used if
3476 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3480 * Changed the default name options in the "ca", "crl", "req" and "x509"
3481 to be "oneline" instead of "compat".
3485 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3486 not aware of clients that still exhibit this bug, and the workaround
3487 hasn't been working properly for a while.
3491 * The return type of BIO_number_read() and BIO_number_written() as well as
3492 the corresponding num_read and num_write members in the BIO structure has
3493 changed from unsigned long to uint64_t. On platforms where an unsigned
3494 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3499 * Given the pervasive nature of TLS extensions it is inadvisable to run
3500 OpenSSL without support for them. It also means that maintaining
3501 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3502 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3506 * Removed support for the two export grade static DH ciphersuites
3507 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3508 were newly added (along with a number of other static DH ciphersuites) to
3509 1.0.2. However the two export ones have *never* worked since they were
3510 introduced. It seems strange in any case to be adding new export
3511 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3515 * Version negotiation has been rewritten. In particular SSLv23_method(),
3516 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3517 and turned into macros which simply call the new preferred function names
3518 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3519 should use the new names instead. Also as part of this change the ssl23.h
3520 header file has been removed.
3524 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3525 code and the associated standard is no longer considered fit-for-purpose.
3529 * RT2547 was closed. When generating a private key, try to make the
3530 output file readable only by the owner. This behavior change might
3531 be noticeable when interacting with other software.
3533 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3538 * Added HTTP GET support to the ocsp command.
3542 * Changed default digest for the dgst and enc commands from MD5 to
3547 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3551 * Added support for TLS extended master secret from
3552 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3553 initial patch which was a great help during development.
3557 * All libssl internal structures have been removed from the public header
3558 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3559 now redundant). Users should not attempt to access internal structures
3560 directly. Instead they should use the provided API functions.
3564 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3565 Access to deprecated functions can be re-enabled by running config with
3566 "enable-deprecated". In addition applications wishing to use deprecated
3567 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3568 will, by default, disable some transitive includes that previously existed
3569 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3573 * Added support for OCB mode. OpenSSL has been granted a patent license
3574 compatible with the OpenSSL license for use of OCB. Details are available
3575 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3576 for OCB can be removed by calling config with no-ocb.
3580 * SSLv2 support has been removed. It still supports receiving a SSLv2
3581 compatible client hello.
3585 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3586 done while fixing the error code for the key-too-small case.
3588 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3590 * CA.sh has been removed; use CA.pl instead.
3594 * Removed old DES API.
3598 * Remove various unsupported platforms:
3604 Sinix/ReliantUNIX RM400
3609 16-bit platforms such as WIN16
3613 * Clean up OPENSSL_NO_xxx #define's
3614 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3615 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3616 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3617 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3618 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3619 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3620 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3621 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3622 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3623 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3627 * Cleaned up dead code
3628 Remove all but one '#ifdef undef' which is to be looked at.
3632 * Clean up calling of xxx_free routines.
3633 Just like free(), fix most of the xxx_free routines to accept
3634 NULL. Remove the non-null checks from callers. Save much code.
3638 * Add secure heap for storage of private keys (when possible).
3639 Add BIO_s_secmem(), CBIGNUM, etc.
3640 Contributed by Akamai Technologies under our Corporate CLA.
3644 * Experimental support for a new, fast, unbiased prime candidate generator,
3645 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3647 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3649 * New output format NSS in the sess_id command line tool. This allows
3650 exporting the session id and the master key in NSS keylog format.
3652 *Martin Kaiser <martin@kaiser.cx>*
3654 * Harmonize version and its documentation. -f flag is used to display
3657 *mancha <mancha1@zoho.com>*
3659 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3660 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3662 *mancha <mancha1@zoho.com>*
3664 * Fix some double frees. These are not thought to be exploitable.
3666 *mancha <mancha1@zoho.com>*
3668 * A missing bounds check in the handling of the TLS heartbeat extension
3669 can be used to reveal up to 64k of memory to a connected client or
3672 Thanks for Neel Mehta of Google Security for discovering this bug and to
3673 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3674 preparing the fix ([CVE-2014-0160])
3676 *Adam Langley, Bodo Moeller*
3678 * Fix for the attack described in the paper "Recovering OpenSSL
3679 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3680 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3681 <http://eprint.iacr.org/2014/140>
3683 Thanks to Yuval Yarom and Naomi Benger for discovering this
3684 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3686 *Yuval Yarom and Naomi Benger*
3688 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3689 this fixes a limitation in previous versions of OpenSSL.
3693 * Experimental encrypt-then-mac support.
3695 Experimental support for encrypt then mac from
3696 draft-gutmann-tls-encrypt-then-mac-02.txt
3698 To enable it set the appropriate extension number (0x42 for the test
3699 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3701 For non-compliant peers (i.e. just about everything) this should have no
3704 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3708 * Add EVP support for key wrapping algorithms, to avoid problems with
3709 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3710 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3711 algorithms and include tests cases.
3715 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3720 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3721 MGF1 digest and OAEP label.
3725 * Make openssl verify return errors.
3727 *Chris Palmer <palmer@google.com> and Ben Laurie*
3729 * New function ASN1_TIME_diff to calculate the difference between two
3730 ASN1_TIME structures or one structure and the current time.
3734 * Update fips_test_suite to support multiple command line options. New
3735 test to induce all self test errors in sequence and check expected
3740 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3741 sign or verify all in one operation.
3745 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3746 test programs and fips_test_suite. Includes functionality to parse
3747 the minimal script output of fipsalgest.pl directly.
3751 * Add authorisation parameter to FIPS_module_mode_set().
3755 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3759 * Use separate DRBG fields for internal and external flags. New function
3760 FIPS_drbg_health_check() to perform on demand health checking. Add
3761 generation tests to fips_test_suite with reduced health check interval to
3762 demonstrate periodic health checking. Add "nodh" option to
3763 fips_test_suite to skip very slow DH test.
3767 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3772 * More extensive health check for DRBG checking many more failure modes.
3773 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3774 combination: call this in fips_test_suite.
3778 * Add support for canonical generation of DSA parameter 'g'. See
3781 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3782 POST to handle HMAC cases.
3786 * Add functions FIPS_module_version() and FIPS_module_version_text()
3787 to return numerical and string versions of the FIPS module number.
3791 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3792 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3793 outside the validated module in the FIPS capable OpenSSL.
3797 * Minor change to DRBG entropy callback semantics. In some cases
3798 there is no multiple of the block length between min_len and
3799 max_len. Allow the callback to return more than max_len bytes
3800 of entropy but discard any extra: it is the callback's responsibility
3801 to ensure that the extra data discarded does not impact the
3802 requested amount of entropy.
3806 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3807 information in FIPS186-3, SP800-57 and SP800-131A.
3811 * CCM support via EVP. Interface is very similar to GCM case except we
3812 must supply all data in one chunk (i.e. no update, final) and the
3813 message length must be supplied if AAD is used. Add algorithm test
3818 * Initial version of POST overhaul. Add POST callback to allow the status
3819 of POST to be monitored and/or failures induced. Modify fips_test_suite
3820 to use callback. Always run all selftests even if one fails.
3824 * XTS support including algorithm test driver in the fips_gcmtest program.
3825 Note: this does increase the maximum key length from 32 to 64 bytes but
3826 there should be no binary compatibility issues as existing applications
3827 will never use XTS mode.
3831 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3832 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3833 performs algorithm blocking for unapproved PRNG types. Also do not
3834 set PRNG type in FIPS_mode_set(): leave this to the application.
3835 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3836 the standard OpenSSL PRNG: set additional data to a date time vector.
3840 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3841 This shouldn't present any incompatibility problems because applications
3842 shouldn't be using these directly and any that are will need to rethink
3843 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3847 * Extensive self tests and health checking required by SP800-90 DRBG.
3848 Remove strength parameter from FIPS_drbg_instantiate and always
3849 instantiate at maximum supported strength.
3853 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3857 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3861 * New function DH_compute_key_padded() to compute a DH key and pad with
3862 leading zeroes if needed: this complies with SP800-56A et al.
3866 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3867 anything, incomplete, subject to change and largely untested at present.
3871 * Modify fipscanisteronly build option to only build the necessary object
3872 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3876 * Add experimental option FIPSSYMS to give all symbols in
3877 fipscanister.o and FIPS or fips prefix. This will avoid
3878 conflicts with future versions of OpenSSL. Add perl script
3879 util/fipsas.pl to preprocess assembly language source files
3880 and rename any affected symbols.
3884 * Add selftest checks and algorithm block of non-fips algorithms in
3885 FIPS mode. Remove DES2 from selftests.
3889 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3890 return internal method without any ENGINE dependencies. Add new
3891 tiny fips sign and verify functions.
3895 * New build option no-ec2m to disable characteristic 2 code.
3899 * New build option "fipscanisteronly". This only builds fipscanister.o
3900 and (currently) associated fips utilities. Uses the file Makefile.fips
3901 instead of Makefile.org as the prototype.
3905 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3906 Update fips_gcmtest to use IV generator.
3910 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3911 setting output buffer to NULL. The `*Final` function must be
3912 called although it will not retrieve any additional data. The tag
3913 can be set or retrieved with a ctrl. The IV length is by default 12
3914 bytes (96 bits) but can be set to an alternative value. If the IV
3915 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3920 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3921 underlying do_cipher function handles all cipher semantics itself
3922 including padding and finalisation. This is useful if (for example)
3923 an ENGINE cipher handles block padding itself. The behaviour of
3924 do_cipher is subtly changed if this flag is set: the return value
3925 is the number of characters written to the output buffer (zero is
3926 no longer an error code) or a negative error code. Also if the
3927 input buffer is NULL and length 0 finalisation should be performed.
3931 * If a candidate issuer certificate is already part of the constructed
3932 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3936 * Improve forward-security support: add functions
3938 void SSL_CTX_set_not_resumable_session_callback(
3939 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3940 void SSL_set_not_resumable_session_callback(
3941 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3943 for use by SSL/TLS servers; the callback function will be called whenever a
3944 new session is created, and gets to decide whether the session may be
3945 cached to make it resumable (return 0) or not (return 1). (As by the
3946 SSL/TLS protocol specifications, the session_id sent by the server will be
3947 empty to indicate that the session is not resumable; also, the server will
3948 not generate RFC 4507 (RFC 5077) session tickets.)
3950 A simple reasonable callback implementation is to return is_forward_secure.
3951 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3952 by the SSL/TLS server library, indicating whether it can provide forward
3955 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3957 * New -verify_name option in command line utilities to set verification
3962 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3963 Add CMAC pkey methods.
3967 * Experimental renegotiation in s_server -www mode. If the client
3968 browses /reneg connection is renegotiated. If /renegcert it is
3969 renegotiated requesting a certificate.
3973 * Add an "external" session cache for debugging purposes to s_server. This
3974 should help trace issues which normally are only apparent in deployed
3975 multi-process servers.
3979 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3980 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3981 BIO_set_cipher() and some obscure PEM functions were changed so they
3982 can now return an error. The RAND changes required a change to the
3983 RAND_METHOD structure.
3987 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3988 a gcc attribute to warn if the result of a function is ignored. This
3989 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3990 whose return value is often ignored.
3994 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3995 These allow SCTs (signed certificate timestamps) to be requested and
3996 validated when establishing a connection.
3998 *Rob Percival <robpercival@google.com>*
4003 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4005 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4006 used even when parsing explicit parameters, when loading a encoded key
4007 or calling `EC_GROUP_new_from_ecpkparameters()`/
4008 `EC_GROUP_new_from_ecparameters()`.
4009 This prevents bypass of security hardening and performance gains,
4010 especially for curves with specialized EC_METHODs.
4011 By default, if a key encoded with explicit parameters is loaded and later
4012 encoded, the output is still encoded with explicit parameters, even if
4013 internally a "named" EC_GROUP is used for computation.
4017 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4018 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4019 NULL. After this change, only the cofactor parameter can be NULL. It also
4020 does some minimal sanity checks on the passed order.
4025 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4026 An attack is simple, if the first CMS_recipientInfo is valid but the
4027 second CMS_recipientInfo is chosen ciphertext. If the second
4028 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4029 encryption key will be replaced by garbage, and the message cannot be
4030 decoded, but if the RSA decryption fails, the correct encryption key is
4031 used and the recipient will not notice the attack.
4032 As a work around for this potential attack the length of the decrypted
4033 key must be equal to the cipher default key length, in case the
4034 certifiate is not given and all recipientInfo are tried out.
4035 The old behaviour can be re-enabled in the CMS code by setting the
4036 CMS_DEBUG_DECRYPT flag.
4041 * Document issue with installation paths in diverse Windows builds
4043 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4044 binaries and run-time config file.
4049 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4051 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4052 This changes the size when using the genpkey app when no size is given. It
4053 fixes an omission in earlier changes that changed all RSA, DSA and DH
4054 generation apps to use 2048 bits by default.
4058 * Add FIPS support for Android Arm 64-bit
4060 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4061 Module in Version 2.0.10. For some reason, the corresponding target
4062 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4063 built with FIPS support on Android Arm 64-bit. This omission has been
4066 *Matthias St. Pierre*
4068 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4070 * 0-byte record padding oracle
4072 If an application encounters a fatal protocol error and then calls
4073 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4074 then OpenSSL can respond differently to the calling application if a 0 byte
4075 record is received with invalid padding compared to if a 0 byte record is
4076 received with an invalid MAC. If the application then behaves differently
4077 based on that in a way that is detectable to the remote peer, then this
4078 amounts to a padding oracle that could be used to decrypt data.
4080 In order for this to be exploitable "non-stitched" ciphersuites must be in
4081 use. Stitched ciphersuites are optimised implementations of certain
4082 commonly used ciphersuites. Also the application must call SSL_shutdown()
4083 twice even if a protocol error has occurred (applications should not do
4084 this but some do anyway).
4086 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4087 Aviram, with additional investigation by Steven Collison and Andrew
4088 Hourselt. It was reported to OpenSSL on 10th December 2018.
4093 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4097 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4099 * Microarchitecture timing vulnerability in ECC scalar multiplication
4101 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4102 shown to be vulnerable to a microarchitecture timing side channel attack.
4103 An attacker with sufficient access to mount local timing attacks during
4104 ECDSA signature generation could recover the private key.
4106 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4107 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4113 * Timing vulnerability in DSA signature generation
4115 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4116 timing side channel attack. An attacker could use variations in the signing
4117 algorithm to recover the private key.
4119 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4124 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4125 Module, accidentally introduced while backporting security fixes from the
4126 development branch and hindering the use of ECC in FIPS mode.
4130 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4132 * Client DoS due to large DH parameter
4134 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4135 malicious server can send a very large prime value to the client. This will
4136 cause the client to spend an unreasonably long period of time generating a
4137 key for this prime resulting in a hang until the client has finished. This
4138 could be exploited in a Denial Of Service attack.
4140 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4145 * Cache timing vulnerability in RSA Key Generation
4147 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4148 a cache timing side channel attack. An attacker with sufficient access to
4149 mount cache timing attacks during the RSA key generation process could
4150 recover the private key.
4152 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4153 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4158 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4159 parameter is no longer accepted, as it leads to a corrupt table. NULL
4160 pem_str is reserved for alias entries only.
4164 * Revert blinding in ECDSA sign and instead make problematic addition
4165 length-invariant. Switch even to fixed-length Montgomery multiplication.
4169 * Change generating and checking of primes so that the error rate of not
4170 being prime depends on the intended use based on the size of the input.
4171 For larger primes this will result in more rounds of Miller-Rabin.
4172 The maximal error rate for primes with more than 1080 bits is lowered
4175 *Kurt Roeckx, Annie Yousar*
4177 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4181 * Add blinding to ECDSA and DSA signatures to protect against side channel
4182 attacks discovered by Keegan Ryan (NCC Group).
4186 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4187 now allow empty (zero character) pass phrases.
4191 * Certificate time validation (X509_cmp_time) enforces stricter
4192 compliance with RFC 5280. Fractional seconds and timezone offsets
4193 are no longer allowed.
4197 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4199 * Constructed ASN.1 types with a recursive definition could exceed the stack
4201 Constructed ASN.1 types with a recursive definition (such as can be found
4202 in PKCS7) could eventually exceed the stack given malicious input with
4203 excessive recursion. This could result in a Denial Of Service attack. There
4204 are no such structures used within SSL/TLS that come from untrusted sources
4205 so this is considered safe.
4207 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4213 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4215 * Read/write after SSL object in error state
4217 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4218 mechanism. The intent was that if a fatal error occurred during a handshake
4219 then OpenSSL would move into the error state and would immediately fail if
4220 you attempted to continue the handshake. This works as designed for the
4221 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4222 SSL_connect()), however due to a bug it does not work correctly if
4223 SSL_read() or SSL_write() is called directly. In that scenario, if the
4224 handshake fails then a fatal error will be returned in the initial function
4225 call. If SSL_read()/SSL_write() is subsequently called by the application
4226 for the same SSL object then it will succeed and the data is passed without
4227 being decrypted/encrypted directly from the SSL/TLS record layer.
4229 In order to exploit this issue an application bug would have to be present
4230 that resulted in a call to SSL_read()/SSL_write() being issued after having
4231 already received a fatal error.
4233 This issue was reported to OpenSSL by David Benjamin (Google).
4238 * rsaz_1024_mul_avx2 overflow bug on x86_64
4240 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4241 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4242 Analysis suggests that attacks against RSA and DSA as a result of this
4243 defect would be very difficult to perform and are not believed likely.
4244 Attacks against DH1024 are considered just feasible, because most of the
4245 work necessary to deduce information about a private key may be performed
4246 offline. The amount of resources required for such an attack would be
4247 significant. However, for an attack on TLS to be meaningful, the server
4248 would have to share the DH1024 private key among multiple clients, which is
4249 no longer an option since CVE-2016-0701.
4251 This only affects processors that support the AVX2 but not ADX extensions
4252 like Intel Haswell (4th generation).
4254 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4255 was originally found via the OSS-Fuzz project.
4260 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4262 * bn_sqrx8x_internal carry bug on x86_64
4264 There is a carry propagating bug in the x86_64 Montgomery squaring
4265 procedure. No EC algorithms are affected. Analysis suggests that attacks
4266 against RSA and DSA as a result of this defect would be very difficult to
4267 perform and are not believed likely. Attacks against DH are considered just
4268 feasible (although very difficult) because most of the work necessary to
4269 deduce information about a private key may be performed offline. The amount
4270 of resources required for such an attack would be very significant and
4271 likely only accessible to a limited number of attackers. An attacker would
4272 additionally need online access to an unpatched system using the target
4273 private key in a scenario with persistent DH parameters and a private
4274 key that is shared between multiple clients.
4276 This only affects processors that support the BMI1, BMI2 and ADX extensions
4277 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4279 This issue was reported to OpenSSL by the OSS-Fuzz project.
4284 * Malformed X.509 IPAddressFamily could cause OOB read
4286 If an X.509 certificate has a malformed IPAddressFamily extension,
4287 OpenSSL could do a one-byte buffer overread. The most likely result
4288 would be an erroneous display of the certificate in text format.
4290 This issue was reported to OpenSSL by the OSS-Fuzz project.
4295 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4297 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4298 platform rather than 'mingw'.
4302 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4304 * Truncated packet could crash via OOB read
4306 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4307 cipher is being used, then a truncated packet can cause that host to
4308 perform an out-of-bounds read, usually resulting in a crash.
4310 This issue was reported to OpenSSL by Robert Święcki of Google.
4315 * BN_mod_exp may produce incorrect results on x86_64
4317 There is a carry propagating bug in the x86_64 Montgomery squaring
4318 procedure. No EC algorithms are affected. Analysis suggests that attacks
4319 against RSA and DSA as a result of this defect would be very difficult to
4320 perform and are not believed likely. Attacks against DH are considered just
4321 feasible (although very difficult) because most of the work necessary to
4322 deduce information about a private key may be performed offline. The amount
4323 of resources required for such an attack would be very significant and
4324 likely only accessible to a limited number of attackers. An attacker would
4325 additionally need online access to an unpatched system using the target
4326 private key in a scenario with persistent DH parameters and a private
4327 key that is shared between multiple clients. For example this can occur by
4328 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4329 similar to CVE-2015-3193 but must be treated as a separate problem.
4331 This issue was reported to OpenSSL by the OSS-Fuzz project.
4336 * Montgomery multiplication may produce incorrect results
4338 There is a carry propagating bug in the Broadwell-specific Montgomery
4339 multiplication procedure that handles input lengths divisible by, but
4340 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4341 and DH private keys are impossible. This is because the subroutine in
4342 question is not used in operations with the private key itself and an input
4343 of the attacker's direct choice. Otherwise the bug can manifest itself as
4344 transient authentication and key negotiation failures or reproducible
4345 erroneous outcome of public-key operations with specially crafted input.
4346 Among EC algorithms only Brainpool P-512 curves are affected and one
4347 presumably can attack ECDH key negotiation. Impact was not analyzed in
4348 detail, because pre-requisites for attack are considered unlikely. Namely
4349 multiple clients have to choose the curve in question and the server has to
4350 share the private key among them, neither of which is default behaviour.
4351 Even then only clients that chose the curve will be affected.
4353 This issue was publicly reported as transient failures and was not
4354 initially recognized as a security issue. Thanks to Richard Morgan for
4355 providing reproducible case.
4360 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4361 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4362 prevent issues where no progress is being made and the peer continually
4363 sends unrecognised record types, using up resources processing them.
4367 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4369 * Missing CRL sanity check
4371 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4372 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4373 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4375 This issue only affects the OpenSSL 1.0.2i
4380 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4382 * OCSP Status Request extension unbounded memory growth
4384 A malicious client can send an excessively large OCSP Status Request
4385 extension. If that client continually requests renegotiation, sending a
4386 large OCSP Status Request extension each time, then there will be unbounded
4387 memory growth on the server. This will eventually lead to a Denial Of
4388 Service attack through memory exhaustion. Servers with a default
4389 configuration are vulnerable even if they do not support OCSP. Builds using
4390 the "no-ocsp" build time option are not affected.
4392 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4397 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4400 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4406 * OOB write in MDC2_Update()
4408 An overflow can occur in MDC2_Update() either if called directly or
4409 through the EVP_DigestUpdate() function using MDC2. If an attacker
4410 is able to supply very large amounts of input data after a previous
4411 call to EVP_EncryptUpdate() with a partial block then a length check
4412 can overflow resulting in a heap corruption.
4414 The amount of data needed is comparable to SIZE_MAX which is impractical
4417 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4422 * Malformed SHA512 ticket DoS
4424 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4425 DoS attack where a malformed ticket will result in an OOB read which will
4428 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4429 a custom server callback and ticket lookup mechanism.
4431 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4436 * OOB write in BN_bn2dec()
4438 The function BN_bn2dec() does not check the return value of BN_div_word().
4439 This can cause an OOB write if an application uses this function with an
4440 overly large BIGNUM. This could be a problem if an overly large certificate
4441 or CRL is printed out from an untrusted source. TLS is not affected because
4442 record limits will reject an oversized certificate before it is parsed.
4444 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4449 * OOB read in TS_OBJ_print_bio()
4451 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4452 the total length the OID text representation would use and not the amount
4453 of data written. This will result in OOB reads when large OIDs are
4456 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4461 * Pointer arithmetic undefined behaviour
4463 Avoid some undefined pointer arithmetic
4465 A common idiom in the codebase is to check limits in the following manner:
4468 Where "p" points to some malloc'd data of SIZE bytes and
4471 "len" here could be from some externally supplied data (e.g. from a TLS
4474 The rules of C pointer arithmetic are such that "p + len" is only well
4475 defined where len <= SIZE. Therefore the above idiom is actually
4476 undefined behaviour.
4478 For example this could cause problems if some malloc implementation
4479 provides an address for "p" such that "p + len" actually overflows for
4480 values of len that are too big and therefore p + len < limit.
4482 This issue was reported to OpenSSL by Guido Vranken
4487 * Constant time flag not preserved in DSA signing
4489 Operations in the DSA signing algorithm should run in constant time in
4490 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4491 implementation means that a non-constant time codepath is followed for
4492 certain operations. This has been demonstrated through a cache-timing
4493 attack to be sufficient for an attacker to recover the private DSA key.
4495 This issue was reported by César Pereida (Aalto University), Billy Brumley
4496 (Tampere University of Technology), and Yuval Yarom (The University of
4497 Adelaide and NICTA).
4502 * DTLS buffered message DoS
4504 In a DTLS connection where handshake messages are delivered out-of-order
4505 those messages that OpenSSL is not yet ready to process will be buffered
4506 for later use. Under certain circumstances, a flaw in the logic means that
4507 those messages do not get removed from the buffer even though the handshake
4508 has been completed. An attacker could force up to approx. 15 messages to
4509 remain in the buffer when they are no longer required. These messages will
4510 be cleared when the DTLS connection is closed. The default maximum size for
4511 a message is 100k. Therefore the attacker could force an additional 1500k
4512 to be consumed per connection. By opening many simulataneous connections an
4513 attacker could cause a DoS attack through memory exhaustion.
4515 This issue was reported to OpenSSL by Quan Luo.
4520 * DTLS replay protection DoS
4522 A flaw in the DTLS replay attack protection mechanism means that records
4523 that arrive for future epochs update the replay protection "window" before
4524 the MAC for the record has been validated. This could be exploited by an
4525 attacker by sending a record for the next epoch (which does not have to
4526 decrypt or have a valid MAC), with a very large sequence number. This means
4527 that all subsequent legitimate packets are dropped causing a denial of
4528 service for a specific DTLS connection.
4530 This issue was reported to OpenSSL by the OCAP audit team.
4535 * Certificate message OOB reads
4537 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4538 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4539 theoretical DoS risk but this has not been observed in practice on common
4542 The messages affected are client certificate, client certificate request
4543 and server certificate. As a result the attack can only be performed
4544 against a client or a server which enables client authentication.
4546 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4551 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4553 * Prevent padding oracle in AES-NI CBC MAC check
4555 A MITM attacker can use a padding oracle attack to decrypt traffic
4556 when the connection uses an AES CBC cipher and the server support
4559 This issue was introduced as part of the fix for Lucky 13 padding
4560 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4561 constant time by making sure that always the same bytes are read and
4562 compared against either the MAC or padding bytes. But it no longer
4563 checked that there was enough data to have both the MAC and padding
4566 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4571 * Fix EVP_EncodeUpdate overflow
4573 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4574 Base64 encoding of binary data. If an attacker is able to supply very large
4575 amounts of input data then a length check can overflow resulting in a heap
4578 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4579 the `PEM_write_bio*` family of functions. These are mainly used within the
4580 OpenSSL command line applications, so any application which processes data
4581 from an untrusted source and outputs it as a PEM file should be considered
4582 vulnerable to this issue. User applications that call these APIs directly
4583 with large amounts of untrusted data may also be vulnerable.
4585 This issue was reported by Guido Vranken.
4590 * Fix EVP_EncryptUpdate overflow
4592 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4593 is able to supply very large amounts of input data after a previous call to
4594 EVP_EncryptUpdate() with a partial block then a length check can overflow
4595 resulting in a heap corruption. Following an analysis of all OpenSSL
4596 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4597 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4598 the first called function after an EVP_EncryptInit(), and therefore that
4599 specific call must be safe. The second form is where the length passed to
4600 EVP_EncryptUpdate() can be seen from the code to be some small value and
4601 therefore there is no possibility of an overflow. Since all instances are
4602 one of these two forms, it is believed that there can be no overflows in
4603 internal code due to this problem. It should be noted that
4604 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4605 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4606 of these calls have also been analysed too and it is believed there are no
4607 instances in internal usage where an overflow could occur.
4609 This issue was reported by Guido Vranken.
4614 * Prevent ASN.1 BIO excessive memory allocation
4616 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4617 a short invalid encoding can cause allocation of large amounts of memory
4618 potentially consuming excessive resources or exhausting memory.
4620 Any application parsing untrusted data through d2i BIO functions is
4621 affected. The memory based functions such as d2i_X509() are *not* affected.
4622 Since the memory based functions are used by the TLS library, TLS
4623 applications are not affected.
4625 This issue was reported by Brian Carpenter.
4632 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4633 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4634 in arbitrary stack data being returned in the buffer.
4636 This issue was reported by Guido Vranken.
4641 * Modify behavior of ALPN to invoke callback after SNI/servername
4642 callback, such that updates to the SSL_CTX affect ALPN.
4646 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4651 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4652 methods are enabled and ssl2 is disabled the methods return NULL.
4656 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4658 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4659 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4660 provide any "EXPORT" or "LOW" strength ciphers.
4664 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4665 is by default disabled at build-time. Builds that are not configured with
4666 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4667 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4668 will need to explicitly call either of:
4670 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4672 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4674 as appropriate. Even if either of those is used, or the application
4675 explicitly uses the version-specific SSLv2_method() or its client and
4676 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4677 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4678 ciphers, and SSLv2 56-bit DES are no longer available.
4683 * Fix a double-free in DSA code
4685 A double free bug was discovered when OpenSSL parses malformed DSA private
4686 keys and could lead to a DoS attack or memory corruption for applications
4687 that receive DSA private keys from untrusted sources. This scenario is
4690 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4696 * Disable SRP fake user seed to address a server memory leak.
4698 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4700 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4701 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4702 was changed to ignore the "fake user" SRP seed, even if the seed
4705 Users should use SRP_VBASE_get1_by_user instead. Note that in
4706 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4707 also that even though configuring the SRP seed attempts to hide
4708 invalid usernames by continuing the handshake with fake
4709 credentials, this behaviour is not constant time and no strong
4710 guarantees are made that the handshake is indistinguishable from
4711 that of a valid user.
4716 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4718 In the BN_hex2bn function the number of hex digits is calculated using an
4719 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4720 large values of `i` this can result in `bn_expand` not allocating any
4721 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4722 field as NULL leading to a subsequent NULL ptr deref. For very large values
4723 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4724 In this case memory is allocated to the internal BIGNUM data field, but it
4725 is insufficiently sized leading to heap corruption. A similar issue exists
4726 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4727 is ever called by user applications with very large untrusted hex/dec data.
4728 This is anticipated to be a rare occurrence.
4730 All OpenSSL internal usage of these functions use data that is not expected
4731 to be untrusted, e.g. config file data or application command line
4732 arguments. If user developed applications generate config file data based
4733 on untrusted data then it is possible that this could also lead to security
4734 consequences. This is also anticipated to be rare.
4736 This issue was reported to OpenSSL by Guido Vranken.
4741 * Fix memory issues in `BIO_*printf` functions
4743 The internal `fmtstr` function used in processing a "%s" format string in
4744 the `BIO_*printf` functions could overflow while calculating the length of a
4745 string and cause an OOB read when printing very long strings.
4747 Additionally the internal `doapr_outch` function can attempt to write to an
4748 OOB memory location (at an offset from the NULL pointer) in the event of a
4749 memory allocation failure. In 1.0.2 and below this could be caused where
4750 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4751 could be in processing a very long "%s" format string. Memory leaks can
4754 The first issue may mask the second issue dependent on compiler behaviour.
4755 These problems could enable attacks where large amounts of untrusted data
4756 is passed to the `BIO_*printf` functions. If applications use these functions
4757 in this way then they could be vulnerable. OpenSSL itself uses these
4758 functions when printing out human-readable dumps of ASN.1 data. Therefore
4759 applications that print this data could be vulnerable if the data is from
4760 untrusted sources. OpenSSL command line applications could also be
4761 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4762 as command line arguments.
4764 Libssl is not considered directly vulnerable. Additionally certificates etc
4765 received via remote connections via libssl are also unlikely to be able to
4766 trigger these issues because of message size limits enforced within libssl.
4768 This issue was reported to OpenSSL Guido Vranken.
4773 * Side channel attack on modular exponentiation
4775 A side-channel attack was found which makes use of cache-bank conflicts on
4776 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4777 of RSA keys. The ability to exploit this issue is limited as it relies on
4778 an attacker who has control of code in a thread running on the same
4779 hyper-threaded core as the victim thread which is performing decryptions.
4781 This issue was reported to OpenSSL by Yuval Yarom, The University of
4782 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4783 Nadia Heninger, University of Pennsylvania with more information at
4784 <http://cachebleed.info>.
4789 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4790 if no keysize is specified with default_bits. This fixes an
4791 omission in an earlier change that changed all RSA/DSA key generation
4792 apps to use 2048 bits by default.
4796 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4798 * DH small subgroups
4800 Historically OpenSSL only ever generated DH parameters based on "safe"
4801 primes. More recently (in version 1.0.2) support was provided for
4802 generating X9.42 style parameter files such as those required for RFC 5114
4803 support. The primes used in such files may not be "safe". Where an
4804 application is using DH configured with parameters based on primes that are
4805 not "safe" then an attacker could use this fact to find a peer's private
4806 DH exponent. This attack requires that the attacker complete multiple
4807 handshakes in which the peer uses the same private DH exponent. For example
4808 this could be used to discover a TLS server's private DH exponent if it's
4809 reusing the private DH exponent or it's using a static DH ciphersuite.
4811 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4812 TLS. It is not on by default. If the option is not set then the server
4813 reuses the same private DH exponent for the life of the server process and
4814 would be vulnerable to this attack. It is believed that many popular
4815 applications do set this option and would therefore not be at risk.
4817 The fix for this issue adds an additional check where a "q" parameter is
4818 available (as is the case in X9.42 based parameters). This detects the
4819 only known attack, and is the only possible defense for static DH
4820 ciphersuites. This could have some performance impact.
4822 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4823 default and cannot be disabled. This could have some performance impact.
4825 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4830 * SSLv2 doesn't block disabled ciphers
4832 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4833 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4834 been disabled, provided that the SSLv2 protocol was not also disabled via
4837 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4838 and Sebastian Schinzel.
4843 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4845 * BN_mod_exp may produce incorrect results on x86_64
4847 There is a carry propagating bug in the x86_64 Montgomery squaring
4848 procedure. No EC algorithms are affected. Analysis suggests that attacks
4849 against RSA and DSA as a result of this defect would be very difficult to
4850 perform and are not believed likely. Attacks against DH are considered just
4851 feasible (although very difficult) because most of the work necessary to
4852 deduce information about a private key may be performed offline. The amount
4853 of resources required for such an attack would be very significant and
4854 likely only accessible to a limited number of attackers. An attacker would
4855 additionally need online access to an unpatched system using the target
4856 private key in a scenario with persistent DH parameters and a private
4857 key that is shared between multiple clients. For example this can occur by
4858 default in OpenSSL DHE based SSL/TLS ciphersuites.
4860 This issue was reported to OpenSSL by Hanno Böck.
4865 * Certificate verify crash with missing PSS parameter
4867 The signature verification routines will crash with a NULL pointer
4868 dereference if presented with an ASN.1 signature using the RSA PSS
4869 algorithm and absent mask generation function parameter. Since these
4870 routines are used to verify certificate signature algorithms this can be
4871 used to crash any certificate verification operation and exploited in a
4872 DoS attack. Any application which performs certificate verification is
4873 vulnerable including OpenSSL clients and servers which enable client
4876 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4881 * X509_ATTRIBUTE memory leak
4883 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4884 memory. This structure is used by the PKCS#7 and CMS routines so any
4885 application which reads PKCS#7 or CMS data from untrusted sources is
4886 affected. SSL/TLS is not affected.
4888 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4894 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4895 This changes the decoding behaviour for some invalid messages,
4896 though the change is mostly in the more lenient direction, and
4897 legacy behaviour is preserved as much as possible.
4901 * In DSA_generate_parameters_ex, if the provided seed is too short,
4904 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4906 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4908 * Alternate chains certificate forgery
4910 During certificate verification, OpenSSL will attempt to find an
4911 alternative certificate chain if the first attempt to build such a chain
4912 fails. An error in the implementation of this logic can mean that an
4913 attacker could cause certain checks on untrusted certificates to be
4914 bypassed, such as the CA flag, enabling them to use a valid leaf
4915 certificate to act as a CA and "issue" an invalid certificate.
4917 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4922 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4924 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4925 incompatibility in the handling of HMAC. The previous ABI has now been
4930 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4932 * Malformed ECParameters causes infinite loop
4934 When processing an ECParameters structure OpenSSL enters an infinite loop
4935 if the curve specified is over a specially malformed binary polynomial
4938 This can be used to perform denial of service against any
4939 system which processes public keys, certificate requests or
4940 certificates. This includes TLS clients and TLS servers with
4941 client authentication enabled.
4943 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4948 * Exploitable out-of-bounds read in X509_cmp_time
4950 X509_cmp_time does not properly check the length of the ASN1_TIME
4951 string and can read a few bytes out of bounds. In addition,
4952 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4955 An attacker can use this to craft malformed certificates and CRLs of
4956 various sizes and potentially cause a segmentation fault, resulting in
4957 a DoS on applications that verify certificates or CRLs. TLS clients
4958 that verify CRLs are affected. TLS clients and servers with client
4959 authentication enabled may be affected if they use custom verification
4962 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4963 independently by Hanno Böck.
4968 * PKCS7 crash with missing EnvelopedContent
4970 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4971 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4972 with missing content and trigger a NULL pointer dereference on parsing.
4974 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4975 structures from untrusted sources are affected. OpenSSL clients and
4976 servers are not affected.
4978 This issue was reported to OpenSSL by Michal Zalewski (Google).
4983 * CMS verify infinite loop with unknown hash function
4985 When verifying a signedData message the CMS code can enter an infinite loop
4986 if presented with an unknown hash function OID. This can be used to perform
4987 denial of service against any system which verifies signedData messages using
4989 This issue was reported to OpenSSL by Johannes Bauer.
4994 * Race condition handling NewSessionTicket
4996 If a NewSessionTicket is received by a multi-threaded client when attempting to
4997 reuse a previous ticket then a race condition can occur potentially leading to
4998 a double free of the ticket data.
5003 * Only support 256-bit or stronger elliptic curves with the
5004 'ecdh_auto' setting (server) or by default (client). Of supported
5005 curves, prefer P-256 (both).
5009 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5011 * ClientHello sigalgs DoS fix
5013 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5014 invalid signature algorithms extension a NULL pointer dereference will
5015 occur. This can be exploited in a DoS attack against the server.
5017 This issue was was reported to OpenSSL by David Ramos of Stanford
5021 *Stephen Henson and Matt Caswell*
5023 * Multiblock corrupted pointer fix
5025 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5026 feature only applies on 64 bit x86 architecture platforms that support AES
5027 NI instructions. A defect in the implementation of "multiblock" can cause
5028 OpenSSL's internal write buffer to become incorrectly set to NULL when
5029 using non-blocking IO. Typically, when the user application is using a
5030 socket BIO for writing, this will only result in a failed connection.
5031 However if some other BIO is used then it is likely that a segmentation
5032 fault will be triggered, thus enabling a potential DoS attack.
5034 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5039 * Segmentation fault in DTLSv1_listen fix
5041 The DTLSv1_listen function is intended to be stateless and processes the
5042 initial ClientHello from many peers. It is common for user code to loop
5043 over the call to DTLSv1_listen until a valid ClientHello is received with
5044 an associated cookie. A defect in the implementation of DTLSv1_listen means
5045 that state is preserved in the SSL object from one invocation to the next
5046 that can lead to a segmentation fault. Errors processing the initial
5047 ClientHello can trigger this scenario. An example of such an error could be
5048 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5051 This issue was reported to OpenSSL by Per Allansson.
5056 * Segmentation fault in ASN1_TYPE_cmp fix
5058 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5059 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5060 certificate signature algorithm consistency this can be used to crash any
5061 certificate verification operation and exploited in a DoS attack. Any
5062 application which performs certificate verification is vulnerable including
5063 OpenSSL clients and servers which enable client authentication.
5068 * Segmentation fault for invalid PSS parameters fix
5070 The signature verification routines will crash with a NULL pointer
5071 dereference if presented with an ASN.1 signature using the RSA PSS
5072 algorithm and invalid parameters. Since these routines are used to verify
5073 certificate signature algorithms this can be used to crash any
5074 certificate verification operation and exploited in a DoS attack. Any
5075 application which performs certificate verification is vulnerable including
5076 OpenSSL clients and servers which enable client authentication.
5078 This issue was was reported to OpenSSL by Brian Carpenter.
5083 * ASN.1 structure reuse memory corruption fix
5085 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5086 memory corruption via an invalid write. Such reuse is and has been
5087 strongly discouraged and is believed to be rare.
5089 Applications that parse structures containing CHOICE or ANY DEFINED BY
5090 components may be affected. Certificate parsing (d2i_X509 and related
5091 functions) are however not affected. OpenSSL clients and servers are
5097 * PKCS7 NULL pointer dereferences fix
5099 The PKCS#7 parsing code does not handle missing outer ContentInfo
5100 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5101 missing content and trigger a NULL pointer dereference on parsing.
5103 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5104 otherwise parse PKCS#7 structures from untrusted sources are
5105 affected. OpenSSL clients and servers are not affected.
5107 This issue was reported to OpenSSL by Michal Zalewski (Google).
5112 * DoS via reachable assert in SSLv2 servers fix
5114 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5115 servers that both support SSLv2 and enable export cipher suites by sending
5116 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5118 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5119 (OpenSSL development team).
5124 * Empty CKE with client auth and DHE fix
5126 If client auth is used then a server can seg fault in the event of a DHE
5127 ciphersuite being selected and a zero length ClientKeyExchange message
5128 being sent by the client. This could be exploited in a DoS attack.
5133 * Handshake with unseeded PRNG fix
5135 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5136 with an unseeded PRNG. The conditions are:
5137 - The client is on a platform where the PRNG has not been seeded
5138 automatically, and the user has not seeded manually
5139 - A protocol specific client method version has been used (i.e. not
5140 SSL_client_methodv23)
5141 - A ciphersuite is used that does not require additional random data from
5142 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5144 If the handshake succeeds then the client random that has been used will
5145 have been generated from a PRNG with insufficient entropy and therefore the
5146 output may be predictable.
5148 For example using the following command with an unseeded openssl will
5149 succeed on an unpatched platform:
5151 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5156 * Use After Free following d2i_ECPrivatekey error fix
5158 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5159 could cause a use after free condition. This, in turn, could cause a double
5160 free in several private key parsing functions (such as d2i_PrivateKey
5161 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5162 for applications that receive EC private keys from untrusted
5163 sources. This scenario is considered rare.
5165 This issue was discovered by the BoringSSL project and fixed in their
5171 * X509_to_X509_REQ NULL pointer deref fix
5173 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5174 the certificate key is invalid. This function is rarely used in practice.
5176 This issue was discovered by Brian Carpenter.
5181 * Removed the export ciphers from the DEFAULT ciphers
5185 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5187 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5188 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5189 So far those who have to target multiple platforms would compromise
5190 and argue that binary targeting say ARMv5 would still execute on
5191 ARMv8. "Universal" build resolves this compromise by providing
5192 near-optimal performance even on newer platforms.
5196 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5197 (other platforms pending).
5199 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5201 * Add support for the SignedCertificateTimestampList certificate and
5202 OCSP response extensions from RFC6962.
5206 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5207 for corner cases. (Certain input points at infinity could lead to
5208 bogus results, with non-infinity inputs mapped to infinity too.)
5212 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5213 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5214 common cases are optimized and there still is room for further
5215 improvements. Vector Permutation AES for Altivec is also added.
5219 * Add support for little-endian ppc64 Linux target.
5221 *Marcelo Cerri (IBM)*
5223 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5224 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5225 are optimized and there still is room for further improvements.
5226 Both 32- and 64-bit modes are supported.
5228 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5230 * Improved ARMv7 NEON support.
5234 * Support for SPARC Architecture 2011 crypto extensions, first
5235 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5236 SHA256/512, MD5, GHASH and modular exponentiation.
5238 *Andy Polyakov, David Miller*
5240 * Accelerated modular exponentiation for Intel processors, a.k.a.
5243 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5245 * Support for new and upcoming Intel processors, including AVX2,
5246 BMI and SHA ISA extensions. This includes additional "stitched"
5247 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5250 This work was sponsored by Intel Corp.
5254 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5255 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5256 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5260 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5261 this fixes a limitation in previous versions of OpenSSL.
5265 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5266 MGF1 digest and OAEP label.
5270 * Add EVP support for key wrapping algorithms, to avoid problems with
5271 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5272 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5273 algorithms and include tests cases.
5277 * Add functions to allocate and set the fields of an ECDSA_METHOD
5280 *Douglas E. Engert, Steve Henson*
5282 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5283 difference in days and seconds between two tm or ASN1_TIME structures.
5287 * Add -rev test option to s_server to just reverse order of characters
5288 received by client and send back to server. Also prints an abbreviated
5289 summary of the connection parameters.
5293 * New option -brief for s_client and s_server to print out a brief summary
5294 of connection parameters.
5298 * Add callbacks for arbitrary TLS extensions.
5300 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5302 * New option -crl_download in several openssl utilities to download CRLs
5303 from CRLDP extension in certificates.
5307 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5311 * New function X509_CRL_diff to generate a delta CRL from the difference
5312 of two full CRLs. Add support to "crl" utility.
5316 * New functions to set lookup_crls function and to retrieve
5317 X509_STORE from X509_STORE_CTX.
5321 * Print out deprecated issuer and subject unique ID fields in
5326 * Extend OCSP I/O functions so they can be used for simple general purpose
5327 HTTP as well as OCSP. New wrapper function which can be used to download
5328 CRLs using the OCSP API.
5332 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5336 * `SSL_CONF*` functions. These provide a common framework for application
5337 configuration using configuration files or command lines.
5341 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5342 message callback and prints the results. Needs compile time option
5343 "enable-ssl-trace". New options to s_client and s_server to enable
5348 * New ctrl and macro to retrieve supported points extensions.
5349 Print out extension in s_server and s_client.
5353 * New functions to retrieve certificate signature and signature
5358 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5363 * New Suite B modes for TLS code. These use and enforce the requirements
5364 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5365 only use Suite B curves. The Suite B modes can be set by using the
5366 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5370 * New chain verification flags for Suite B levels of security. Check
5371 algorithms are acceptable when flags are set in X509_verify_cert.
5375 * Make tls1_check_chain return a set of flags indicating checks passed
5376 by a certificate chain. Add additional tests to handle client
5377 certificates: checks for matching certificate type and issuer name
5382 * If an attempt is made to use a signature algorithm not in the peer
5383 preference list abort the handshake. If client has no suitable
5384 signature algorithms in response to a certificate request do not
5385 use the certificate.
5389 * If server EC tmp key is not in client preference list abort handshake.
5393 * Add support for certificate stores in CERT structure. This makes it
5394 possible to have different stores per SSL structure or one store in
5395 the parent SSL_CTX. Include distinct stores for certificate chain
5396 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5397 to build and store a certificate chain in CERT structure: returning
5398 an error if the chain cannot be built: this will allow applications
5399 to test if a chain is correctly configured.
5401 Note: if the CERT based stores are not set then the parent SSL_CTX
5402 store is used to retain compatibility with existing behaviour.
5406 * New function ssl_set_client_disabled to set a ciphersuite disabled
5407 mask based on the current session, check mask when sending client
5408 hello and checking the requested ciphersuite.
5412 * New ctrls to retrieve and set certificate types in a certificate
5413 request message. Print out received values in s_client. If certificate
5414 types is not set with custom values set sensible values based on
5415 supported signature algorithms.
5419 * Support for distinct client and server supported signature algorithms.
5423 * Add certificate callback. If set this is called whenever a certificate
5424 is required by client or server. An application can decide which
5425 certificate chain to present based on arbitrary criteria: for example
5426 supported signature algorithms. Add very simple example to s_server.
5427 This fixes many of the problems and restrictions of the existing client
5428 certificate callback: for example you can now clear an existing
5429 certificate and specify the whole chain.
5433 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5434 the certificate can be used for (if anything). Set valid_flags field
5435 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5436 to have similar checks in it.
5438 Add new "cert_flags" field to CERT structure and include a "strict mode".
5439 This enforces some TLS certificate requirements (such as only permitting
5440 certificate signature algorithms contained in the supported algorithms
5441 extension) which some implementations ignore: this option should be used
5442 with caution as it could cause interoperability issues.
5446 * Update and tidy signature algorithm extension processing. Work out
5447 shared signature algorithms based on preferences and peer algorithms
5448 and print them out in s_client and s_server. Abort handshake if no
5449 shared signature algorithms.
5453 * Add new functions to allow customised supported signature algorithms
5454 for SSL and SSL_CTX structures. Add options to s_client and s_server
5459 * New function SSL_certs_clear() to delete all references to certificates
5460 from an SSL structure. Before this once a certificate had been added
5461 it couldn't be removed.
5465 * Integrate hostname, email address and IP address checking with certificate
5466 verification. New verify options supporting checking in openssl utility.
5470 * Fixes and wildcard matching support to hostname and email checking
5471 functions. Add manual page.
5473 *Florian Weimer (Red Hat Product Security Team)*
5475 * New functions to check a hostname email or IP address against a
5476 certificate. Add options x509 utility to print results of checks against
5481 * Fix OCSP checking.
5483 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5485 * Initial experimental support for explicitly trusted non-root CAs.
5486 OpenSSL still tries to build a complete chain to a root but if an
5487 intermediate CA has a trust setting included that is used. The first
5488 setting is used: whether to trust (e.g., -addtrust option to the x509
5493 * Add -trusted_first option which attempts to find certificates in the
5494 trusted store even if an untrusted chain is also supplied.
5498 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5499 platform support for Linux and Android.
5503 * Support for linux-x32, ILP32 environment in x86_64 framework.
5507 * Experimental multi-implementation support for FIPS capable OpenSSL.
5508 When in FIPS mode the approved implementations are used as normal,
5509 when not in FIPS mode the internal unapproved versions are used instead.
5510 This means that the FIPS capable OpenSSL isn't forced to use the
5511 (often lower performance) FIPS implementations outside FIPS mode.
5515 * Transparently support X9.42 DH parameters when calling
5516 PEM_read_bio_DHparameters. This means existing applications can handle
5517 the new parameter format automatically.
5521 * Initial experimental support for X9.42 DH parameter format: mainly
5522 to support use of 'q' parameter for RFC5114 parameters.
5526 * Add DH parameters from RFC5114 including test data to dhtest.
5530 * Support for automatic EC temporary key parameter selection. If enabled
5531 the most preferred EC parameters are automatically used instead of
5532 hardcoded fixed parameters. Now a server just has to call:
5533 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5534 support ECDH and use the most appropriate parameters.
5538 * Enhance and tidy EC curve and point format TLS extension code. Use
5539 static structures instead of allocation if default values are used.
5540 New ctrls to set curves we wish to support and to retrieve shared curves.
5541 Print out shared curves in s_server. New options to s_server and s_client
5542 to set list of supported curves.
5546 * New ctrls to retrieve supported signature algorithms and
5547 supported curve values as an array of NIDs. Extend openssl utility
5548 to print out received values.
5552 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5553 between NIDs and the more common NIST names such as "P-256". Enhance
5554 ecparam utility and ECC method to recognise the NIST names for curves.
5558 * Enhance SSL/TLS certificate chain handling to support different
5559 chains for each certificate instead of one chain in the parent SSL_CTX.
5563 * Support for fixed DH ciphersuite client authentication: where both
5564 server and client use DH certificates with common parameters.
5568 * Support for fixed DH ciphersuites: those requiring DH server
5573 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5575 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5576 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5577 X509_CINF_get_signature were reverted post internal team review.
5582 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5584 * OCSP Status Request extension unbounded memory growth
5586 A malicious client can send an excessively large OCSP Status Request
5587 extension. If that client continually requests renegotiation, sending a
5588 large OCSP Status Request extension each time, then there will be unbounded
5589 memory growth on the server. This will eventually lead to a Denial Of
5590 Service attack through memory exhaustion. Servers with a default
5591 configuration are vulnerable even if they do not support OCSP. Builds using
5592 the "no-ocsp" build time option are not affected.
5594 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5599 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5602 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5608 * OOB write in MDC2_Update()
5610 An overflow can occur in MDC2_Update() either if called directly or
5611 through the EVP_DigestUpdate() function using MDC2. If an attacker
5612 is able to supply very large amounts of input data after a previous
5613 call to EVP_EncryptUpdate() with a partial block then a length check
5614 can overflow resulting in a heap corruption.
5616 The amount of data needed is comparable to SIZE_MAX which is impractical
5619 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5624 * Malformed SHA512 ticket DoS
5626 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5627 DoS attack where a malformed ticket will result in an OOB read which will
5630 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5631 a custom server callback and ticket lookup mechanism.
5633 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5638 * OOB write in BN_bn2dec()
5640 The function BN_bn2dec() does not check the return value of BN_div_word().
5641 This can cause an OOB write if an application uses this function with an
5642 overly large BIGNUM. This could be a problem if an overly large certificate
5643 or CRL is printed out from an untrusted source. TLS is not affected because
5644 record limits will reject an oversized certificate before it is parsed.
5646 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5651 * OOB read in TS_OBJ_print_bio()
5653 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5654 the total length the OID text representation would use and not the amount
5655 of data written. This will result in OOB reads when large OIDs are
5658 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5663 * Pointer arithmetic undefined behaviour
5665 Avoid some undefined pointer arithmetic
5667 A common idiom in the codebase is to check limits in the following manner:
5670 Where "p" points to some malloc'd data of SIZE bytes and
5673 "len" here could be from some externally supplied data (e.g. from a TLS
5676 The rules of C pointer arithmetic are such that "p + len" is only well
5677 defined where len <= SIZE. Therefore the above idiom is actually
5678 undefined behaviour.
5680 For example this could cause problems if some malloc implementation
5681 provides an address for "p" such that "p + len" actually overflows for
5682 values of len that are too big and therefore p + len < limit.
5684 This issue was reported to OpenSSL by Guido Vranken
5689 * Constant time flag not preserved in DSA signing
5691 Operations in the DSA signing algorithm should run in constant time in
5692 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5693 implementation means that a non-constant time codepath is followed for
5694 certain operations. This has been demonstrated through a cache-timing
5695 attack to be sufficient for an attacker to recover the private DSA key.
5697 This issue was reported by César Pereida (Aalto University), Billy Brumley
5698 (Tampere University of Technology), and Yuval Yarom (The University of
5699 Adelaide and NICTA).
5704 * DTLS buffered message DoS
5706 In a DTLS connection where handshake messages are delivered out-of-order
5707 those messages that OpenSSL is not yet ready to process will be buffered
5708 for later use. Under certain circumstances, a flaw in the logic means that
5709 those messages do not get removed from the buffer even though the handshake
5710 has been completed. An attacker could force up to approx. 15 messages to
5711 remain in the buffer when they are no longer required. These messages will
5712 be cleared when the DTLS connection is closed. The default maximum size for
5713 a message is 100k. Therefore the attacker could force an additional 1500k
5714 to be consumed per connection. By opening many simulataneous connections an
5715 attacker could cause a DoS attack through memory exhaustion.
5717 This issue was reported to OpenSSL by Quan Luo.
5722 * DTLS replay protection DoS
5724 A flaw in the DTLS replay attack protection mechanism means that records
5725 that arrive for future epochs update the replay protection "window" before
5726 the MAC for the record has been validated. This could be exploited by an
5727 attacker by sending a record for the next epoch (which does not have to
5728 decrypt or have a valid MAC), with a very large sequence number. This means
5729 that all subsequent legitimate packets are dropped causing a denial of
5730 service for a specific DTLS connection.
5732 This issue was reported to OpenSSL by the OCAP audit team.
5737 * Certificate message OOB reads
5739 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5740 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5741 theoretical DoS risk but this has not been observed in practice on common
5744 The messages affected are client certificate, client certificate request
5745 and server certificate. As a result the attack can only be performed
5746 against a client or a server which enables client authentication.
5748 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5753 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5755 * Prevent padding oracle in AES-NI CBC MAC check
5757 A MITM attacker can use a padding oracle attack to decrypt traffic
5758 when the connection uses an AES CBC cipher and the server support
5761 This issue was introduced as part of the fix for Lucky 13 padding
5762 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5763 constant time by making sure that always the same bytes are read and
5764 compared against either the MAC or padding bytes. But it no longer
5765 checked that there was enough data to have both the MAC and padding
5768 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5773 * Fix EVP_EncodeUpdate overflow
5775 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5776 Base64 encoding of binary data. If an attacker is able to supply very large
5777 amounts of input data then a length check can overflow resulting in a heap
5780 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5781 the `PEM_write_bio*` family of functions. These are mainly used within the
5782 OpenSSL command line applications, so any application which processes data
5783 from an untrusted source and outputs it as a PEM file should be considered
5784 vulnerable to this issue. User applications that call these APIs directly
5785 with large amounts of untrusted data may also be vulnerable.
5787 This issue was reported by Guido Vranken.
5792 * Fix EVP_EncryptUpdate overflow
5794 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5795 is able to supply very large amounts of input data after a previous call to
5796 EVP_EncryptUpdate() with a partial block then a length check can overflow
5797 resulting in a heap corruption. Following an analysis of all OpenSSL
5798 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5799 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5800 the first called function after an EVP_EncryptInit(), and therefore that
5801 specific call must be safe. The second form is where the length passed to
5802 EVP_EncryptUpdate() can be seen from the code to be some small value and
5803 therefore there is no possibility of an overflow. Since all instances are
5804 one of these two forms, it is believed that there can be no overflows in
5805 internal code due to this problem. It should be noted that
5806 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5807 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5808 of these calls have also been analysed too and it is believed there are no
5809 instances in internal usage where an overflow could occur.
5811 This issue was reported by Guido Vranken.
5816 * Prevent ASN.1 BIO excessive memory allocation
5818 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5819 a short invalid encoding can casuse allocation of large amounts of memory
5820 potentially consuming excessive resources or exhausting memory.
5822 Any application parsing untrusted data through d2i BIO functions is
5823 affected. The memory based functions such as d2i_X509() are *not* affected.
5824 Since the memory based functions are used by the TLS library, TLS
5825 applications are not affected.
5827 This issue was reported by Brian Carpenter.
5834 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5835 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5836 in arbitrary stack data being returned in the buffer.
5838 This issue was reported by Guido Vranken.
5843 * Modify behavior of ALPN to invoke callback after SNI/servername
5844 callback, such that updates to the SSL_CTX affect ALPN.
5848 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5853 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5854 methods are enabled and ssl2 is disabled the methods return NULL.
5858 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5860 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5861 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5862 provide any "EXPORT" or "LOW" strength ciphers.
5866 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5867 is by default disabled at build-time. Builds that are not configured with
5868 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5869 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5870 will need to explicitly call either of:
5872 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5874 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5876 as appropriate. Even if either of those is used, or the application
5877 explicitly uses the version-specific SSLv2_method() or its client and
5878 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5879 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5880 ciphers, and SSLv2 56-bit DES are no longer available.
5885 * Fix a double-free in DSA code
5887 A double free bug was discovered when OpenSSL parses malformed DSA private
5888 keys and could lead to a DoS attack or memory corruption for applications
5889 that receive DSA private keys from untrusted sources. This scenario is
5892 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5898 * Disable SRP fake user seed to address a server memory leak.
5900 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5902 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5903 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5904 was changed to ignore the "fake user" SRP seed, even if the seed
5907 Users should use SRP_VBASE_get1_by_user instead. Note that in
5908 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5909 also that even though configuring the SRP seed attempts to hide
5910 invalid usernames by continuing the handshake with fake
5911 credentials, this behaviour is not constant time and no strong
5912 guarantees are made that the handshake is indistinguishable from
5913 that of a valid user.
5918 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5920 In the BN_hex2bn function the number of hex digits is calculated using an
5921 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5922 large values of `i` this can result in `bn_expand` not allocating any
5923 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5924 field as NULL leading to a subsequent NULL ptr deref. For very large values
5925 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5926 In this case memory is allocated to the internal BIGNUM data field, but it
5927 is insufficiently sized leading to heap corruption. A similar issue exists
5928 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5929 is ever called by user applications with very large untrusted hex/dec data.
5930 This is anticipated to be a rare occurrence.
5932 All OpenSSL internal usage of these functions use data that is not expected
5933 to be untrusted, e.g. config file data or application command line
5934 arguments. If user developed applications generate config file data based
5935 on untrusted data then it is possible that this could also lead to security
5936 consequences. This is also anticipated to be rare.
5938 This issue was reported to OpenSSL by Guido Vranken.
5943 * Fix memory issues in `BIO_*printf` functions
5945 The internal `fmtstr` function used in processing a "%s" format string in
5946 the `BIO_*printf` functions could overflow while calculating the length of a
5947 string and cause an OOB read when printing very long strings.
5949 Additionally the internal `doapr_outch` function can attempt to write to an
5950 OOB memory location (at an offset from the NULL pointer) in the event of a
5951 memory allocation failure. In 1.0.2 and below this could be caused where
5952 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5953 could be in processing a very long "%s" format string. Memory leaks can
5956 The first issue may mask the second issue dependent on compiler behaviour.
5957 These problems could enable attacks where large amounts of untrusted data
5958 is passed to the `BIO_*printf` functions. If applications use these functions
5959 in this way then they could be vulnerable. OpenSSL itself uses these
5960 functions when printing out human-readable dumps of ASN.1 data. Therefore
5961 applications that print this data could be vulnerable if the data is from
5962 untrusted sources. OpenSSL command line applications could also be
5963 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5964 as command line arguments.
5966 Libssl is not considered directly vulnerable. Additionally certificates etc
5967 received via remote connections via libssl are also unlikely to be able to
5968 trigger these issues because of message size limits enforced within libssl.
5970 This issue was reported to OpenSSL Guido Vranken.
5975 * Side channel attack on modular exponentiation
5977 A side-channel attack was found which makes use of cache-bank conflicts on
5978 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5979 of RSA keys. The ability to exploit this issue is limited as it relies on
5980 an attacker who has control of code in a thread running on the same
5981 hyper-threaded core as the victim thread which is performing decryptions.
5983 This issue was reported to OpenSSL by Yuval Yarom, The University of
5984 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5985 Nadia Heninger, University of Pennsylvania with more information at
5986 <http://cachebleed.info>.
5991 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5992 if no keysize is specified with default_bits. This fixes an
5993 omission in an earlier change that changed all RSA/DSA key generation
5994 apps to use 2048 bits by default.
5998 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6000 * Protection for DH small subgroup attacks
6002 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6003 switched on by default and cannot be disabled. This could have some
6008 * SSLv2 doesn't block disabled ciphers
6010 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6011 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6012 been disabled, provided that the SSLv2 protocol was not also disabled via
6015 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6016 and Sebastian Schinzel.
6021 * Reject DH handshakes with parameters shorter than 1024 bits.
6025 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6027 * Certificate verify crash with missing PSS parameter
6029 The signature verification routines will crash with a NULL pointer
6030 dereference if presented with an ASN.1 signature using the RSA PSS
6031 algorithm and absent mask generation function parameter. Since these
6032 routines are used to verify certificate signature algorithms this can be
6033 used to crash any certificate verification operation and exploited in a
6034 DoS attack. Any application which performs certificate verification is
6035 vulnerable including OpenSSL clients and servers which enable client
6038 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6043 * X509_ATTRIBUTE memory leak
6045 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6046 memory. This structure is used by the PKCS#7 and CMS routines so any
6047 application which reads PKCS#7 or CMS data from untrusted sources is
6048 affected. SSL/TLS is not affected.
6050 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6056 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6057 This changes the decoding behaviour for some invalid messages,
6058 though the change is mostly in the more lenient direction, and
6059 legacy behaviour is preserved as much as possible.
6063 * In DSA_generate_parameters_ex, if the provided seed is too short,
6064 use a random seed, as already documented.
6066 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6068 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6070 * Alternate chains certificate forgery
6072 During certificate verfification, OpenSSL will attempt to find an
6073 alternative certificate chain if the first attempt to build such a chain
6074 fails. An error in the implementation of this logic can mean that an
6075 attacker could cause certain checks on untrusted certificates to be
6076 bypassed, such as the CA flag, enabling them to use a valid leaf
6077 certificate to act as a CA and "issue" an invalid certificate.
6079 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6085 * Race condition handling PSK identify hint
6087 If PSK identity hints are received by a multi-threaded client then
6088 the values are wrongly updated in the parent SSL_CTX structure. This can
6089 result in a race condition potentially leading to a double free of the
6095 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6097 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6098 incompatibility in the handling of HMAC. The previous ABI has now been
6101 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6103 * Malformed ECParameters causes infinite loop
6105 When processing an ECParameters structure OpenSSL enters an infinite loop
6106 if the curve specified is over a specially malformed binary polynomial
6109 This can be used to perform denial of service against any
6110 system which processes public keys, certificate requests or
6111 certificates. This includes TLS clients and TLS servers with
6112 client authentication enabled.
6114 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6119 * Exploitable out-of-bounds read in X509_cmp_time
6121 X509_cmp_time does not properly check the length of the ASN1_TIME
6122 string and can read a few bytes out of bounds. In addition,
6123 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6126 An attacker can use this to craft malformed certificates and CRLs of
6127 various sizes and potentially cause a segmentation fault, resulting in
6128 a DoS on applications that verify certificates or CRLs. TLS clients
6129 that verify CRLs are affected. TLS clients and servers with client
6130 authentication enabled may be affected if they use custom verification
6133 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6134 independently by Hanno Böck.
6139 * PKCS7 crash with missing EnvelopedContent
6141 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6142 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6143 with missing content and trigger a NULL pointer dereference on parsing.
6145 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6146 structures from untrusted sources are affected. OpenSSL clients and
6147 servers are not affected.
6149 This issue was reported to OpenSSL by Michal Zalewski (Google).
6154 * CMS verify infinite loop with unknown hash function
6156 When verifying a signedData message the CMS code can enter an infinite loop
6157 if presented with an unknown hash function OID. This can be used to perform
6158 denial of service against any system which verifies signedData messages using
6160 This issue was reported to OpenSSL by Johannes Bauer.
6165 * Race condition handling NewSessionTicket
6167 If a NewSessionTicket is received by a multi-threaded client when attempting to
6168 reuse a previous ticket then a race condition can occur potentially leading to
6169 a double free of the ticket data.
6174 * Reject DH handshakes with parameters shorter than 768 bits.
6176 *Kurt Roeckx and Emilia Kasper*
6178 * dhparam: generate 2048-bit parameters by default.
6180 *Kurt Roeckx and Emilia Kasper*
6182 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6184 * Segmentation fault in ASN1_TYPE_cmp fix
6186 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6187 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6188 certificate signature algorithm consistency this can be used to crash any
6189 certificate verification operation and exploited in a DoS attack. Any
6190 application which performs certificate verification is vulnerable including
6191 OpenSSL clients and servers which enable client authentication.
6196 * ASN.1 structure reuse memory corruption fix
6198 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6199 memory corruption via an invalid write. Such reuse is and has been
6200 strongly discouraged and is believed to be rare.
6202 Applications that parse structures containing CHOICE or ANY DEFINED BY
6203 components may be affected. Certificate parsing (d2i_X509 and related
6204 functions) are however not affected. OpenSSL clients and servers are
6210 * PKCS7 NULL pointer dereferences fix
6212 The PKCS#7 parsing code does not handle missing outer ContentInfo
6213 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6214 missing content and trigger a NULL pointer dereference on parsing.
6216 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6217 otherwise parse PKCS#7 structures from untrusted sources are
6218 affected. OpenSSL clients and servers are not affected.
6220 This issue was reported to OpenSSL by Michal Zalewski (Google).
6225 * DoS via reachable assert in SSLv2 servers fix
6227 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6228 servers that both support SSLv2 and enable export cipher suites by sending
6229 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6231 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6232 (OpenSSL development team).
6237 * Use After Free following d2i_ECPrivatekey error fix
6239 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6240 could cause a use after free condition. This, in turn, could cause a double
6241 free in several private key parsing functions (such as d2i_PrivateKey
6242 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6243 for applications that receive EC private keys from untrusted
6244 sources. This scenario is considered rare.
6246 This issue was discovered by the BoringSSL project and fixed in their
6252 * X509_to_X509_REQ NULL pointer deref fix
6254 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6255 the certificate key is invalid. This function is rarely used in practice.
6257 This issue was discovered by Brian Carpenter.
6262 * Removed the export ciphers from the DEFAULT ciphers
6266 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6268 * Build fixes for the Windows and OpenVMS platforms
6270 *Matt Caswell and Richard Levitte*
6272 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6274 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6275 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6276 dereference. This could lead to a Denial Of Service attack. Thanks to
6277 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6282 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6283 dtls1_buffer_record function under certain conditions. In particular this
6284 could occur if an attacker sent repeated DTLS records with the same
6285 sequence number but for the next epoch. The memory leak could be exploited
6286 by an attacker in a Denial of Service attack through memory exhaustion.
6287 Thanks to Chris Mueller for reporting this issue.
6292 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6293 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6294 method would be set to NULL which could later result in a NULL pointer
6295 dereference. Thanks to Frank Schmirler for reporting this issue.
6300 * Abort handshake if server key exchange message is omitted for ephemeral
6303 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6304 reporting this issue.
6309 * Remove non-export ephemeral RSA code on client and server. This code
6310 violated the TLS standard by allowing the use of temporary RSA keys in
6311 non-export ciphersuites and could be used by a server to effectively
6312 downgrade the RSA key length used to a value smaller than the server
6313 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6314 INRIA or reporting this issue.
6319 * Fixed issue where DH client certificates are accepted without verification.
6320 An OpenSSL server will accept a DH certificate for client authentication
6321 without the certificate verify message. This effectively allows a client to
6322 authenticate without the use of a private key. This only affects servers
6323 which trust a client certificate authority which issues certificates
6324 containing DH keys: these are extremely rare and hardly ever encountered.
6325 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6331 * Ensure that the session ID context of an SSL is updated when its
6332 SSL_CTX is updated via SSL_set_SSL_CTX.
6334 The session ID context is typically set from the parent SSL_CTX,
6335 and can vary with the CTX.
6339 * Fix various certificate fingerprint issues.
6341 By using non-DER or invalid encodings outside the signed portion of a
6342 certificate the fingerprint can be changed without breaking the signature.
6343 Although no details of the signed portion of the certificate can be changed
6344 this can cause problems with some applications: e.g. those using the
6345 certificate fingerprint for blacklists.
6347 1. Reject signatures with non zero unused bits.
6349 If the BIT STRING containing the signature has non zero unused bits reject
6350 the signature. All current signature algorithms require zero unused bits.
6352 2. Check certificate algorithm consistency.
6354 Check the AlgorithmIdentifier inside TBS matches the one in the
6355 certificate signature. NB: this will result in signature failure
6356 errors for some broken certificates.
6358 Thanks to Konrad Kraszewski from Google for reporting this issue.
6360 3. Check DSA/ECDSA signatures use DER.
6362 Re-encode DSA/ECDSA signatures and compare with the original received
6363 signature. Return an error if there is a mismatch.
6365 This will reject various cases including garbage after signature
6366 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6367 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6368 (negative or with leading zeroes).
6370 Further analysis was conducted and fixes were developed by Stephen Henson
6371 of the OpenSSL core team.
6377 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6378 results on some platforms, including x86_64. This bug occurs at random
6379 with a very low probability, and is not known to be exploitable in any
6380 way, though its exact impact is difficult to determine. Thanks to Pieter
6381 Wuille (Blockstream) who reported this issue and also suggested an initial
6382 fix. Further analysis was conducted by the OpenSSL development team and
6383 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6384 the OpenSSL core team.
6389 * Do not resume sessions on the server if the negotiated protocol
6390 version does not match the session's version. Resuming with a different
6391 version, while not strictly forbidden by the RFC, is of questionable
6392 sanity and breaks all known clients.
6394 *David Benjamin, Emilia Käsper*
6396 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6397 early CCS messages during renegotiation. (Note that because
6398 renegotiation is encrypted, this early CCS was not exploitable.)
6402 * Tighten client-side session ticket handling during renegotiation:
6403 ensure that the client only accepts a session ticket if the server sends
6404 the extension anew in the ServerHello. Previously, a TLS client would
6405 reuse the old extension state and thus accept a session ticket if one was
6406 announced in the initial ServerHello.
6408 Similarly, ensure that the client requires a session ticket if one
6409 was advertised in the ServerHello. Previously, a TLS client would
6410 ignore a missing NewSessionTicket message.
6414 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6418 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6419 sends a carefully crafted handshake message, to cause OpenSSL to fail
6420 to free up to 64k of memory causing a memory leak. This could be
6421 exploited in a Denial Of Service attack. This issue affects OpenSSL
6422 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6423 whether SRTP is used or configured. Implementations of OpenSSL that
6424 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6426 The fix was developed by the OpenSSL team.
6431 * Session Ticket Memory Leak.
6433 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6434 integrity of that ticket is first verified. In the event of a session
6435 ticket integrity check failing, OpenSSL will fail to free memory
6436 causing a memory leak. By sending a large number of invalid session
6437 tickets an attacker could exploit this issue in a Denial Of Service
6443 * Build option no-ssl3 is incomplete.
6445 When OpenSSL is configured with "no-ssl3" as a build option, servers
6446 could accept and complete a SSL 3.0 handshake, and clients could be
6447 configured to send them.
6450 *Akamai and the OpenSSL team*
6452 * Add support for TLS_FALLBACK_SCSV.
6453 Client applications doing fallback retries should call
6454 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6457 *Adam Langley, Bodo Moeller*
6459 * Add additional DigestInfo checks.
6461 Re-encode DigestInto in DER and check against the original when
6462 verifying RSA signature: this will reject any improperly encoded
6463 DigestInfo structures.
6465 Note: this is a precautionary measure and no attacks are currently known.
6469 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6471 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6472 SRP code can be overrun an internal buffer. Add sanity check that
6473 g, A, B < N to SRP code.
6475 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6476 Group for discovering this issue.
6481 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6482 TLS 1.0 instead of higher protocol versions when the ClientHello message
6483 is badly fragmented. This allows a man-in-the-middle attacker to force a
6484 downgrade to TLS 1.0 even if both the server and the client support a
6485 higher protocol version, by modifying the client's TLS records.
6487 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6488 researching this issue.
6493 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6494 to a denial of service attack. A malicious server can crash the client
6495 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6496 ciphersuite and sending carefully crafted handshake messages.
6498 Thanks to Felix Gröbert (Google) for discovering and researching this
6504 * By sending carefully crafted DTLS packets an attacker could cause openssl
6505 to leak memory. This can be exploited through a Denial of Service attack.
6506 Thanks to Adam Langley for discovering and researching this issue.
6511 * An attacker can force openssl to consume large amounts of memory whilst
6512 processing DTLS handshake messages. This can be exploited through a
6513 Denial of Service attack.
6514 Thanks to Adam Langley for discovering and researching this issue.
6519 * An attacker can force an error condition which causes openssl to crash
6520 whilst processing DTLS packets due to memory being freed twice. This
6521 can be exploited through a Denial of Service attack.
6522 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6528 * If a multithreaded client connects to a malicious server using a resumed
6529 session and the server sends an ec point format extension it could write
6530 up to 255 bytes to freed memory.
6532 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6538 * A malicious server can crash an OpenSSL client with a null pointer
6539 dereference (read) by specifying an SRP ciphersuite even though it was not
6540 properly negotiated with the client. This can be exploited through a
6541 Denial of Service attack.
6543 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6544 discovering and researching this issue.
6549 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6550 X509_name_oneline, X509_name_print_ex et al. to leak some information
6551 from the stack. Applications may be affected if they echo pretty printing
6552 output to the attacker.
6554 Thanks to Ivan Fratric (Google) for discovering this issue.
6557 *Emilia Käsper, and Steve Henson*
6559 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6560 for corner cases. (Certain input points at infinity could lead to
6561 bogus results, with non-infinity inputs mapped to infinity too.)
6565 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6567 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6568 handshake can force the use of weak keying material in OpenSSL
6569 SSL/TLS clients and servers.
6571 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6572 researching this issue. ([CVE-2014-0224])
6574 *KIKUCHI Masashi, Steve Henson*
6576 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6577 OpenSSL DTLS client the code can be made to recurse eventually crashing
6580 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6583 *Imre Rad, Steve Henson*
6585 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6586 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6587 client or server. This is potentially exploitable to run arbitrary
6588 code on a vulnerable client or server.
6590 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6592 *Jüri Aedla, Steve Henson*
6594 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6595 are subject to a denial of service attack.
6597 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6598 this issue. ([CVE-2014-3470])
6600 *Felix Gröbert, Ivan Fratric, Steve Henson*
6602 * Harmonize version and its documentation. -f flag is used to display
6605 *mancha <mancha1@zoho.com>*
6607 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6608 in i2d_ECPrivateKey.
6610 *mancha <mancha1@zoho.com>*
6612 * Fix some double frees. These are not thought to be exploitable.
6614 *mancha <mancha1@zoho.com>*
6616 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6618 * A missing bounds check in the handling of the TLS heartbeat extension
6619 can be used to reveal up to 64k of memory to a connected client or
6622 Thanks for Neel Mehta of Google Security for discovering this bug and to
6623 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6624 preparing the fix ([CVE-2014-0160])
6626 *Adam Langley, Bodo Moeller*
6628 * Fix for the attack described in the paper "Recovering OpenSSL
6629 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6630 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6631 <http://eprint.iacr.org/2014/140>
6633 Thanks to Yuval Yarom and Naomi Benger for discovering this
6634 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6636 *Yuval Yarom and Naomi Benger*
6638 * TLS pad extension: draft-agl-tls-padding-03
6640 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6641 TLS client Hello record length value would otherwise be > 255 and
6642 less that 512 pad with a dummy extension containing zeroes so it
6643 is at least 512 bytes long.
6645 *Adam Langley, Steve Henson*
6647 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6649 * Fix for TLS record tampering bug. A carefully crafted invalid
6650 handshake could crash OpenSSL with a NULL pointer exception.
6651 Thanks to Anton Johansson for reporting this issues.
6654 * Keep original DTLS digest and encryption contexts in retransmission
6655 structures so we can use the previous session parameters if they need
6656 to be resent. ([CVE-2013-6450])
6660 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6661 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6662 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6663 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6664 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6665 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6667 *Rob Stradling, Adam Langley*
6669 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6671 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6672 supporting platforms or when small records were transferred.
6674 *Andy Polyakov, Steve Henson*
6676 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6678 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6680 This addresses the flaw in CBC record processing discovered by
6681 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6682 at: <http://www.isg.rhul.ac.uk/tls/>
6684 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6685 Security Group at Royal Holloway, University of London
6686 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6687 Emilia Käsper for the initial patch.
6690 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6692 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6693 ciphersuites which can be exploited in a denial of service attack.
6694 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6695 and detecting this bug and to Wolfgang Ettlinger
6696 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6701 * Return an error when checking OCSP signatures when key is NULL.
6702 This fixes a DoS attack. ([CVE-2013-0166])
6706 * Make openssl verify return errors.
6708 *Chris Palmer <palmer@google.com> and Ben Laurie*
6710 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6711 the right response is stapled. Also change SSL_get_certificate()
6712 so it returns the certificate actually sent.
6713 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6715 *Rob Stradling <rob.stradling@comodo.com>*
6717 * Fix possible deadlock when decoding public keys.
6721 * Don't use TLS 1.0 record version number in initial client hello
6726 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6728 * Sanity check record length before skipping explicit IV in TLS
6729 1.2, 1.1 and DTLS to fix DoS attack.
6731 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6732 fuzzing as a service testing platform.
6737 * Initialise tkeylen properly when encrypting CMS messages.
6738 Thanks to Solar Designer of Openwall for reporting this issue.
6742 * In FIPS mode don't try to use composite ciphers as they are not
6747 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6749 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6750 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6751 mean any application compiled against OpenSSL 1.0.0 headers setting
6752 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6753 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6754 0x10000000L Any application which was previously compiled against
6755 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6756 will need to be recompiled as a result. Letting be results in
6757 inability to disable specifically TLS 1.1 and in client context,
6758 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6762 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6763 disable just protocol X, but all protocols above X *if* there are
6764 protocols *below* X still enabled. In more practical terms it means
6765 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6766 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6767 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6772 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6774 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6775 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6776 in CRYPTO_realloc_clean.
6778 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6779 issue and to Adam Langley <agl@chromium.org> for fixing it.
6782 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6784 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6788 * Workarounds for some broken servers that "hang" if a client hello
6789 record length exceeds 255 bytes.
6791 1. Do not use record version number > TLS 1.0 in initial client
6792 hello: some (but not all) hanging servers will now work.
6793 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6794 the number of ciphers sent in the client hello. This should be
6795 set to an even number, such as 50, for example by passing:
6796 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6797 Most broken servers should now work.
6798 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6799 TLS 1.2 client support entirely.
6803 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6807 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6809 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6810 STRING form instead of a DigestInfo.
6814 * The format used for MDC2 RSA signatures is inconsistent between EVP
6815 and the RSA_sign/RSA_verify functions. This was made more apparent when
6816 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6817 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6818 the correct format in RSA_verify so both forms transparently work.
6822 * Some servers which support TLS 1.0 can choke if we initially indicate
6823 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6824 encrypted premaster secret. As a workaround use the maximum permitted
6825 client version in client hello, this should keep such servers happy
6826 and still work with previous versions of OpenSSL.
6830 * Add support for TLS/DTLS heartbeats.
6832 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6834 * Add support for SCTP.
6836 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6838 * Improved PRNG seeding for VOS.
6840 *Paul Green <Paul.Green@stratus.com>*
6842 * Extensive assembler packs updates, most notably:
6844 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6845 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6846 - x86_64: bit-sliced AES implementation;
6847 - ARM: NEON support, contemporary platforms optimizations;
6848 - s390x: z196 support;
6849 - `*`: GHASH and GF(2^m) multiplication implementations;
6853 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6854 (removal of unnecessary code)
6856 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6858 * Add TLS key material exporter from RFC 5705.
6862 * Add DTLS-SRTP negotiation from RFC 5764.
6866 * Add Next Protocol Negotiation,
6867 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6868 disabled with a no-npn flag to config or Configure. Code donated
6871 *Adam Langley <agl@google.com> and Ben Laurie*
6873 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6874 NIST-P256, NIST-P521, with constant-time single point multiplication on
6875 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6876 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6877 Code made available under Apache License version 2.0.
6879 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6880 line to include this in your build of OpenSSL, and run "make depend" (or
6881 "make update"). This enables the following EC_METHODs:
6883 EC_GFp_nistp224_method()
6884 EC_GFp_nistp256_method()
6885 EC_GFp_nistp521_method()
6887 EC_GROUP_new_by_curve_name() will automatically use these (while
6888 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6891 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6893 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6894 all platforms. Move ssize_t definition from e_os.h to the public
6895 header file e_os2.h as it now appears in public header file cms.h
6899 * New -sigopt option to the ca, req and x509 utilities. Additional
6900 signature parameters can be passed using this option and in
6905 * Add RSA PSS signing function. This will generate and set the
6906 appropriate AlgorithmIdentifiers for PSS based on those in the
6907 corresponding EVP_MD_CTX structure. No application support yet.
6911 * Support for companion algorithm specific ASN1 signing routines.
6912 New function ASN1_item_sign_ctx() signs a pre-initialised
6913 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6914 the appropriate parameters.
6918 * Add new algorithm specific ASN1 verification initialisation function
6919 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6920 handling will be the same no matter what EVP_PKEY_METHOD is used.
6921 Add a PSS handler to support verification of PSS signatures: checked
6922 against a number of sample certificates.
6926 * Add signature printing for PSS. Add PSS OIDs.
6928 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6930 * Add algorithm specific signature printing. An individual ASN1 method
6931 can now print out signatures instead of the standard hex dump.
6933 More complex signatures (e.g. PSS) can print out more meaningful
6934 information. Include DSA version that prints out the signature
6939 * Password based recipient info support for CMS library: implementing
6944 * Split password based encryption into PBES2 and PBKDF2 functions. This
6945 neatly separates the code into cipher and PBE sections and is required
6946 for some algorithms that split PBES2 into separate pieces (such as
6947 password based CMS).
6951 * Session-handling fixes:
6952 - Fix handling of connections that are resuming with a session ID,
6953 but also support Session Tickets.
6954 - Fix a bug that suppressed issuing of a new ticket if the client
6955 presented a ticket with an expired session.
6956 - Try to set the ticket lifetime hint to something reasonable.
6957 - Make tickets shorter by excluding irrelevant information.
6958 - On the client side, don't ignore renewed tickets.
6960 *Adam Langley, Bodo Moeller (Google)*
6962 * Fix PSK session representation.
6966 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6968 This work was sponsored by Intel.
6972 * Add GCM support to TLS library. Some custom code is needed to split
6973 the IV between the fixed (from PRF) and explicit (from TLS record)
6974 portions. This adds all GCM ciphersuites supported by RFC5288 and
6975 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6976 add a special AESGCM string for GCM only.
6980 * Expand range of ctrls for AES GCM. Permit setting invocation
6981 field on decrypt and retrieval of invocation field only on encrypt.
6985 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6986 As required by RFC5289 these ciphersuites cannot be used if for
6987 versions of TLS earlier than 1.2.
6991 * For FIPS capable OpenSSL interpret a NULL default public key method
6992 as unset and return the appropriate default but do *not* set the default.
6993 This means we can return the appropriate method in applications that
6994 switch between FIPS and non-FIPS modes.
6998 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6999 ENGINE is used then we cannot handle that in the FIPS module so we
7000 keep original code iff non-FIPS operations are allowed.
7004 * Add -attime option to openssl utilities.
7006 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7008 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7012 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7013 FIPS EC methods unconditionally for now.
7017 * New build option no-ec2m to disable characteristic 2 code.
7021 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7022 all cases can be covered as some introduce binary incompatibilities.
7026 * Redirect RSA operations to FIPS module including keygen,
7027 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7031 * Add similar low level API blocking to ciphers.
7035 * Low level digest APIs are not approved in FIPS mode: any attempt
7036 to use these will cause a fatal error. Applications that *really* want
7037 to use them can use the `private_*` version instead.
7041 * Redirect cipher operations to FIPS module for FIPS builds.
7045 * Redirect digest operations to FIPS module for FIPS builds.
7049 * Update build system to add "fips" flag which will link in fipscanister.o
7050 for static and shared library builds embedding a signature if needed.
7054 * Output TLS supported curves in preference order instead of numerical
7055 order. This is currently hardcoded for the highest order curves first.
7056 This should be configurable so applications can judge speed vs strength.
7060 * Add TLS v1.2 server support for client authentication.
7064 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7069 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7070 FIPS modules versions.
7074 * Add TLS v1.2 client side support for client authentication. Keep cache
7075 of handshake records longer as we don't know the hash algorithm to use
7076 until after the certificate request message is received.
7080 * Initial TLS v1.2 client support. Add a default signature algorithms
7081 extension including all the algorithms we support. Parse new signature
7082 format in client key exchange. Relax some ECC signing restrictions for
7083 TLS v1.2 as indicated in RFC5246.
7087 * Add server support for TLS v1.2 signature algorithms extension. Switch
7088 to new signature format when needed using client digest preference.
7089 All server ciphersuites should now work correctly in TLS v1.2. No client
7090 support yet and no support for client certificates.
7094 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7095 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7096 ciphersuites. At present only RSA key exchange ciphersuites work with
7097 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7098 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7099 and version checking.
7103 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7104 with this defined it will not be affected by any changes to ssl internal
7105 structures. Add several utility functions to allow openssl application
7106 to work with OPENSSL_NO_SSL_INTERN defined.
7110 * A long standing patch to add support for SRP from EdelWeb (Peter
7111 Sylvester and Christophe Renou) was integrated.
7112 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7113 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7116 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7120 * Permit abbreviated handshakes when renegotiating using the function
7121 SSL_renegotiate_abbreviated().
7123 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7125 * Add call to ENGINE_register_all_complete() to
7126 ENGINE_load_builtin_engines(), so some implementations get used
7127 automatically instead of needing explicit application support.
7131 * Add support for TLS key exporter as described in RFC5705.
7133 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7135 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7136 a few changes are required:
7138 Add SSL_OP_NO_TLSv1_1 flag.
7139 Add TLSv1_1 methods.
7140 Update version checking logic to handle version 1.1.
7141 Add explicit IV handling (ported from DTLS code).
7142 Add command line options to s_client/s_server.
7149 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7151 * X509_ATTRIBUTE memory leak
7153 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7154 memory. This structure is used by the PKCS#7 and CMS routines so any
7155 application which reads PKCS#7 or CMS data from untrusted sources is
7156 affected. SSL/TLS is not affected.
7158 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7164 * Race condition handling PSK identify hint
7166 If PSK identity hints are received by a multi-threaded client then
7167 the values are wrongly updated in the parent SSL_CTX structure. This can
7168 result in a race condition potentially leading to a double free of the
7174 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7176 * Malformed ECParameters causes infinite loop
7178 When processing an ECParameters structure OpenSSL enters an infinite loop
7179 if the curve specified is over a specially malformed binary polynomial
7182 This can be used to perform denial of service against any
7183 system which processes public keys, certificate requests or
7184 certificates. This includes TLS clients and TLS servers with
7185 client authentication enabled.
7187 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7192 * Exploitable out-of-bounds read in X509_cmp_time
7194 X509_cmp_time does not properly check the length of the ASN1_TIME
7195 string and can read a few bytes out of bounds. In addition,
7196 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7199 An attacker can use this to craft malformed certificates and CRLs of
7200 various sizes and potentially cause a segmentation fault, resulting in
7201 a DoS on applications that verify certificates or CRLs. TLS clients
7202 that verify CRLs are affected. TLS clients and servers with client
7203 authentication enabled may be affected if they use custom verification
7206 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7207 independently by Hanno Böck.
7212 * PKCS7 crash with missing EnvelopedContent
7214 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7215 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7216 with missing content and trigger a NULL pointer dereference on parsing.
7218 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7219 structures from untrusted sources are affected. OpenSSL clients and
7220 servers are not affected.
7222 This issue was reported to OpenSSL by Michal Zalewski (Google).
7227 * CMS verify infinite loop with unknown hash function
7229 When verifying a signedData message the CMS code can enter an infinite loop
7230 if presented with an unknown hash function OID. This can be used to perform
7231 denial of service against any system which verifies signedData messages using
7233 This issue was reported to OpenSSL by Johannes Bauer.
7238 * Race condition handling NewSessionTicket
7240 If a NewSessionTicket is received by a multi-threaded client when attempting to
7241 reuse a previous ticket then a race condition can occur potentially leading to
7242 a double free of the ticket data.
7247 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7249 * Segmentation fault in ASN1_TYPE_cmp fix
7251 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7252 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7253 certificate signature algorithm consistency this can be used to crash any
7254 certificate verification operation and exploited in a DoS attack. Any
7255 application which performs certificate verification is vulnerable including
7256 OpenSSL clients and servers which enable client authentication.
7261 * ASN.1 structure reuse memory corruption fix
7263 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7264 memory corruption via an invalid write. Such reuse is and has been
7265 strongly discouraged and is believed to be rare.
7267 Applications that parse structures containing CHOICE or ANY DEFINED BY
7268 components may be affected. Certificate parsing (d2i_X509 and related
7269 functions) are however not affected. OpenSSL clients and servers are
7275 * PKCS7 NULL pointer dereferences fix
7277 The PKCS#7 parsing code does not handle missing outer ContentInfo
7278 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7279 missing content and trigger a NULL pointer dereference on parsing.
7281 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7282 otherwise parse PKCS#7 structures from untrusted sources are
7283 affected. OpenSSL clients and servers are not affected.
7285 This issue was reported to OpenSSL by Michal Zalewski (Google).
7290 * DoS via reachable assert in SSLv2 servers fix
7292 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7293 servers that both support SSLv2 and enable export cipher suites by sending
7294 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7296 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7297 (OpenSSL development team).
7302 * Use After Free following d2i_ECPrivatekey error fix
7304 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7305 could cause a use after free condition. This, in turn, could cause a double
7306 free in several private key parsing functions (such as d2i_PrivateKey
7307 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7308 for applications that receive EC private keys from untrusted
7309 sources. This scenario is considered rare.
7311 This issue was discovered by the BoringSSL project and fixed in their
7317 * X509_to_X509_REQ NULL pointer deref fix
7319 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7320 the certificate key is invalid. This function is rarely used in practice.
7322 This issue was discovered by Brian Carpenter.
7327 * Removed the export ciphers from the DEFAULT ciphers
7331 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7333 * Build fixes for the Windows and OpenVMS platforms
7335 *Matt Caswell and Richard Levitte*
7337 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7339 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7340 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7341 dereference. This could lead to a Denial Of Service attack. Thanks to
7342 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7347 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7348 dtls1_buffer_record function under certain conditions. In particular this
7349 could occur if an attacker sent repeated DTLS records with the same
7350 sequence number but for the next epoch. The memory leak could be exploited
7351 by an attacker in a Denial of Service attack through memory exhaustion.
7352 Thanks to Chris Mueller for reporting this issue.
7357 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7358 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7359 method would be set to NULL which could later result in a NULL pointer
7360 dereference. Thanks to Frank Schmirler for reporting this issue.
7365 * Abort handshake if server key exchange message is omitted for ephemeral
7368 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7369 reporting this issue.
7374 * Remove non-export ephemeral RSA code on client and server. This code
7375 violated the TLS standard by allowing the use of temporary RSA keys in
7376 non-export ciphersuites and could be used by a server to effectively
7377 downgrade the RSA key length used to a value smaller than the server
7378 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7379 INRIA or reporting this issue.
7384 * Fixed issue where DH client certificates are accepted without verification.
7385 An OpenSSL server will accept a DH certificate for client authentication
7386 without the certificate verify message. This effectively allows a client to
7387 authenticate without the use of a private key. This only affects servers
7388 which trust a client certificate authority which issues certificates
7389 containing DH keys: these are extremely rare and hardly ever encountered.
7390 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7396 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7397 results on some platforms, including x86_64. This bug occurs at random
7398 with a very low probability, and is not known to be exploitable in any
7399 way, though its exact impact is difficult to determine. Thanks to Pieter
7400 Wuille (Blockstream) who reported this issue and also suggested an initial
7401 fix. Further analysis was conducted by the OpenSSL development team and
7402 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7403 the OpenSSL core team.
7408 * Fix various certificate fingerprint issues.
7410 By using non-DER or invalid encodings outside the signed portion of a
7411 certificate the fingerprint can be changed without breaking the signature.
7412 Although no details of the signed portion of the certificate can be changed
7413 this can cause problems with some applications: e.g. those using the
7414 certificate fingerprint for blacklists.
7416 1. Reject signatures with non zero unused bits.
7418 If the BIT STRING containing the signature has non zero unused bits reject
7419 the signature. All current signature algorithms require zero unused bits.
7421 2. Check certificate algorithm consistency.
7423 Check the AlgorithmIdentifier inside TBS matches the one in the
7424 certificate signature. NB: this will result in signature failure
7425 errors for some broken certificates.
7427 Thanks to Konrad Kraszewski from Google for reporting this issue.
7429 3. Check DSA/ECDSA signatures use DER.
7431 Reencode DSA/ECDSA signatures and compare with the original received
7432 signature. Return an error if there is a mismatch.
7434 This will reject various cases including garbage after signature
7435 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7436 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7437 (negative or with leading zeroes).
7439 Further analysis was conducted and fixes were developed by Stephen Henson
7440 of the OpenSSL core team.
7446 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7448 * Session Ticket Memory Leak.
7450 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7451 integrity of that ticket is first verified. In the event of a session
7452 ticket integrity check failing, OpenSSL will fail to free memory
7453 causing a memory leak. By sending a large number of invalid session
7454 tickets an attacker could exploit this issue in a Denial Of Service
7460 * Build option no-ssl3 is incomplete.
7462 When OpenSSL is configured with "no-ssl3" as a build option, servers
7463 could accept and complete a SSL 3.0 handshake, and clients could be
7464 configured to send them.
7467 *Akamai and the OpenSSL team*
7469 * Add support for TLS_FALLBACK_SCSV.
7470 Client applications doing fallback retries should call
7471 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7474 *Adam Langley, Bodo Moeller*
7476 * Add additional DigestInfo checks.
7478 Reencode DigestInto in DER and check against the original when
7479 verifying RSA signature: this will reject any improperly encoded
7480 DigestInfo structures.
7482 Note: this is a precautionary measure and no attacks are currently known.
7486 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7488 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7489 to a denial of service attack. A malicious server can crash the client
7490 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7491 ciphersuite and sending carefully crafted handshake messages.
7493 Thanks to Felix Gröbert (Google) for discovering and researching this
7499 * By sending carefully crafted DTLS packets an attacker could cause openssl
7500 to leak memory. This can be exploited through a Denial of Service attack.
7501 Thanks to Adam Langley for discovering and researching this issue.
7506 * An attacker can force openssl to consume large amounts of memory whilst
7507 processing DTLS handshake messages. This can be exploited through a
7508 Denial of Service attack.
7509 Thanks to Adam Langley for discovering and researching this issue.
7514 * An attacker can force an error condition which causes openssl to crash
7515 whilst processing DTLS packets due to memory being freed twice. This
7516 can be exploited through a Denial of Service attack.
7517 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7523 * If a multithreaded client connects to a malicious server using a resumed
7524 session and the server sends an ec point format extension it could write
7525 up to 255 bytes to freed memory.
7527 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7533 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7534 X509_name_oneline, X509_name_print_ex et al. to leak some information
7535 from the stack. Applications may be affected if they echo pretty printing
7536 output to the attacker.
7538 Thanks to Ivan Fratric (Google) for discovering this issue.
7541 *Emilia Käsper, and Steve Henson*
7543 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7544 for corner cases. (Certain input points at infinity could lead to
7545 bogus results, with non-infinity inputs mapped to infinity too.)
7549 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7551 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7552 handshake can force the use of weak keying material in OpenSSL
7553 SSL/TLS clients and servers.
7555 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7556 researching this issue. ([CVE-2014-0224])
7558 *KIKUCHI Masashi, Steve Henson*
7560 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7561 OpenSSL DTLS client the code can be made to recurse eventually crashing
7564 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7567 *Imre Rad, Steve Henson*
7569 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7570 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7571 client or server. This is potentially exploitable to run arbitrary
7572 code on a vulnerable client or server.
7574 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7576 *Jüri Aedla, Steve Henson*
7578 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7579 are subject to a denial of service attack.
7581 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7582 this issue. ([CVE-2014-3470])
7584 *Felix Gröbert, Ivan Fratric, Steve Henson*
7586 * Harmonize version and its documentation. -f flag is used to display
7589 *mancha <mancha1@zoho.com>*
7591 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7592 in i2d_ECPrivateKey.
7594 *mancha <mancha1@zoho.com>*
7596 * Fix some double frees. These are not thought to be exploitable.
7598 *mancha <mancha1@zoho.com>*
7600 * Fix for the attack described in the paper "Recovering OpenSSL
7601 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7602 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7603 <http://eprint.iacr.org/2014/140>
7605 Thanks to Yuval Yarom and Naomi Benger for discovering this
7606 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7608 *Yuval Yarom and Naomi Benger*
7610 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7612 * Keep original DTLS digest and encryption contexts in retransmission
7613 structures so we can use the previous session parameters if they need
7614 to be resent. ([CVE-2013-6450])
7618 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7619 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7620 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7621 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7622 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7623 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7625 *Rob Stradling, Adam Langley*
7627 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7629 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7631 This addresses the flaw in CBC record processing discovered by
7632 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7633 at: <http://www.isg.rhul.ac.uk/tls/>
7635 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7636 Security Group at Royal Holloway, University of London
7637 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7638 Emilia Käsper for the initial patch.
7641 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7643 * Return an error when checking OCSP signatures when key is NULL.
7644 This fixes a DoS attack. ([CVE-2013-0166])
7648 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7649 the right response is stapled. Also change SSL_get_certificate()
7650 so it returns the certificate actually sent.
7651 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7652 (This is a backport)
7654 *Rob Stradling <rob.stradling@comodo.com>*
7656 * Fix possible deadlock when decoding public keys.
7660 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7662 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7665 * Sanity check record length before skipping explicit IV in DTLS
7668 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7669 fuzzing as a service testing platform.
7674 * Initialise tkeylen properly when encrypting CMS messages.
7675 Thanks to Solar Designer of Openwall for reporting this issue.
7679 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7681 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7682 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7683 in CRYPTO_realloc_clean.
7685 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7686 issue and to Adam Langley <agl@chromium.org> for fixing it.
7689 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7691 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7693 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7694 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7695 content decryption and always return the same error. Note: this attack
7696 needs on average 2^20 messages so it only affects automated senders. The
7697 old behaviour can be re-enabled in the CMS code by setting the
7698 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7699 an MMA defence is not necessary.
7700 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7701 this issue. ([CVE-2012-0884])
7705 * Fix CVE-2011-4619: make sure we really are receiving a
7706 client hello before rejecting multiple SGC restarts. Thanks to
7707 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7711 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7713 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7714 Thanks to Antonio Martin, Enterprise Secure Access Research and
7715 Development, Cisco Systems, Inc. for discovering this bug and
7716 preparing a fix. ([CVE-2012-0050])
7720 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7722 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7723 of the Vaudenay padding oracle attack on CBC mode encryption
7724 which enables an efficient plaintext recovery attack against
7725 the OpenSSL implementation of DTLS. Their attack exploits timing
7726 differences arising during decryption processing. A research
7727 paper describing this attack can be found at:
7728 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7729 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7730 Security Group at Royal Holloway, University of London
7731 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7732 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7733 for preparing the fix. ([CVE-2011-4108])
7735 *Robin Seggelmann, Michael Tuexen*
7737 * Clear bytes used for block padding of SSL 3.0 records.
7740 *Adam Langley (Google)*
7742 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7743 Kadianakis <desnacked@gmail.com> for discovering this issue and
7744 Adam Langley for preparing the fix. ([CVE-2011-4619])
7746 *Adam Langley (Google)*
7748 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7750 *Andrey Kulikov <amdeich@gmail.com>*
7752 * Prevent malformed RFC3779 data triggering an assertion failure.
7753 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7754 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7756 *Rob Austein <sra@hactrn.net>*
7758 * Improved PRNG seeding for VOS.
7760 *Paul Green <Paul.Green@stratus.com>*
7762 * Fix ssl_ciph.c set-up race.
7764 *Adam Langley (Google)*
7766 * Fix spurious failures in ecdsatest.c.
7768 *Emilia Käsper (Google)*
7770 * Fix the BIO_f_buffer() implementation (which was mixing different
7771 interpretations of the `..._len` fields).
7773 *Adam Langley (Google)*
7775 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7776 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7777 threads won't reuse the same blinding coefficients.
7779 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7780 lock to call BN_BLINDING_invert_ex, and avoids one use of
7781 BN_BLINDING_update for each BN_BLINDING structure (previously,
7782 the last update always remained unused).
7784 *Emilia Käsper (Google)*
7786 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7788 *Bob Buckholz (Google)*
7790 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7792 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7793 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7795 *Kaspar Brand <ossl@velox.ch>*
7797 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7798 for multi-threaded use of ECDH. ([CVE-2011-3210])
7800 *Adam Langley (Google)*
7802 * Fix x509_name_ex_d2i memory leak on bad inputs.
7806 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7807 signature public key algorithm by using OID xref utilities instead.
7808 Before this you could only use some ECC ciphersuites with SHA1 only.
7812 * Add protection against ECDSA timing attacks as mentioned in the paper
7813 by Billy Bob Brumley and Nicola Tuveri, see:
7814 <http://eprint.iacr.org/2011/232.pdf>
7816 *Billy Bob Brumley and Nicola Tuveri*
7818 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7820 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7822 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7824 * Fix bug in string printing code: if *any* escaping is enabled we must
7825 escape the escape character (backslash) or the resulting string is
7830 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7832 * Disable code workaround for ancient and obsolete Netscape browsers
7833 and servers: an attacker can use it in a ciphersuite downgrade attack.
7834 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7838 * Fixed J-PAKE implementation error, originally discovered by
7839 Sebastien Martini, further info and confirmation from Stefan
7840 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7844 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7846 * Fix extension code to avoid race conditions which can result in a buffer
7847 overrun vulnerability: resumed sessions must not be modified as they can
7848 be shared by multiple threads. CVE-2010-3864
7852 * Fix WIN32 build system to correctly link an ENGINE directory into
7857 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7859 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7862 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7864 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7866 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7867 context. The operation can be customised via the ctrl mechanism in
7868 case ENGINEs want to include additional functionality.
7872 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7876 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7877 output hashes compatible with older versions of OpenSSL.
7879 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7881 * Fix compression algorithm handling: if resuming a session use the
7882 compression algorithm of the resumed session instead of determining
7883 it from client hello again. Don't allow server to change algorithm.
7887 * Add load_crls() function to apps tidying load_certs() too. Add option
7888 to verify utility to allow additional CRLs to be included.
7892 * Update OCSP request code to permit adding custom headers to the request:
7893 some responders need this.
7897 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7900 *Julia Lawall <julia@diku.dk>*
7902 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7903 needlessly dereferenced structures, used obsolete functions and
7904 didn't handle all updated verify codes correctly.
7908 * Disable MD2 in the default configuration.
7912 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7913 indicate the initial BIO being pushed or popped. This makes it possible
7914 to determine whether the BIO is the one explicitly called or as a result
7915 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7916 it handles reference counts correctly and doesn't zero out the I/O bio
7917 when it is not being explicitly popped. WARNING: applications which
7918 included workarounds for the old buggy behaviour will need to be modified
7919 or they could free up already freed BIOs.
7923 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7924 renaming to all platforms (within the 0.9.8 branch, this was
7925 done conditionally on Netware platforms to avoid a name clash).
7927 *Guenter <lists@gknw.net>*
7929 * Add ECDHE and PSK support to DTLS.
7931 *Michael Tuexen <tuexen@fh-muenster.de>*
7933 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7938 * Add "missing" function EVP_MD_flags() (without this the only way to
7939 retrieve a digest flags is by accessing the structure directly. Update
7940 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7941 or cipher is registered as in the "from" argument. Print out all
7942 registered digests in the dgst usage message instead of manually
7943 attempting to work them out.
7947 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7948 this allows the use of compression and extensions. Change default cipher
7949 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7950 by default unless an application cipher string requests it.
7954 * Alter match criteria in PKCS12_parse(). It used to try to use local
7955 key ids to find matching certificates and keys but some PKCS#12 files
7956 don't follow the (somewhat unwritten) rules and this strategy fails.
7957 Now just gather all certificates together and the first private key
7958 then look for the first certificate that matches the key.
7962 * Support use of registered digest and cipher names for dgst and cipher
7963 commands instead of having to add each one as a special case. So now
7970 openssl dgst -sha256 foo
7972 and this works for ENGINE based algorithms too.
7976 * Update Gost ENGINE to support parameter files.
7978 *Victor B. Wagner <vitus@cryptocom.ru>*
7980 * Support GeneralizedTime in ca utility.
7982 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7984 * Enhance the hash format used for certificate directory links. The new
7985 form uses the canonical encoding (meaning equivalent names will work
7986 even if they aren't identical) and uses SHA1 instead of MD5. This form
7987 is incompatible with the older format and as a result c_rehash should
7988 be used to rebuild symbolic links.
7992 * Make PKCS#8 the default write format for private keys, replacing the
7993 traditional format. This form is standardised, more secure and doesn't
7994 include an implicit MD5 dependency.
7998 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7999 committed to OpenSSL should pass this lot as a minimum.
8003 * Add session ticket override functionality for use by EAP-FAST.
8005 *Jouni Malinen <j@w1.fi>*
8007 * Modify HMAC functions to return a value. Since these can be implemented
8008 in an ENGINE errors can occur.
8012 * Type-checked OBJ_bsearch_ex.
8016 * Type-checked OBJ_bsearch. Also some constification necessitated
8017 by type-checking. Still to come: TXT_DB, bsearch(?),
8018 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8023 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8024 seconds to a tm structure directly, instead of going through OS
8025 specific date routines. This avoids any issues with OS routines such
8026 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8027 and X509_time_adj_ex() to cover the extended range. The existing
8028 X509_time_adj() is still usable and will no longer have any date issues.
8032 * Delta CRL support. New use deltas option which will attempt to locate
8033 and search any appropriate delta CRLs available.
8035 This work was sponsored by Google.
8039 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8040 code and add additional score elements. Validate alternate CRL paths
8041 as part of the CRL checking and indicate a new error "CRL path validation
8042 error" in this case. Applications wanting additional details can use
8043 the verify callback and check the new "parent" field. If this is not
8044 NULL CRL path validation is taking place. Existing applications won't
8045 see this because it requires extended CRL support which is off by
8048 This work was sponsored by Google.
8052 * Support for freshest CRL extension.
8054 This work was sponsored by Google.
8058 * Initial indirect CRL support. Currently only supported in the CRLs
8059 passed directly and not via lookup. Process certificate issuer
8060 CRL entry extension and lookup CRL entries by bother issuer name
8061 and serial number. Check and process CRL issuer entry in IDP extension.
8063 This work was sponsored by Google.
8067 * Add support for distinct certificate and CRL paths. The CRL issuer
8068 certificate is validated separately in this case. Only enabled if
8069 an extended CRL support flag is set: this flag will enable additional
8070 CRL functionality in future.
8072 This work was sponsored by Google.
8076 * Add support for policy mappings extension.
8078 This work was sponsored by Google.
8082 * Fixes to pathlength constraint, self issued certificate handling,
8083 policy processing to align with RFC3280 and PKITS tests.
8085 This work was sponsored by Google.
8089 * Support for name constraints certificate extension. DN, email, DNS
8090 and URI types are currently supported.
8092 This work was sponsored by Google.
8096 * To cater for systems that provide a pointer-based thread ID rather
8097 than numeric, deprecate the current numeric thread ID mechanism and
8098 replace it with a structure and associated callback type. This
8099 mechanism allows a numeric "hash" to be extracted from a thread ID in
8100 either case, and on platforms where pointers are larger than 'long',
8101 mixing is done to help ensure the numeric 'hash' is usable even if it
8102 can't be guaranteed unique. The default mechanism is to use "&errno"
8103 as a pointer-based thread ID to distinguish between threads.
8105 Applications that want to provide their own thread IDs should now use
8106 CRYPTO_THREADID_set_callback() to register a callback that will call
8107 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8109 Note that ERR_remove_state() is now deprecated, because it is tied
8110 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8111 to free the current thread's error state should be replaced by
8112 ERR_remove_thread_state(NULL).
8114 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8115 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8116 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8117 application was previously providing a numeric thread callback that
8118 was inappropriate for distinguishing threads, then uniqueness might
8119 have been obtained with &errno that happened immediately in the
8120 intermediate development versions of OpenSSL; this is no longer the
8121 case, the numeric thread callback will now override the automatic use
8124 *Geoff Thorpe, with help from Bodo Moeller*
8126 * Initial support for different CRL issuing certificates. This covers a
8127 simple case where the self issued certificates in the chain exist and
8128 the real CRL issuer is higher in the existing chain.
8130 This work was sponsored by Google.
8134 * Removed effectively defunct crypto/store from the build.
8138 * Revamp of STACK to provide stronger type-checking. Still to come:
8139 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8140 ASN1_STRING, CONF_VALUE.
8144 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8145 RAM on SSL connections. This option can save about 34k per idle SSL.
8149 * Revamp of LHASH to provide stronger type-checking. Still to come:
8150 STACK, TXT_DB, bsearch, qsort.
8154 * Initial support for Cryptographic Message Syntax (aka CMS) based
8155 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8156 support for data, signedData, compressedData, digestedData and
8157 encryptedData, envelopedData types included. Scripts to check against
8158 RFC4134 examples draft and interop and consistency checks of many
8159 content types and variants.
8163 * Add options to enc utility to support use of zlib compression BIO.
8167 * Extend mk1mf to support importing of options and assembly language
8168 files from Configure script, currently only included in VC-WIN32.
8169 The assembly language rules can now optionally generate the source
8170 files from the associated perl scripts.
8174 * Implement remaining functionality needed to support GOST ciphersuites.
8175 Interop testing has been performed using CryptoPro implementations.
8177 *Victor B. Wagner <vitus@cryptocom.ru>*
8179 * s390x assembler pack.
8183 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8188 * Implement Opaque PRF Input TLS extension as specified in
8189 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8190 official specification yet and no extension type assignment by
8191 IANA exists, this extension (for now) will have to be explicitly
8192 enabled when building OpenSSL by providing the extension number
8193 to use. For example, specify an option
8195 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8197 to the "config" or "Configure" script to enable the extension,
8198 assuming extension number 0x9527 (which is a completely arbitrary
8199 and unofficial assignment based on the MD5 hash of the Internet
8200 Draft). Note that by doing so, you potentially lose
8201 interoperability with other TLS implementations since these might
8202 be using the same extension number for other purposes.
8204 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8205 opaque PRF input value to use in the handshake. This will create
8206 an internal copy of the length-'len' string at 'src', and will
8207 return non-zero for success.
8209 To get more control and flexibility, provide a callback function
8212 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8213 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8217 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8220 Callback function 'cb' will be called in handshakes, and is
8221 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8222 Argument 'arg' is for application purposes (the value as given to
8223 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8224 be provided to the callback function). The callback function
8225 has to return non-zero to report success: usually 1 to use opaque
8226 PRF input just if possible, or 2 to enforce use of the opaque PRF
8227 input. In the latter case, the library will abort the handshake
8228 if opaque PRF input is not successfully negotiated.
8230 Arguments 'peerinput' and 'len' given to the callback function
8231 will always be NULL and 0 in the case of a client. A server will
8232 see the client's opaque PRF input through these variables if
8233 available (NULL and 0 otherwise). Note that if the server
8234 provides an opaque PRF input, the length must be the same as the
8235 length of the client's opaque PRF input.
8237 Note that the callback function will only be called when creating
8238 a new session (session resumption can resume whatever was
8239 previously negotiated), and will not be called in SSL 2.0
8240 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8241 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8242 for applications that need to enforce opaque PRF input.
8246 * Update ssl code to support digests other than SHA1+MD5 for handshake
8249 *Victor B. Wagner <vitus@cryptocom.ru>*
8251 * Add RFC4507 support to OpenSSL. This includes the corrections in
8252 RFC4507bis. The encrypted ticket format is an encrypted encoded
8253 SSL_SESSION structure, that way new session features are automatically
8256 If a client application caches session in an SSL_SESSION structure
8257 support is transparent because tickets are now stored in the encoded
8260 The SSL_CTX structure automatically generates keys for ticket
8261 protection in servers so again support should be possible
8262 with no application modification.
8264 If a client or server wishes to disable RFC4507 support then the option
8265 SSL_OP_NO_TICKET can be set.
8267 Add a TLS extension debugging callback to allow the contents of any client
8268 or server extensions to be examined.
8270 This work was sponsored by Google.
8274 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8275 OpenSSL should now compile cleanly on gcc 4.2
8277 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8279 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8280 support including streaming MAC support: this is required for GOST
8281 ciphersuite support.
8283 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8285 * Add option -stream to use PKCS#7 streaming in smime utility. New
8286 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8287 to output in BER and PEM format.
8291 * Experimental support for use of HMAC via EVP_PKEY interface. This
8292 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8293 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8294 ENGINE support for HMAC keys which are unextractable. New -mac and
8295 -macopt options to dgst utility.
8299 * New option -sigopt to dgst utility. Update dgst to use
8300 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8301 alternative signing parameters such as X9.31 or PSS in the dgst
8306 * Change ssl_cipher_apply_rule(), the internal function that does
8307 the work each time a ciphersuite string requests enabling
8308 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8309 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8310 the order of disabled ciphersuites such that those ciphersuites
8311 that most recently went from enabled to disabled not only stay
8312 in order with respect to each other, but also have higher priority
8313 than other disabled ciphersuites the next time ciphersuites are
8316 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8317 the same ciphersuites as with "HIGH" alone, but in a specific
8318 order where the PSK ciphersuites come first (since they are the
8319 most recently disabled ciphersuites when "HIGH" is parsed).
8321 Also, change ssl_create_cipher_list() (using this new
8322 functionality) such that between otherwise identical
8323 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8328 * Change ssl_create_cipher_list() so that it automatically
8329 arranges the ciphersuites in reasonable order before starting
8330 to process the rule string. Thus, the definition for "DEFAULT"
8331 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8332 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8333 This makes it much easier to arrive at a reasonable default order
8334 in applications for which anonymous ciphers are OK (meaning
8335 that you can't actually use DEFAULT).
8337 *Bodo Moeller; suggested by Victor Duchovni*
8339 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8340 processing) into multiple integers instead of setting
8341 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8342 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8343 (These masks as well as the individual bit definitions are hidden
8344 away into the non-exported interface ssl/ssl_locl.h, so this
8345 change to the definition of the SSL_CIPHER structure shouldn't
8346 affect applications.) This give us more bits for each of these
8347 categories, so there is no longer a need to coagulate AES128 and
8348 AES256 into a single algorithm bit, and to coagulate Camellia128
8349 and Camellia256 into a single algorithm bit, which has led to all
8352 Thus, among other things, the kludge introduced in 0.9.7m and
8353 0.9.8e for masking out AES256 independently of AES128 or masking
8354 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8356 With the change, we also introduce new ciphersuite aliases that
8357 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8362 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8363 Use the leftmost N bytes of the signature input if the input is
8364 larger than the prime q (with N being the size in bytes of q).
8368 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8369 it yet and it is largely untested.
8373 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8377 * Initial incomplete changes to avoid need for function casts in OpenSSL
8378 some compilers (gcc 4.2 and later) reject their use. Safestack is
8379 reimplemented. Update ASN1 to avoid use of legacy functions.
8383 * Win32/64 targets are linked with Winsock2.
8387 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8388 to external functions. This can be used to increase CRL handling
8389 efficiency especially when CRLs are very large by (for example) storing
8390 the CRL revoked certificates in a database.
8394 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8395 new CRLs added to a directory can be used. New command line option
8396 -verify_return_error to s_client and s_server. This causes real errors
8397 to be returned by the verify callback instead of carrying on no matter
8398 what. This reflects the way a "real world" verify callback would behave.
8402 * GOST engine, supporting several GOST algorithms and public key formats.
8403 Kindly donated by Cryptocom.
8407 * Partial support for Issuing Distribution Point CRL extension. CRLs
8408 partitioned by DP are handled but no indirect CRL or reason partitioning
8409 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8410 selected via a scoring technique which handles IDP and AKID in CRLs.
8414 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8415 will ultimately be used for all verify operations: this will remove the
8416 X509_STORE dependency on certificate verification and allow alternative
8417 lookup methods. X509_STORE based implementations of these two callbacks.
8421 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8422 Modify get_crl() to find a valid (unexpired) CRL if possible.
8426 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8427 this would be called X509_CRL_cmp() but that name is already used by
8428 a function that just compares CRL issuer names. Cache several CRL
8429 extensions in X509_CRL structure and cache CRLDP in X509.
8433 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8434 this maps equivalent X509_NAME structures into a consistent structure.
8435 Name comparison can then be performed rapidly using memcmp().
8439 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8444 * Allow digests to supply their own micalg string for S/MIME type using
8445 the ctrl EVP_MD_CTRL_MICALG.
8449 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8450 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8451 ctrl. It can then customise the structure before and/or after signing
8456 * New function OBJ_add_sigid() to allow application defined signature OIDs
8457 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8458 to free up any added signature OIDs.
8462 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8463 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8464 digest and cipher tables. New options added to openssl utility:
8465 list-message-digest-algorithms and list-cipher-algorithms.
8469 * Change the array representation of binary polynomials: the list
8470 of degrees of non-zero coefficients is now terminated with -1.
8471 Previously it was terminated with 0, which was also part of the
8472 value; thus, the array representation was not applicable to
8473 polynomials where t^0 has coefficient zero. This change makes
8474 the array representation useful in a more general context.
8478 * Various modifications and fixes to SSL/TLS cipher string
8479 handling. For ECC, the code now distinguishes between fixed ECDH
8480 with RSA certificates on the one hand and with ECDSA certificates
8481 on the other hand, since these are separate ciphersuites. The
8482 unused code for Fortezza ciphersuites has been removed.
8484 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8485 (not "ECDHE"). For consistency with the code for DH
8486 certificates, use of ECDH certificates is now considered ECDH
8487 authentication, not RSA or ECDSA authentication (the latter is
8488 merely the CA's signing algorithm and not actively used in the
8491 The temporary ciphersuite alias "ECCdraft" is no longer
8492 available, and ECC ciphersuites are no longer excluded from "ALL"
8493 and "DEFAULT". The following aliases now exist for RFC 4492
8494 ciphersuites, most of these by analogy with the DH case:
8496 kECDHr - ECDH cert, signed with RSA
8497 kECDHe - ECDH cert, signed with ECDSA
8498 kECDH - ECDH cert (signed with either RSA or ECDSA)
8499 kEECDH - ephemeral ECDH
8500 ECDH - ECDH cert or ephemeral ECDH
8506 AECDH - anonymous ECDH
8507 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8511 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8512 Use correct micalg parameters depending on digest(s) in signed message.
8516 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8517 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8521 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8522 an engine to register a method. Add ENGINE lookups for methods and
8523 functional reference processing.
8527 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8528 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8533 * New -resign option to smime utility. This adds one or more signers
8534 to an existing PKCS#7 signedData structure. Also -md option to use an
8535 alternative message digest algorithm for signing.
8539 * Tidy up PKCS#7 routines and add new functions to make it easier to
8540 create PKCS7 structures containing multiple signers. Update smime
8541 application to support multiple signers.
8545 * New -macalg option to pkcs12 utility to allow setting of an alternative
8550 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8551 Reorganize PBE internals to lookup from a static table using NIDs,
8552 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8553 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8554 PRF which will be automatically used with PBES2.
8558 * Replace the algorithm specific calls to generate keys in "req" with the
8563 * Update PKCS#7 enveloped data routines to use new API. This is now
8564 supported by any public key method supporting the encrypt operation. A
8565 ctrl is added to allow the public key algorithm to examine or modify
8566 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8571 * Add a ctrl to asn1 method to allow a public key algorithm to express
8572 a default digest type to use. In most cases this will be SHA1 but some
8573 algorithms (such as GOST) need to specify an alternative digest. The
8574 return value indicates how strong the preference is 1 means optional and
8575 2 is mandatory (that is it is the only supported type). Modify
8576 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8577 use the default md. Update openssl utilities to use the default digest
8578 type for signing if it is not explicitly indicated.
8582 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8583 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8584 signing method from the key type. This effectively removes the link
8585 between digests and public key types.
8589 * Add an OID cross reference table and utility functions. Its purpose is to
8590 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8591 rsaEncryption. This will allow some of the algorithm specific hackery
8592 needed to use the correct OID to be removed.
8596 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8597 structures for PKCS7_sign(). They are now set up by the relevant public
8602 * Add provisional EC pkey method with support for ECDSA and ECDH.
8606 * Add support for key derivation (agreement) in the API, DH method and
8611 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8612 public and private key formats. As a side effect these add additional
8613 command line functionality not previously available: DSA signatures can be
8614 generated and verified using pkeyutl and DH key support and generation in
8621 *Oliver Tappe <zooey@hirschkaefer.de>*
8623 * New make target "install_html_docs" installs HTML renditions of the
8626 *Oliver Tappe <zooey@hirschkaefer.de>*
8628 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8629 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8630 support key and parameter generation and add initial key generation
8631 functionality for RSA.
8635 * Add functions for main EVP_PKEY_method operations. The undocumented
8636 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8637 `EVP_PKEY_{encrypt,decrypt}_old`.
8641 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8642 key API, doesn't do much yet.
8646 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8647 public key algorithms. New option to openssl utility:
8648 "list-public-key-algorithms" to print out info.
8652 * Implement the Supported Elliptic Curves Extension for
8653 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8657 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8658 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8662 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8663 utilities such as rsa, dsa, dsaparam etc except they process any key
8668 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8669 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8670 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8675 * Initial support for pluggable public key ASN1.
8676 De-spaghettify the public key ASN1 handling. Move public and private
8677 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8678 algorithm specific handling to a single module within the relevant
8679 algorithm directory. Add functions to allow (near) opaque processing
8680 of public and private key structures.
8684 * Implement the Supported Point Formats Extension for
8685 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8689 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8690 for the psk identity [hint] and the psk callback functions to the
8691 SSL_SESSION, SSL and SSL_CTX structure.
8694 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8698 SSL_CTX_use_psk_identity_hint
8699 SSL_get_psk_identity_hint
8700 SSL_get_psk_identity
8701 SSL_use_psk_identity_hint
8703 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8705 * Add RFC 3161 compliant time stamp request creation, response generation
8706 and response verification functionality.
8708 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8710 * Add initial support for TLS extensions, specifically for the server_name
8711 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8712 have new members for a host name. The SSL data structure has an
8713 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8714 stored in that context to allow for session resumption, even after the
8715 SSL has been switched to a new SSL_CTX in reaction to a client's
8716 server_name extension.
8718 New functions (subject to change):
8720 SSL_get_servername()
8721 SSL_get_servername_type()
8724 New CTRL codes and macros (subject to change):
8726 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8727 - SSL_CTX_set_tlsext_servername_callback()
8728 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8729 - SSL_CTX_set_tlsext_servername_arg()
8730 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8732 openssl s_client has a new '-servername ...' option.
8734 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8735 '-key2 ...', '-servername_fatal' (subject to change). This allows
8736 testing the HostName extension for a specific single host name ('-cert'
8737 and '-key' remain fallbacks for handshakes without HostName
8738 negotiation). If the unrecognized_name alert has to be sent, this by
8739 default is a warning; it becomes fatal with the '-servername_fatal'
8742 *Peter Sylvester, Remy Allais, Christophe Renou*
8744 * Whirlpool hash implementation is added.
8748 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8749 bn(64,32). Because of instruction set limitations it doesn't have
8750 any negative impact on performance. This was done mostly in order
8751 to make it possible to share assembler modules, such as bn_mul_mont
8752 implementations, between 32- and 64-bit builds without hassle.
8756 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8757 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8762 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8763 dedicated Montgomery multiplication procedure, is introduced.
8764 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8765 "64-bit" performance on certain 32-bit targets.
8769 * New option SSL_OP_NO_COMP to disable use of compression selectively
8770 in SSL structures. New SSL ctrl to set maximum send fragment size.
8771 Save memory by setting the I/O buffer sizes dynamically instead of
8772 using the maximum available value.
8776 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8777 in addition to the text details.
8781 * Very, very preliminary EXPERIMENTAL support for printing of general
8782 ASN1 structures. This currently produces rather ugly output and doesn't
8783 handle several customised structures at all.
8787 * Integrated support for PVK file format and some related formats such
8788 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8789 these in the 'rsa' and 'dsa' utilities.
8793 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8797 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8798 place for the (very old) "NETSCAPE" format certificates which are now
8799 handled using new ASN1 code equivalents.
8803 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8804 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8805 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8809 * Modify CRL distribution points extension code to print out previously
8810 unsupported fields. Enhance extension setting code to allow setting of
8815 * Add print and set support for Issuing Distribution Point CRL extension.
8819 * Change 'Configure' script to enable Camellia by default.
8826 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8828 * When rejecting SSL/TLS records due to an incorrect version number, never
8829 update s->server with a new major version number. As of
8830 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8831 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8832 the previous behavior could result in a read attempt at NULL when
8833 receiving specific incorrect SSL/TLS records once record payload
8834 protection is active. ([CVE-2010-0740])
8836 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8838 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8839 could be crashed if the relevant tables were not present (e.g. chrooted).
8841 *Tomas Hoger <thoger@redhat.com>*
8843 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8845 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8847 *Martin Olsson, Neel Mehta*
8849 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8850 accommodate for stack sorting, always a write lock!).
8854 * On some versions of WIN32 Heap32Next is very slow. This can cause
8855 excessive delays in the RAND_poll(): over a minute. As a workaround
8856 include a time check in the inner Heap32Next loop too.
8860 * The code that handled flushing of data in SSL/TLS originally used the
8861 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8862 the problem outlined in PR#1949. The fix suggested there however can
8863 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8864 of Apache). So instead simplify the code to flush unconditionally.
8865 This should be fine since flushing with no data to flush is a no op.
8869 * Handle TLS versions 2.0 and later properly and correctly use the
8870 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8871 off ancient servers have a habit of sticking around for a while...
8875 * Modify compression code so it frees up structures without using the
8876 ex_data callbacks. This works around a problem where some applications
8877 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8878 restarting) then use compression (e.g. SSL with compression) later.
8879 This results in significant per-connection memory leaks and
8880 has caused some security issues including CVE-2008-1678 and
8885 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8886 change when encrypting or decrypting.
8890 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8891 connect and renegotiate with servers which do not support RI.
8892 Until RI is more widely deployed this option is enabled by default.
8896 * Add "missing" ssl ctrls to clear options and mode.
8900 * If client attempts to renegotiate and doesn't support RI respond with
8901 a no_renegotiation alert as required by RFC5746. Some renegotiating
8902 TLS clients will continue a connection gracefully when they receive
8903 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8904 waiting for a server hello which it will never receive. Now we treat a
8905 received no_renegotiation alert as a fatal error. This is because
8906 applications requesting a renegotiation might well expect it to succeed
8907 and would have no code in place to handle the server denying it so the
8908 only safe thing to do is to terminate the connection.
8912 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8913 peer supports secure renegotiation and 0 otherwise. Print out peer
8914 renegotiation support in s_client/s_server.
8918 * Replace the highly broken and deprecated SPKAC certification method with
8919 the updated NID creation version. This should correctly handle UTF8.
8923 * Implement RFC5746. Re-enable renegotiation but require the extension
8924 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8925 turns out to be a bad idea. It has been replaced by
8926 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8927 SSL_CTX_set_options(). This is really not recommended unless you
8928 know what you are doing.
8930 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8932 * Fixes to stateless session resumption handling. Use initial_ctx when
8933 issuing and attempting to decrypt tickets in case it has changed during
8934 servername handling. Use a non-zero length session ID when attempting
8935 stateless session resumption: this makes it possible to determine if
8936 a resumption has occurred immediately after receiving server hello
8937 (several places in OpenSSL subtly assume this) instead of later in
8942 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8943 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8944 fixes for a few places where the return code is not checked
8947 *Julia Lawall <julia@diku.dk>*
8949 * Add --strict-warnings option to Configure script to include devteam
8950 warnings in other configurations.
8954 * Add support for --libdir option and LIBDIR variable in makefiles. This
8955 makes it possible to install openssl libraries in locations which
8956 have names other than "lib", for example "/usr/lib64" which some
8959 *Steve Henson, based on patch from Jeremy Utley*
8961 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8962 X690 8.9.12 and can produce some misleading textual output of OIDs.
8964 *Steve Henson, reported by Dan Kaminsky*
8966 * Delete MD2 from algorithm tables. This follows the recommendation in
8967 several standards that it is not used in new applications due to
8968 several cryptographic weaknesses. For binary compatibility reasons
8969 the MD2 API is still compiled in by default.
8973 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8978 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8979 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8982 *Guenter <lists@gknw.net>*
8984 * Fix the server certificate chain building code to use X509_verify_cert(),
8985 it used to have an ad-hoc builder which was unable to cope with anything
8986 other than a simple chain.
8988 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8990 * Don't check self signed certificate signatures in X509_verify_cert()
8991 by default (a flag can override this): it just wastes time without
8992 adding any security. As a useful side effect self signed root CAs
8993 with non-FIPS digests are now usable in FIPS mode.
8997 * In dtls1_process_out_of_seq_message() the check if the current message
8998 is already buffered was missing. For every new message was memory
8999 allocated, allowing an attacker to perform an denial of service attack
9000 with sending out of seq handshake messages until there is no memory
9001 left. Additionally every future message was buffered, even if the
9002 sequence number made no sense and would be part of another handshake.
9003 So only messages with sequence numbers less than 10 in advance will be
9004 buffered. ([CVE-2009-1378])
9006 *Robin Seggelmann, discovered by Daniel Mentz*
9008 * Records are buffered if they arrive with a future epoch to be
9009 processed after finishing the corresponding handshake. There is
9010 currently no limitation to this buffer allowing an attacker to perform
9011 a DOS attack with sending records with future epochs until there is no
9012 memory left. This patch adds the pqueue_size() function to determine
9013 the size of a buffer and limits the record buffer to 100 entries.
9016 *Robin Seggelmann, discovered by Daniel Mentz*
9018 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9019 parent structure is freed. ([CVE-2009-1379])
9023 * Handle non-blocking I/O properly in SSL_shutdown() call.
9025 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9027 * Add `2.5.4.*` OIDs
9029 *Ilya O. <vrghost@gmail.com>*
9031 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9033 * Disable renegotiation completely - this fixes a severe security
9034 problem ([CVE-2009-3555]) at the cost of breaking all
9035 renegotiation. Renegotiation can be re-enabled by setting
9036 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9037 run-time. This is really not recommended unless you know what
9042 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9044 * Don't set val to NULL when freeing up structures, it is freed up by
9045 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9046 zeroing past the valid field. ([CVE-2009-0789])
9048 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9050 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9051 checked correctly. This would allow some invalid signed attributes to
9052 appear to verify correctly. ([CVE-2009-0591])
9054 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9056 * Reject UniversalString and BMPString types with invalid lengths. This
9057 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9058 a legal length. ([CVE-2009-0590])
9062 * Set S/MIME signing as the default purpose rather than setting it
9063 unconditionally. This allows applications to override it at the store
9068 * Permit restricted recursion of ASN1 strings. This is needed in practice
9069 to handle some structures.
9073 * Improve efficiency of mem_gets: don't search whole buffer each time
9076 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9078 * New -hex option for openssl rand.
9082 * Print out UTF8String and NumericString when parsing ASN1.
9086 * Support NumericString type for name components.
9090 * Allow CC in the environment to override the automatically chosen
9091 compiler. Note that nothing is done to ensure flags work with the
9096 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9098 * Properly check EVP_VerifyFinal() and similar return values
9101 *Ben Laurie, Bodo Moeller, Google Security Team*
9103 * Enable TLS extensions by default.
9107 * Allow the CHIL engine to be loaded, whether the application is
9108 multithreaded or not. (This does not release the developer from the
9109 obligation to set up the dynamic locking callbacks.)
9111 *Sander Temme <sander@temme.net>*
9113 * Use correct exit code if there is an error in dgst command.
9115 *Steve Henson; problem pointed out by Roland Dirlewanger*
9117 * Tweak Configure so that you need to say "experimental-jpake" to enable
9118 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9122 * Add experimental JPAKE support, including demo authentication in
9123 s_client and s_server.
9127 * Set the comparison function in v3_addr_canonize().
9129 *Rob Austein <sra@hactrn.net>*
9131 * Add support for XMPP STARTTLS in s_client.
9133 *Philip Paeps <philip@freebsd.org>*
9135 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9136 to ensure that even with this option, only ciphersuites in the
9137 server's preference list will be accepted. (Note that the option
9138 applies only when resuming a session, so the earlier behavior was
9139 just about the algorithm choice for symmetric cryptography.)
9143 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9145 * Fix NULL pointer dereference if a DTLS server received
9146 ChangeCipherSpec as first record ([CVE-2009-1386]).
9150 * Fix a state transition in s3_srvr.c and d1_srvr.c
9151 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9155 * The fix in 0.9.8c that supposedly got rid of unsafe
9156 double-checked locking was incomplete for RSA blinding,
9157 addressing just one layer of what turns out to have been
9158 doubly unsafe triple-checked locking.
9160 So now fix this for real by retiring the MONT_HELPER macro
9161 in crypto/rsa/rsa_eay.c.
9163 *Bodo Moeller; problem pointed out by Marius Schilder*
9165 * Various precautionary measures:
9167 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9169 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9170 (NB: This would require knowledge of the secret session ticket key
9171 to exploit, in which case you'd be SOL either way.)
9173 - Change bn_nist.c so that it will properly handle input BIGNUMs
9174 outside the expected range.
9176 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9179 *Neel Mehta, Bodo Moeller*
9181 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9182 the load fails. Useful for distros.
9184 *Ben Laurie and the FreeBSD team*
9186 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9190 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9194 * Expand ENGINE to support engine supplied SSL client certificate functions.
9196 This work was sponsored by Logica.
9200 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9201 keystores. Support for SSL/TLS client authentication too.
9202 Not compiled unless enable-capieng specified to Configure.
9204 This work was sponsored by Logica.
9208 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9209 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9210 attribute creation routines such as certificate requests and PKCS#12
9215 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9217 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9218 handshake which could lead to a client crash as found using the
9219 Codenomicon TLS test suite ([CVE-2008-1672])
9221 *Steve Henson, Mark Cox*
9223 * Fix double free in TLS server name extensions which could lead to
9224 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9228 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9230 Clear the error queue to ensure that error entries left from
9231 older function calls do not interfere with the correct operation.
9233 *Lutz Jaenicke, Erik de Castro Lopo*
9235 * Remove root CA certificates of commercial CAs:
9237 The OpenSSL project does not recommend any specific CA and does not
9238 have any policy with respect to including or excluding any CA.
9239 Therefore it does not make any sense to ship an arbitrary selection
9240 of root CA certificates with the OpenSSL software.
9244 * RSA OAEP patches to fix two separate invalid memory reads.
9245 The first one involves inputs when 'lzero' is greater than
9246 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9247 before the beginning of from). The second one involves inputs where
9248 the 'db' section contains nothing but zeroes (there is a one-byte
9249 invalid read after the end of 'db').
9251 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9253 * Partial backport from 0.9.9-dev:
9255 Introduce bn_mul_mont (dedicated Montgomery multiplication
9256 procedure) as a candidate for BIGNUM assembler implementation.
9257 While 0.9.9-dev uses assembler for various architectures, only
9258 x86_64 is available by default here in the 0.9.8 branch, and
9259 32-bit x86 is available through a compile-time setting.
9261 To try the 32-bit x86 assembler implementation, use Configure
9262 option "enable-montasm" (which exists only for this backport).
9264 As "enable-montasm" for 32-bit x86 disclaims code stability
9265 anyway, in this constellation we activate additional code
9266 backported from 0.9.9-dev for further performance improvements,
9267 namely BN_from_montgomery_word. (To enable this otherwise,
9268 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9270 *Andy Polyakov (backport partially by Bodo Moeller)*
9272 * Add TLS session ticket callback. This allows an application to set
9273 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9274 values. This is useful for key rollover for example where several key
9275 sets may exist with different names.
9279 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9280 This was broken until now in 0.9.8 releases, such that the only way
9281 a registered ENGINE could be used (assuming it initialises
9282 successfully on the host) was to explicitly set it as the default
9283 for the relevant algorithms. This is in contradiction with 0.9.7
9284 behaviour and the documentation. With this fix, when an ENGINE is
9285 registered into a given algorithm's table of implementations, the
9286 'uptodate' flag is reset so that auto-discovery will be used next
9287 time a new context for that algorithm attempts to select an
9290 *Ian Lister (tweaked by Geoff Thorpe)*
9292 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9293 implementation in the following ways:
9295 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9298 Lack of BER streaming support means one pass streaming processing is
9299 only supported if data is detached: setting the streaming flag is
9300 ignored for embedded content.
9302 CMS support is disabled by default and must be explicitly enabled
9303 with the enable-cms configuration option.
9307 * Update the GMP engine glue to do direct copies between BIGNUM and
9308 mpz_t when openssl and GMP use the same limb size. Otherwise the
9309 existing "conversion via a text string export" trick is still used.
9311 *Paul Sheer <paulsheer@gmail.com>*
9313 * Zlib compression BIO. This is a filter BIO which compressed and
9314 uncompresses any data passed through it.
9318 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9319 RFC3394 compatible AES key wrapping.
9323 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9324 sets string data without copying. X509_ALGOR_set0() and
9325 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9326 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9327 from an X509_ATTRIBUTE structure optionally checking it occurs only
9328 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9333 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9334 to get the expected BN_FLG_CONSTTIME behavior.
9336 *Bodo Moeller (Google)*
9340 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9341 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9342 - added some more tests to do_tests.pl
9343 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9344 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9345 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9346 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9347 - various changes to netware.pl to enable gcc-cross builds on Win32
9349 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9350 - various changes to fix missing prototype warnings
9351 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9352 - added AES, WHIRLPOOL and CPUID assembler code to build files
9353 - added missing AES assembler make rules to mk1mf.pl
9354 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9356 *Guenter Knauf <eflash@gmx.net>*
9358 * Implement certificate status request TLS extension defined in RFC3546.
9359 A client can set the appropriate parameters and receive the encoded
9360 OCSP response via a callback. A server can query the supplied parameters
9361 and set the encoded OCSP response in the callback. Add simplified examples
9362 to s_client and s_server.
9366 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9369 + Binary incompatibility of ssl_ctx_st structure
9370 + DTLS interoperation with non-compliant servers
9371 + Don't call get_session_cb() without proposed session
9372 + Fix ia64 assembler code
9374 *Andy Polyakov, Steve Henson*
9376 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9378 * DTLS Handshake overhaul. There were longstanding issues with
9379 OpenSSL DTLS implementation, which were making it impossible for
9380 RFC 4347 compliant client to communicate with OpenSSL server.
9381 Unfortunately just fixing these incompatibilities would "cut off"
9382 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9383 server keeps tolerating non RFC compliant syntax. The opposite is
9384 not true, 0.9.8f client can not communicate with earlier server.
9385 This update even addresses CVE-2007-4995.
9389 * Changes to avoid need for function casts in OpenSSL: some compilers
9390 (gcc 4.2 and later) reject their use.
9391 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9394 * Add RFC4507 support to OpenSSL. This includes the corrections in
9395 RFC4507bis. The encrypted ticket format is an encrypted encoded
9396 SSL_SESSION structure, that way new session features are automatically
9399 If a client application caches session in an SSL_SESSION structure
9400 support is transparent because tickets are now stored in the encoded
9403 The SSL_CTX structure automatically generates keys for ticket
9404 protection in servers so again support should be possible
9405 with no application modification.
9407 If a client or server wishes to disable RFC4507 support then the option
9408 SSL_OP_NO_TICKET can be set.
9410 Add a TLS extension debugging callback to allow the contents of any client
9411 or server extensions to be examined.
9413 This work was sponsored by Google.
9417 * Add initial support for TLS extensions, specifically for the server_name
9418 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9419 have new members for a host name. The SSL data structure has an
9420 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9421 stored in that context to allow for session resumption, even after the
9422 SSL has been switched to a new SSL_CTX in reaction to a client's
9423 server_name extension.
9425 New functions (subject to change):
9427 SSL_get_servername()
9428 SSL_get_servername_type()
9431 New CTRL codes and macros (subject to change):
9433 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9434 - SSL_CTX_set_tlsext_servername_callback()
9435 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9436 - SSL_CTX_set_tlsext_servername_arg()
9437 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9439 openssl s_client has a new '-servername ...' option.
9441 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9442 '-key2 ...', '-servername_fatal' (subject to change). This allows
9443 testing the HostName extension for a specific single host name ('-cert'
9444 and '-key' remain fallbacks for handshakes without HostName
9445 negotiation). If the unrecognized_name alert has to be sent, this by
9446 default is a warning; it becomes fatal with the '-servername_fatal'
9449 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9451 * Add AES and SSE2 assembly language support to VC++ build.
9455 * Mitigate attack on final subtraction in Montgomery reduction.
9459 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9460 (which previously caused an internal error).
9464 * Squeeze another 10% out of IGE mode when in != out.
9468 * AES IGE mode speedup.
9470 *Dean Gaudet (Google)*
9472 * Add the Korean symmetric 128-bit cipher SEED (see
9473 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9474 add SEED ciphersuites from RFC 4162:
9476 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9477 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9478 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9479 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9481 To minimize changes between patchlevels in the OpenSSL 0.9.8
9482 series, SEED remains excluded from compilation unless OpenSSL
9483 is configured with 'enable-seed'.
9485 *KISA, Bodo Moeller*
9487 * Mitigate branch prediction attacks, which can be practical if a
9488 single processor is shared, allowing a spy process to extract
9489 information. For detailed background information, see
9490 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9491 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9492 and Necessary Software Countermeasures"). The core of the change
9493 are new versions BN_div_no_branch() and
9494 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9495 respectively, which are slower, but avoid the security-relevant
9496 conditional branches. These are automatically called by BN_div()
9497 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9498 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9499 remove a conditional branch.
9501 BN_FLG_CONSTTIME is the new name for the previous
9502 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9503 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9504 in the exponent causes BN_mod_exp_mont() to use the alternative
9505 implementation in BN_mod_exp_mont_consttime().) The old name
9506 remains as a deprecated alias.
9508 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9509 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9510 constant-time implementations for more than just exponentiation.
9511 Here too the old name is kept as a deprecated alias.
9513 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9514 the BN_BLINDING structure gets an independent copy of the
9515 modulus. This means that the previous `BIGNUM *m` argument to
9516 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9517 essentially becomes `const BIGNUM *m`, although we can't actually
9518 change this in the header file before 0.9.9. It allows
9519 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9520 enable BN_FLG_CONSTTIME.
9522 *Matthew D Wood (Intel Corp)*
9524 * In the SSL/TLS server implementation, be strict about session ID
9525 context matching (which matters if an application uses a single
9526 external cache for different purposes). Previously,
9527 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9528 set. This did ensure strict client verification, but meant that,
9529 with applications using a single external cache for quite
9530 different requirements, clients could circumvent ciphersuite
9531 restrictions for a given session ID context by starting a session
9532 in a different context.
9536 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9537 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9538 authentication-only ciphersuites.
9542 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9543 not complete and could lead to a possible single byte overflow
9544 ([CVE-2007-5135]) [Ben Laurie]
9546 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9548 * Since AES128 and AES256 (and similarly Camellia128 and
9549 Camellia256) share a single mask bit in the logic of
9550 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9551 kludge to work properly if AES128 is available and AES256 isn't
9552 (or if Camellia128 is available and Camellia256 isn't).
9556 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9557 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9558 When a point or a seed is encoded in a BIT STRING, we need to
9559 prevent the removal of trailing zero bits to get the proper DER
9560 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9561 of a NamedBitList, for which trailing 0 bits need to be removed.)
9565 * Have SSL/TLS server implementation tolerate "mismatched" record
9566 protocol version while receiving ClientHello even if the
9567 ClientHello is fragmented. (The server can't insist on the
9568 particular protocol version it has chosen before the ServerHello
9569 message has informed the client about his choice.)
9573 * Add RFC 3779 support.
9575 *Rob Austein for ARIN, Ben Laurie*
9577 * Load error codes if they are not already present instead of using a
9578 static variable. This allows them to be cleanly unloaded and reloaded.
9579 Improve header file function name parsing.
9583 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9584 or CAPABILITY handshake as required by RFCs.
9588 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9590 * Introduce limits to prevent malicious keys being able to
9591 cause a denial of service. ([CVE-2006-2940])
9593 *Steve Henson, Bodo Moeller*
9595 * Fix ASN.1 parsing of certain invalid structures that can result
9596 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9598 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9599 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9601 * Fix SSL client code which could crash if connecting to a
9602 malicious SSLv2 server. ([CVE-2006-4343])
9604 *Tavis Ormandy and Will Drewry, Google Security Team*
9606 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9607 match only those. Before that, "AES256-SHA" would be interpreted
9608 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9609 the same strength classification in 0.9.7h) as we currently only
9610 have a single AES bit in the ciphersuite description bitmap.
9611 That change, however, also applied to ciphersuite strings such as
9612 "RC4-MD5" that intentionally matched multiple ciphersuites --
9613 namely, SSL 2.0 ciphersuites in addition to the more common ones
9614 from SSL 3.0/TLS 1.0.
9616 So we change the selection algorithm again: Naming an explicit
9617 ciphersuite selects this one ciphersuite, and any other similar
9618 ciphersuite (same bitmap) from *other* protocol versions.
9619 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9620 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9622 Since SSL 2.0 does not have any ciphersuites for which the
9623 128/256 bit distinction would be relevant, this works for now.
9624 The proper fix will be to use different bits for AES128 and
9625 AES256, which would have avoided the problems from the beginning;
9626 however, bits are scarce, so we can only do this in a new release
9627 (not just a patchlevel) when we can change the SSL_CIPHER
9628 definition to split the single 'unsigned long mask' bitmap into
9629 multiple values to extend the available space.
9633 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9635 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9636 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9638 * Add AES IGE and biIGE modes.
9642 * Change the Unix randomness entropy gathering to use poll() when
9643 possible instead of select(), since the latter has some
9644 undesirable limitations.
9646 *Darryl Miles via Richard Levitte and Bodo Moeller*
9648 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9649 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9650 cannot be implicitly activated as part of, e.g., the "AES" alias.
9651 However, please upgrade to OpenSSL 0.9.9[-dev] for
9652 non-experimental use of the ECC ciphersuites to get TLS extension
9653 support, which is required for curve and point format negotiation
9654 to avoid potential handshake problems.
9658 * Disable rogue ciphersuites:
9660 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9661 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9662 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9664 The latter two were purportedly from
9665 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9668 Also deactivate the remaining ciphersuites from
9669 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9670 unofficial, and the ID has long expired.
9674 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9675 dual-core machines) and other potential thread-safety issues.
9679 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9680 versions), which is now available for royalty-free use
9681 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9682 Also, add Camellia TLS ciphersuites from RFC 4132.
9684 To minimize changes between patchlevels in the OpenSSL 0.9.8
9685 series, Camellia remains excluded from compilation unless OpenSSL
9686 is configured with 'enable-camellia'.
9690 * Disable the padding bug check when compression is in use. The padding
9691 bug check assumes the first packet is of even length, this is not
9692 necessarily true if compression is enabled and can result in false
9693 positives causing handshake failure. The actual bug test is ancient
9694 code so it is hoped that implementations will either have fixed it by
9695 now or any which still have the bug do not support compression.
9699 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9701 * When applying a cipher rule check to see if string match is an explicit
9702 cipher suite and only match that one cipher suite if it is.
9706 * Link in manifests for VC++ if needed.
9708 *Austin Ziegler <halostatue@gmail.com>*
9710 * Update support for ECC-based TLS ciphersuites according to
9711 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9712 TLS extensions, which are supported starting with the 0.9.9
9713 branch, not in the OpenSSL 0.9.8 branch).
9717 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9718 opaque EVP_CIPHER_CTX handling.
9722 * Fixes and enhancements to zlib compression code. We now only use
9723 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9724 to conform with the standards mentioned here:
9725 <http://www.zlib.net/DLL_FAQ.txt>
9726 Static zlib linking now works on Windows and the new --with-zlib-include
9727 --with-zlib-lib options to Configure can be used to supply the location
9728 of the headers and library. Gracefully handle case where zlib library
9733 * Several fixes and enhancements to the OID generation code. The old code
9734 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9735 handle numbers larger than ULONG_MAX, truncated printing and had a
9736 non standard OBJ_obj2txt() behaviour.
9740 * Add support for building of engines under engine/ as shared libraries
9741 under VC++ build system.
9745 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9746 Hopefully, we will not see any false combination of paths any more.
9750 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9752 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9753 (part of SSL_OP_ALL). This option used to disable the
9754 countermeasure against man-in-the-middle protocol-version
9755 rollback in the SSL 2.0 server implementation, which is a bad
9756 idea. ([CVE-2005-2969])
9758 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9759 for Information Security, National Institute of Advanced Industrial
9760 Science and Technology [AIST], Japan)*
9762 * Add two function to clear and return the verify parameter flags.
9766 * Keep cipherlists sorted in the source instead of sorting them at
9767 runtime, thus removing the need for a lock.
9771 * Avoid some small subgroup attacks in Diffie-Hellman.
9773 *Nick Mathewson and Ben Laurie*
9775 * Add functions for well-known primes.
9779 * Extended Windows CE support.
9781 *Satoshi Nakamura and Andy Polyakov*
9783 * Initialize SSL_METHOD structures at compile time instead of during
9784 runtime, thus removing the need for a lock.
9788 * Make PKCS7_decrypt() work even if no certificate is supplied by
9789 attempting to decrypt each encrypted key in turn. Add support to
9794 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9796 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9799 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9803 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9804 key into the same file any more.
9808 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9812 * Add -utf8 command line and config file option to 'ca'.
9814 *Stefan <stf@udoma.org*
9816 * Removed the macro des_crypt(), as it seems to conflict with some
9817 libraries. Use DES_crypt().
9821 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9822 involves renaming the source and generated shared-libs for
9823 both. The engines will accept the corrected or legacy ids
9824 ('ncipher' and '4758_cca' respectively) when binding. NB,
9825 this only applies when building 'shared'.
9827 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9829 * Add attribute functions to EVP_PKEY structure. Modify
9830 PKCS12_create() to recognize a CSP name attribute and
9831 use it. Make -CSP option work again in pkcs12 utility.
9835 * Add new functionality to the bn blinding code:
9836 - automatic re-creation of the BN_BLINDING parameters after
9837 a fixed number of uses (currently 32)
9838 - add new function for parameter creation
9839 - introduce flags to control the update behaviour of the
9840 BN_BLINDING parameters
9841 - hide BN_BLINDING structure
9842 Add a second BN_BLINDING slot to the RSA structure to improve
9843 performance when a single RSA object is shared among several
9848 * Add support for DTLS.
9850 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9852 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9853 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9857 * Remove buggy and incomplete DH cert support from
9858 ssl/ssl_rsa.c and ssl/s3_both.c
9862 * Use SHA-1 instead of MD5 as the default digest algorithm for
9863 the apps/openssl applications.
9867 * Compile clean with "-Wall -Wmissing-prototypes
9868 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9869 DEBUG_SAFESTACK must also be set.
9873 * Change ./Configure so that certain algorithms can be disabled by default.
9874 The new counterpiece to "no-xxx" is "enable-xxx".
9876 The patented RC5 and MDC2 algorithms will now be disabled unless
9877 "enable-rc5" and "enable-mdc2", respectively, are specified.
9879 (IDEA remains enabled despite being patented. This is because IDEA
9880 is frequently required for interoperability, and there is no license
9881 fee for non-commercial use. As before, "no-idea" can be used to
9882 avoid this algorithm.)
9886 * Add processing of proxy certificates (see RFC 3820). This work was
9887 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9888 EGEE (Enabling Grids for E-science in Europe).
9892 * RC4 performance overhaul on modern architectures/implementations, such
9893 as Intel P4, IA-64 and AMD64.
9897 * New utility extract-section.pl. This can be used specify an alternative
9898 section number in a pod file instead of having to treat each file as
9899 a separate case in Makefile. This can be done by adding two lines to the
9902 =for comment openssl_section:XXX
9904 The blank line is mandatory.
9908 * New arguments -certform, -keyform and -pass for s_client and s_server
9909 to allow alternative format key and certificate files and passphrase
9914 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9915 update associated structures and add various utility functions.
9917 Add new policy related verify parameters, include policy checking in
9918 standard verify code. Enhance 'smime' application with extra parameters
9919 to support policy checking and print out.
9923 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9924 Nehemiah processors. These extensions support AES encryption in hardware
9925 as well as RNG (though RNG support is currently disabled).
9927 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9929 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9933 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9935 *Andy Polyakov and a number of other people*
9937 * Improved PowerPC platform support. Most notably BIGNUM assembler
9938 implementation contributed by IBM.
9940 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9942 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9943 exponent rather than 'unsigned long'. There is a corresponding change to
9944 the new 'rsa_keygen' element of the RSA_METHOD structure.
9946 *Jelte Jansen, Geoff Thorpe*
9948 * Functionality for creating the initial serial number file is now
9949 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9951 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9952 number file to 1, which is bound to cause problems. To avoid
9953 the problems while respecting compatibility between different 0.9.7
9954 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9955 CA.pl for serial number initialization. With the new release 0.9.8,
9956 we can fix the problem directly in the 'ca' utility.)
9960 * Reduced header interdependencies by declaring more opaque objects in
9961 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9962 give fewer recursive includes, which could break lazy source code - so
9963 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9964 developers should define this symbol when building and using openssl to
9965 ensure they track the recommended behaviour, interfaces, [etc], but
9966 backwards-compatible behaviour prevails when this isn't defined.
9970 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9974 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9975 This will generate a random key of the appropriate length based on the
9976 cipher context. The EVP_CIPHER can provide its own random key generation
9977 routine to support keys of a specific form. This is used in the des and
9978 3des routines to generate a key of the correct parity. Update S/MIME
9979 code to use new functions and hence generate correct parity DES keys.
9980 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9981 valid (weak or incorrect parity).
9985 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9986 as looking them up. This is useful when the verified structure may contain
9987 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9988 present unless the new PKCS7_NO_CRL flag is asserted.
9992 * Extend ASN1 oid configuration module. It now additionally accepts the
9995 shortName = some long name, 1.2.3.4
9999 * Reimplemented the BN_CTX implementation. There is now no more static
10000 limitation on the number of variables it can handle nor the depth of the
10001 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10002 information can now expand as required, and rather than having a single
10003 static array of bignums, BN_CTX now uses a linked-list of such arrays
10004 allowing it to expand on demand whilst maintaining the usefulness of
10005 BN_CTX's "bundling".
10009 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10010 to allow all RSA operations to function using a single BN_CTX.
10014 * Preliminary support for certificate policy evaluation and checking. This
10015 is initially intended to pass the tests outlined in "Conformance Testing
10016 of Relying Party Client Certificate Path Processing Logic" v1.07.
10020 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10021 remained unused and not that useful. A variety of other little bignum
10022 tweaks and fixes have also been made continuing on from the audit (see
10027 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10028 associated ASN1, EVP and SSL functions and old ASN1 macros.
10032 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10033 and this should never fail. So the return value from the use of
10034 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10035 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10039 * BN_CTX_get() should return zero-valued bignums, providing the same
10040 initialised value as BN_new().
10042 *Geoff Thorpe, suggested by Ulf Möller*
10044 * Support for inhibitAnyPolicy certificate extension.
10048 * An audit of the BIGNUM code is underway, for which debugging code is
10049 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10050 is considered valid when processing BIGNUMs, and causes execution to
10051 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10052 further steps are taken to deliberately pollute unused data in BIGNUM
10053 structures to try and expose faulty code further on. For now, openssl will
10054 (in its default mode of operation) continue to tolerate the inconsistent
10055 forms that it has tolerated in the past, but authors and packagers should
10056 consider trying openssl and their own applications when compiled with
10057 these debugging symbols defined. It will help highlight potential bugs in
10058 their own code, and will improve the test coverage for OpenSSL itself. At
10059 some point, these tighter rules will become openssl's default to improve
10060 maintainability, though the assert()s and other overheads will remain only
10061 in debugging configurations. See bn.h for more details.
10063 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10065 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10066 that can only be obtained through BN_CTX_new() (which implicitly
10067 initialises it). The presence of this function only made it possible
10068 to overwrite an existing structure (and cause memory leaks).
10072 * Because of the callback-based approach for implementing LHASH as a
10073 template type, lh_insert() adds opaque objects to hash-tables and
10074 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10075 to clean up those corresponding objects before destroying the hash table
10076 (and losing the object pointers). So some over-zealous constifications in
10077 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10078 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10079 prototyped to have "const" restrictions on the object pointers they are
10080 given (and so aren't required to cast them away any more).
10084 * The tmdiff.h API was so ugly and minimal that our own timing utility
10085 (speed) prefers to use its own implementation. The two implementations
10086 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10087 its object type properly exposed (MS_TM) instead of casting to/from
10088 `char *`. This may still change yet if someone realises MS_TM and
10090 aren't necessarily the greatest nomenclatures - but this is what was used
10091 internally to the implementation so I've used that for now.
10095 * Ensure that deprecated functions do not get compiled when
10096 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10097 the self-tests were still using deprecated key-generation functions so
10098 these have been updated also.
10102 * Reorganise PKCS#7 code to separate the digest location functionality
10103 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10104 New function PKCS7_set_digest() to set the digest type for PKCS#7
10105 digestedData type. Add additional code to correctly generate the
10106 digestedData type and add support for this type in PKCS7 initialization
10111 * New function PKCS7_set0_type_other() this initializes a PKCS7
10112 structure of type "other".
10116 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10117 sure the loop does correctly stop and breaking ("division by zero")
10118 modulus operations are not performed. The (pre-generated) prime
10119 table crypto/bn/bn_prime.h was already correct, but it could not be
10120 re-generated on some platforms because of the "division by zero"
10121 situation in the script.
10123 *Ralf S. Engelschall*
10125 * Update support for ECC-based TLS ciphersuites according to
10126 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10127 SHA-1 now is only used for "small" curves (where the
10128 representation of a field element takes up to 24 bytes); for
10129 larger curves, the field element resulting from ECDH is directly
10130 used as premaster secret.
10132 *Douglas Stebila (Sun Microsystems Laboratories)*
10134 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10135 curve secp160r1 to the tests.
10137 *Douglas Stebila (Sun Microsystems Laboratories)*
10139 * Add the possibility to load symbols globally with DSO.
10141 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10143 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10144 control of the error stack.
10148 * Add support for STORE in ENGINE.
10152 * Add the STORE type. The intention is to provide a common interface
10153 to certificate and key stores, be they simple file-based stores, or
10154 HSM-type store, or LDAP stores, or...
10155 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10159 * Add a generic structure called OPENSSL_ITEM. This can be used to
10160 pass a list of arguments to any function as well as provide a way
10161 for a function to pass data back to the caller.
10165 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10166 works like BUF_strdup() but can be used to duplicate a portion of
10167 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10172 * Add the function sk_find_ex() which works like sk_find(), but will
10173 return an index to an element even if an exact match couldn't be
10174 found. The index is guaranteed to point at the element where the
10175 searched-for key would be inserted to preserve sorting order.
10179 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10180 takes an extra flags argument for optional functionality. Currently,
10181 the following flags are defined:
10183 OBJ_BSEARCH_VALUE_ON_NOMATCH
10184 This one gets OBJ_bsearch_ex() to return a pointer to the first
10185 element where the comparing function returns a negative or zero
10188 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10189 This one gets OBJ_bsearch_ex() to return a pointer to the first
10190 element where the comparing function returns zero. This is useful
10191 if there are more than one element where the comparing function
10196 * Make it possible to create self-signed certificates with 'openssl ca'
10197 in such a way that the self-signed certificate becomes part of the
10198 CA database and uses the same mechanisms for serial number generation
10199 as all other certificate signing. The new flag '-selfsign' enables
10200 this functionality. Adapt CA.sh and CA.pl.in.
10204 * Add functionality to check the public key of a certificate request
10205 against a given private. This is useful to check that a certificate
10206 request can be signed by that key (self-signing).
10210 * Make it possible to have multiple active certificates with the same
10211 subject in the CA index file. This is done only if the keyword
10212 'unique_subject' is set to 'no' in the main CA section (default
10213 if 'CA_default') of the configuration file. The value is saved
10214 with the database itself in a separate index attribute file,
10215 named like the index file with '.attr' appended to the name.
10219 * Generate multi-valued AVAs using '+' notation in config files for
10224 * Support for nameConstraints certificate extension.
10228 * Support for policyConstraints certificate extension.
10232 * Support for policyMappings certificate extension.
10236 * Make sure the default DSA_METHOD implementation only uses its
10237 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10238 and change its own handlers to be NULL so as to remove unnecessary
10239 indirection. This lets alternative implementations fallback to the
10240 default implementation more easily.
10244 * Support for directoryName in GeneralName related extensions
10249 * Make it possible to link applications using Makefile.shared.
10250 Make that possible even when linking against static libraries!
10254 * Support for single pass processing for S/MIME signing. This now
10255 means that S/MIME signing can be done from a pipe, in addition
10256 cleartext signing (multipart/signed type) is effectively streaming
10257 and the signed data does not need to be all held in memory.
10259 This is done with a new flag PKCS7_STREAM. When this flag is set
10260 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10261 is done after the data is output (and digests calculated) in
10262 SMIME_write_PKCS7().
10266 * Add full support for -rpath/-R, both in shared libraries and
10267 applications, at least on the platforms where it's known how
10272 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10273 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10274 will now compute a table of multiples of the generator that
10275 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10276 faster (notably in the case of a single point multiplication,
10277 scalar * generator).
10279 *Nils Larsch, Bodo Moeller*
10281 * IPv6 support for certificate extensions. The various extensions
10282 which use the IP:a.b.c.d can now take IPv6 addresses using the
10283 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10288 * Added an ENGINE that implements RSA by performing private key
10289 exponentiations with the GMP library. The conversions to and from
10290 GMP's mpz_t format aren't optimised nor are any montgomery forms
10291 cached, and on x86 it appears OpenSSL's own performance has caught up.
10292 However there are likely to be other architectures where GMP could
10293 provide a boost. This ENGINE is not built in by default, but it can be
10294 specified at Configure time and should be accompanied by the necessary
10295 linker additions, eg;
10296 ./config -DOPENSSL_USE_GMP -lgmp
10300 * "openssl engine" will not display ENGINE/DSO load failure errors when
10301 testing availability of engines with "-t" - the old behaviour is
10302 produced by increasing the feature's verbosity with "-tt".
10306 * ECDSA routines: under certain error conditions uninitialized BN objects
10307 could be freed. Solution: make sure initialization is performed early
10308 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10313 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10314 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10315 software implementations. For DSA and DH, parameter generation can
10316 also be overridden by providing the appropriate method callbacks.
10320 * Change the "progress" mechanism used in key-generation and
10321 primality testing to functions that take a new BN_GENCB pointer in
10322 place of callback/argument pairs. The new API functions have `_ex`
10323 postfixes and the older functions are reimplemented as wrappers for
10324 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10325 declarations of the old functions to help (graceful) attempts to
10326 migrate to the new functions. Also, the new key-generation API
10327 functions operate on a caller-supplied key-structure and return
10328 success/failure rather than returning a key or NULL - this is to
10329 help make "keygen" another member function of RSA_METHOD etc.
10331 Example for using the new callback interface:
10333 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10334 void *my_arg = ...;
10337 BN_GENCB_set(&my_cb, my_callback, my_arg);
10339 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10340 /* For the meaning of a, b in calls to my_callback(), see the
10341 * documentation of the function that calls the callback.
10342 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10343 * my_callback should return 1 if it wants BN_is_prime_ex()
10344 * to continue, or 0 to stop.
10349 * Change the ZLIB compression method to be stateful, and make it
10350 available to TLS with the number defined in
10351 draft-ietf-tls-compression-04.txt.
10355 * Add the ASN.1 structures and functions for CertificatePair, which
10356 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10358 CertificatePair ::= SEQUENCE {
10359 forward [0] Certificate OPTIONAL,
10360 reverse [1] Certificate OPTIONAL,
10361 -- at least one of the pair shall be present -- }
10363 Also implement the PEM functions to read and write certificate
10364 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10366 This needed to be defined, mostly for the sake of the LDAP
10367 attribute crossCertificatePair, but may prove useful elsewhere as
10372 * Make it possible to inhibit symlinking of shared libraries in
10373 Makefile.shared, for Cygwin's sake.
10377 * Extend the BIGNUM API by creating a function
10378 void BN_set_negative(BIGNUM *a, int neg);
10379 and a macro that behave like
10380 int BN_is_negative(const BIGNUM *a);
10382 to avoid the need to access 'a->neg' directly in applications.
10386 * Implement fast modular reduction for pseudo-Mersenne primes
10387 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10388 EC_GROUP_new_curve_GFp() will now automatically use this
10391 *Nils Larsch <nla@trustcenter.de>*
10393 * Add new lock type (CRYPTO_LOCK_BN).
10397 * Change the ENGINE framework to automatically load engines
10398 dynamically from specific directories unless they could be
10399 found to already be built in or loaded. Move all the
10400 current engines except for the cryptodev one to a new
10401 directory engines/.
10402 The engines in engines/ are built as shared libraries if
10403 the "shared" options was given to ./Configure or ./config.
10404 Otherwise, they are inserted in libcrypto.a.
10405 /usr/local/ssl/engines is the default directory for dynamic
10406 engines, but that can be overridden at configure time through
10407 the usual use of --prefix and/or --openssldir, and at run
10408 time with the environment variable OPENSSL_ENGINES.
10410 *Geoff Thorpe and Richard Levitte*
10412 * Add Makefile.shared, a helper makefile to build shared
10413 libraries. Adapt Makefile.org.
10417 * Add version info to Win32 DLLs.
10419 *Peter 'Luna' Runestig" <peter@runestig.com>*
10421 * Add new 'medium level' PKCS#12 API. Certificates and keys
10422 can be added using this API to created arbitrary PKCS#12
10423 files while avoiding the low level API.
10425 New options to PKCS12_create(), key or cert can be NULL and
10426 will then be omitted from the output file. The encryption
10427 algorithm NIDs can be set to -1 for no encryption, the mac
10428 iteration count can be set to 0 to omit the mac.
10430 Enhance pkcs12 utility by making the -nokeys and -nocerts
10431 options work when creating a PKCS#12 file. New option -nomac
10432 to omit the mac, NONE can be set for an encryption algorithm.
10433 New code is modified to use the enhanced PKCS12_create()
10434 instead of the low level API.
10438 * Extend ASN1 encoder to support indefinite length constructed
10439 encoding. This can output sequences tags and octet strings in
10440 this form. Modify pk7_asn1.c to support indefinite length
10441 encoding. This is experimental and needs additional code to
10442 be useful, such as an ASN1 bio and some enhanced streaming
10445 Extend template encode functionality so that tagging is passed
10446 down to the template encoder.
10450 * Let 'openssl req' fail if an argument to '-newkey' is not
10451 recognized instead of using RSA as a default.
10455 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10456 As these are not official, they are not included in "ALL";
10457 the "ECCdraft" ciphersuite group alias can be used to select them.
10459 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10461 * Add ECDH engine support.
10463 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10465 * Add ECDH in new directory crypto/ecdh/.
10467 *Douglas Stebila (Sun Microsystems Laboratories)*
10469 * Let BN_rand_range() abort with an error after 100 iterations
10470 without success (which indicates a broken PRNG).
10474 * Change BN_mod_sqrt() so that it verifies that the input value
10475 is really the square of the return value. (Previously,
10476 BN_mod_sqrt would show GIGO behaviour.)
10480 * Add named elliptic curves over binary fields from X9.62, SECG,
10481 and WAP/WTLS; add OIDs that were still missing.
10483 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10485 * Extend the EC library for elliptic curves over binary fields
10486 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10489 EC_GF2m_simple_method
10493 EC_GROUP_new_curve_GF2m
10494 EC_GROUP_set_curve_GF2m
10495 EC_GROUP_get_curve_GF2m
10496 EC_POINT_set_affine_coordinates_GF2m
10497 EC_POINT_get_affine_coordinates_GF2m
10498 EC_POINT_set_compressed_coordinates_GF2m
10500 Point compression for binary fields is disabled by default for
10501 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10504 As binary polynomials are represented as BIGNUMs, various members
10505 of the EC_GROUP and EC_POINT data structures can be shared
10506 between the implementations for prime fields and binary fields;
10507 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10508 are essentially identical to their `..._GFp` counterparts.
10509 (For simplicity, the `..._GFp` prefix has been dropped from
10510 various internal method names.)
10512 An internal 'field_div' method (similar to 'field_mul' and
10513 'field_sqr') has been added; this is used only for binary fields.
10515 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10517 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10518 through methods ('mul', 'precompute_mult').
10520 The generic implementations (now internally called 'ec_wNAF_mul'
10521 and 'ec_wNAF_precomputed_mult') remain the default if these
10522 methods are undefined.
10524 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10526 * New function EC_GROUP_get_degree, which is defined through
10527 EC_METHOD. For curves over prime fields, this returns the bit
10528 length of the modulus.
10530 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10532 * New functions EC_GROUP_dup, EC_POINT_dup.
10533 (These simply call ..._new and ..._copy).
10535 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10537 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10538 Polynomials are represented as BIGNUMs (where the sign bit is not
10539 used) in the following functions [macros]:
10542 BN_GF2m_sub [= BN_GF2m_add]
10543 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10544 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10545 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10547 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10548 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10549 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10550 BN_GF2m_cmp [= BN_ucmp]
10552 (Note that only the 'mod' functions are actually for fields GF(2^m).
10553 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10555 For some functions, an the irreducible polynomial defining a
10556 field can be given as an 'unsigned int[]' with strictly
10557 decreasing elements giving the indices of those bits that are set;
10558 i.e., p[] represents the polynomial
10559 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10561 p[0] > p[1] > ... > p[k] = 0.
10562 This applies to the following functions:
10565 BN_GF2m_mod_mul_arr
10566 BN_GF2m_mod_sqr_arr
10567 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10568 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10569 BN_GF2m_mod_exp_arr
10570 BN_GF2m_mod_sqrt_arr
10571 BN_GF2m_mod_solve_quad_arr
10575 Conversion can be performed by the following functions:
10580 bntest.c has additional tests for binary polynomial arithmetic.
10582 Two implementations for BN_GF2m_mod_div() are available.
10583 The default algorithm simply uses BN_GF2m_mod_inv() and
10584 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10585 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10586 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10588 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10590 * Add new error code 'ERR_R_DISABLED' that can be used when some
10591 functionality is disabled at compile-time.
10593 *Douglas Stebila <douglas.stebila@sun.com>*
10595 * Change default behaviour of 'openssl asn1parse' so that more
10596 information is visible when viewing, e.g., a certificate:
10598 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10599 mode the content of non-printable OCTET STRINGs is output in a
10600 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10601 avoid the appearance of a printable string.
10603 *Nils Larsch <nla@trustcenter.de>*
10605 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10607 EC_GROUP_set_asn1_flag()
10608 EC_GROUP_get_asn1_flag()
10609 EC_GROUP_set_point_conversion_form()
10610 EC_GROUP_get_point_conversion_form()
10611 These control ASN1 encoding details:
10612 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10613 has been set to OPENSSL_EC_NAMED_CURVE.
10614 - Points are encoded in uncompressed form by default; options for
10615 asn1_for are as for point2oct, namely
10616 POINT_CONVERSION_COMPRESSED
10617 POINT_CONVERSION_UNCOMPRESSED
10618 POINT_CONVERSION_HYBRID
10620 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10622 EC_GROUP_set_seed()
10623 EC_GROUP_get0_seed()
10624 EC_GROUP_get_seed_len()
10625 This is used only for ASN1 purposes (so far).
10627 *Nils Larsch <nla@trustcenter.de>*
10629 * Add 'field_type' member to EC_METHOD, which holds the NID
10630 of the appropriate field type OID. The new function
10631 EC_METHOD_get_field_type() returns this value.
10633 *Nils Larsch <nla@trustcenter.de>*
10636 EC_POINT_point2bn()
10637 EC_POINT_bn2point()
10638 EC_POINT_point2hex()
10639 EC_POINT_hex2point()
10640 providing useful interfaces to EC_POINT_point2oct() and
10641 EC_POINT_oct2point().
10643 *Nils Larsch <nla@trustcenter.de>*
10645 * Change internals of the EC library so that the functions
10646 EC_GROUP_set_generator()
10647 EC_GROUP_get_generator()
10648 EC_GROUP_get_order()
10649 EC_GROUP_get_cofactor()
10650 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10651 to methods, which would lead to unnecessary code duplication when
10652 adding different types of curves.
10654 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10656 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10657 arithmetic, and such that modified wNAFs are generated
10658 (which avoid length expansion in many cases).
10662 * Add a function EC_GROUP_check_discriminant() (defined via
10663 EC_METHOD) that verifies that the curve discriminant is non-zero.
10665 Add a function EC_GROUP_check() that makes some sanity tests
10666 on a EC_GROUP, its generator and order. This includes
10667 EC_GROUP_check_discriminant().
10669 *Nils Larsch <nla@trustcenter.de>*
10671 * Add ECDSA in new directory crypto/ecdsa/.
10673 Add applications 'openssl ecparam' and 'openssl ecdsa'
10674 (these are based on 'openssl dsaparam' and 'openssl dsa').
10676 ECDSA support is also included in various other files across the
10677 library. Most notably,
10678 - 'openssl req' now has a '-newkey ecdsa:file' option;
10679 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10680 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10681 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10682 them suitable for ECDSA where domain parameters must be
10683 extracted before the specific public key;
10684 - ECDSA engine support has been added.
10686 *Nils Larsch <nla@trustcenter.de>*
10688 * Include some named elliptic curves, and add OIDs from X9.62,
10689 SECG, and WAP/WTLS. Each curve can be obtained from the new
10691 EC_GROUP_new_by_curve_name(),
10692 and the list of available named curves can be obtained with
10693 EC_get_builtin_curves().
10694 Also add a 'curve_name' member to EC_GROUP objects, which can be
10696 EC_GROUP_set_curve_name()
10697 EC_GROUP_get_curve_name()
10699 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10701 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10702 was actually never needed) and in BN_mul(). The removal in BN_mul()
10703 required a small change in bn_mul_part_recursive() and the addition
10704 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10705 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10706 bn_sub_words() and bn_add_words() except they take arrays with
10711 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10713 * Cleanse PEM buffers before freeing them since they may contain
10716 *Benjamin Bennett <ben@psc.edu>*
10718 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10719 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10720 authentication-only ciphersuites.
10724 * Since AES128 and AES256 share a single mask bit in the logic of
10725 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10726 kludge to work properly if AES128 is available and AES256 isn't.
10730 * Expand security boundary to match 1.1.1 module.
10734 * Remove redundant features: hash file source, editing of test vectors
10735 modify fipsld to use external fips_premain.c signature.
10739 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10740 run algorithm test programs.
10744 * Make algorithm test programs more tolerant of whitespace.
10748 * Have SSL/TLS server implementation tolerate "mismatched" record
10749 protocol version while receiving ClientHello even if the
10750 ClientHello is fragmented. (The server can't insist on the
10751 particular protocol version it has chosen before the ServerHello
10752 message has informed the client about his choice.)
10756 * Load error codes if they are not already present instead of using a
10757 static variable. This allows them to be cleanly unloaded and reloaded.
10761 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10763 * Introduce limits to prevent malicious keys being able to
10764 cause a denial of service. ([CVE-2006-2940])
10766 *Steve Henson, Bodo Moeller*
10768 * Fix ASN.1 parsing of certain invalid structures that can result
10769 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10771 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10772 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10774 * Fix SSL client code which could crash if connecting to a
10775 malicious SSLv2 server. ([CVE-2006-4343])
10777 *Tavis Ormandy and Will Drewry, Google Security Team*
10779 * Change ciphersuite string processing so that an explicit
10780 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10781 will no longer include "AES128-SHA"), and any other similar
10782 ciphersuite (same bitmap) from *other* protocol versions (so that
10783 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10784 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10785 changes from 0.9.8b and 0.9.8d.
10789 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10791 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10792 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10794 * Change the Unix randomness entropy gathering to use poll() when
10795 possible instead of select(), since the latter has some
10796 undesirable limitations.
10798 *Darryl Miles via Richard Levitte and Bodo Moeller*
10800 * Disable rogue ciphersuites:
10802 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10803 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10804 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10806 The latter two were purportedly from
10807 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10810 Also deactivate the remaining ciphersuites from
10811 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10812 unofficial, and the ID has long expired.
10816 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10817 dual-core machines) and other potential thread-safety issues.
10821 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10823 * Adapt fipsld and the build system to link against the validated FIPS
10824 module in FIPS mode.
10828 * Fixes for VC++ 2005 build under Windows.
10832 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10833 from a Windows bash shell such as MSYS. It is autodetected from the
10834 "config" script when run from a VC++ environment. Modify standard VC++
10835 build to use fipscanister.o from the GNU make build.
10839 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10841 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10842 The value now differs depending on if you build for FIPS or not.
10843 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10844 safely run with a non-FIPSed libcrypto, as it may crash because of
10845 the difference induced by this change.
10849 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10851 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10852 (part of SSL_OP_ALL). This option used to disable the
10853 countermeasure against man-in-the-middle protocol-version
10854 rollback in the SSL 2.0 server implementation, which is a bad
10855 idea. ([CVE-2005-2969])
10857 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10858 for Information Security, National Institute of Advanced Industrial
10859 Science and Technology [AIST, Japan)]*
10861 * Minimal support for X9.31 signatures and PSS padding modes. This is
10862 mainly for FIPS compliance and not fully integrated at this stage.
10866 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10867 the exponentiation using a fixed-length exponent. (Otherwise,
10868 the information leaked through timing could expose the secret key
10869 after many signatures; cf. Bleichenbacher's attack on DSA with
10874 * Make a new fixed-window mod_exp implementation the default for
10875 RSA, DSA, and DH private-key operations so that the sequence of
10876 squares and multiplies and the memory access pattern are
10877 independent of the particular secret key. This will mitigate
10878 cache-timing and potential related attacks.
10880 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10881 and this is automatically used by BN_mod_exp_mont() if the new flag
10882 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10883 will use this BN flag for private exponents unless the flag
10884 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10885 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10887 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10889 * Change the client implementation for SSLv23_method() and
10890 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10891 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10892 (Previously, the SSL 2.0 backwards compatible Client Hello
10893 message format would be used even with SSL_OP_NO_SSLv2.)
10897 * Add support for smime-type MIME parameter in S/MIME messages which some
10902 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10903 a threadsafe manner. Modify rsa code to use new function and add calls
10904 to dsa and dh code (which had race conditions before).
10908 * Include the fixed error library code in the C error file definitions
10909 instead of fixing them up at runtime. This keeps the error code
10910 structures constant.
10914 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10916 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10919 * Fixes for newer kerberos headers. NB: the casts are needed because
10920 the 'length' field is signed on one version and unsigned on another
10921 with no (?) obvious way to tell the difference, without these VC++
10922 complains. Also the "definition" of FAR (blank) is no longer included
10923 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10924 some needed definitions.
10928 * Undo Cygwin change.
10932 * Added support for proxy certificates according to RFC 3820.
10933 Because they may be a security thread to unaware applications,
10934 they must be explicitly allowed in run-time. See
10935 docs/HOWTO/proxy_certificates.txt for further information.
10939 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10941 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10942 server and client random values. Previously
10943 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10944 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10946 This change has negligible security impact because:
10948 1. Server and client random values still have 24 bytes of pseudo random
10951 2. Server and client random values are sent in the clear in the initial
10954 3. The master secret is derived using the premaster secret (48 bytes in
10955 size for static RSA ciphersuites) as well as client server and random
10958 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10961 *Stephen Henson, reported by UK NISCC*
10963 * Use Windows randomness collection on Cygwin.
10967 * Fix hang in EGD/PRNGD query when communication socket is closed
10968 prematurely by EGD/PRNGD.
10970 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10972 * Prompt for pass phrases when appropriate for PKCS12 input format.
10976 * Back-port of selected performance improvements from development
10977 branch, as well as improved support for PowerPC platforms.
10981 * Add lots of checks for memory allocation failure, error codes to indicate
10982 failure and freeing up memory if a failure occurs.
10984 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10986 * Add new -passin argument to dgst.
10990 * Perform some character comparisons of different types in X509_NAME_cmp:
10991 this is needed for some certificates that re-encode DNs into UTF8Strings
10992 (in violation of RFC3280) and can't or won't issue name rollover
10997 * Make an explicit check during certificate validation to see that
10998 the CA setting in each certificate on the chain is correct. As a
10999 side effect always do the following basic checks on extensions,
11000 not just when there's an associated purpose to the check:
11002 - if there is an unhandled critical extension (unless the user
11003 has chosen to ignore this fault)
11004 - if the path length has been exceeded (if one is set at all)
11005 - that certain extensions fit the associated purpose (if one has
11010 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11012 * Avoid a race condition when CRLs are checked in a multi threaded
11013 environment. This would happen due to the reordering of the revoked
11014 entries during signature checking and serial number lookup. Now the
11015 encoding is cached and the serial number sort performed under a lock.
11016 Add new STACK function sk_is_sorted().
11020 * Add Delta CRL to the extension code.
11024 * Various fixes to s3_pkt.c so alerts are sent properly.
11026 *David Holmes <d.holmes@f5.com>*
11028 * Reduce the chances of duplicate issuer name and serial numbers (in
11029 violation of RFC3280) using the OpenSSL certificate creation utilities.
11030 This is done by creating a random 64 bit value for the initial serial
11031 number when a serial number file is created or when a self signed
11032 certificate is created using 'openssl req -x509'. The initial serial
11033 number file is created using 'openssl x509 -next_serial' in CA.pl
11034 rather than being initialized to 1.
11038 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11040 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11041 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11043 *Joe Orton, Steve Henson*
11045 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11048 *Joe Orton, Steve Henson*
11050 * Make it possible to have multiple active certificates with the same
11051 subject in the CA index file. This is done only if the keyword
11052 'unique_subject' is set to 'no' in the main CA section (default
11053 if 'CA_default') of the configuration file. The value is saved
11054 with the database itself in a separate index attribute file,
11055 named like the index file with '.attr' appended to the name.
11059 * X509 verify fixes. Disable broken certificate workarounds when
11060 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11061 keyUsage extension present. Don't accept CRLs with unhandled critical
11062 extensions: since verify currently doesn't process CRL extensions this
11063 rejects a CRL with *any* critical extensions. Add new verify error codes
11068 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11069 A clarification of RFC2560 will require the use of OCTET STRINGs and
11070 some implementations cannot handle the current raw format. Since OpenSSL
11071 copies and compares OCSP nonces as opaque blobs without any attempt at
11072 parsing them this should not create any compatibility issues.
11076 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11077 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11078 this HMAC (and other) operations are several times slower than OpenSSL
11083 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11085 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11087 * Use the correct content when signing type "other".
11091 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11093 * Fix various bugs revealed by running the NISCC test suite:
11095 Stop out of bounds reads in the ASN1 code when presented with
11096 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11098 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11100 If verify callback ignores invalid public key errors don't try to check
11101 certificate signature with the NULL public key.
11105 * New -ignore_err option in ocsp application to stop the server
11106 exiting on the first error in a request.
11110 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11111 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11116 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11117 extra data after the compression methods not only for TLS 1.0
11118 but also for SSL 3.0 (as required by the specification).
11120 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11122 * Change X509_certificate_type() to mark the key as exported/exportable
11123 when it's 512 *bits* long, not 512 bytes.
11127 * Change AES_cbc_encrypt() so it outputs exact multiple of
11128 blocks during encryption.
11132 * Various fixes to base64 BIO and non blocking I/O. On write
11133 flushes were not handled properly if the BIO retried. On read
11134 data was not being buffered properly and had various logic bugs.
11135 This also affects blocking I/O when the data being decoded is a
11140 * Various S/MIME bugfixes and compatibility changes:
11141 output correct application/pkcs7 MIME type if
11142 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11143 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11144 of files as .eml work). Correctly handle very long lines in MIME
11149 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11151 * Countermeasure against the Klima-Pokorny-Rosa extension of
11152 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11153 a protocol version number mismatch like a decryption error
11154 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11158 * Turn on RSA blinding by default in the default implementation
11159 to avoid a timing attack. Applications that don't want it can call
11160 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11161 They would be ill-advised to do so in most cases.
11163 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11165 * Change RSA blinding code so that it works when the PRNG is not
11166 seeded (in this case, the secret RSA exponent is abused as
11167 an unpredictable seed -- if it is not unpredictable, there
11168 is no point in blinding anyway). Make RSA blinding thread-safe
11169 by remembering the creator's thread ID in rsa->blinding and
11170 having all other threads use local one-time blinding factors
11171 (this requires more computation than sharing rsa->blinding, but
11172 avoids excessive locking; and if an RSA object is not shared
11173 between threads, blinding will still be very fast).
11177 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11178 ENGINE as defaults for all supported algorithms irrespective of
11179 the 'flags' parameter. 'flags' is now honoured, so applications
11180 should make sure they are passing it correctly.
11184 * Target "mingw" now allows native Windows code to be generated in
11185 the Cygwin environment as well as with the MinGW compiler.
11189 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11191 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11192 via timing by performing a MAC computation even if incorrect
11193 block cipher padding has been found. This is a countermeasure
11194 against active attacks where the attacker has to distinguish
11195 between bad padding and a MAC verification error. ([CVE-2003-0078])
11197 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11198 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11199 Martin Vuagnoux (EPFL, Ilion)*
11201 * Make the no-err option work as intended. The intention with no-err
11202 is not to have the whole error stack handling routines removed from
11203 libcrypto, it's only intended to remove all the function name and
11204 reason texts, thereby removing some of the footprint that may not
11205 be interesting if those errors aren't displayed anyway.
11207 NOTE: it's still possible for any application or module to have its
11208 own set of error texts inserted. The routines are there, just not
11209 used by default when no-err is given.
11213 * Add support for FreeBSD on IA64.
11215 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11217 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11218 Kerberos function mit_des_cbc_cksum(). Before this change,
11219 the value returned by DES_cbc_cksum() was like the one from
11220 mit_des_cbc_cksum(), except the bytes were swapped.
11222 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11224 * Allow an application to disable the automatic SSL chain building.
11225 Before this a rather primitive chain build was always performed in
11226 ssl3_output_cert_chain(): an application had no way to send the
11227 correct chain if the automatic operation produced an incorrect result.
11229 Now the chain builder is disabled if either:
11231 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11233 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11235 The reasoning behind this is that an application would not want the
11236 auto chain building to take place if extra chain certificates are
11237 present and it might also want a means of sending no additional
11238 certificates (for example the chain has two certificates and the
11243 * Add the possibility to build without the ENGINE framework.
11245 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11247 * Under Win32 gmtime() can return NULL: check return value in
11248 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11252 * DSA routines: under certain error conditions uninitialized BN objects
11253 could be freed. Solution: make sure initialization is performed early
11254 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11255 Nils Larsch <nla@trustcenter.de> via PR#459)
11259 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11260 checked on reconnect on the client side, therefore session resumption
11261 could still fail with a "ssl session id is different" error. This
11262 behaviour is masked when SSL_OP_ALL is used due to
11263 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11264 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11265 followup to PR #377.
11269 * IA-32 assembler support enhancements: unified ELF targets, support
11270 for SCO/Caldera platforms, fix for Cygwin shared build.
11274 * Add support for FreeBSD on sparc64. As a consequence, support for
11275 FreeBSD on non-x86 processors is separate from x86 processors on
11276 the config script, much like the NetBSD support.
11278 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11280 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11282 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11285 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11286 code (06) was taken as the first octet of the session ID and the last
11287 octet was ignored consequently. As a result SSLv2 client side session
11288 caching could not have worked due to the session ID mismatch between
11290 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11295 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11296 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11301 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11302 seems that in spite of existing for more than a year, many application
11303 author have done nothing to provide the necessary callbacks, which
11304 means that this particular engine will not work properly anywhere.
11305 This is a very unfortunate situation which forces us, in the name
11306 of usability, to give the hw_ncipher.c a static lock, which is part
11308 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11309 appear in 0.9.8 or later. We EXPECT application authors to have
11310 dealt properly with this when 0.9.8 is released (unless we actually
11311 make such changes in the libcrypto locking code that changes will
11312 have to be made anyway).
11316 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11317 octets have been read, EOF or an error occurs. Without this change
11318 some truncated ASN1 structures will not produce an error.
11322 * Disable Heimdal support, since it hasn't been fully implemented.
11323 Still give the possibility to force the use of Heimdal, but with
11324 warnings and a request that patches get sent to openssl-dev.
11328 * Add the VC-CE target, introduce the WINCE sysname, and add
11329 INSTALL.WCE and appropriate conditionals to make it build.
11331 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11333 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11334 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11335 edit numbers of the version.
11337 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11339 * Introduce safe string copy and catenation functions
11340 (BUF_strlcpy() and BUF_strlcat()).
11342 *Ben Laurie (CHATS) and Richard Levitte*
11344 * Avoid using fixed-size buffers for one-line DNs.
11346 *Ben Laurie (CHATS)*
11348 * Add BUF_MEM_grow_clean() to avoid information leakage when
11349 resizing buffers containing secrets, and use where appropriate.
11351 *Ben Laurie (CHATS)*
11353 * Avoid using fixed size buffers for configuration file location.
11355 *Ben Laurie (CHATS)*
11357 * Avoid filename truncation for various CA files.
11359 *Ben Laurie (CHATS)*
11361 * Use sizeof in preference to magic numbers.
11363 *Ben Laurie (CHATS)*
11365 * Avoid filename truncation in cert requests.
11367 *Ben Laurie (CHATS)*
11369 * Add assertions to check for (supposedly impossible) buffer
11372 *Ben Laurie (CHATS)*
11374 * Don't cache truncated DNS entries in the local cache (this could
11375 potentially lead to a spoofing attack).
11377 *Ben Laurie (CHATS)*
11379 * Fix various buffers to be large enough for hex/decimal
11380 representations in a platform independent manner.
11382 *Ben Laurie (CHATS)*
11384 * Add CRYPTO_realloc_clean() to avoid information leakage when
11385 resizing buffers containing secrets, and use where appropriate.
11387 *Ben Laurie (CHATS)*
11389 * Add BIO_indent() to avoid much slightly worrying code to do
11392 *Ben Laurie (CHATS)*
11394 * Convert sprintf()/BIO_puts() to BIO_printf().
11396 *Ben Laurie (CHATS)*
11398 * buffer_gets() could terminate with the buffer only half
11401 *Ben Laurie (CHATS)*
11403 * Add assertions to prevent user-supplied crypto functions from
11404 overflowing internal buffers by having large block sizes, etc.
11406 *Ben Laurie (CHATS)*
11408 * New OPENSSL_assert() macro (similar to assert(), but enabled
11411 *Ben Laurie (CHATS)*
11413 * Eliminate unused copy of key in RC4.
11415 *Ben Laurie (CHATS)*
11417 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11419 *Ben Laurie (CHATS)*
11421 * Fix off-by-one error in EGD path.
11423 *Ben Laurie (CHATS)*
11425 * If RANDFILE path is too long, ignore instead of truncating.
11427 *Ben Laurie (CHATS)*
11429 * Eliminate unused and incorrectly sized X.509 structure
11432 *Ben Laurie (CHATS)*
11434 * Eliminate unused and dangerous function knumber().
11436 *Ben Laurie (CHATS)*
11438 * Eliminate unused and dangerous structure, KSSL_ERR.
11440 *Ben Laurie (CHATS)*
11442 * Protect against overlong session ID context length in an encoded
11443 session object. Since these are local, this does not appear to be
11446 *Ben Laurie (CHATS)*
11448 * Change from security patch (see 0.9.6e below) that did not affect
11449 the 0.9.6 release series:
11451 Remote buffer overflow in SSL3 protocol - an attacker could
11452 supply an oversized master key in Kerberos-enabled versions.
11455 *Ben Laurie (CHATS)*
11457 * Change the SSL kerb5 codes to match RFC 2712.
11461 * Make -nameopt work fully for req and add -reqopt switch.
11463 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11465 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11467 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11469 * Make sure tests can be performed even if the corresponding algorithms
11470 have been removed entirely. This was also the last step to make
11471 OpenSSL compilable with DJGPP under all reasonable conditions.
11473 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11475 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11476 to allow version independent disabling of normally unselected ciphers,
11477 which may be activated as a side-effect of selecting a single cipher.
11479 (E.g., cipher list string "RSA" enables ciphersuites that are left
11480 out of "ALL" because they do not provide symmetric encryption.
11481 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11483 *Lutz Jaenicke, Bodo Moeller*
11485 * Add appropriate support for separate platform-dependent build
11486 directories. The recommended way to make a platform-dependent
11487 build directory is the following (tested on Linux), maybe with
11490 # Place yourself outside of the OpenSSL source tree. In
11491 # this example, the environment variable OPENSSL_SOURCE
11492 # is assumed to contain the absolute OpenSSL source directory.
11493 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11494 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11495 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11496 mkdir -p `dirname $F`
11497 ln -s $OPENSSL_SOURCE/$F $F
11500 To be absolutely sure not to disturb the source tree, a "make clean"
11501 is a good thing. If it isn't successful, don't worry about it,
11502 it probably means the source directory is very clean.
11506 * Make sure any ENGINE control commands make local copies of string
11507 pointers passed to them whenever necessary. Otherwise it is possible
11508 the caller may have overwritten (or deallocated) the original string
11509 data when a later ENGINE operation tries to use the stored values.
11511 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11513 * Improve diagnostics in file reading and command-line digests.
11515 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11517 * Add AES modes CFB and OFB to the object database. Correct an
11518 error in AES-CFB decryption.
11522 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11523 allows existing EVP_CIPHER_CTX structures to be reused after
11524 calling `EVP_*Final()`. This behaviour is used by encryption
11525 BIOs and some applications. This has the side effect that
11526 applications must explicitly clean up cipher contexts with
11527 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11531 * Check the values of dna and dnb in bn_mul_recursive before calling
11532 bn_mul_comba (a non zero value means the a or b arrays do not contain
11533 n2 elements) and fallback to bn_mul_normal if either is not zero.
11537 * Fix escaping of non-ASCII characters when using the -subj option
11538 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11542 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11543 form for "surname", serialNumber has no short form.
11544 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11545 therefore remove "mail" short name for "internet 7".
11546 The OID for unique identifiers in X509 certificates is
11547 x500UniqueIdentifier, not uniqueIdentifier.
11548 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11552 * Add an "init" command to the ENGINE config module and auto initialize
11553 ENGINEs. Without any "init" command the ENGINE will be initialized
11554 after all ctrl commands have been executed on it. If init=1 the
11555 ENGINE is initialized at that point (ctrls before that point are run
11556 on the uninitialized ENGINE and after on the initialized one). If
11557 init=0 then the ENGINE will not be initialized at all.
11561 * Fix the 'app_verify_callback' interface so that the user-defined
11562 argument is actually passed to the callback: In the
11563 SSL_CTX_set_cert_verify_callback() prototype, the callback
11564 declaration has been changed from
11567 int (*cb)(X509_STORE_CTX *,void *);
11568 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11569 i=s->ctx->app_verify_callback(&ctx)
11570 has been changed into
11571 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11573 To update applications using SSL_CTX_set_cert_verify_callback(),
11574 a dummy argument can be added to their callback functions.
11576 *D. K. Smetters <smetters@parc.xerox.com>*
11578 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11580 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11582 * Add and OPENSSL_LOAD_CONF define which will cause
11583 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11584 This allows older applications to transparently support certain
11585 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11586 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11587 load the config file and OPENSSL_add_all_algorithms_conf() which will
11588 always load it have also been added.
11592 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11593 Adjust NIDs and EVP layer.
11595 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11597 * Config modules support in openssl utility.
11599 Most commands now load modules from the config file,
11600 though in a few (such as version) this isn't done
11601 because it couldn't be used for anything.
11603 In the case of ca and req the config file used is
11604 the same as the utility itself: that is the -config
11605 command line option can be used to specify an
11610 * Move default behaviour from OPENSSL_config(). If appname is NULL
11611 use "openssl_conf" if filename is NULL use default openssl config file.
11615 * Add an argument to OPENSSL_config() to allow the use of an alternative
11616 config section name. Add a new flag to tolerate a missing config file
11617 and move code to CONF_modules_load_file().
11621 * Support for crypto accelerator cards from Accelerated Encryption
11622 Processing, www.aep.ie. (Use engine 'aep')
11623 The support was copied from 0.9.6c [engine] and adapted/corrected
11624 to work with the new engine framework.
11626 *AEP Inc. and Richard Levitte*
11628 * Support for SureWare crypto accelerator cards from Baltimore
11629 Technologies. (Use engine 'sureware')
11630 The support was copied from 0.9.6c [engine] and adapted
11631 to work with the new engine framework.
11635 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11636 make the newer ENGINE framework commands for the CHIL engine work.
11638 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11640 * Make it possible to produce shared libraries on ReliantUNIX.
11642 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11644 * Add the configuration target debug-linux-ppro.
11645 Make 'openssl rsa' use the general key loading routines
11646 implemented in apps.c, and make those routines able to
11647 handle the key format FORMAT_NETSCAPE and the variant
11650 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11652 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11654 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11656 * Add -keyform to rsautl, and document -engine.
11658 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11660 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11661 BIO_R_NO_SUCH_FILE error code rather than the generic
11662 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11666 * Add new functions
11667 ERR_peek_last_error
11668 ERR_peek_last_error_line
11669 ERR_peek_last_error_line_data.
11670 These are similar to
11672 ERR_peek_error_line
11673 ERR_peek_error_line_data,
11674 but report on the latest error recorded rather than the first one
11675 still in the error queue.
11677 *Ben Laurie, Bodo Moeller*
11679 * default_algorithms option in ENGINE config module. This allows things
11681 default_algorithms = ALL
11682 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11686 * Preliminary ENGINE config module.
11690 * New experimental application configuration code.
11694 * Change the AES code to follow the same name structure as all other
11695 symmetric ciphers, and behave the same way. Move everything to
11696 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11698 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11700 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11702 *Ben Laurie and Theo de Raadt*
11704 * Add option to output public keys in req command.
11706 *Massimiliano Pala madwolf@openca.org*
11708 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11709 (up to about 10% better than before for P-192 and P-224).
11713 * New functions/macros
11715 SSL_CTX_set_msg_callback(ctx, cb)
11716 SSL_CTX_set_msg_callback_arg(ctx, arg)
11717 SSL_set_msg_callback(ssl, cb)
11718 SSL_set_msg_callback_arg(ssl, arg)
11720 to request calling a callback function
11722 void cb(int write_p, int version, int content_type,
11723 const void *buf, size_t len, SSL *ssl, void *arg)
11725 whenever a protocol message has been completely received
11726 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11727 protocol version according to which the SSL library interprets
11728 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11729 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11730 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11731 specification (change_cipher_spec(20), alert(21), handshake(22)).
11732 'buf' and 'len' point to the actual message, 'ssl' to the
11733 SSL object, and 'arg' is the application-defined value set by
11734 SSL[_CTX]_set_msg_callback_arg().
11736 'openssl s_client' and 'openssl s_server' have new '-msg' options
11737 to enable a callback that displays all protocol messages.
11741 * Change the shared library support so shared libraries are built as
11742 soon as the corresponding static library is finished, and thereby get
11743 openssl and the test programs linked against the shared library.
11744 This still only happens when the keyword "shard" has been given to
11745 the configuration scripts.
11747 NOTE: shared library support is still an experimental thing, and
11748 backward binary compatibility is still not guaranteed.
11750 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11752 * Add support for Subject Information Access extension.
11754 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11756 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11757 additional bytes when new memory had to be allocated, not just
11758 when reusing an existing buffer.
11762 * New command line and configuration option 'utf8' for the req command.
11763 This allows field values to be specified as UTF8 strings.
11767 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11768 runs for the former and machine-readable output for the latter.
11772 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11773 of the e-mail address in the DN (i.e., it will go into a certificate
11774 extension only). The new configuration file option 'email_in_dn = no'
11775 has the same effect.
11777 *Massimiliano Pala madwolf@openca.org*
11779 * Change all functions with names starting with `des_` to be starting
11780 with `DES_` instead. Add wrappers that are compatible with libdes,
11781 but are named `_ossl_old_des_*`. Finally, add macros that map the
11782 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11783 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11784 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11787 Since we provide two compatibility mappings, the user needs to
11788 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11789 compatibility is desired. The default (i.e., when that macro
11790 isn't defined) is OpenSSL 0.9.6c compatibility.
11792 There are also macros that enable and disable the support of old
11793 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11794 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11795 are defined, the default will apply: to support the old des routines.
11797 In either case, one must include openssl/des.h to get the correct
11798 definitions. Do not try to just include openssl/des_old.h, that
11801 NOTE: This is a major break of an old API into a new one. Software
11802 authors are encouraged to switch to the `DES_` style functions. Some
11803 time in the future, des_old.h and the libdes compatibility functions
11804 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11805 default), and then completely removed.
11809 * Test for certificates which contain unsupported critical extensions.
11810 If such a certificate is found during a verify operation it is
11811 rejected by default: this behaviour can be overridden by either
11812 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11813 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11814 X509_supported_extension() has also been added which returns 1 if a
11815 particular extension is supported.
11819 * Modify the behaviour of EVP cipher functions in similar way to digests
11820 to retain compatibility with existing code.
11824 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11825 compatibility with existing code. In particular the 'ctx' parameter does
11826 not have to be to be initialized before the call to EVP_DigestInit() and
11827 it is tidied up after a call to EVP_DigestFinal(). New function
11828 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11829 EVP_MD_CTX_copy() changed to not require the destination to be
11830 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11831 requires the destination to be valid.
11833 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11834 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11838 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11839 so that complete 'Handshake' protocol structures are kept in memory
11840 instead of overwriting 'msg_type' and 'length' with 'body' data.
11844 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11846 *Massimo Santin via Richard Levitte*
11848 * Major restructuring to the underlying ENGINE code. This includes
11849 reduction of linker bloat, separation of pure "ENGINE" manipulation
11850 (initialisation, etc) from functionality dealing with implementations
11851 of specific crypto interfaces. This change also introduces integrated
11852 support for symmetric ciphers and digest implementations - so ENGINEs
11853 can now accelerate these by providing EVP_CIPHER and EVP_MD
11854 implementations of their own. This is detailed in
11855 [crypto/engine/README.md](crypto/engine/README.md)
11856 as it couldn't be adequately described here. However, there are a few
11857 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11858 were changed in the original introduction of ENGINE code have now
11859 reverted back - the hooking from this code to ENGINE is now a good
11860 deal more passive and at run-time, operations deal directly with
11861 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11862 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11863 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11864 they were not being used by the framework as there is no concept of a
11865 BIGNUM_METHOD and they could not be generalised to the new
11866 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11867 ENGINE_cpy() has been removed as it cannot be consistently defined in
11872 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11876 * Change mkdef.pl to sort symbols that get the same entry number,
11877 and make sure the automatically generated functions `ERR_load_*`
11878 become part of libeay.num as well.
11882 * New function SSL_renegotiate_pending(). This returns true once
11883 renegotiation has been requested (either SSL_renegotiate() call
11884 or HelloRequest/ClientHello received from the peer) and becomes
11885 false once a handshake has been completed.
11886 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11887 sends a HelloRequest, but does not ensure that a handshake takes
11888 place. SSL_renegotiate_pending() is useful for checking if the
11889 client has followed the request.)
11893 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11894 By default, clients may request session resumption even during
11895 renegotiation (if session ID contexts permit); with this option,
11896 session resumption is possible only in the first handshake.
11898 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11899 more bits available for options that should not be part of
11900 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11904 * Add some demos for certificate and certificate request creation.
11908 * Make maximum certificate chain size accepted from the peer application
11909 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11910 "Douglas E. Engert" <deengert@anl.gov>.
11914 * Add support for shared libraries for Unixware-7
11915 (Boyd Lynn Gerber <gerberb@zenez.com>).
11919 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11920 be done prior to destruction. Use this to unload error strings from
11921 ENGINEs that load their own error strings. NB: This adds two new API
11922 functions to "get" and "set" this destroy handler in an ENGINE.
11926 * Alter all existing ENGINE implementations (except "openssl" and
11927 "openbsd") to dynamically instantiate their own error strings. This
11928 makes them more flexible to be built both as statically-linked ENGINEs
11929 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11930 Also, add stub code to each that makes building them as self-contained
11931 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11935 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11936 implementations into applications that are completely implemented in
11937 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11938 commands that can be used to configure what shared-library to load and
11939 to control aspects of the way it is handled. Also, made an update to
11940 the [README-Engine.md](README-Engine.md) file
11941 that brings its information up-to-date and
11942 provides some information and instructions on the "dynamic" ENGINE
11943 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11947 * Make it possible to unload ranges of ERR strings with a new
11948 "ERR_unload_strings" function.
11952 * Add a copy() function to EVP_MD.
11956 * Make EVP_MD routines take a context pointer instead of just the
11957 md_data void pointer.
11961 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11962 that the digest can only process a single chunk of data
11963 (typically because it is provided by a piece of
11964 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11965 is only going to provide a single chunk of data, and hence the
11966 framework needn't accumulate the data for oneshot drivers.
11970 * As with "ERR", make it possible to replace the underlying "ex_data"
11971 functions. This change also alters the storage and management of global
11972 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11973 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11974 index counters. The API functions that use this state have been changed
11975 to take a "class_index" rather than pointers to the class's local STACK
11976 and counter, and there is now an API function to dynamically create new
11977 classes. This centralisation allows us to (a) plug a lot of the
11978 thread-safety problems that existed, and (b) makes it possible to clean
11979 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11980 such data would previously have always leaked in application code and
11981 workarounds were in place to make the memory debugging turn a blind eye
11982 to it. Application code that doesn't use this new function will still
11983 leak as before, but their memory debugging output will announce it now
11984 rather than letting it slide.
11986 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11987 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11988 has a return value to indicate success or failure.
11992 * Make it possible to replace the underlying "ERR" functions such that the
11993 global state (2 LHASH tables and 2 locks) is only used by the "default"
11994 implementation. This change also adds two functions to "get" and "set"
11995 the implementation prior to it being automatically set the first time
11996 any other ERR function takes place. Ie. an application can call "get",
11997 pass the return value to a module it has just loaded, and that module
11998 can call its own "set" function using that value. This means the
11999 module's "ERR" operations will use (and modify) the error state in the
12000 application and not in its own statically linked copy of OpenSSL code.
12004 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12005 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12006 the operation, and provides a more encapsulated way for external code
12007 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12008 to use these functions rather than manually incrementing the counts.
12010 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12014 * Add EVP test program.
12018 * Add symmetric cipher support to ENGINE. Expect the API to change!
12022 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12023 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12024 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12025 These allow a CRL to be built without having to access X509_CRL fields
12026 directly. Modify 'ca' application to use new functions.
12030 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12031 bug workarounds. Rollback attack detection is a security feature.
12032 The problem will only arise on OpenSSL servers when TLSv1 is not
12033 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12034 Software authors not wanting to support TLSv1 will have special reasons
12035 for their choice and can explicitly enable this option.
12037 *Bodo Moeller, Lutz Jaenicke*
12039 * Rationalise EVP so it can be extended: don't include a union of
12040 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12041 (similar to those existing for EVP_CIPHER_CTX).
12046 EVP_MD_CTX_init(&md); /* new function call */
12047 EVP_DigestInit(&md, EVP_sha1());
12048 EVP_DigestUpdate(&md, in, len);
12049 EVP_DigestFinal(&md, out, NULL);
12050 EVP_MD_CTX_cleanup(&md); /* new function call */
12054 * Make DES key schedule conform to the usual scheme, as well as
12055 correcting its structure. This means that calls to DES functions
12056 now have to pass a pointer to a des_key_schedule instead of a
12057 plain des_key_schedule (which was actually always a pointer
12060 des_key_schedule ks;
12062 des_set_key_checked(..., &ks);
12063 des_ncbc_encrypt(..., &ks, ...);
12065 (Note that a later change renames 'des_...' into 'DES_...'.)
12069 * Initial reduction of linker bloat: the use of some functions, such as
12070 PEM causes large amounts of unused functions to be linked in due to
12071 poor organisation. For example pem_all.c contains every PEM function
12072 which has a knock on effect of linking in large amounts of (unused)
12073 ASN1 code. Grouping together similar functions and splitting unrelated
12074 functions prevents this.
12078 * Cleanup of EVP macros.
12082 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12083 correct `_ecb suffix`.
12087 * Add initial OCSP responder support to ocsp application. The
12088 revocation information is handled using the text based index
12089 use by the ca application. The responder can either handle
12090 requests generated internally, supplied in files (for example
12091 via a CGI script) or using an internal minimal server.
12095 * Add configuration choices to get zlib compression for TLS.
12099 * Changes to Kerberos SSL for RFC 2712 compliance:
12100 1. Implemented real KerberosWrapper, instead of just using
12101 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12102 2. Implemented optional authenticator field of KerberosWrapper.
12104 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12105 and authenticator structs; see crypto/krb5/.
12107 Generalized Kerberos calls to support multiple Kerberos libraries.
12108 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12109 via Richard Levitte*
12111 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12112 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12113 values for each of the key sizes rather than having just
12114 parameters (and 'speed' generating keys each time).
12118 * Speed up EVP routines.
12121 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12122 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12123 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12124 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12126 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12127 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12128 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12131 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12133 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12137 * Added the OS2-EMX target.
12139 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12141 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12142 to support NCONF routines in extension code. New function CONF_set_nconf()
12143 to allow functions which take an NCONF to also handle the old LHASH
12144 structure: this means that the old CONF compatible routines can be
12145 retained (in particular wrt extensions) without having to duplicate the
12146 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12150 * Enhance the general user interface with mechanisms for inner control
12151 and with possibilities to have yes/no kind of prompts.
12155 * Change all calls to low level digest routines in the library and
12156 applications to use EVP. Add missing calls to HMAC_cleanup() and
12157 don't assume HMAC_CTX can be copied using memcpy().
12159 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12161 * Add the possibility to control engines through control names but with
12162 arbitrary arguments instead of just a string.
12163 Change the key loaders to take a UI_METHOD instead of a callback
12164 function pointer. NOTE: this breaks binary compatibility with earlier
12165 versions of OpenSSL [engine].
12166 Adapt the nCipher code for these new conditions and add a card insertion
12171 * Enhance the general user interface with mechanisms to better support
12172 dialog box interfaces, application-defined prompts, the possibility
12173 to use defaults (for example default passwords from somewhere else)
12174 and interrupts/cancellations.
12178 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12179 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12183 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12184 tidy up some unnecessarily weird code in 'sk_new()').
12186 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12188 * Change the key loading routines for ENGINEs to use the same kind
12189 callback (pem_password_cb) as all other routines that need this
12194 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12195 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12196 than this minimum value is recommended.
12200 * New random seeder for OpenVMS, using the system process statistics
12201 that are easily reachable.
12205 * Windows apparently can't transparently handle global
12206 variables defined in DLLs. Initialisations such as:
12208 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12210 won't compile. This is used by the any applications that need to
12211 declare their own ASN1 modules. This was fixed by adding the option
12212 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12213 needed for static libraries under Win32.
12217 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12218 setting of purpose and trust fields. New X509_STORE trust and
12219 purpose functions and tidy up setting in other SSL functions.
12223 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12224 structure. These are inherited by X509_STORE_CTX when it is
12225 initialised. This allows various defaults to be set in the
12226 X509_STORE structure (such as flags for CRL checking and custom
12227 purpose or trust settings) for functions which only use X509_STORE_CTX
12228 internally such as S/MIME.
12230 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12231 trust settings if they are not set in X509_STORE. This allows X509_STORE
12232 purposes and trust (in S/MIME for example) to override any set by default.
12234 Add command line options for CRL checking to smime, s_client and s_server
12239 * Initial CRL based revocation checking. If the CRL checking flag(s)
12240 are set then the CRL is looked up in the X509_STORE structure and
12241 its validity and signature checked, then if the certificate is found
12242 in the CRL the verify fails with a revoked error.
12244 Various new CRL related callbacks added to X509_STORE_CTX structure.
12246 Command line options added to 'verify' application to support this.
12248 This needs some additional work, such as being able to handle multiple
12249 CRLs with different times, extension based lookup (rather than just
12250 by subject name) and ultimately more complete V2 CRL extension
12255 * Add a general user interface API (crypto/ui/). This is designed
12256 to replace things like des_read_password and friends (backward
12257 compatibility functions using this new API are provided).
12258 The purpose is to remove prompting functions from the DES code
12259 section as well as provide for prompting through dialog boxes in
12260 a window system and the like.
12264 * Add "ex_data" support to ENGINE so implementations can add state at a
12265 per-structure level rather than having to store it globally.
12269 * Make it possible for ENGINE structures to be copied when retrieved by
12270 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12271 This causes the "original" ENGINE structure to act like a template,
12272 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12273 operational state can be localised to each ENGINE structure, despite the
12274 fact they all share the same "methods". New ENGINE structures returned in
12275 this case have no functional references and the return value is the single
12276 structural reference. This matches the single structural reference returned
12277 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12282 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12283 needs to match any other type at all we need to manually clear the
12288 * Changes to the "openssl engine" utility to include;
12289 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12290 about an ENGINE's available control commands.
12291 - executing control commands from command line arguments using the
12292 '-pre' and '-post' switches. '-post' is only used if '-t' is
12293 specified and the ENGINE is successfully initialised. The syntax for
12294 the individual commands are colon-separated, for example;
12295 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12299 * New dynamic control command support for ENGINEs. ENGINEs can now
12300 declare their own commands (numbers), names (strings), descriptions,
12301 and input types for run-time discovery by calling applications. A
12302 subset of these commands are implicitly classed as "executable"
12303 depending on their input type, and only these can be invoked through
12304 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12305 can be based on user input, config files, etc). The distinction is
12306 that "executable" commands cannot return anything other than a boolean
12307 result and can only support numeric or string input, whereas some
12308 discoverable commands may only be for direct use through
12309 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12310 pointers, or other custom uses. The "executable" commands are to
12311 support parameterisations of ENGINE behaviour that can be
12312 unambiguously defined by ENGINEs and used consistently across any
12313 OpenSSL-based application. Commands have been added to all the
12314 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12315 control over shared-library paths without source code alterations.
12319 * Changed all ENGINE implementations to dynamically allocate their
12320 ENGINEs rather than declaring them statically. Apart from this being
12321 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12322 this also allows the implementations to compile without using the
12323 internal engine_int.h header.
12327 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12328 'const' value. Any code that should be able to modify a RAND_METHOD
12329 should already have non-const pointers to it (ie. they should only
12330 modify their own ones).
12334 * Made a variety of little tweaks to the ENGINE code.
12335 - "atalla" and "ubsec" string definitions were moved from header files
12336 to C code. "nuron" string definitions were placed in variables
12337 rather than hard-coded - allowing parameterisation of these values
12338 later on via ctrl() commands.
12339 - Removed unused "#if 0"'d code.
12340 - Fixed engine list iteration code so it uses ENGINE_free() to release
12341 structural references.
12342 - Constified the RAND_METHOD element of ENGINE structures.
12343 - Constified various get/set functions as appropriate and added
12344 missing functions (including a catch-all ENGINE_cpy that duplicates
12345 all ENGINE values onto a new ENGINE except reference counts/state).
12346 - Removed NULL parameter checks in get/set functions. Setting a method
12347 or function to NULL is a way of cancelling out a previously set
12348 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12349 and doesn't justify the extra error symbols and code.
12350 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12351 flags from engine_int.h to engine.h.
12352 - Changed prototypes for ENGINE handler functions (init(), finish(),
12353 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12357 * Implement binary inversion algorithm for BN_mod_inverse in addition
12358 to the algorithm using long division. The binary algorithm can be
12359 used only if the modulus is odd. On 32-bit systems, it is faster
12360 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12361 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12362 up to 450 bits. In 64-bit environments, the binary algorithm
12363 appears to be advantageous for much longer moduli; here we use it
12364 for moduli up to 2048 bits.
12368 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12369 could not support the combine flag in choice fields.
12373 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12374 extensions from a certificate request to the certificate.
12378 * Allow multiple 'certopt' and 'nameopt' options to be separated
12379 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12380 file: this allows the display of the certificate about to be
12381 signed to be customised, to allow certain fields to be included
12382 or excluded and extension details. The old system didn't display
12383 multicharacter strings properly, omitted fields not in the policy
12384 and couldn't display additional details such as extensions.
12388 * Function EC_POINTs_mul for multiple scalar multiplication
12389 of an arbitrary number of elliptic curve points
12390 \sum scalars[i]*points[i],
12391 optionally including the generator defined for the EC_GROUP:
12392 scalar*generator + \sum scalars[i]*points[i].
12394 EC_POINT_mul is a simple wrapper function for the typical case
12395 that the point list has just one item (besides the optional
12400 * First EC_METHODs for curves over GF(p):
12402 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12403 operations and provides various method functions that can also
12404 operate with faster implementations of modular arithmetic.
12406 EC_GFp_mont_method() reuses most functions that are part of
12407 EC_GFp_simple_method, but uses Montgomery arithmetic.
12409 *Bodo Moeller; point addition and point doubling
12410 implementation directly derived from source code provided by
12411 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12413 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12414 crypto/ec/ec_lib.c):
12416 Curves are EC_GROUP objects (with an optional group generator)
12417 based on EC_METHODs that are built into the library.
12419 Points are EC_POINT objects based on EC_GROUP objects.
12421 Most of the framework would be able to handle curves over arbitrary
12422 finite fields, but as there are no obvious types for fields other
12423 than GF(p), some functions are limited to that for now.
12427 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12428 that the file contains a complete HTTP response.
12432 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12433 change the def and num file printf format specifier from "%-40sXXX"
12434 to "%-39s XXX". The latter will always guarantee a space after the
12435 field while the former will cause them to run together if the field
12436 is 40 of more characters long.
12440 * Constify the cipher and digest 'method' functions and structures
12441 and modify related functions to take constant EVP_MD and EVP_CIPHER
12446 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12447 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12451 * Modify `EVP_Digest*()` routines so they now return values. Although the
12452 internal software routines can never fail additional hardware versions
12457 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12459 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12460 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12463 ERR_R_NESTED_ASN1_ERROR
12465 ERR_R_MISSING_ASN1_EOS
12466 were 4 .. 9, conflicting with
12467 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12469 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12470 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12472 Add new error code 'ERR_R_INTERNAL_ERROR'.
12476 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12481 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12482 sets the subject name for a new request or supersedes the
12483 subject name in a given request. Formats that can be parsed are
12484 'CN=Some Name, OU=myOU, C=IT'
12486 'CN=Some Name/OU=myOU/C=IT'.
12488 Add options '-batch' and '-verbose' to 'openssl req'.
12490 *Massimiliano Pala <madwolf@hackmasters.net>*
12492 * Introduce the possibility to access global variables through
12493 functions on platform were that's the best way to handle exporting
12494 global variables in shared libraries. To enable this functionality,
12495 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12496 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12497 is normally done by Configure or something similar).
12499 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12500 in the source file (foo.c) like this:
12502 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12503 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12505 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12506 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12508 OPENSSL_DECLARE_GLOBAL(int,foo);
12509 #define foo OPENSSL_GLOBAL_REF(foo)
12510 OPENSSL_DECLARE_GLOBAL(double,bar);
12511 #define bar OPENSSL_GLOBAL_REF(bar)
12513 The #defines are very important, and therefore so is including the
12514 header file everywhere where the defined globals are used.
12516 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12517 of ASN.1 items, but that structure is a bit different.
12519 The largest change is in util/mkdef.pl which has been enhanced with
12520 better and easier to understand logic to choose which symbols should
12521 go into the Windows .def files as well as a number of fixes and code
12522 cleanup (among others, algorithm keywords are now sorted
12523 lexicographically to avoid constant rewrites).
12527 * In BN_div() keep a copy of the sign of 'num' before writing the
12528 result to 'rm' because if rm==num the value will be overwritten
12529 and produce the wrong result if 'num' is negative: this caused
12530 problems with BN_mod() and BN_nnmod().
12534 * Function OCSP_request_verify(). This checks the signature on an
12535 OCSP request and verifies the signer certificate. The signer
12536 certificate is just checked for a generic purpose and OCSP request
12541 * Add OCSP_check_validity() function to check the validity of OCSP
12542 responses. OCSP responses are prepared in real time and may only
12543 be a few seconds old. Simply checking that the current time lies
12544 between thisUpdate and nextUpdate max reject otherwise valid responses
12545 caused by either OCSP responder or client clock inaccuracy. Instead
12546 we allow thisUpdate and nextUpdate to fall within a certain period of
12547 the current time. The age of the response can also optionally be
12548 checked. Two new options -validity_period and -status_age added to
12553 * If signature or public key algorithm is unrecognized print out its
12554 OID rather that just UNKNOWN.
12558 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12559 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12560 ID to be generated from the issuer certificate alone which can then be
12561 passed to OCSP_id_issuer_cmp().
12565 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12566 ASN1 modules to export functions returning ASN1_ITEM pointers
12567 instead of the ASN1_ITEM structures themselves. This adds several
12568 new macros which allow the underlying ASN1 function/structure to
12569 be accessed transparently. As a result code should not use ASN1_ITEM
12570 references directly (such as &X509_it) but instead use the relevant
12571 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12572 use of the new ASN1 code on platforms where exporting structures
12573 is problematical (for example in shared libraries) but exporting
12574 functions returning pointers to structures is not.
12578 * Add support for overriding the generation of SSL/TLS session IDs.
12579 These callbacks can be registered either in an SSL_CTX or per SSL.
12580 The purpose of this is to allow applications to control, if they wish,
12581 the arbitrary values chosen for use as session IDs, particularly as it
12582 can be useful for session caching in multiple-server environments. A
12583 command-line switch for testing this (and any client code that wishes
12584 to use such a feature) has been added to "s_server".
12586 *Geoff Thorpe, Lutz Jaenicke*
12588 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12589 of the form `#if defined(...) || defined(...) || ...` and
12590 `#if !defined(...) && !defined(...) && ...`. This also avoids
12591 the growing number of special cases it was previously handling.
12595 * Make all configuration macros available for application by making
12596 sure they are available in opensslconf.h, by giving them names starting
12597 with `OPENSSL_` to avoid conflicts with other packages and by making
12598 sure e_os2.h will cover all platform-specific cases together with
12600 Additionally, it is now possible to define configuration/platform-
12601 specific names (called "system identities"). In the C code, these
12602 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12603 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12604 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12609 * New option -set_serial to 'req' and 'x509' this allows the serial
12610 number to use to be specified on the command line. Previously self
12611 signed certificates were hard coded with serial number 0 and the
12612 CA options of 'x509' had to use a serial number in a file which was
12617 * New options to 'ca' utility to support V2 CRL entry extensions.
12618 Currently CRL reason, invalidity date and hold instruction are
12619 supported. Add new CRL extensions to V3 code and some new objects.
12623 * New function EVP_CIPHER_CTX_set_padding() this is used to
12624 disable standard block padding (aka PKCS#5 padding) in the EVP
12625 API, which was previously mandatory. This means that the data is
12626 not padded in any way and so the total length much be a multiple
12627 of the block size, otherwise an error occurs.
12631 * Initial (incomplete) OCSP SSL support.
12635 * New function OCSP_parse_url(). This splits up a URL into its host,
12636 port and path components: primarily to parse OCSP URLs. New -url
12637 option to ocsp utility.
12641 * New nonce behavior. The return value of OCSP_check_nonce() now
12642 reflects the various checks performed. Applications can decide
12643 whether to tolerate certain situations such as an absent nonce
12644 in a response when one was present in a request: the ocsp application
12645 just prints out a warning. New function OCSP_add1_basic_nonce()
12646 this is to allow responders to include a nonce in a response even if
12647 the request is nonce-less.
12651 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12652 skipped when using openssl x509 multiple times on a single input file,
12653 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12657 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12658 set string type: to handle setting ASN1_TIME structures. Fix ca
12659 utility to correctly initialize revocation date of CRLs.
12663 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12664 the clients preferred ciphersuites and rather use its own preferences.
12665 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12666 Internet Explorer by ensuring unchanged hash method during stepup.
12667 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12671 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12672 to aes and add a new 'exist' option to print out symbols that don't
12677 * Additional options to ocsp utility to allow flags to be set and
12678 additional certificates supplied.
12682 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12683 OCSP client a number of certificate to only verify the response
12688 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12689 handle the new API. Currently only ECB, CBC modes supported. Add new
12692 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12693 Encryption Standard (AES) Ciphersuites for Transport Layer
12694 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12695 not enabled by default and were not part of the "ALL" ciphersuite
12696 alias because they were not yet official; they could be
12697 explicitly requested by specifying the "AESdraft" ciphersuite
12698 group alias. In the final release of OpenSSL 0.9.7, the group
12699 alias is called "AES" and is part of "ALL".)
12701 *Ben Laurie, Steve Henson, Bodo Moeller*
12703 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12704 request to response.
12708 * Functions for OCSP responders. OCSP_request_onereq_count(),
12709 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12710 extract information from a certificate request. OCSP_response_create()
12711 creates a response and optionally adds a basic response structure.
12712 OCSP_basic_add1_status() adds a complete single response to a basic
12713 response and returns the OCSP_SINGLERESP structure just added (to allow
12714 extensions to be included for example). OCSP_basic_add1_cert() adds a
12715 certificate to a basic response and OCSP_basic_sign() signs a basic
12716 response with various flags. New helper functions ASN1_TIME_check()
12717 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12718 (converts ASN1_TIME to GeneralizedTime).
12722 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12723 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12724 structure from a certificate. X509_pubkey_digest() digests the public_key
12725 contents: this is used in various key identifiers.
12729 * Make sk_sort() tolerate a NULL argument.
12731 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12733 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12734 passed by the function are trusted implicitly. If any of them signed the
12735 response then it is assumed to be valid and is not verified.
12739 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12740 to data. This was previously part of the PKCS7 ASN1 code. This
12741 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12742 *Steve Henson, reported by Kenneth R. Robinette
12743 <support@securenetterm.com>*
12745 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12746 routines: without these tracing memory leaks is very painful.
12747 Fix leaks in PKCS12 and PKCS7 routines.
12751 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12752 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12753 effectively meant GeneralizedTime would never be used. Now it
12754 is initialised to -1 but X509_time_adj() now has to check the value
12755 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12756 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12757 *Steve Henson, reported by Kenneth R. Robinette
12758 <support@securenetterm.com>*
12760 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12761 result in a zero length in the ASN1_INTEGER structure which was
12762 not consistent with the structure when d2i_ASN1_INTEGER() was used
12763 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12764 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12765 where it did not print out a minus for negative ASN1_INTEGER.
12769 * Add summary printout to ocsp utility. The various functions which
12770 convert status values to strings have been renamed to:
12771 OCSP_response_status_str(), OCSP_cert_status_str() and
12772 OCSP_crl_reason_str() and are no longer static. New options
12773 to verify nonce values and to disable verification. OCSP response
12774 printout format cleaned up.
12778 * Add additional OCSP certificate checks. These are those specified
12779 in RFC2560. This consists of two separate checks: the CA of the
12780 certificate being checked must either be the OCSP signer certificate
12781 or the issuer of the OCSP signer certificate. In the latter case the
12782 OCSP signer certificate must contain the OCSP signing extended key
12783 usage. This check is performed by attempting to match the OCSP
12784 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12785 in the OCSP_CERTID structures of the response.
12789 * Initial OCSP certificate verification added to OCSP_basic_verify()
12790 and related routines. This uses the standard OpenSSL certificate
12791 verify routines to perform initial checks (just CA validity) and
12792 to obtain the certificate chain. Then additional checks will be
12793 performed on the chain. Currently the root CA is checked to see
12794 if it is explicitly trusted for OCSP signing. This is used to set
12795 a root CA as a global signing root: that is any certificate that
12796 chains to that CA is an acceptable OCSP signing certificate.
12800 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12801 extensions from a separate configuration file.
12802 As when reading extensions from the main configuration file,
12803 the '-extensions ...' option may be used for specifying the
12806 *Massimiliano Pala <madwolf@comune.modena.it>*
12808 * New OCSP utility. Allows OCSP requests to be generated or
12809 read. The request can be sent to a responder and the output
12810 parsed, outputted or printed in text form. Not complete yet:
12811 still needs to check the OCSP response validity.
12815 * New subcommands for 'openssl ca':
12816 `openssl ca -status <serial>` prints the status of the cert with
12817 the given serial number (according to the index file).
12818 `openssl ca -updatedb` updates the expiry status of certificates
12821 *Massimiliano Pala <madwolf@comune.modena.it>*
12823 * New '-newreq-nodes' command option to CA.pl. This is like
12824 '-newreq', but calls 'openssl req' with the '-nodes' option
12825 so that the resulting key is not encrypted.
12827 *Damien Miller <djm@mindrot.org>*
12829 * New configuration for the GNU Hurd.
12831 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12833 * Initial code to implement OCSP basic response verify. This
12834 is currently incomplete. Currently just finds the signer's
12835 certificate and verifies the signature on the response.
12839 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12840 value of OPENSSLDIR. This is available via the new '-d' option
12841 to 'openssl version', and is also included in 'openssl version -a'.
12845 * Allowing defining memory allocation callbacks that will be given
12846 file name and line number information in additional arguments
12847 (a `const char*` and an int). The basic functionality remains, as
12848 well as the original possibility to just replace malloc(),
12849 realloc() and free() by functions that do not know about these
12850 additional arguments. To register and find out the current
12851 settings for extended allocation functions, the following
12852 functions are provided:
12854 CRYPTO_set_mem_ex_functions
12855 CRYPTO_set_locked_mem_ex_functions
12856 CRYPTO_get_mem_ex_functions
12857 CRYPTO_get_locked_mem_ex_functions
12859 These work the same way as CRYPTO_set_mem_functions and friends.
12860 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12861 extended allocation function is enabled.
12862 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12863 a conventional allocation function is enabled.
12865 *Richard Levitte, Bodo Moeller*
12867 * Finish off removing the remaining LHASH function pointer casts.
12868 There should no longer be any prototype-casting required when using
12869 the LHASH abstraction, and any casts that remain are "bugs". See
12870 the callback types and macros at the head of lhash.h for details
12871 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12875 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12876 If /dev/[u]random devices are not available or do not return enough
12877 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12879 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12880 /etc/entropy will be queried once each in this sequence, querying stops
12881 when enough entropy was collected without querying more sockets.
12885 * Change the Unix RAND_poll() variant to be able to poll several
12886 random devices, as specified by DEVRANDOM, until a sufficient amount
12887 of data has been collected. We spend at most 10 ms on each file
12888 (select timeout) and read in non-blocking mode. DEVRANDOM now
12889 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12890 (previously it was just the string "/dev/urandom"), so on typical
12891 platforms the 10 ms delay will never occur.
12892 Also separate out the Unix variant to its own file, rand_unix.c.
12893 For VMS, there's a currently-empty rand_vms.c.
12897 * Move OCSP client related routines to ocsp_cl.c. These
12898 provide utility functions which an application needing
12899 to issue a request to an OCSP responder and analyse the
12900 response will typically need: as opposed to those which an
12901 OCSP responder itself would need which will be added later.
12903 OCSP_request_sign() signs an OCSP request with an API similar
12904 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12905 response. OCSP_response_get1_basic() extracts basic response
12906 from response. OCSP_resp_find_status(): finds and extracts status
12907 information from an OCSP_CERTID structure (which will be created
12908 when the request structure is built). These are built from lower
12909 level functions which work on OCSP_SINGLERESP structures but
12910 won't normally be used unless the application wishes to examine
12911 extensions in the OCSP response for example.
12913 Replace nonce routines with a pair of functions.
12914 OCSP_request_add1_nonce() adds a nonce value and optionally
12915 generates a random value. OCSP_check_nonce() checks the
12916 validity of the nonce in an OCSP response.
12920 * Change function OCSP_request_add() to OCSP_request_add0_id().
12921 This doesn't copy the supplied OCSP_CERTID and avoids the
12922 need to free up the newly created id. Change return type
12923 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12924 This can then be used to add extensions to the request.
12925 Deleted OCSP_request_new(), since most of its functionality
12926 is now in OCSP_REQUEST_new() (and the case insensitive name
12927 clash) apart from the ability to set the request name which
12928 will be added elsewhere.
12932 * Update OCSP API. Remove obsolete extensions argument from
12933 various functions. Extensions are now handled using the new
12934 OCSP extension code. New simple OCSP HTTP function which
12935 can be used to send requests and parse the response.
12939 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12940 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12941 uses the special reorder version of SET OF to sort the attributes
12942 and reorder them to match the encoded order. This resolves a long
12943 standing problem: a verify on a PKCS7 structure just after signing
12944 it used to fail because the attribute order did not match the
12945 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12946 it uses the received order. This is necessary to tolerate some broken
12947 software that does not order SET OF. This is handled by encoding
12948 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12949 to produce the required SET OF.
12953 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12954 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12955 files to get correct declarations of the ASN.1 item variables.
12959 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12960 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12961 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12962 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12963 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12964 ASN1_ITEM and no wrapper functions.
12968 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12969 replace the old function pointer based I/O routines. Change most of
12970 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12974 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12975 lines, recognize more "algorithms" that can be deselected, and make
12976 it complain about algorithm deselection that isn't recognised.
12980 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12981 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12982 to use new functions. Add NO_ASN1_OLD which can be set to remove
12983 some old style ASN1 functions: this can be used to determine if old
12984 code will still work when these eventually go away.
12988 * New extension functions for OCSP structures, these follow the
12989 same conventions as certificates and CRLs.
12993 * New function X509V3_add1_i2d(). This automatically encodes and
12994 adds an extension. Its behaviour can be customised with various
12995 flags to append, replace or delete. Various wrappers added for
12996 certificates and CRLs.
13000 * Fix to avoid calling the underlying ASN1 print routine when
13001 an extension cannot be parsed. Correct a typo in the
13002 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13006 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13007 entries for variables.
13011 * Add functionality to apps/openssl.c for detecting locking
13012 problems: As the program is single-threaded, all we have
13013 to do is register a locking callback using an array for
13014 storing which locks are currently held by the program.
13018 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13019 SSL_get_ex_data_X509_STORE_idx(), which is used in
13020 ssl_verify_cert_chain() and thus can be called at any time
13021 during TLS/SSL handshakes so that thread-safety is essential.
13022 Unfortunately, the ex_data design is not at all suited
13023 for multi-threaded use, so it probably should be abolished.
13027 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13029 *Broadcom, tweaked and integrated by Geoff Thorpe*
13031 * Move common extension printing code to new function
13032 X509V3_print_extensions(). Reorganise OCSP print routines and
13033 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13037 * New function X509_signature_print() to remove duplication in some
13042 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13043 set (this was treated exactly the same as SET OF previously). This
13044 is used to reorder the STACK representing the structure to match the
13045 encoding. This will be used to get round a problem where a PKCS7
13046 structure which was signed could not be verified because the STACK
13047 order did not reflect the encoded order.
13051 * Reimplement the OCSP ASN1 module using the new code.
13055 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13056 for its ASN1 operations. The old style function pointers still exist
13057 for now but they will eventually go away.
13061 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13062 completely replaces the old ASN1 functionality with a table driven
13063 encoder and decoder which interprets an ASN1_ITEM structure describing
13064 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13065 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13066 has also been converted to the new form.
13070 * Change BN_mod_exp_recp so that negative moduli are tolerated
13071 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13072 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13073 for negative moduli.
13077 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13078 of not touching the result's sign bit.
13082 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13087 * Changed the LHASH code to use prototypes for callbacks, and created
13088 macros to declare and implement thin (optionally static) functions
13089 that provide type-safety and avoid function pointer casting for the
13090 type-specific callbacks.
13094 * Added Kerberos Cipher Suites to be used with TLS, as written in
13096 *Veers Staats <staatsvr@asc.hpc.mil>,
13097 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13099 * Reformat the FAQ so the different questions and answers can be divided
13100 in sections depending on the subject.
13104 * Have the zlib compression code load ZLIB.DLL dynamically under
13109 * New function BN_mod_sqrt for computing square roots modulo a prime
13110 (using the probabilistic Tonelli-Shanks algorithm unless
13111 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13112 be handled deterministically).
13114 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13116 * Make BN_mod_inverse faster by explicitly handling small quotients
13117 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13118 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13122 * New function BN_kronecker.
13126 * Fix BN_gcd so that it works on negative inputs; the result is
13127 positive unless both parameters are zero.
13128 Previously something reasonably close to an infinite loop was
13129 possible because numbers could be growing instead of shrinking
13130 in the implementation of Euclid's algorithm.
13134 * Fix BN_is_word() and BN_is_one() macros to take into account the
13135 sign of the number in question.
13137 Fix BN_is_word(a,w) to work correctly for w == 0.
13139 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13140 because its test if the absolute value of 'a' equals 'w'.
13141 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13142 it exists mostly for use in the implementations of BN_is_zero(),
13143 BN_is_one(), and BN_is_word().
13147 * New function BN_swap.
13151 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13152 the exponentiation functions are more likely to produce reasonable
13153 results on negative inputs.
13157 * Change BN_mod_mul so that the result is always non-negative.
13158 Previously, it could be negative if one of the factors was negative;
13159 I don't think anyone really wanted that behaviour.
13163 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13164 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13165 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13166 and add new functions:
13175 BN_mod_lshift1_quick
13177 BN_mod_lshift_quick
13179 These functions always generate non-negative results.
13181 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13182 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13184 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13185 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13186 be reduced modulo `m`.
13188 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13191 The following entry accidentally appeared in the CHANGES file
13192 distributed with OpenSSL 0.9.7. The modifications described in
13193 it do *not* apply to OpenSSL 0.9.7.
13195 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13196 was actually never needed) and in BN_mul(). The removal in BN_mul()
13197 required a small change in bn_mul_part_recursive() and the addition
13198 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13199 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13200 bn_sub_words() and bn_add_words() except they take arrays with
13206 * In 'openssl passwd', verify passwords read from the terminal
13207 unless the '-salt' option is used (which usually means that
13208 verification would just waste user's time since the resulting
13209 hash is going to be compared with some given password hash)
13210 or the new '-noverify' option is used.
13212 This is an incompatible change, but it does not affect
13213 non-interactive use of 'openssl passwd' (passwords on the command
13214 line, '-stdin' option, '-in ...' option) and thus should not
13215 cause any problems.
13219 * Remove all references to RSAref, since there's no more need for it.
13223 * Make DSO load along a path given through an environment variable
13224 (SHLIB_PATH) with shl_load().
13228 * Constify the ENGINE code as a result of BIGNUM constification.
13229 Also constify the RSA code and most things related to it. In a
13230 few places, most notable in the depth of the ASN.1 code, ugly
13231 casts back to non-const were required (to be solved at a later
13236 * Make it so the openssl application has all engines loaded by default.
13240 * Constify the BIGNUM routines a little more.
13244 * Add the following functions:
13246 ENGINE_load_cswift()
13248 ENGINE_load_atalla()
13249 ENGINE_load_nuron()
13250 ENGINE_load_builtin_engines()
13252 That way, an application can itself choose if external engines that
13253 are built-in in OpenSSL shall ever be used or not. The benefit is
13254 that applications won't have to be linked with libdl or other dso
13255 libraries unless it's really needed.
13257 Changed 'openssl engine' to load all engines on demand.
13258 Changed the engine header files to avoid the duplication of some
13259 declarations (they differed!).
13263 * 'openssl engine' can now list capabilities.
13267 * Better error reporting in 'openssl engine'.
13271 * Never call load_dh_param(NULL) in s_server.
13275 * Add engine application. It can currently list engines by name and
13276 identity, and test if they are actually available.
13280 * Improve RPM specification file by forcing symbolic linking and making
13281 sure the installed documentation is also owned by root.root.
13283 *Damien Miller <djm@mindrot.org>*
13285 * Give the OpenSSL applications more possibilities to make use of
13286 keys (public as well as private) handled by engines.
13290 * Add OCSP code that comes from CertCo.
13294 * Add VMS support for the Rijndael code.
13298 * Added untested support for Nuron crypto accelerator.
13302 * Add support for external cryptographic devices. This code was
13303 previously distributed separately as the "engine" branch.
13305 *Geoff Thorpe, Richard Levitte*
13307 * Rework the filename-translation in the DSO code. It is now possible to
13308 have far greater control over how a "name" is turned into a filename
13309 depending on the operating environment and any oddities about the
13310 different shared library filenames on each system.
13314 * Support threads on FreeBSD-elf in Configure.
13318 * Fix for SHA1 assembly problem with MASM: it produces
13319 warnings about corrupt line number information when assembling
13320 with debugging information. This is caused by the overlapping
13323 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13326 NCONF_get_number() has no error checking at all. As a replacement,
13327 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13328 promoted strongly. The old NCONF_get_number is kept around for
13329 binary backward compatibility.
13330 Make it possible for methods to load from something other than a BIO,
13331 by providing a function pointer that is given a name instead of a BIO.
13332 For example, this could be used to load configuration data from an
13337 * Fix for non blocking accept BIOs. Added new I/O special reason
13338 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13339 with non blocking I/O was not possible because no retry code was
13340 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13345 * Added the beginnings of Rijndael support.
13349 * Fix for bug in DirectoryString mask setting. Add support for
13350 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13351 to allow certificate printing to more controllable, additional
13352 'certopt' option to 'x509' to allow new printing options to be
13357 * Clean old EAY MD5 hack from e_os.h.
13361 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13363 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13364 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13366 *Joe Orton, Steve Henson*
13368 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13370 * Fix additional bug revealed by the NISCC test suite:
13372 Stop bug triggering large recursion when presented with
13373 certain ASN.1 tags ([CVE-2003-0851])
13377 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13379 * Fix various bugs revealed by running the NISCC test suite:
13381 Stop out of bounds reads in the ASN1 code when presented with
13382 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13384 If verify callback ignores invalid public key errors don't try to check
13385 certificate signature with the NULL public key.
13389 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13390 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13395 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13396 extra data after the compression methods not only for TLS 1.0
13397 but also for SSL 3.0 (as required by the specification).
13399 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13401 * Change X509_certificate_type() to mark the key as exported/exportable
13402 when it's 512 *bits* long, not 512 bytes.
13406 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13408 * Countermeasure against the Klima-Pokorny-Rosa extension of
13409 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13410 a protocol version number mismatch like a decryption error
13411 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13415 * Turn on RSA blinding by default in the default implementation
13416 to avoid a timing attack. Applications that don't want it can call
13417 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13418 They would be ill-advised to do so in most cases.
13420 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13422 * Change RSA blinding code so that it works when the PRNG is not
13423 seeded (in this case, the secret RSA exponent is abused as
13424 an unpredictable seed -- if it is not unpredictable, there
13425 is no point in blinding anyway). Make RSA blinding thread-safe
13426 by remembering the creator's thread ID in rsa->blinding and
13427 having all other threads use local one-time blinding factors
13428 (this requires more computation than sharing rsa->blinding, but
13429 avoids excessive locking; and if an RSA object is not shared
13430 between threads, blinding will still be very fast).
13434 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13436 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13437 via timing by performing a MAC computation even if incorrect
13438 block cipher padding has been found. This is a countermeasure
13439 against active attacks where the attacker has to distinguish
13440 between bad padding and a MAC verification error. ([CVE-2003-0078])
13442 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13443 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13444 Martin Vuagnoux (EPFL, Ilion)*
13446 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13448 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13449 memory from its contents. This is done with a counter that will
13450 place alternating values in each byte. This can be used to solve
13451 two issues: 1) the removal of calls to memset() by highly optimizing
13452 compilers, and 2) cleansing with other values than 0, since those can
13453 be read through on certain media, for example a swap space on disk.
13457 * Bugfix: client side session caching did not work with external caching,
13458 because the session->cipher setting was not restored when reloading
13459 from the external cache. This problem was masked, when
13460 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13461 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13465 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13466 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13468 *Zeev Lieber <zeev-l@yahoo.com>*
13470 * Undo an undocumented change introduced in 0.9.6e which caused
13471 repeated calls to OpenSSL_add_all_ciphers() and
13472 OpenSSL_add_all_digests() to be ignored, even after calling
13477 * Change the default configuration reader to deal with last line not
13478 being properly terminated.
13482 * Change X509_NAME_cmp() so it applies the special rules on handling
13483 DN values that are of type PrintableString, as well as RDNs of type
13484 emailAddress where the value has the type ia5String.
13486 *stefank@valicert.com via Richard Levitte*
13488 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13489 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13490 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13491 the bitwise-OR of the two for use by the majority of applications
13492 wanting this behaviour, and update the docs. The documented
13493 behaviour and actual behaviour were inconsistent and had been
13494 changing anyway, so this is more a bug-fix than a behavioural
13497 *Geoff Thorpe, diagnosed by Nadav Har'El*
13499 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13500 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13504 * Fix initialization code race conditions in
13505 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13506 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13507 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13508 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13509 ssl2_get_cipher_by_char(),
13510 ssl3_get_cipher_by_char().
13512 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13514 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13515 the cached sessions are flushed, as the remove_cb() might use ex_data
13516 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13517 (see [openssl.org #212]).
13519 *Geoff Thorpe, Lutz Jaenicke*
13521 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13522 length, instead of the encoding length to d2i_ASN1_OBJECT.
13526 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13528 * [In 0.9.6g-engine release:]
13529 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13531 *Lynn Gazis <lgazis@rainbow.com>*
13533 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13535 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13536 and get fix the header length calculation.
13537 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13538 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13540 * Use proper error handling instead of 'assertions' in buffer
13541 overflow checks added in 0.9.6e. This prevents DoS (the
13542 assertions could call abort()).
13544 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13546 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13548 * Add various sanity checks to asn1_get_length() to reject
13549 the ASN1 length bytes if they exceed sizeof(long), will appear
13550 negative or the content length exceeds the length of the
13553 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13555 * Fix cipher selection routines: ciphers without encryption had no flags
13556 for the cipher strength set and where therefore not handled correctly
13557 by the selection routines (PR #130).
13561 * Fix EVP_dsa_sha macro.
13566 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13567 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13568 that was added in OpenSSL 0.9.6d.
13570 As the countermeasure turned out to be incompatible with some
13571 broken SSL implementations, the new option is part of SSL_OP_ALL.
13572 SSL_OP_ALL is usually employed when compatibility with weird SSL
13573 implementations is desired (e.g. '-bugs' option to 's_client' and
13574 's_server'), so the new option is automatically set in many
13579 * Changes in security patch:
13581 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13582 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13583 Air Force Materiel Command, USAF, under agreement number
13586 * Add various sanity checks to asn1_get_length() to reject
13587 the ASN1 length bytes if they exceed sizeof(long), will appear
13588 negative or the content length exceeds the length of the
13589 supplied buffer. ([CVE-2002-0659])
13591 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13593 * Assertions for various potential buffer overflows, not known to
13594 happen in practice.
13596 *Ben Laurie (CHATS)*
13598 * Various temporary buffers to hold ASCII versions of integers were
13599 too small for 64 bit platforms. ([CVE-2002-0655])
13600 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13602 * Remote buffer overflow in SSL3 protocol - an attacker could
13603 supply an oversized session ID to a client. ([CVE-2002-0656])
13605 *Ben Laurie (CHATS)*
13607 * Remote buffer overflow in SSL2 protocol - an attacker could
13608 supply an oversized client master key. ([CVE-2002-0656])
13610 *Ben Laurie (CHATS)*
13612 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13614 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13615 encoded as NULL) with id-dsa-with-sha1.
13617 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13619 * Check various `X509_...()` return values in apps/req.c.
13621 *Nils Larsch <nla@trustcenter.de>*
13623 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13624 an end-of-file condition would erroneously be flagged, when the CRLF
13625 was just at the end of a processed block. The bug was discovered when
13626 processing data through a buffering memory BIO handing the data to a
13627 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13628 <ptsekov@syntrex.com> and Nedelcho Stanev.
13632 * Implement a countermeasure against a vulnerability recently found
13633 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13634 before application data chunks to avoid the use of known IVs
13635 with data potentially chosen by the attacker.
13639 * Fix length checks in ssl3_get_client_hello().
13643 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13644 to prevent ssl3_read_internal() from incorrectly assuming that
13645 ssl3_read_bytes() found application data while handshake
13646 processing was enabled when in fact s->s3->in_read_app_data was
13647 merely automatically cleared during the initial handshake.
13649 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13651 * Fix object definitions for Private and Enterprise: they were not
13652 recognized in their shortname (=lowercase) representation. Extend
13653 obj_dat.pl to issue an error when using undefined keywords instead
13654 of silently ignoring the problem (Svenning Sorensen
13655 <sss@sss.dnsalias.net>).
13659 * Fix DH_generate_parameters() so that it works for 'non-standard'
13660 generators, i.e. generators other than 2 and 5. (Previously, the
13661 code did not properly initialise the 'add' and 'rem' values to
13662 BN_generate_prime().)
13664 In the new general case, we do not insist that 'generator' is
13665 actually a primitive root: This requirement is rather pointless;
13666 a generator of the order-q subgroup is just as good, if not
13671 * Map new X509 verification errors to alerts. Discovered and submitted by
13672 Tom Wu <tom@arcot.com>.
13676 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13677 returning non-zero before the data has been completely received
13678 when using non-blocking I/O.
13680 *Bodo Moeller; problem pointed out by John Hughes*
13682 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13684 *Ben Laurie, Lutz Jaenicke*
13686 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13687 Yoram Zahavi <YoramZ@gilian.com>).
13691 * Add information about CygWin 1.3 and on, and preserve proper
13692 configuration for the versions before that.
13694 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13696 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13697 check whether we deal with a copy of a session and do not delete from
13698 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13699 <izhar@checkpoint.com>.
13703 * Do not store session data into the internal session cache, if it
13704 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13705 flag is set). Proposed by Aslam <aslam@funk.com>.
13709 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13714 * [In 0.9.6d-engine release:]
13715 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13717 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13719 * Add the configuration target linux-s390x.
13721 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13723 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13724 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13725 variable as an indication that a ClientHello message has been
13726 received. As the flag value will be lost between multiple
13727 invocations of ssl3_accept when using non-blocking I/O, the
13728 function may not be aware that a handshake has actually taken
13729 place, thus preventing a new session from being added to the
13732 To avoid this problem, we now set s->new_session to 2 instead of
13733 using a local variable.
13735 *Lutz Jaenicke, Bodo Moeller*
13737 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13738 if the SSL_R_LENGTH_MISMATCH error is detected.
13740 *Geoff Thorpe, Bodo Moeller*
13742 * New 'shared_ldflag' column in Configure platform table.
13746 * Fix EVP_CIPHER_mode macro.
13748 *"Dan S. Camper" <dan@bti.net>*
13750 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13751 type, we must throw them away by setting rr->length to 0.
13753 *D P Chang <dpc@qualys.com>*
13755 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13757 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13758 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13759 worked incorrectly for those cases where range = `10..._2` and
13760 `3*range` is two bits longer than range.)
13764 * Only add signing time to PKCS7 structures if it is not already
13769 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13770 OBJ_ld_ce should be OBJ_id_ce.
13771 Also some ip-pda OIDs in crypto/objects/objects.txt were
13772 incorrect (cf. RFC 3039).
13774 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13776 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13777 returns early because it has nothing to do.
13779 *Andy Schneider <andy.schneider@bjss.co.uk>*
13781 * [In 0.9.6c-engine release:]
13782 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13784 *Andy Schneider <andy.schneider@bjss.co.uk>*
13786 * [In 0.9.6c-engine release:]
13787 Add support for Cryptographic Appliance's keyserver technology.
13788 (Use engine 'keyclient')
13790 *Cryptographic Appliances and Geoff Thorpe*
13792 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13793 is called via tools/c89.sh because arguments have to be
13794 rearranged (all '-L' options must appear before the first object
13797 *Richard Shapiro <rshapiro@abinitio.com>*
13799 * [In 0.9.6c-engine release:]
13800 Add support for Broadcom crypto accelerator cards, backported
13803 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13805 * [In 0.9.6c-engine release:]
13806 Add support for SureWare crypto accelerator cards from
13807 Baltimore Technologies. (Use engine 'sureware')
13809 *Baltimore Technologies and Mark Cox*
13811 * [In 0.9.6c-engine release:]
13812 Add support for crypto accelerator cards from Accelerated
13813 Encryption Processing, www.aep.ie. (Use engine 'aep')
13815 *AEP Inc. and Mark Cox*
13817 * Add a configuration entry for gcc on UnixWare.
13819 *Gary Benson <gbenson@redhat.com>*
13821 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13822 messages are stored in a single piece (fixed-length part and
13823 variable-length part combined) and fix various bugs found on the way.
13827 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13828 instead. BIO_gethostbyname() does not know what timeouts are
13829 appropriate, so entries would stay in cache even when they have
13831 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13833 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13834 faced with a pathologically small ClientHello fragment that does
13835 not contain client_version: Instead of aborting with an error,
13836 simply choose the highest available protocol version (i.e.,
13837 TLS 1.0 unless it is disabled). In practice, ClientHello
13838 messages are never sent like this, but this change gives us
13839 strictly correct behaviour at least for TLS.
13843 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13844 never resets s->method to s->ctx->method when called from within
13845 one of the SSL handshake functions.
13847 *Bodo Moeller; problem pointed out by Niko Baric*
13849 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13850 (sent using the client's version number) if client_version is
13851 smaller than the protocol version in use. Also change
13852 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13853 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13854 the client will at least see that alert.
13858 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13863 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13864 client receives HelloRequest while in a handshake.
13866 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13868 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13869 should end in 'break', not 'goto end' which circumvents various
13870 cleanups done in state SSL_ST_OK. But session related stuff
13871 must be disabled for SSL_ST_OK in the case that we just sent a
13874 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13875 before just sending a HelloRequest.
13877 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13879 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13880 reveal whether illegal block cipher padding was found or a MAC
13881 verification error occurred. (Neither SSLerr() codes nor alerts
13882 are directly visible to potential attackers, but the information
13883 may leak via logfiles.)
13885 Similar changes are not required for the SSL 2.0 implementation
13886 because the number of padding bytes is sent in clear for SSL 2.0,
13887 and the extra bytes are just ignored. However ssl/s2_pkt.c
13888 failed to verify that the purported number of padding bytes is in
13893 * Add OpenUNIX-8 support including shared libraries
13894 (Boyd Lynn Gerber <gerberb@zenez.com>).
13898 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13899 'wristwatch attack' using huge encoding parameters (cf.
13900 James H. Manger's CRYPTO 2001 paper). Note that the
13901 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13902 encoding parameters and hence was not vulnerable.
13906 * BN_sqr() bug fix.
13908 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13910 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13911 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13912 followed by modular reduction.
13914 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13916 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13917 equivalent based on BN_pseudo_rand() instead of BN_rand().
13921 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13922 This function was broken, as the check for a new client hello message
13923 to handle SGC did not allow these large messages.
13924 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13928 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13932 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13933 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13937 * Rework the configuration and shared library support for Tru64 Unix.
13938 The configuration part makes use of modern compiler features and
13939 still retains old compiler behavior for those that run older versions
13940 of the OS. The shared library support part includes a variant that
13941 uses the RPATH feature, and is available through the special
13942 configuration target "alpha-cc-rpath", which will never be selected
13945 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13947 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13948 with the same message size as in ssl3_get_certificate_request().
13949 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13950 messages might inadvertently be reject as too long.
13952 *Petr Lampa <lampa@fee.vutbr.cz>*
13954 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13958 * Modified SSL library such that the verify_callback that has been set
13959 specifically for an SSL object with SSL_set_verify() is actually being
13960 used. Before the change, a verify_callback set with this function was
13961 ignored and the verify_callback() set in the SSL_CTX at the time of
13962 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13963 to allow the necessary settings.
13967 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13968 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13969 done automatically (in contradiction to the requirements of the C
13970 standard). This made problems when used from OpenSSH.
13974 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13975 dh->length and always used
13977 BN_rand_range(priv_key, dh->p).
13979 BN_rand_range() is not necessary for Diffie-Hellman, and this
13980 specific range makes Diffie-Hellman unnecessarily inefficient if
13981 dh->length (recommended exponent length) is much smaller than the
13982 length of dh->p. We could use BN_rand_range() if the order of
13983 the subgroup was stored in the DH structure, but we only have
13988 BN_rand(priv_key, l, ...)
13990 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13997 RSA_eay_public_encrypt
13998 RSA_eay_private_decrypt
13999 RSA_eay_private_encrypt (signing)
14000 RSA_eay_public_decrypt (signature verification)
14002 (default implementations for RSA_public_encrypt,
14003 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14004 always reject numbers >= n.
14008 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14009 to synchronize access to 'locking_thread'. This is necessary on
14010 systems where access to 'locking_thread' (an 'unsigned long'
14011 variable) is not atomic.
14015 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14016 *before* setting the 'crypto_lock_rand' flag. The previous code had
14017 a race condition if 0 is a valid thread ID.
14019 *Travis Vitek <vitek@roguewave.com>*
14021 * Add support for shared libraries under Irix.
14023 *Albert Chin-A-Young <china@thewrittenword.com>*
14025 * Add configuration option to build on Linux on both big-endian and
14026 little-endian MIPS.
14028 *Ralf Baechle <ralf@uni-koblenz.de>*
14030 * Add the possibility to create shared libraries on HP-UX.
14034 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14036 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14037 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14038 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14039 PRNG state recovery was possible based on the output of
14040 one PRNG request appropriately sized to gain knowledge on
14041 'md' followed by enough consecutive 1-byte PRNG requests
14042 to traverse all of 'state'.
14044 1. When updating 'md_local' (the current thread's copy of 'md')
14045 during PRNG output generation, hash all of the previous
14046 'md_local' value, not just the half used for PRNG output.
14048 2. Make the number of bytes from 'state' included into the hash
14049 independent from the number of PRNG bytes requested.
14051 The first measure alone would be sufficient to avoid
14052 Markku-Juhani's attack. (Actually it had never occurred
14053 to me that the half of 'md_local' used for chaining was the
14054 half from which PRNG output bytes were taken -- I had always
14055 assumed that the secret half would be used.) The second
14056 measure makes sure that additional data from 'state' is never
14057 mixed into 'md_local' in small portions; this heuristically
14058 further strengthens the PRNG.
14062 * Fix crypto/bn/asm/mips3.s.
14066 * When only the key is given to "enc", the IV is undefined. Print out
14067 an error message in this case.
14071 * Handle special case when X509_NAME is empty in X509 printing routines.
14075 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14076 positive and less than q.
14080 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14081 used: it isn't thread safe and the add_lock_callback should handle
14084 *Paul Rose <Paul.Rose@bridge.com>*
14086 * Verify that incoming data obeys the block size in
14087 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14093 *Ulf Möller, Bodo Möller*
14095 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14096 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14097 when fixing the server behaviour for backwards-compatible 'client
14098 hello' messages. (Note that the attack is impractical against
14099 SSL 3.0 and TLS 1.0 anyway because length and version checking
14100 means that the probability of guessing a valid ciphertext is
14101 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14104 Before 0.9.5, the countermeasure (hide the error by generating a
14105 random 'decryption result') did not work properly because
14106 ERR_clear_error() was missing, meaning that SSL_get_error() would
14107 detect the supposedly ignored error.
14109 Both problems are now fixed.
14113 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14114 (previously it was 1024).
14118 * Fix for compatibility mode trust settings: ignore trust settings
14119 unless some valid trust or reject settings are present.
14123 * Fix for blowfish EVP: its a variable length cipher.
14127 * Fix various bugs related to DSA S/MIME verification. Handle missing
14128 parameters in DSA public key structures and return an error in the
14129 DSA routines if parameters are absent.
14133 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14134 in the current directory if neither $RANDFILE nor $HOME was set.
14135 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14136 caused some confusion to Windows users who haven't defined $HOME.
14137 Thus RAND_file_name() is changed again: e_os.h can define a
14138 DEFAULT_HOME, which will be used if $HOME is not set.
14139 For Windows, we use "C:"; on other platforms, we still require
14140 environment variables.
14142 * Move 'if (!initialized) RAND_poll()' into regions protected by
14143 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14144 having multiple threads call RAND_poll() concurrently.
14148 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14149 combination of a flag and a thread ID variable.
14150 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14151 flag), *other* threads can enter ssleay_add_bytes without obeying
14152 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14153 that they do not hold after the first thread unsets add_do_not_lock).
14157 * Change bctest again: '-x' expressions are not available in all
14158 versions of 'test'.
14162 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14164 * Fix a couple of memory leaks in PKCS7_dataDecode()
14166 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14168 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14169 the default extension for executables, if any. Also, make the perl
14170 scripts that use symlink() to test if it really exists and use "cp"
14171 if it doesn't. All this made OpenSSL compilable and installable in
14176 * Fix for asn1_GetSequence() for indefinite length constructed data.
14177 If SEQUENCE is length is indefinite just set c->slen to the total
14178 amount of data available.
14180 *Steve Henson, reported by shige@FreeBSD.org*
14182 *This change does not apply to 0.9.7.*
14184 * Change bctest to avoid here-documents inside command substitution
14185 (workaround for FreeBSD /bin/sh bug).
14186 For compatibility with Ultrix, avoid shell functions (introduced
14187 in the bctest version that searches along $PATH).
14191 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14192 with des_encrypt() defined on some operating systems, like Solaris
14197 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14198 On the Importance of Eliminating Errors in Cryptographic
14199 Computations, J. Cryptology 14 (2001) 2, 101-119,
14200 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14204 * MIPS assembler BIGNUM division bug fix.
14208 * Disabled incorrect Alpha assembler code.
14212 * Fix PKCS#7 decode routines so they correctly update the length
14213 after reading an EOC for the EXPLICIT tag.
14217 *This change does not apply to 0.9.7.*
14219 * Fix bug in PKCS#12 key generation routines. This was triggered
14220 if a 3DES key was generated with a 0 initial byte. Include
14221 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14222 (but broken) behaviour.
14226 * Enhance bctest to search for a working bc along $PATH and print
14229 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14231 * Fix memory leaks in err.c: free err_data string if necessary;
14232 don't write to the wrong index in ERR_set_error_data.
14236 * Implement ssl23_peek (analogous to ssl23_read), which previously
14241 * Replace rdtsc with `_emit` statements for VC++ version 5.
14243 *Jeremy Cooper <jeremy@baymoo.org>*
14245 * Make it possible to reuse SSLv2 sessions.
14249 * In copy_email() check for >= 0 as a return value for
14250 X509_NAME_get_index_by_NID() since 0 is a valid index.
14252 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14254 * Avoid coredump with unsupported or invalid public keys by checking if
14255 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14256 PKCS7_verify() fails with non detached data.
14260 * Don't use getenv in library functions when run as setuid/setgid.
14261 New function OPENSSL_issetugid().
14265 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14266 due to incorrect handling of multi-threading:
14268 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14270 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14272 3. Count how many times MemCheck_off() has been called so that
14273 nested use can be treated correctly. This also avoids
14274 inband-signalling in the previous code (which relied on the
14275 assumption that thread ID 0 is impossible).
14279 * Add "-rand" option also to s_client and s_server.
14283 * Fix CPU detection on Irix 6.x.
14284 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14285 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14287 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14292 *This change does not apply to 0.9.7.*
14294 * Use the cached encoding of an X509_NAME structure rather than
14295 copying it. This is apparently the reason for the libsafe "errors"
14296 but the code is actually correct.
14300 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14301 Bleichenbacher's DSA attack.
14302 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14303 to be set and top=0 forces the highest bit to be set; top=-1 is new
14304 and leaves the highest bit random.
14306 *Ulf Moeller, Bodo Moeller*
14308 * In the `NCONF_...`-based implementations for `CONF_...` queries
14309 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14310 a temporary CONF structure with the data component set to NULL
14311 (which gives segmentation faults in lh_retrieve).
14312 Instead, use NULL for the CONF pointer in CONF_get_string and
14313 CONF_get_number (which may use environment variables) and directly
14314 return NULL from CONF_get_section.
14318 * Fix potential buffer overrun for EBCDIC.
14322 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14323 keyUsage if basicConstraints absent for a CA.
14327 * Make SMIME_write_PKCS7() write mail header values with a format that
14328 is more generally accepted (no spaces before the semicolon), since
14329 some programs can't parse those values properly otherwise. Also make
14330 sure BIO's that break lines after each write do not create invalid
14335 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14336 macros previously used would not encode an empty SEQUENCE OF
14337 and break the signature.
14341 *This change does not apply to 0.9.7.*
14343 * Zero the premaster secret after deriving the master secret in
14348 * Add some EVP_add_digest_alias registrations (as found in
14349 OpenSSL_add_all_digests()) to SSL_library_init()
14350 aka OpenSSL_add_ssl_algorithms(). This provides improved
14351 compatibility with peers using X.509 certificates
14352 with unconventional AlgorithmIdentifier OIDs.
14356 * Fix for Irix with NO_ASM.
14358 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14360 * ./config script fixes.
14362 *Ulf Moeller, Richard Levitte*
14364 * Fix 'openssl passwd -1'.
14368 * Change PKCS12_key_gen_asc() so it can cope with non null
14369 terminated strings whose length is passed in the passlen
14370 parameter, for example from PEM callbacks. This was done
14371 by adding an extra length parameter to asc2uni().
14373 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14375 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14376 call failed, free the DSA structure.
14380 * Fix to uni2asc() to cope with zero length Unicode strings.
14381 These are present in some PKCS#12 files.
14385 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14386 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14387 when writing a 32767 byte record.
14389 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14391 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14392 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14394 (RSA objects have a reference count access to which is protected
14395 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14396 so they are meant to be shared between threads.)
14397 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14398 "Reddie, Steven" <Steven.Reddie@ca.com>*
14400 * Fix a deadlock in CRYPTO_mem_leaks().
14404 * Use better test patterns in bntest.
14408 * rand_win.c fix for Borland C.
14412 * BN_rshift bugfix for n == 0.
14416 * Add a 'bctest' script that checks for some known 'bc' bugs
14417 so that 'make test' does not abort just because 'bc' is broken.
14421 * Store verify_result within SSL_SESSION also for client side to
14422 avoid potential security hole. (Re-used sessions on the client side
14423 always resulted in verify_result==X509_V_OK, not using the original
14424 result of the server certificate verification.)
14428 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14429 SSL3_RT_APPLICATION_DATA, return 0.
14430 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14435 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14436 releases, have been re-implemented by renaming the previous
14437 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14438 and ssl3_read_internal, respectively, and adding 'peek' parameters
14439 to them. The new ssl[23]_{read,peek} functions are calls to
14440 ssl[23]_read_internal with the 'peek' flag set appropriately.
14441 A 'peek' parameter has also been added to ssl3_read_bytes, which
14442 does the actual work for ssl3_read_internal.
14446 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14447 the method-specific "init()" handler. Also clean up ex_data after
14448 calling the method-specific "finish()" handler. Previously, this was
14449 happening the other way round.
14453 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14454 The previous value, 12, was not always sufficient for BN_mod_exp().
14458 * Make sure that shared libraries get the internal name engine with
14459 the full version number and not just 0. This should mark the
14460 shared libraries as not backward compatible. Of course, this should
14461 be changed again when we can guarantee backward binary compatibility.
14465 * Fix typo in get_cert_by_subject() in by_dir.c
14467 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14469 * Rework the system to generate shared libraries:
14471 - Make note of the expected extension for the shared libraries and
14472 if there is a need for symbolic links from for example libcrypto.so.0
14473 to libcrypto.so.0.9.7. There is extended info in Configure for
14476 - Make as few rebuilds of the shared libraries as possible.
14478 - Still avoid linking the OpenSSL programs with the shared libraries.
14480 - When installing, install the shared libraries separately from the
14485 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14487 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14488 and not in SSL_clear because the latter is also used by the
14489 accept/connect functions; previously, the settings made by
14490 SSL_set_read_ahead would be lost during the handshake.
14492 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14494 * Correct util/mkdef.pl to be selective about disabled algorithms.
14495 Previously, it would create entries for disabled algorithms no
14500 * Added several new manual pages for SSL_* function.
14504 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14506 * In ssl23_get_client_hello, generate an error message when faced
14507 with an initial SSL 3.0/TLS record that is too small to contain the
14508 first two bytes of the ClientHello message, i.e. client_version.
14509 (Note that this is a pathologic case that probably has never happened
14510 in real life.) The previous approach was to use the version number
14511 from the record header as a substitute; but our protocol choice
14512 should not depend on that one because it is not authenticated
14513 by the Finished messages.
14517 * More robust randomness gathering functions for Windows.
14519 *Jeffrey Altman <jaltman@columbia.edu>*
14521 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14522 not set then we don't setup the error code for issuer check errors
14523 to avoid possibly overwriting other errors which the callback does
14524 handle. If an application does set the flag then we assume it knows
14525 what it is doing and can handle the new informational codes
14530 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14531 a general "ANY" type, as such it should be able to decode anything
14532 including tagged types. However it didn't check the class so it would
14533 wrongly interpret tagged types in the same way as their universal
14534 counterpart and unknown types were just rejected. Changed so that the
14535 tagged and unknown types are handled in the same way as a SEQUENCE:
14536 that is the encoding is stored intact. There is also a new type
14537 "V_ASN1_OTHER" which is used when the class is not universal, in this
14538 case we have no idea what the actual type is so we just lump them all
14543 * On VMS, stdout may very well lead to a file that is written to
14544 in a record-oriented fashion. That means that every write() will
14545 write a separate record, which will be read separately by the
14546 programs trying to read from it. This can be very confusing.
14548 The solution is to put a BIO filter in the way that will buffer
14549 text until a linefeed is reached, and then write everything a
14550 line at a time, so every record written will be an actual line,
14551 not chunks of lines and not (usually doesn't happen, but I've
14552 seen it once) several lines in one record. BIO_f_linebuffer() is
14555 Currently, it's a VMS-only method, because that's where it has
14556 been tested well enough.
14560 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14561 it can return incorrect results.
14562 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14563 but it was in 0.9.6-beta[12].)
14567 * Disable the check for content being present when verifying detached
14568 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14569 include zero length content when signing messages.
14573 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14574 BIO_ctrl (for BIO pairs).
14578 * Add DSO method for VMS.
14582 * Bug fix: Montgomery multiplication could produce results with the
14587 * Add RPM specification openssl.spec and modify it to build three
14588 packages. The default package contains applications, application
14589 documentation and run-time libraries. The devel package contains
14590 include files, static libraries and function documentation. The
14591 doc package contains the contents of the doc directory. The original
14592 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14596 * Add a large number of documentation files for many SSL routines.
14598 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14600 * Add a configuration entry for Sony News 4.
14602 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14604 * Don't set the two most significant bits to one when generating a
14605 random number < q in the DSA library.
14609 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14610 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14611 the underlying transport is blocking) if a handshake took place.
14612 (The default behaviour is needed by applications such as s_client
14613 and s_server that use select() to determine when to use SSL_read;
14614 but for applications that know in advance when to expect data, it
14615 just makes things more complicated.)
14619 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14624 * Add a few more EBCDIC conditionals that make `req` and `x509`
14625 work better on such systems.
14627 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14629 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14630 Update PKCS12_parse() so it copies the friendlyName and the
14631 keyid to the certificates aux info.
14635 * Fix bug in PKCS7_verify() which caused an infinite loop
14636 if there was more than one signature.
14638 *Sven Uszpelkat <su@celocom.de>*
14640 * Major change in util/mkdef.pl to include extra information
14641 about each symbol, as well as presenting variables as well
14642 as functions. This change means that there's n more need
14643 to rebuild the .num files when some algorithms are excluded.
14647 * Allow the verify time to be set by an application,
14648 rather than always using the current time.
14652 * Phase 2 verify code reorganisation. The certificate
14653 verify code now looks up an issuer certificate by a
14654 number of criteria: subject name, authority key id
14655 and key usage. It also verifies self signed certificates
14656 by the same criteria. The main comparison function is
14657 X509_check_issued() which performs these checks.
14659 Lot of changes were necessary in order to support this
14660 without completely rewriting the lookup code.
14662 Authority and subject key identifier are now cached.
14664 The LHASH 'certs' is X509_STORE has now been replaced
14665 by a STACK_OF(X509_OBJECT). This is mainly because an
14666 LHASH can't store or retrieve multiple objects with
14667 the same hash value.
14669 As a result various functions (which were all internal
14670 use only) have changed to handle the new X509_STORE
14671 structure. This will break anything that messed round
14672 with X509_STORE internally.
14674 The functions X509_STORE_add_cert() now checks for an
14675 exact match, rather than just subject name.
14677 The X509_STORE API doesn't directly support the retrieval
14678 of multiple certificates matching a given criteria, however
14679 this can be worked round by performing a lookup first
14680 (which will fill the cache with candidate certificates)
14681 and then examining the cache for matches. This is probably
14682 the best we can do without throwing out X509_LOOKUP
14683 entirely (maybe later...).
14685 The X509_VERIFY_CTX structure has been enhanced considerably.
14687 All certificate lookup operations now go via a get_issuer()
14688 callback. Although this currently uses an X509_STORE it
14689 can be replaced by custom lookups. This is a simple way
14690 to bypass the X509_STORE hackery necessary to make this
14691 work and makes it possible to use more efficient techniques
14692 in future. A very simple version which uses a simple
14693 STACK for its trusted certificate store is also provided
14694 using X509_STORE_CTX_trusted_stack().
14696 The verify_cb() and verify() callbacks now have equivalents
14697 in the X509_STORE_CTX structure.
14699 X509_STORE_CTX also has a 'flags' field which can be used
14700 to customise the verify behaviour.
14704 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14705 excludes S/MIME capabilities.
14709 * When a certificate request is read in keep a copy of the
14710 original encoding of the signed data and use it when outputting
14711 again. Signatures then use the original encoding rather than
14712 a decoded, encoded version which may cause problems if the
14713 request is improperly encoded.
14717 * For consistency with other BIO_puts implementations, call
14718 buffer_write(b, ...) directly in buffer_puts instead of calling
14721 In BIO_puts, increment b->num_write as in BIO_write.
14723 *Peter.Sylvester@EdelWeb.fr*
14725 * Fix BN_mul_word for the case where the word is 0. (We have to use
14726 BN_zero, we may not return a BIGNUM with an array consisting of
14727 words set to zero.)
14731 * Avoid calling abort() from within the library when problems are
14732 detected, except if preprocessor symbols have been defined
14733 (such as REF_CHECK, BN_DEBUG etc.).
14737 * New openssl application 'rsautl'. This utility can be
14738 used for low level RSA operations. DER public key
14739 BIO/fp routines also added.
14743 * New Configure entry and patches for compiling on QNX 4.
14745 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14747 * A demo state-machine implementation was sponsored by
14748 Nuron (<http://www.nuron.com/>) and is now available in
14749 demos/state_machine.
14753 * New options added to the 'dgst' utility for signature
14754 generation and verification.
14758 * Unrecognized PKCS#7 content types are now handled via a
14759 catch all ASN1_TYPE structure. This allows unsupported
14760 types to be stored as a "blob" and an application can
14761 encode and decode it manually.
14765 * Fix various signed/unsigned issues to make a_strex.c
14766 compile under VC++.
14768 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14770 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14771 length if passed a buffer. ASN1_INTEGER_to_BN failed
14772 if passed a NULL BN and its argument was negative.
14774 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14776 * Modification to PKCS#7 encoding routines to output definite
14777 length encoding. Since currently the whole structures are in
14778 memory there's not real point in using indefinite length
14779 constructed encoding. However if OpenSSL is compiled with
14780 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14784 * Added BIO_vprintf() and BIO_vsnprintf().
14788 * Added more prefixes to parse for in the strings written
14789 through a logging bio, to cover all the levels that are available
14790 through syslog. The prefixes are now:
14792 PANIC, EMERG, EMR => LOG_EMERG
14793 ALERT, ALR => LOG_ALERT
14794 CRIT, CRI => LOG_CRIT
14795 ERROR, ERR => LOG_ERR
14796 WARNING, WARN, WAR => LOG_WARNING
14797 NOTICE, NOTE, NOT => LOG_NOTICE
14798 INFO, INF => LOG_INFO
14799 DEBUG, DBG => LOG_DEBUG
14801 and as before, if none of those prefixes are present at the
14802 beginning of the string, LOG_ERR is chosen.
14804 On Win32, the `LOG_*` levels are mapped according to this:
14806 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14807 LOG_WARNING => EVENTLOG_WARNING_TYPE
14808 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14812 * Made it possible to reconfigure with just the configuration
14813 argument "reconf" or "reconfigure". The command line arguments
14814 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14815 and are retrieved from there when reconfiguring.
14821 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14823 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14827 * The obj_dat.pl script was messing up the sorting of object
14828 names. The reason was that it compared the quoted version
14829 of strings as a result "OCSP" > "OCSP Signing" because
14830 " > SPACE. Changed script to store unquoted versions of
14831 names and add quotes on output. It was also omitting some
14832 names from the lookup table if they were given a default
14833 value (that is if SN is missing it is given the same
14834 value as LN and vice versa), these are now added on the
14835 grounds that if an object has a name we should be able to
14836 look it up. Finally added warning output when duplicate
14837 short or long names are found.
14841 * Changes needed for Tandem NSK.
14843 *Scott Uroff <scott@xypro.com>*
14845 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14846 RSA_padding_check_SSLv23(), special padding was never detected
14847 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14848 version rollback attacks was not effective.
14850 In s23_clnt.c, don't use special rollback-attack detection padding
14851 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14852 client; similarly, in s23_srvr.c, don't do the rollback check if
14853 SSL 2.0 is the only protocol enabled in the server.
14857 * Make it possible to get hexdumps of unprintable data with 'openssl
14858 asn1parse'. By implication, the functions ASN1_parse_dump() and
14859 BIO_dump_indent() are added.
14863 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14864 these print out strings and name structures based on various
14865 flags including RFC2253 support and proper handling of
14866 multibyte characters. Added options to the 'x509' utility
14867 to allow the various flags to be set.
14871 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14872 Also change the functions X509_cmp_current_time() and
14873 X509_gmtime_adj() work with an ASN1_TIME structure,
14874 this will enable certificates using GeneralizedTime in validity
14875 dates to be checked.
14879 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14880 negative public key encodings) on by default,
14881 NO_NEG_PUBKEY_BUG can be set to disable it.
14885 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14886 content octets. An i2c_ASN1_OBJECT is unnecessary because
14887 the encoding can be trivially obtained from the structure.
14891 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14892 not read locks (`CRYPTO_r_[un]lock`).
14896 * A first attempt at creating official support for shared
14897 libraries through configuration. I've kept it so the
14898 default is static libraries only, and the OpenSSL programs
14899 are always statically linked for now, but there are
14900 preparations for dynamic linking in place.
14901 This has been tested on Linux and Tru64.
14905 * Randomness polling function for Win9x, as described in:
14906 Peter Gutmann, Software Generation of Practically Strong
14911 * Fix so PRNG is seeded in req if using an already existing
14916 * New options to smime application. -inform and -outform
14917 allow alternative formats for the S/MIME message including
14918 PEM and DER. The -content option allows the content to be
14919 specified separately. This should allow things like Netscape
14920 form signing output easier to verify.
14924 * Fix the ASN1 encoding of tags using the 'long form'.
14928 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14929 STRING types. These convert content octets to and from the
14930 underlying type. The actual tag and length octets are
14931 already assumed to have been read in and checked. These
14932 are needed because all other string types have virtually
14933 identical handling apart from the tag. By having versions
14934 of the ASN1 functions that just operate on content octets
14935 IMPLICIT tagging can be handled properly. It also allows
14936 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14937 and ASN1_INTEGER are identical apart from the tag.
14941 * Change the handling of OID objects as follows:
14943 - New object identifiers are inserted in objects.txt, following
14944 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14945 - objects.pl is used to process obj_mac.num and create a new
14947 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14950 This is currently kind of a hack, and the perl code in objects.pl
14951 isn't very elegant, but it works as I intended. The simplest way
14952 to check that it worked correctly is to look in obj_dat.h and
14953 check the array nid_objs and make sure the objects haven't moved
14954 around (this is important!). Additions are OK, as well as
14955 consistent name changes.
14959 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14963 * Addition of the command line parameter '-rand file' to 'openssl req'.
14964 The given file adds to whatever has already been seeded into the
14965 random pool through the RANDFILE configuration file option or
14966 environment variable, or the default random state file.
14970 * mkstack.pl now sorts each macro group into lexical order.
14971 Previously the output order depended on the order the files
14972 appeared in the directory, resulting in needless rewriting
14977 * Patches to make OpenSSL compile under Win32 again. Mostly
14978 work arounds for the VC++ problem that it treats func() as
14979 func(void). Also stripped out the parts of mkdef.pl that
14980 added extra typesafe functions: these no longer exist.
14984 * Reorganisation of the stack code. The macros are now all
14985 collected in safestack.h . Each macro is defined in terms of
14986 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14987 DEBUG_SAFESTACK is now handled in terms of function casts,
14988 this has the advantage of retaining type safety without the
14989 use of additional functions. If DEBUG_SAFESTACK is not defined
14990 then the non typesafe macros are used instead. Also modified the
14991 mkstack.pl script to handle the new form. Needs testing to see
14992 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14993 the default if no major problems. Similar behaviour for ASN1_SET_OF
14994 and PKCS12_STACK_OF.
14998 * When some versions of IIS use the 'NET' form of private key the
14999 key derivation algorithm is different. Normally MD5(password) is
15000 used as a 128 bit RC4 key. In the modified case
15001 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15002 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15003 as the old Netscape_RSA functions except they have an additional
15004 'sgckey' parameter which uses the modified algorithm. Also added
15005 an -sgckey command line option to the rsa utility. Thanks to
15006 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15007 algorithm to openssl-dev.
15011 * The evp_local.h macros were using 'c.##kname' which resulted in
15012 invalid expansion on some systems (SCO 5.0.5 for example).
15013 Corrected to 'c.kname'.
15015 *Phillip Porch <root@theporch.com>*
15017 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15018 a STACK of email addresses from a certificate or request, these look
15019 in the subject name and the subject alternative name extensions and
15020 omit any duplicate addresses.
15024 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15025 This makes DSA verification about 2 % faster.
15029 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15030 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15031 plus overhead for 1024 bit moduli).
15032 This makes exponentiations about 0.5 % faster for 1024 bit
15033 exponents (as measured by "openssl speed rsa2048").
15037 * Rename memory handling macros to avoid conflicts with other
15039 Malloc => OPENSSL_malloc
15040 Malloc_locked => OPENSSL_malloc_locked
15041 Realloc => OPENSSL_realloc
15042 Free => OPENSSL_free
15046 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15047 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15051 * CygWin32 support.
15053 *John Jarvie <jjarvie@newsguy.com>*
15055 * The type-safe stack code has been rejigged. It is now only compiled
15056 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15057 by default all type-specific stack functions are "#define"d back to
15058 standard stack functions. This results in more streamlined output
15059 but retains the type-safety checking possibilities of the original
15064 * The STACK code has been cleaned up, and certain type declarations
15065 that didn't make a lot of sense have been brought in line. This has
15066 also involved a cleanup of sorts in safestack.h to more correctly
15067 map type-safe stack functions onto their plain stack counterparts.
15068 This work has also resulted in a variety of "const"ifications of
15069 lots of the code, especially `_cmp` operations which should normally
15070 be prototyped with "const" parameters anyway.
15074 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15075 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15076 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15077 where all of 'md' is used each time the PRNG is used, but 'state'
15078 is used only indexed by a cyclic counter. As entropy may not be
15079 well distributed from the beginning, 'md' is important as a
15080 chaining variable. However, the output function chains only half
15081 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15082 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15083 in all of 'state' being rewritten, with the new values depending
15084 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15088 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15089 the handshake is continued after ssl_verify_cert_chain();
15090 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15091 can lead to 'unexplainable' connection aborts later.
15093 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15095 * Major EVP API cipher revision.
15096 Add hooks for extra EVP features. This allows various cipher
15097 parameters to be set in the EVP interface. Support added for variable
15098 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15099 setting of RC2 and RC5 parameters.
15101 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15104 Remove lots of duplicated code from the EVP library. For example *every*
15105 cipher init() function handles the 'iv' in the same way according to the
15106 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15107 for CFB and OFB modes they zero ctx->num.
15109 New functionality allows removal of S/MIME code RC2 hack.
15111 Most of the routines have the same form and so can be declared in terms
15114 By shifting this to the top level EVP_CipherInit() it can be removed from
15115 all individual ciphers. If the cipher wants to handle IVs or keys
15116 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15119 Change lots of functions like EVP_EncryptUpdate() to now return a
15120 value: although software versions of the algorithms cannot fail
15121 any installed hardware versions can.
15125 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15126 this option is set, tolerate broken clients that send the negotiated
15127 protocol version number instead of the requested protocol version
15132 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15133 i.e. non-zero for export ciphersuites, zero otherwise.
15134 Previous versions had this flag inverted, inconsistent with
15135 rsa_tmp_cb (..._TMP_RSA_CB).
15137 *Bodo Moeller; problem reported by Amit Chopra*
15139 * Add missing DSA library text string. Work around for some IIS
15140 key files with invalid SEQUENCE encoding.
15144 * Add a document (doc/standards.txt) that list all kinds of standards
15145 and so on that are implemented in OpenSSL.
15149 * Enhance c_rehash script. Old version would mishandle certificates
15150 with the same subject name hash and wouldn't handle CRLs at all.
15151 Added -fingerprint option to crl utility, to support new c_rehash
15156 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15160 * Fix for SSL server purpose checking. Server checking was
15161 rejecting certificates which had extended key usage present
15162 but no ssl client purpose.
15164 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15166 * Make PKCS#12 code work with no password. The PKCS#12 spec
15167 is a little unclear about how a blank password is handled.
15168 Since the password in encoded as a BMPString with terminating
15169 double NULL a zero length password would end up as just the
15170 double NULL. However no password at all is different and is
15171 handled differently in the PKCS#12 key generation code. NS
15172 treats a blank password as zero length. MSIE treats it as no
15173 password on export: but it will try both on import. We now do
15174 the same: PKCS12_parse() tries zero length and no password if
15175 the password is set to "" or NULL (NULL is now a valid password:
15176 it wasn't before) as does the pkcs12 application.
15180 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15181 perror when PEM_read_bio_X509_REQ fails, the error message must
15182 be obtained from the error queue.
15186 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15187 it in ERR_remove_state if appropriate, and change ERR_get_state
15188 accordingly to avoid race conditions (this is necessary because
15189 thread_hash is no longer constant once set).
15193 * Bugfix for linux-elf makefile.one.
15197 * RSA_get_default_method() will now cause a default
15198 RSA_METHOD to be chosen if one doesn't exist already.
15199 Previously this was only set during a call to RSA_new()
15200 or RSA_new_method(NULL) meaning it was possible for
15201 RSA_get_default_method() to return NULL.
15205 * Added native name translation to the existing DSO code
15206 that will convert (if the flag to do so is set) filenames
15207 that are sufficiently small and have no path information
15208 into a canonical native form. Eg. "blah" converted to
15209 "libblah.so" or "blah.dll" etc.
15213 * New function ERR_error_string_n(e, buf, len) which is like
15214 ERR_error_string(e, buf), but writes at most 'len' bytes
15215 including the 0 terminator. For ERR_error_string_n, 'buf'
15218 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15220 * CONF library reworked to become more general. A new CONF
15221 configuration file reader "class" is implemented as well as a
15222 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15223 old `CONF_*` functions are still there, but are reimplemented to
15224 work in terms of the new functions. Also, a set of functions
15225 to handle the internal storage of the configuration data is
15226 provided to make it easier to write new configuration file
15227 reader "classes" (I can definitely see something reading a
15228 configuration file in XML format, for example), called `_CONF_*`,
15229 or "the configuration storage API"...
15231 The new configuration file reading functions are:
15233 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15234 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15236 NCONF_default, NCONF_WIN32
15238 NCONF_dump_fp, NCONF_dump_bio
15240 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15241 NCONF_new creates a new CONF object. This works in the same way
15242 as other interfaces in OpenSSL, like the BIO interface.
15243 `NCONF_dump_*` dump the internal storage of the configuration file,
15244 which is useful for debugging. All other functions take the same
15245 arguments as the old `CONF_*` functions with the exception of the
15246 first that must be a `CONF *` instead of a `LHASH *`.
15248 To make it easier to use the new classes with the old `CONF_*` functions,
15249 the function CONF_set_default_method is provided.
15253 * Add '-tls1' option to 'openssl ciphers', which was already
15254 mentioned in the documentation but had not been implemented.
15255 (This option is not yet really useful because even the additional
15256 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15260 * Initial DSO code added into libcrypto for letting OpenSSL (and
15261 OpenSSL-based applications) load shared libraries and bind to
15262 them in a portable way.
15264 *Geoff Thorpe, with contributions from Richard Levitte*
15266 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15268 * Make sure _lrotl and _lrotr are only used with MSVC.
15270 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15271 (the default implementation of RAND_status).
15273 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15274 to '-clrext' (= clear extensions), as intended and documented.
15275 *Bodo Moeller; inconsistency pointed out by Michael Attili
15276 <attili@amaxo.com>*
15278 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15279 was larger than the MD block size.
15281 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15283 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15284 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15285 using the passed key: if the passed key was a private key the result
15286 of X509_print(), for example, would be to print out all the private key
15291 * des_quad_cksum() byte order bug fix.
15292 *Ulf Möller, using the problem description in krb4-0.9.7, where
15293 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15295 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15298 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15300 * For easily testing in shell scripts whether some command
15301 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15302 returns with exit code 0 iff no command of the given name is available.
15303 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15304 the output goes to stdout and nothing is printed to stderr.
15305 Additional arguments are always ignored.
15307 Since for each cipher there is a command of the same name,
15308 the 'no-cipher' compilation switches can be tested this way.
15310 ('openssl no-XXX' is not able to detect pseudo-commands such
15311 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15315 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15319 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15320 is set; it will be thrown away anyway because each handshake creates
15322 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15323 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15324 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15325 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15329 * New s_client option -ign_eof: EOF at stdin is ignored, and
15330 'Q' and 'R' lose their special meanings (quit/renegotiate).
15331 This is part of what -quiet does; unlike -quiet, -ign_eof
15332 does not suppress any output.
15336 * Add compatibility options to the purpose and trust code. The
15337 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15338 accepts a certificate or CA, this was the previous behaviour,
15339 with all the associated security issues.
15341 X509_TRUST_COMPAT is the old trust behaviour: only and
15342 automatically trust self signed roots in certificate store. A
15343 new trust setting X509_TRUST_DEFAULT is used to specify that
15344 a purpose has no associated trust setting and it should instead
15345 use the value in the default purpose.
15349 * Fix the PKCS#8 DSA private key code so it decodes keys again
15350 and fix a memory leak.
15354 * In util/mkerr.pl (which implements 'make errors'), preserve
15355 reason strings from the previous version of the .c file, as
15356 the default to have only downcase letters (and digits) in
15357 automatically generated reasons codes is not always appropriate.
15361 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15362 using strerror. Previously, ERR_reason_error_string() returned
15363 library names as reason strings for SYSerr; but SYSerr is a special
15364 case where small numbers are errno values, not library numbers.
15368 * Add '-dsaparam' option to 'openssl dhparam' application. This
15369 converts DSA parameters into DH parameters. (When creating parameters,
15370 DSA_generate_parameters is used.)
15374 * Include 'length' (recommended exponent length) in C code generated
15375 by 'openssl dhparam -C'.
15379 * The second argument to set_label in perlasm was already being used
15380 so couldn't be used as a "file scope" flag. Moved to third argument
15385 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15386 instead of RAND_bytes for encryption IVs and salts.
15390 * Include RAND_status() into RAND_METHOD instead of implementing
15391 it only for md_rand.c Otherwise replacing the PRNG by calling
15392 RAND_set_rand_method would be impossible.
15396 * Don't let DSA_generate_key() enter an infinite loop if the random
15397 number generation fails.
15401 * New 'rand' application for creating pseudo-random output.
15405 * Added configuration support for Linux/IA64
15407 *Rolf Haberrecker <rolf@suse.de>*
15409 * Assembler module support for Mingw32.
15413 * Shared library support for HPUX (in shlib/).
15415 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15417 * Shared library support for Solaris gcc.
15419 *Lutz Behnke <behnke@trustcenter.de>*
15421 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15423 * PKCS7_encrypt() was adding text MIME headers twice because they
15424 were added manually and by SMIME_crlf_copy().
15428 * In bntest.c don't call BN_rand with zero bits argument.
15430 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15432 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15433 case was implemented. This caused BN_div_recp() to fail occasionally.
15437 * Add an optional second argument to the set_label() in the perl
15438 assembly language builder. If this argument exists and is set
15439 to 1 it signals that the assembler should use a symbol whose
15440 scope is the entire file, not just the current function. This
15441 is needed with MASM which uses the format label:: for this scope.
15443 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15445 * Change the ASN1 types so they are typedefs by default. Before
15446 almost all types were #define'd to ASN1_STRING which was causing
15447 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15452 * Change names of new functions to the new get1/get0 naming
15453 convention: After 'get1', the caller owns a reference count
15454 and has to call `..._free`; 'get0' returns a pointer to some
15455 data structure without incrementing reference counters.
15456 (Some of the existing 'get' functions increment a reference
15457 counter, some don't.)
15458 Similarly, 'set1' and 'add1' functions increase reference
15459 counters or duplicate objects.
15463 * Allow for the possibility of temp RSA key generation failure:
15464 the code used to assume it always worked and crashed on failure.
15468 * Fix potential buffer overrun problem in BIO_printf().
15469 *Ulf Möller, using public domain code by Patrick Powell; problem
15470 pointed out by David Sacerdote <das33@cornell.edu>*
15472 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15473 RAND_egd() and RAND_status(). In the command line application,
15474 the EGD socket can be specified like a seed file using RANDFILE
15479 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15480 Some CAs (e.g. Verisign) distribute certificates in this form.
15484 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15485 list to exclude them. This means that no special compilation option
15486 is needed to use anonymous DH: it just needs to be included in the
15491 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15492 EVP_MD_type. The old functionality is available in a new macro called
15493 EVP_MD_md(). Change code that uses it and update docs.
15497 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15498 where the `void *` argument is replaced by a function pointer argument.
15499 Previously `void *` was abused to point to functions, which works on
15500 many platforms, but is not correct. As these functions are usually
15501 called by macros defined in OpenSSL header files, most source code
15502 should work without changes.
15506 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15507 sections with information on -D... compiler switches used for
15508 compiling the library so that applications can see them. To enable
15509 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15510 must be defined. E.g.,
15511 #define OPENSSL_ALGORITHM_DEFINES
15512 #include <openssl/opensslconf.h>
15513 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15515 *Richard Levitte, Ulf and Bodo Möller*
15517 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15522 * Change the 'other' type in certificate aux info to a STACK_OF
15523 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15524 the required ASN1 format: arbitrary types determined by an OID.
15528 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15529 argument to 'req'. This is not because the function is newer or
15530 better than others it just uses the work 'NEW' in the certificate
15531 request header lines. Some software needs this.
15535 * Reorganise password command line arguments: now passwords can be
15536 obtained from various sources. Delete the PEM_cb function and make
15537 it the default behaviour: i.e. if the callback is NULL and the
15538 usrdata argument is not NULL interpret it as a null terminated pass
15539 phrase. If usrdata and the callback are NULL then the pass phrase
15540 is prompted for as usual.
15544 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15545 the support is automatically enabled. The resulting binaries will
15546 autodetect the card and use it if present.
15548 *Ben Laurie and Compaq Inc.*
15550 * Work around for Netscape hang bug. This sends certificate request
15551 and server done in one record. Since this is perfectly legal in the
15552 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15553 the bugs/SSLv3 entry for more info.
15557 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15561 * Add -rand argument to smime and pkcs12 applications and read/write
15566 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15570 * Add command line password options to the remaining applications.
15574 * Bug fix for BN_div_recp() for numerators with an even number of
15579 * More tests in bntest.c, and changed test_bn output.
15583 * ./config recognizes MacOS X now.
15587 * Bug fix for BN_div() when the first words of num and divisor are
15588 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15592 * Add support for various broken PKCS#8 formats, and command line
15593 options to produce them.
15597 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15598 get temporary BIGNUMs from a BN_CTX.
15602 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15607 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15608 include a #define from the old name to the new. The original intent
15609 was that statically linked binaries could for example just call
15610 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15611 link with digests. This never worked because SSLeay_add_all_digests()
15612 and SSLeay_add_all_ciphers() were in the same source file so calling
15613 one would link with the other. They are now in separate source files.
15617 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15621 * Use a less unusual form of the Miller-Rabin primality test (it used
15622 a binary algorithm for exponentiation integrated into the Miller-Rabin
15623 loop, our standard modexp algorithms are faster).
15627 * Support for the EBCDIC character set completed.
15629 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15631 * Source code cleanups: use const where appropriate, eliminate casts,
15632 use `void *` instead of `char *` in lhash.
15636 * Bugfix: ssl3_send_server_key_exchange was not restartable
15637 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15638 this the server could overwrite ephemeral keys that the client
15643 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15644 using 50 iterations of the Rabin-Miller test.
15646 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15647 iterations of the Rabin-Miller test as required by the appendix
15648 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15649 As BN_is_prime_fasttest includes trial division, DSA parameter
15650 generation becomes much faster.
15652 This implies a change for the callback functions in DSA_is_prime
15653 and DSA_generate_parameters: The callback function is called once
15654 for each positive witness in the Rabin-Miller test, not just
15655 occasionally in the inner loop; and the parameters to the
15656 callback function now provide an iteration count for the outer
15657 loop rather than for the current invocation of the inner loop.
15658 DSA_generate_parameters additionally can call the callback
15659 function with an 'iteration count' of -1, meaning that a
15660 candidate has passed the trial division test (when q is generated
15661 from an application-provided seed, trial division is skipped).
15665 * New function BN_is_prime_fasttest that optionally does trial
15666 division before starting the Rabin-Miller test and has
15667 an additional BN_CTX * argument (whereas BN_is_prime always
15668 has to allocate at least one BN_CTX).
15669 'callback(1, -1, cb_arg)' is called when a number has passed the
15670 trial division stage.
15674 * Fix for bug in CRL encoding. The validity dates weren't being handled
15679 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15683 * New function BN_pseudo_rand().
15687 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15688 bignum version of BN_from_montgomery() with the working code from
15689 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15694 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15695 made it impossible to use the same SSL_SESSION data structure in
15696 SSL2 clients in multiple threads.
15700 * The return value of RAND_load_file() no longer counts bytes obtained
15701 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15702 to seed the PRNG (previously an explicit byte count was required).
15704 *Ulf Möller, Bodo Möller*
15706 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15707 used `char *` instead of `void *` and had casts all over the place.
15711 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15715 * Retain source code compatibility for BN_prime_checks macro:
15716 BN_is_prime(..., BN_prime_checks, ...) now uses
15717 BN_prime_checks_for_size to determine the appropriate number of
15718 Rabin-Miller iterations.
15722 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15723 DH_CHECK_P_NOT_SAFE_PRIME.
15724 (Check if this is true? OpenPGP calls them "strong".)
15728 * Merge the functionality of "dh" and "gendh" programs into a new program
15729 "dhparam". The old programs are retained for now but will handle DH keys
15730 (instead of parameters) in future.
15734 * Make the ciphers, s_server and s_client programs check the return values
15735 when a new cipher list is set.
15739 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15740 ciphers. Before when the 56bit ciphers were enabled the sorting was
15743 The syntax for the cipher sorting has been extended to support sorting by
15744 cipher-strength (using the strength_bits hard coded in the tables).
15745 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15747 Fix a bug in the cipher-command parser: when supplying a cipher command
15748 string with an "undefined" symbol (neither command nor alphanumeric
15749 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15750 an error is flagged.
15752 Due to the strength-sorting extension, the code of the
15753 ssl_create_cipher_list() function was completely rearranged. I hope that
15754 the readability was also increased :-)
15756 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15758 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15759 for the first serial number and places 2 in the serial number file. This
15760 avoids problems when the root CA is created with serial number zero and
15761 the first user certificate has the same issuer name and serial number
15766 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15767 the new code. Add documentation for this stuff.
15771 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15772 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15773 structures and behave in an analogous way to the X509v3 functions:
15774 they shouldn't be called directly but wrapper functions should be used
15777 So we also now have some wrapper functions that call the X509at functions
15778 when passed certificate requests. (TO DO: similar things can be done with
15779 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15780 things. Some of these need some d2i or i2d and print functionality
15781 because they handle more complex structures.)
15785 * Add missing #ifndefs that caused missing symbols when building libssl
15786 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15787 NO_RSA in `ssl/s2*.c`.
15789 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15791 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15792 has a return value which indicates the quality of the random data
15793 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15794 error queue. New function RAND_pseudo_bytes() generates output that is
15795 guaranteed to be unique but not unpredictable. RAND_add is like
15796 RAND_seed, but takes an extra argument for an entropy estimate
15797 (RAND_seed always assumes full entropy).
15801 * Do more iterations of Rabin-Miller probable prime test (specifically,
15802 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15803 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15804 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15805 false-positive rate of at most 2^-80 for random input.
15809 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15813 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15814 in the 0.9.5 release), this returns the chain
15815 from an X509_CTX structure with a dup of the stack and all
15816 the X509 reference counts upped: so the stack will exist
15817 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15820 Also make SSL_SESSION_print() print out the verify return
15825 * Add manpage for the pkcs12 command. Also change the default
15826 behaviour so MAC iteration counts are used unless the new
15827 -nomaciter option is used. This improves file security and
15828 only older versions of MSIE (4.0 for example) need it.
15832 * Honor the no-xxx Configure options when creating .DEF files.
15836 * Add PKCS#10 attributes to field table: challengePassword,
15837 unstructuredName and unstructuredAddress. These are taken from
15838 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15839 international characters are used.
15841 More changes to X509_ATTRIBUTE code: allow the setting of types
15842 based on strings. Remove the 'loc' parameter when adding
15843 attributes because these will be a SET OF encoding which is sorted
15848 * Initial changes to the 'req' utility to allow request generation
15849 automation. This will allow an application to just generate a template
15850 file containing all the field values and have req construct the
15853 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15854 used all over the place including certificate requests and PKCS#7
15855 structures. They are currently handled manually where necessary with
15856 some primitive wrappers for PKCS#7. The new functions behave in a
15857 manner analogous to the X509 extension functions: they allow
15858 attributes to be looked up by NID and added.
15860 Later something similar to the X509V3 code would be desirable to
15861 automatically handle the encoding, decoding and printing of the
15862 more complex types. The string types like challengePassword can
15863 be handled by the string table functions.
15865 Also modified the multi byte string table handling. Now there is
15866 a 'global mask' which masks out certain types. The table itself
15867 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15868 is useful when for example there is only one permissible type
15869 (as in countryName) and using the mask might result in no valid
15874 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15875 SSL_get_peer_finished to allow applications to obtain the latest
15876 Finished messages sent to the peer or expected from the peer,
15877 respectively. (SSL_get_peer_finished is usually the Finished message
15878 actually received from the peer, otherwise the protocol will be aborted.)
15880 As the Finished message are message digests of the complete handshake
15881 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15882 be used for external authentication procedures when the authentication
15883 provided by SSL/TLS is not desired or is not enough.
15887 * Enhanced support for Alpha Linux is added. Now ./config checks if
15888 the host supports BWX extension and if Compaq C is present on the
15889 $PATH. Just exploiting of the BWX extension results in 20-30%
15890 performance kick for some algorithms, e.g. DES and RC4 to mention
15891 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15896 * Add support for MS "fast SGC". This is arguably a violation of the
15897 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15898 weak crypto and after checking the certificate is SGC a second one
15899 with strong crypto. MS SGC stops the first handshake after receiving
15900 the server certificate message and sends a second client hello. Since
15901 a server will typically do all the time consuming operations before
15902 expecting any further messages from the client (server key exchange
15903 is the most expensive) there is little difference between the two.
15905 To get OpenSSL to support MS SGC we have to permit a second client
15906 hello message after we have sent server done. In addition we have to
15907 reset the MAC if we do get this second client hello.
15911 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15912 if a DER encoded private key is RSA or DSA traditional format. Changed
15913 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15914 format DER encoded private key. Newer code should use PKCS#8 format which
15915 has the key type encoded in the ASN1 structure. Added DER private key
15916 support to pkcs8 application.
15920 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15921 ciphersuites has been selected (as required by the SSL 3/TLS 1
15922 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15923 is set, we interpret this as a request to violate the specification
15924 (the worst that can happen is a handshake failure, and 'correct'
15925 behaviour would result in a handshake failure anyway).
15929 * In SSL_CTX_add_session, take into account that there might be multiple
15930 SSL_SESSION structures with the same session ID (e.g. when two threads
15931 concurrently obtain them from an external cache).
15932 The internal cache can handle only one SSL_SESSION with a given ID,
15933 so if there's a conflict, we now throw out the old one to achieve
15938 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15939 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15940 some routines that use cipher OIDs: some ciphers do not have OIDs
15941 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15946 * Simplify the trust setting structure and code. Now we just have
15947 two sequences of OIDs for trusted and rejected settings. These will
15948 typically have values the same as the extended key usage extension
15949 and any application specific purposes.
15951 The trust checking code now has a default behaviour: it will just
15952 check for an object with the same NID as the passed id. Functions can
15953 be provided to override either the default behaviour or the behaviour
15954 for a given id. SSL client, server and email already have functions
15955 in place for compatibility: they check the NID and also return "trusted"
15956 if the certificate is self signed.
15960 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15961 traditional format into an EVP_PKEY structure.
15965 * Add a password callback function PEM_cb() which either prompts for
15966 a password if usr_data is NULL or otherwise assumes it is a null
15967 terminated password. Allow passwords to be passed on command line
15968 environment or config files in a few more utilities.
15972 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15973 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15974 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15975 Update documentation.
15979 * Support for ASN1 "NULL" type. This could be handled before by using
15980 ASN1_TYPE but there wasn't any function that would try to read a NULL
15981 and produce an error if it couldn't. For compatibility we also have
15982 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15983 don't allocate anything because they don't need to.
15987 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15990 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15992 * Rebuild of the memory allocation routines used by OpenSSL code and
15993 possibly others as well. The purpose is to make an interface that
15994 provide hooks so anyone can build a separate set of allocation and
15995 deallocation routines to be used by OpenSSL, for example memory
15996 pool implementations, or something else, which was previously hard
15997 since Malloc(), Realloc() and Free() were defined as macros having
15998 the values malloc, realloc and free, respectively (except for Win32
15999 compilations). The same is provided for memory debugging code.
16000 OpenSSL already comes with functionality to find memory leaks, but
16001 this gives people a chance to debug other memory problems.
16003 With these changes, a new set of functions and macros have appeared:
16005 CRYPTO_set_mem_debug_functions() [F]
16006 CRYPTO_get_mem_debug_functions() [F]
16007 CRYPTO_dbg_set_options() [F]
16008 CRYPTO_dbg_get_options() [F]
16009 CRYPTO_malloc_debug_init() [M]
16011 The memory debug functions are NULL by default, unless the library
16012 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16013 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16014 gives the standard debugging functions that come with OpenSSL) or
16015 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16016 provided by the library user) must be used. When the standard
16017 debugging functions are used, CRYPTO_dbg_set_options can be used to
16018 request additional information:
16019 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16020 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16022 Also, things like CRYPTO_set_mem_functions will always give the
16023 expected result (the new set of functions is used for allocation
16024 and deallocation) at all times, regardless of platform and compiler
16027 To finish it up, some functions that were never use in any other
16028 way than through macros have a new API and new semantic:
16030 CRYPTO_dbg_malloc()
16031 CRYPTO_dbg_realloc()
16034 All macros of value have retained their old syntax.
16036 *Richard Levitte and Bodo Moeller*
16038 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16039 ordering of SMIMECapabilities wasn't in "strength order" and there
16040 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16045 * Some ASN1 types with illegal zero length encoding (INTEGER,
16046 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16048 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16050 * Merge in my S/MIME library for OpenSSL. This provides a simple
16051 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16052 functionality to handle multipart/signed properly) and a utility
16053 called 'smime' to call all this stuff. This is based on code I
16054 originally wrote for Celo who have kindly allowed it to be
16055 included in OpenSSL.
16059 * Add variants des_set_key_checked and des_set_key_unchecked of
16060 des_set_key (aka des_key_sched). Global variable des_check_key
16061 decides which of these is called by des_set_key; this way
16062 des_check_key behaves as it always did, but applications and
16063 the library itself, which was buggy for des_check_key == 1,
16064 have a cleaner way to pick the version they need.
16068 * New function PKCS12_newpass() which changes the password of a
16073 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16074 dynamic mix. In both cases the ids can be used as an index into the
16075 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16076 functions so they accept a list of the field values and the
16077 application doesn't need to directly manipulate the X509_TRUST
16082 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16087 * Modify the way the V3 extension code looks up extensions. This now
16088 works in a similar way to the object code: we have some "standard"
16089 extensions in a static table which is searched with OBJ_bsearch()
16090 and the application can add dynamic ones if needed. The file
16091 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16092 updated whenever a new extension is added to the core code and kept
16093 in ext_nid order. There is a simple program 'tabtest.c' which checks
16094 this. New extensions are not added too often so this file can readily
16095 be maintained manually.
16097 There are two big advantages in doing things this way. The extensions
16098 can be looked up immediately and no longer need to be "added" using
16099 X509V3_add_standard_extensions(): this function now does nothing.
16100 Side note: I get *lots* of email saying the extension code doesn't
16101 work because people forget to call this function.
16102 Also no dynamic allocation is done unless new extensions are added:
16103 so if we don't add custom extensions there is no need to call
16104 X509V3_EXT_cleanup().
16108 * Modify enc utility's salting as follows: make salting the default. Add a
16109 magic header, so unsalted files fail gracefully instead of just decrypting
16110 to garbage. This is because not salting is a big security hole, so people
16111 should be discouraged from doing it.
16115 * Fixes and enhancements to the 'x509' utility. It allowed a message
16116 digest to be passed on the command line but it only used this
16117 parameter when signing a certificate. Modified so all relevant
16118 operations are affected by the digest parameter including the
16119 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16120 DSA key was used because it didn't fix the digest.
16124 * Initial certificate chain verify code. Currently tests the untrusted
16125 certificates for consistency with the verify purpose (which is set
16126 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16128 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16129 this is because it will reject chains with invalid extensions whereas
16130 every previous version of OpenSSL and SSLeay made no checks at all.
16132 Trust code: checks the root CA for the relevant trust settings. Trust
16133 settings have an initial value consistent with the verify purpose: e.g.
16134 if the verify purpose is for SSL client use it expects the CA to be
16135 trusted for SSL client use. However the default value can be changed to
16136 permit custom trust settings: one example of this would be to only trust
16137 certificates from a specific "secure" set of CAs.
16139 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16140 which should be used for version portability: especially since the
16141 verify structure is likely to change more often now.
16143 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16144 to set them. If not set then assume SSL clients will verify SSL servers
16147 Two new options to the verify program: -untrusted allows a set of
16148 untrusted certificates to be passed in and -purpose which sets the
16149 intended purpose of the certificate. If a purpose is set then the
16150 new chain verify code is used to check extension consistency.
16154 * Support for the authority information access extension.
16158 * Modify RSA and DSA PEM read routines to transparently handle
16159 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16160 public keys in a format compatible with certificate
16161 SubjectPublicKeyInfo structures. Unfortunately there were already
16162 functions called *_PublicKey_* which used various odd formats so
16163 these are retained for compatibility: however the DSA variants were
16164 never in a public release so they have been deleted. Changed dsa/rsa
16165 utilities to handle the new format: note no releases ever handled public
16166 keys so we should be OK.
16168 The primary motivation for this change is to avoid the same fiasco
16169 that dogs private keys: there are several incompatible private key
16170 formats some of which are standard and some OpenSSL specific and
16171 require various evil hacks to allow partial transparent handling and
16172 even then it doesn't work with DER formats. Given the option anything
16173 other than PKCS#8 should be dumped: but the other formats have to
16174 stay in the name of compatibility.
16176 With public keys and the benefit of hindsight one standard format
16177 is used which works with EVP_PKEY, RSA or DSA structures: though
16178 it clearly returns an error if you try to read the wrong kind of key.
16180 Added a -pubkey option to the 'x509' utility to output the public key.
16181 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16182 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16183 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16184 that do the same as the `EVP_PKEY_assign_*()` except they up the
16185 reference count of the added key (they don't "swallow" the
16190 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16191 CRLs would fail if the file contained no certificates or no CRLs:
16192 added a new function to read in both types and return the number
16193 read: this means that if none are read it will be an error. The
16194 DER versions of the certificate and CRL reader would always fail
16195 because it isn't possible to mix certificates and CRLs in DER format
16196 without choking one or the other routine. Changed this to just read
16197 a certificate: this is the best we can do. Also modified the code
16198 in apps/verify.c to take notice of return codes: it was previously
16199 attempting to read in certificates from NULL pointers and ignoring
16200 any errors: this is one reason why the cert and CRL reader seemed
16201 to work. It doesn't check return codes from the default certificate
16202 routines: these may well fail if the certificates aren't installed.
16206 * Code to support otherName option in GeneralName.
16210 * First update to verify code. Change the verify utility
16211 so it warns if it is passed a self signed certificate:
16212 for consistency with the normal behaviour. X509_verify
16213 has been modified to it will now verify a self signed
16214 certificate if *exactly* the same certificate appears
16215 in the store: it was previously impossible to trust a
16216 single self signed certificate. This means that:
16217 openssl verify ss.pem
16218 now gives a warning about a self signed certificate but
16219 openssl verify -CAfile ss.pem ss.pem
16224 * For servers, store verify_result in SSL_SESSION data structure
16225 (and add it to external session representation).
16226 This is needed when client certificate verifications fails,
16227 but an application-provided verification callback (set by
16228 SSL_CTX_set_cert_verify_callback) allows accepting the session
16229 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16230 but returns 1): When the session is reused, we have to set
16231 ssl->verify_result to the appropriate error code to avoid
16234 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16236 * Fix a bug in the new PKCS#7 code: it didn't consider the
16237 case in PKCS7_dataInit() where the signed PKCS7 structure
16238 didn't contain any existing data because it was being created.
16240 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16242 * Add a salt to the key derivation routines in enc.c. This
16243 forms the first 8 bytes of the encrypted file. Also add a
16244 -S option to allow a salt to be input on the command line.
16248 * New function X509_cmp(). Oddly enough there wasn't a function
16249 to compare two certificates. We do this by working out the SHA1
16250 hash and comparing that. X509_cmp() will be needed by the trust
16255 * SSL_get1_session() is like SSL_get_session(), but increments
16256 the reference count in the SSL_SESSION returned.
16258 *Geoff Thorpe <geoff@eu.c2.net>*
16260 * Fix for 'req': it was adding a null to request attributes.
16261 Also change the X509_LOOKUP and X509_INFO code to handle
16262 certificate auxiliary information.
16266 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16271 * Add the possibility to add extra information to the memory leak
16272 detecting output, to form tracebacks, showing from where each
16273 allocation was originated: CRYPTO_push_info("constant string") adds
16274 the string plus current file name and line number to a per-thread
16275 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16276 is like calling CYRPTO_pop_info() until the stack is empty.
16277 Also updated memory leak detection code to be multi-thread-safe.
16281 * Add options -text and -noout to pkcs7 utility and delete the
16282 encryption options which never did anything. Update docs.
16286 * Add options to some of the utilities to allow the pass phrase
16287 to be included on either the command line (not recommended on
16288 OSes like Unix) or read from the environment. Update the
16289 manpages and fix a few bugs.
16293 * Add a few manpages for some of the openssl commands.
16297 * Fix the -revoke option in ca. It was freeing up memory twice,
16298 leaking and not finding already revoked certificates.
16302 * Extensive changes to support certificate auxiliary information.
16303 This involves the use of X509_CERT_AUX structure and X509_AUX
16304 functions. An X509_AUX function such as PEM_read_X509_AUX()
16305 can still read in a certificate file in the usual way but it
16306 will also read in any additional "auxiliary information". By
16307 doing things this way a fair degree of compatibility can be
16308 retained: existing certificates can have this information added
16309 using the new 'x509' options.
16311 Current auxiliary information includes an "alias" and some trust
16312 settings. The trust settings will ultimately be used in enhanced
16313 certificate chain verification routines: currently a certificate
16314 can only be trusted if it is self signed and then it is trusted
16319 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16320 The problem was that one of the replacement routines had not been working
16321 since SSLeay releases. For now the offending routine has been replaced
16322 with non-optimised assembler. Even so, this now gives around 95%
16323 performance improvement for 1024 bit RSA signs.
16327 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16328 handling. Most clients have the effective key size in bits equal to
16329 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16330 A few however don't do this and instead use the size of the decrypted key
16331 to determine the RC2 key length and the AlgorithmIdentifier to determine
16332 the effective key length. In this case the effective key length can still
16333 be 40 bits but the key length can be 168 bits for example. This is fixed
16334 by manually forcing an RC2 key into the EVP_PKEY structure because the
16335 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16336 the key length and effective key length are equal.
16340 * Add a bunch of functions that should simplify the creation of
16341 X509_NAME structures. Now you should be able to do:
16342 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16343 and have it automatically work out the correct field type and fill in
16344 the structures. The more adventurous can try:
16345 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16346 and it will (hopefully) work out the correct multibyte encoding.
16350 * Change the 'req' utility to use the new field handling and multibyte
16351 copy routines. Before the DN field creation was handled in an ad hoc
16352 way in req, ca, and x509 which was rather broken and didn't support
16353 BMPStrings or UTF8Strings. Since some software doesn't implement
16354 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16355 using the dirstring_type option. See the new comment in the default
16356 openssl.cnf for more info.
16360 * Make crypto/rand/md_rand.c more robust:
16361 - Assure unique random numbers after fork().
16362 - Make sure that concurrent threads access the global counter and
16363 md serializably so that we never lose entropy in them
16364 or use exactly the same state in multiple threads.
16365 Access to the large state is not always serializable because
16366 the additional locking could be a performance killer, and
16367 md should be large enough anyway.
16371 * New file apps/app_rand.c with commonly needed functionality
16372 for handling the random seed file.
16374 Use the random seed file in some applications that previously did not:
16376 dsaparam -genkey (which also ignored its '-rand' option),
16379 x509 (when signing).
16380 Except on systems with /dev/urandom, it is crucial to have a random
16381 seed file at least for key creation, DSA signing, and for DH exchanges;
16382 for RSA signatures we could do without one.
16384 gendh and gendsa (unlike genrsa) used to read only the first byte
16385 of each file listed in the '-rand' option. The function as previously
16386 found in genrsa is now in app_rand.c and is used by all programs
16387 that support '-rand'.
16391 * In RAND_write_file, use mode 0600 for creating files;
16392 don't just chmod when it may be too late.
16396 * Report an error from X509_STORE_load_locations
16397 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16401 * New function ASN1_mbstring_copy() this copies a string in either
16402 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16403 into an ASN1_STRING type. A mask of permissible types is passed
16404 and it chooses the "minimal" type to use or an error if not type
16409 * Add function equivalents to the various macros in asn1.h. The old
16410 macros are retained with an `M_` prefix. Code inside the library can
16411 use the `M_` macros. External code (including the openssl utility)
16412 should *NOT* in order to be "shared library friendly".
16416 * Add various functions that can check a certificate's extensions
16417 to see if it usable for various purposes such as SSL client,
16418 server or S/MIME and CAs of these types. This is currently
16419 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16420 verification. Also added a -purpose flag to x509 utility to
16421 print out all the purposes.
16425 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16430 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16431 for, obtain and decode and extension and obtain its critical flag.
16432 This allows all the necessary extension code to be handled in a
16433 single function call.
16437 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16438 platforms. See crypto/rc4/rc4_enc.c for further details.
16442 * New -noout option to asn1parse. This causes no output to be produced
16443 its main use is when combined with -strparse and -out to extract data
16444 from a file (which may not be in ASN.1 format).
16448 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16449 when producing the local key id.
16451 *Richard Levitte <levitte@stacken.kth.se>*
16453 * New option -dhparam in s_server. This allows a DH parameter file to be
16454 stated explicitly. If it is not stated then it tries the first server
16455 certificate file. The previous behaviour hard coded the filename
16460 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16461 a public key to be input or output. For example:
16462 openssl rsa -in key.pem -pubout -out pubkey.pem
16463 Also added necessary DSA public key functions to handle this.
16467 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16468 in the message. This was handled by allowing
16469 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16471 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16473 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16474 to the end of the strings whereas this didn't. This would cause problems
16475 if strings read with d2i_ASN1_bytes() were later modified.
16477 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16479 * Fix for base64 decode bug. When a base64 bio reads only one line of
16480 data and it contains EOF it will end up returning an error. This is
16481 caused by input 46 bytes long. The cause is due to the way base64
16482 BIOs find the start of base64 encoded data. They do this by trying a
16483 trial decode on each line until they find one that works. When they
16484 do a flag is set and it starts again knowing it can pass all the
16485 data directly through the decoder. Unfortunately it doesn't reset
16486 the context it uses. This means that if EOF is reached an attempt
16487 is made to pass two EOFs through the context and this causes the
16488 resulting error. This can also cause other problems as well. As is
16489 usual with these problems it takes *ages* to find and the fix is
16490 trivial: move one line.
16492 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16494 * Ugly workaround to get s_client and s_server working under Windows. The
16495 old code wouldn't work because it needed to select() on sockets and the
16496 tty (for keypresses and to see if data could be written). Win32 only
16497 supports select() on sockets so we select() with a 1s timeout on the
16498 sockets and then see if any characters are waiting to be read, if none
16499 are present then we retry, we also assume we can always write data to
16500 the tty. This isn't nice because the code then blocks until we've
16501 received a complete line of data and it is effectively polling the
16502 keyboard at 1s intervals: however it's quite a bit better than not
16503 working at all :-) A dedicated Windows application might handle this
16504 with an event loop for example.
16508 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16509 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16510 will be called when RSA_sign() and RSA_verify() are used. This is useful
16511 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16512 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16513 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16514 This necessitated the support of an extra signature type NID_md5_sha1
16515 for SSL signatures and modifications to the SSL library to use it instead
16516 of calling RSA_public_decrypt() and RSA_private_encrypt().
16520 * Add new -verify -CAfile and -CApath options to the crl program, these
16521 will lookup a CRL issuers certificate and verify the signature in a
16522 similar way to the verify program. Tidy up the crl program so it
16523 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16524 less strict. It will now permit CRL extensions even if it is not
16525 a V2 CRL: this will allow it to tolerate some broken CRLs.
16529 * Initialize all non-automatic variables each time one of the openssl
16530 sub-programs is started (this is necessary as they may be started
16531 multiple times from the "OpenSSL>" prompt).
16533 *Lennart Bang, Bodo Moeller*
16535 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16536 removing all other RSA functionality (this is what NO_RSA does). This
16537 is so (for example) those in the US can disable those operations covered
16538 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16543 * Non-copying interface to BIO pairs.
16544 (still largely untested)
16548 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16549 ASCII string. This was handled independently in various places before.
16553 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16554 UTF8 strings a character at a time.
16558 * Use client_version from client hello to select the protocol
16559 (s23_srvr.c) and for RSA client key exchange verification
16560 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16564 * Add various utility functions to handle SPKACs, these were previously
16565 handled by poking round in the structure internals. Added new function
16566 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16567 print, verify and generate SPKACs. Based on an original idea from
16568 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16572 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16576 * Allow the config file extension section to be overwritten on the
16577 command line. Based on an original idea from Massimiliano Pala
16578 <madwolf@comune.modena.it>. The new option is called -extensions
16579 and can be applied to ca, req and x509. Also -reqexts to override
16580 the request extensions in req and -crlexts to override the crl extensions
16585 * Add new feature to the SPKAC handling in ca. Now you can include
16586 the same field multiple times by preceding it by "XXXX." for example:
16589 this is the same syntax as used in the req config file.
16593 * Allow certificate extensions to be added to certificate requests. These
16594 are specified in a 'req_extensions' option of the req section of the
16595 config file. They can be printed out with the -text option to req but
16596 are otherwise ignored at present.
16600 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16601 data read consists of only the final block it would not decrypted because
16602 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16603 A misplaced 'break' also meant the decrypted final block might not be
16604 copied until the next read.
16608 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16609 a few extra parameters to the DH structure: these will be useful if
16610 for example we want the value of 'q' or implement X9.42 DH.
16614 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16615 provides hooks that allow the default DSA functions or functions on a
16616 "per key" basis to be replaced. This allows hardware acceleration and
16617 hardware key storage to be handled without major modification to the
16618 library. Also added low level modexp hooks and CRYPTO_EX structure and
16619 associated functions.
16623 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16624 as "read only": it can't be written to and the buffer it points to will
16625 not be freed. Reading from a read only BIO is much more efficient than
16626 a normal memory BIO. This was added because there are several times when
16627 an area of memory needs to be read from a BIO. The previous method was
16628 to create a memory BIO and write the data to it, this results in two
16629 copies of the data and an O(n^2) reading algorithm. There is a new
16630 function BIO_new_mem_buf() which creates a read only memory BIO from
16631 an area of memory. Also modified the PKCS#7 routines to use read only
16636 * Bugfix: ssl23_get_client_hello did not work properly when called in
16637 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16638 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16639 but a retry condition occurred while trying to read the rest.
16643 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16644 NID_pkcs7_encrypted by default: this was wrong since this should almost
16645 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16646 the encrypted data type: this is a more sensible place to put it and it
16647 allows the PKCS#12 code to be tidied up that duplicated this
16652 * Changed obj_dat.pl script so it takes its input and output files on
16653 the command line. This should avoid shell escape redirection problems
16658 * Initial support for certificate extension requests, these are included
16659 in things like Xenroll certificate requests. Included functions to allow
16660 extensions to be obtained and added.
16664 * -crlf option to s_client and s_server for sending newlines as
16665 CRLF (as required by many protocols).
16669 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16671 * Install libRSAglue.a when OpenSSL is built with RSAref.
16673 *Ralf S. Engelschall*
16675 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16677 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16679 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16684 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16685 DH parameters/keys (q is lost during that conversion, but the resulting
16686 DH parameters contain its length).
16688 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16689 much faster than DH_generate_parameters (which creates parameters
16690 where `p = 2*q + 1`), and also the smaller q makes DH computations
16691 much more efficient (160-bit exponentiation instead of 1024-bit
16692 exponentiation); so this provides a convenient way to support DHE
16693 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16694 utter importance to use
16695 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16697 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16698 when such DH parameters are used, because otherwise small subgroup
16699 attacks may become possible!
16703 * Avoid memory leak in i2d_DHparams.
16707 * Allow the -k option to be used more than once in the enc program:
16708 this allows the same encrypted message to be read by multiple recipients.
16712 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16713 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16714 it will always use the numerical form of the OID, even if it has a short
16719 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16720 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16721 otherwise bn_mod_exp was called. In the case of hardware keys for example
16722 no private key components need be present and it might store extra data
16723 in the RSA structure, which cannot be accessed from bn_mod_exp.
16724 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16725 private key operations.
16729 * Added support for SPARC Linux.
16733 * pem_password_cb function type incompatibly changed from
16734 typedef int pem_password_cb(char *buf, int size, int rwflag);
16736 ....(char *buf, int size, int rwflag, void *userdata);
16737 so that applications can pass data to their callbacks:
16738 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16739 additional void * argument, which is just handed through whenever
16740 the password callback is called.
16742 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16744 New function SSL_CTX_set_default_passwd_cb_userdata.
16746 Compatibility note: As many C implementations push function arguments
16747 onto the stack in reverse order, the new library version is likely to
16748 interoperate with programs that have been compiled with the old
16749 pem_password_cb definition (PEM_whatever takes some data that
16750 happens to be on the stack as its last argument, and the callback
16751 just ignores this garbage); but there is no guarantee whatsoever that
16754 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16755 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16756 problems not only on Windows, but also on some Unix platforms.
16757 To avoid problematic command lines, these definitions are now in an
16758 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16759 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16763 * MIPS III/IV assembler module is reimplemented.
16767 * More DES library cleanups: remove references to srand/rand and
16768 delete an unused file.
16772 * Add support for the free Netwide assembler (NASM) under Win32,
16773 since not many people have MASM (ml) and it can be hard to obtain.
16774 This is currently experimental but it seems to work OK and pass all
16775 the tests. Check out INSTALL.W32 for info.
16779 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16780 without temporary keys kept an extra copy of the server key,
16781 and connections with temporary keys did not free everything in case
16786 * New function RSA_check_key and new openssl rsa option -check
16787 for verifying the consistency of RSA keys.
16789 *Ulf Moeller, Bodo Moeller*
16791 * Various changes to make Win32 compile work:
16792 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16793 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16794 comparison" warnings.
16795 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16799 * Add a debugging option to PKCS#5 v2 key generation function: when
16800 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16801 derived keys are printed to stderr.
16805 * Copy the flags in ASN1_STRING_dup().
16807 *Roman E. Pavlov <pre@mo.msk.ru>*
16809 * The x509 application mishandled signing requests containing DSA
16810 keys when the signing key was also DSA and the parameters didn't match.
16812 It was supposed to omit the parameters when they matched the signing key:
16813 the verifying software was then supposed to automatically use the CA's
16814 parameters if they were absent from the end user certificate.
16816 Omitting parameters is no longer recommended. The test was also
16817 the wrong way round! This was probably due to unusual behaviour in
16818 EVP_cmp_parameters() which returns 1 if the parameters match.
16819 This meant that parameters were omitted when they *didn't* match and
16820 the certificate was useless. Certificates signed with 'ca' didn't have
16823 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16825 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16826 The interface is as follows:
16827 Applications can use
16828 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16829 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16830 "off" is now the default.
16831 The library internally uses
16832 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16833 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16834 to disable memory-checking temporarily.
16836 Some inconsistent states that previously were possible (and were
16837 even the default) are now avoided.
16839 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16840 with each memory chunk allocated; this is occasionally more helpful
16841 than just having a counter.
16843 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16845 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16850 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16851 which largely parallels "options", but is for changing API behaviour,
16852 whereas "options" are about protocol behaviour.
16853 Initial "mode" flags are:
16855 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16856 a single record has been written.
16857 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16858 retries use the same buffer location.
16859 (But all of the contents must be
16864 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16867 * Fix problems with no-hmac etc.
16869 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16871 * New functions RSA_get_default_method(), RSA_set_method() and
16872 RSA_get_method(). These allows replacement of RSA_METHODs without having
16873 to mess around with the internals of an RSA structure.
16877 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16878 Also really enable memory leak checks in openssl.c and in some
16881 *Chad C. Mulligan, Bodo Moeller*
16883 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16884 up the length of negative integers. This has now been simplified to just
16885 store the length when it is first determined and use it later, rather
16886 than trying to keep track of where data is copied and updating it to
16888 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16890 * Add a new function PKCS7_signatureVerify. This allows the verification
16891 of a PKCS#7 signature but with the signing certificate passed to the
16892 function itself. This contrasts with PKCS7_dataVerify which assumes the
16893 certificate is present in the PKCS#7 structure. This isn't always the
16894 case: certificates can be omitted from a PKCS#7 structure and be
16895 distributed by "out of band" means (such as a certificate database).
16899 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16900 function prototypes in pem.h, also change util/mkdef.pl to add the
16901 necessary function names.
16905 * mk1mf.pl (used by Windows builds) did not properly read the
16906 options set by Configure in the top level Makefile, and Configure
16907 was not even able to write more than one option correctly.
16908 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16912 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16913 file to be loaded from a BIO or FILE pointer. The BIO version will
16914 for example allow memory BIOs to contain config info.
16918 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16919 Whoever hopes to achieve shared-library compatibility across versions
16920 must use this, not the compile-time macro.
16921 (Exercise 0.9.4: Which is the minimum library version required by
16923 Note: All this applies only to multi-threaded programs, others don't
16928 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16929 through a BIO pair triggered the default case, i.e.
16930 SSLerr(...,SSL_R_UNKNOWN_STATE).
16934 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16935 can use the SSL library even if none of the specific BIOs is
16940 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16941 for the encoded length.
16943 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16945 * Add initial documentation of the X509V3 functions.
16949 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16950 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16951 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16952 secure PKCS#8 private key format with a high iteration count.
16956 * Fix determination of Perl interpreter: A perl or perl5
16957 *directory* in $PATH was also accepted as the interpreter.
16959 *Ralf S. Engelschall*
16961 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16962 wrong with it but it was very old and did things like calling
16963 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16964 unusual formatting.
16968 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16969 to use the new extension code.
16973 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16974 with macros. This should make it easier to change their form, add extra
16975 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16980 * Add to configuration table a new entry that can specify an alternative
16981 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16982 according to Mark Crispin <MRC@Panda.COM>.
16986 * DES CBC did not update the IV. Weird.
16990 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16991 Changing the behaviour of the former might break existing programs --
16992 where IV updating is needed, des_ncbc_encrypt can be used.
16995 * When bntest is run from "make test" it drives bc to check its
16996 calculations, as well as internally checking them. If an internal check
16997 fails, it needs to cause bc to give a non-zero result or make test carries
16998 on without noticing the failure. Fixed.
17002 * DES library cleanups.
17006 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17007 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17008 ciphers. NOTE: although the key derivation function has been verified
17009 against some published test vectors it has not been extensively tested
17010 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17015 * Instead of "mkdir -p", which is not fully portable, use new
17016 Perl script "util/mkdir-p.pl".
17020 * Rewrite the way password based encryption (PBE) is handled. It used to
17021 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17022 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17023 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17024 the 'parameter' field of the AlgorithmIdentifier is passed to the
17025 underlying key generation function so it must do its own ASN1 parsing.
17026 This has also changed the EVP_PBE_CipherInit() function which now has a
17027 'parameter' argument instead of literal salt and iteration count values
17028 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17032 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17033 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17034 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17035 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17036 value was just used as a "magic string" and not used directly its
17037 value doesn't matter.
17041 * Introduce some semblance of const correctness to BN. Shame C doesn't
17046 * "linux-sparc64" configuration (ultrapenguin).
17048 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17049 "linux-sparc" configuration.
17051 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17053 * config now generates no-xxx options for missing ciphers.
17057 * Support the EBCDIC character set (work in progress).
17058 File ebcdic.c not yet included because it has a different license.
17060 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17062 * Support BS2000/OSD-POSIX.
17064 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17066 * Make callbacks for key generation use `void *` instead of `char *`.
17070 * Make S/MIME samples compile (not yet tested).
17074 * Additional typesafe stacks.
17078 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17082 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17084 * New configuration variant "sco5-gcc".
17086 * Updated some demos.
17088 *Sean O Riordain, Wade Scholine*
17090 * Add missing BIO_free at exit of pkcs12 application.
17094 * Fix memory leak in conf.c.
17098 * Updates for Win32 to assembler version of MD5.
17102 * Set #! path to perl in apps/der_chop to where we found it
17103 instead of using a fixed path.
17107 * SHA library changes for irix64-mips4-cc.
17111 * Improvements for VMS support.
17115 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17117 * Bignum library bug fix. IRIX 6 passes "make test" now!
17118 This also avoids the problems with SC4.2 and unpatched SC5.
17120 *Andy Polyakov <appro@fy.chalmers.se>*
17122 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17123 These are required because of the typesafe stack would otherwise break
17124 existing code. If old code used a structure member which used to be STACK
17125 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17126 sk_num or sk_value it would produce an error because the num, data members
17127 are not present in STACK_OF. Now it just produces a warning. sk_set
17128 replaces the old method of assigning a value to sk_value
17129 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17130 that does this will no longer work (and should use sk_set instead) but
17131 this could be regarded as a "questionable" behaviour anyway.
17135 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17136 correctly handle encrypted S/MIME data.
17140 * Change type of various DES function arguments from des_cblock
17141 (which means, in function argument declarations, pointer to char)
17142 to des_cblock * (meaning pointer to array with 8 char elements),
17143 which allows the compiler to do more typechecking; it was like
17144 that back in SSLeay, but with lots of ugly casts.
17146 Introduce new type const_des_cblock.
17150 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17151 problems: find RecipientInfo structure that matches recipient certificate
17152 and initialise the ASN1 structures properly based on passed cipher.
17156 * Belatedly make the BN tests actually check the results.
17160 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17161 to and from BNs: it was completely broken. New compilation option
17162 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17163 key elements as negative integers.
17167 * Reorganize and speed up MD5.
17169 *Andy Polyakov <appro@fy.chalmers.se>*
17173 *Richard Levitte <richard@levitte.org>*
17175 * New option -out to asn1parse to allow the parsed structure to be
17176 output to a file. This is most useful when combined with the -strparse
17177 option to examine the output of things like OCTET STRINGS.
17181 * Make SSL library a little more fool-proof by not requiring any longer
17182 that `SSL_set_{accept,connect}_state` be called before
17183 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17184 in many applications because usually everything *appeared* to work as
17185 intended anyway -- now it really works as intended).
17189 * Move openssl.cnf out of lib/.
17193 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17194 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17195 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17197 *Ralf S. Engelschall*
17199 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17200 handle PKCS#7 enveloped data properly.
17202 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17204 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17205 copying pointers. The cert_st handling is changed by this in
17206 various ways (and thus what used to be known as ctx->default_cert
17207 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17208 any longer when s->cert does not give us what we need).
17209 ssl_cert_instantiate becomes obsolete by this change.
17210 As soon as we've got the new code right (possibly it already is?),
17211 we have solved a couple of bugs of the earlier code where s->cert
17212 was used as if it could not have been shared with other SSL structures.
17214 Note that using the SSL API in certain dirty ways now will result
17215 in different behaviour than observed with earlier library versions:
17216 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17217 does not influence s as it used to.
17219 In order to clean up things more thoroughly, inside SSL_SESSION
17220 we don't use CERT any longer, but a new structure SESS_CERT
17221 that holds per-session data (if available); currently, this is
17222 the peer's certificate chain and, for clients, the server's certificate
17223 and temporary key. CERT holds only those values that can have
17224 meaningful defaults in an SSL_CTX.
17228 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17229 from the internal representation. Various PKCS#7 fixes: remove some
17230 evil casts and set the enc_dig_alg field properly based on the signing
17235 * Allow PKCS#12 password to be set from the command line or the
17236 environment. Let 'ca' get its config file name from the environment
17237 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17242 * Allow certificate policies extension to use an IA5STRING for the
17243 organization field. This is contrary to the PKIX definition but
17244 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17249 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17250 without disallowing inline assembler and the like for non-pedantic builds.
17254 * Support Borland C++ builder.
17256 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17262 * SHA-1 cleanups and performance enhancements.
17264 *Andy Polyakov <appro@fy.chalmers.se>*
17266 * Sparc v8plus assembler for the bignum library.
17268 *Andy Polyakov <appro@fy.chalmers.se>*
17270 * Accept any -xxx and +xxx compiler options in Configure.
17274 * Update HPUX configuration.
17278 * Add missing `sk_<type>_unshift()` function to safestack.h
17280 *Ralf S. Engelschall*
17282 * New function SSL_CTX_use_certificate_chain_file that sets the
17283 "extra_cert"s in addition to the certificate. (This makes sense
17284 only for "PEM" format files, as chains as a whole are not
17289 * Support verify_depth from the SSL API.
17290 x509_vfy.c had what can be considered an off-by-one-error:
17291 Its depth (which was not part of the external interface)
17292 was actually counting the number of certificates in a chain;
17293 now it really counts the depth.
17297 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17298 instead of X509err, which often resulted in confusing error
17299 messages since the error codes are not globally unique
17300 (e.g. an alleged error in ssl3_accept when a certificate
17301 didn't match the private key).
17303 * New function SSL_CTX_set_session_id_context that allows to set a default
17304 value (so that you don't need SSL_set_session_id_context for each
17305 connection using the SSL_CTX).
17309 * OAEP decoding bug fix.
17313 * Support INSTALL_PREFIX for package builders, as proposed by
17318 * New Configure options "threads" and "no-threads". For systems
17319 where the proper compiler options are known (currently Solaris
17320 and Linux), "threads" is the default.
17324 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17328 * Install various scripts to $(OPENSSLDIR)/misc, not to
17329 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17330 such as /usr/local/bin.
17334 * "make linux-shared" to build shared libraries.
17336 *Niels Poppe <niels@netbox.org>*
17338 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17342 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17343 extension adding in x509 utility.
17347 * Remove NOPROTO sections and error code comments.
17351 * Partial rewrite of the DEF file generator to now parse the ANSI
17356 * New Configure options --prefix=DIR and --openssldir=DIR.
17360 * Complete rewrite of the error code script(s). It is all now handled
17361 by one script at the top level which handles error code gathering,
17362 header rewriting and C source file generation. It should be much better
17363 than the old method: it now uses a modified version of Ulf's parser to
17364 read the ANSI prototypes in all header files (thus the old K&R definitions
17365 aren't needed for error creation any more) and do a better job of
17366 translating function codes into names. The old 'ASN1 error code embedded
17367 in a comment' is no longer necessary and it doesn't use .err files which
17368 have now been deleted. Also the error code call doesn't have to appear all
17369 on one line (which resulted in some large lines...).
17373 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17377 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17378 0 (which usually indicates a closed connection), but continue reading.
17382 * Fix some race conditions.
17386 * Add support for CRL distribution points extension. Add Certificate
17387 Policies and CRL distribution points documentation.
17391 * Move the autogenerated header file parts to crypto/opensslconf.h.
17395 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17396 8 of keying material. Merlin has also confirmed interop with this fix
17397 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17399 *Merlin Hughes <merlin@baltimore.ie>*
17401 * Fix lots of warnings.
17403 *Richard Levitte <levitte@stacken.kth.se>*
17405 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17406 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17408 *Richard Levitte <levitte@stacken.kth.se>*
17410 * Fix problems with sizeof(long) == 8.
17412 *Andy Polyakov <appro@fy.chalmers.se>*
17414 * Change functions to ANSI C.
17418 * Fix typos in error codes.
17420 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17422 * Remove defunct assembler files from Configure.
17426 * SPARC v8 assembler BIGNUM implementation.
17428 *Andy Polyakov <appro@fy.chalmers.se>*
17430 * Support for Certificate Policies extension: both print and set.
17431 Various additions to support the r2i method this uses.
17435 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17436 return a const string when you are expecting an allocated buffer.
17440 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17441 types DirectoryString and DisplayText.
17445 * Add code to allow r2i extensions to access the configuration database,
17446 add an LHASH database driver and add several ctx helper functions.
17450 * Fix an evil bug in bn_expand2() which caused various BN functions to
17451 fail when they extended the size of a BIGNUM.
17455 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17456 support typesafe stack.
17460 * Fix typo in SSL_[gs]et_options().
17462 *Nils Frostberg <nils@medcom.se>*
17464 * Delete various functions and files that belonged to the (now obsolete)
17465 old X509V3 handling code.
17469 * New Configure option "rsaref".
17473 * Don't auto-generate pem.h.
17477 * Introduce type-safe ASN.1 SETs.
17481 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17483 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17485 * Introduce type-safe STACKs. This will almost certainly break lots of code
17486 that links with OpenSSL (well at least cause lots of warnings), but fear
17487 not: the conversion is trivial, and it eliminates loads of evil casts. A
17488 few STACKed things have been converted already. Feel free to convert more.
17489 In the fullness of time, I'll do away with the STACK type altogether.
17493 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17494 specified in `<certfile>` by updating the entry in the index.txt file.
17495 This way one no longer has to edit the index.txt file manually for
17496 revoking a certificate. The -revoke option does the gory details now.
17498 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17500 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17501 `-text` option at all and this way the `-noout -text` combination was
17502 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17504 *Ralf S. Engelschall*
17506 * Make sure a corresponding plain text error message exists for the
17507 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17508 verify callback function determined that a certificate was revoked.
17510 *Ralf S. Engelschall*
17512 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17513 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17514 all available ciphers including rc5, which was forgotten until now.
17515 In order to let the testing shell script know which algorithms
17516 are available, a new (up to now undocumented) command
17517 `openssl list-cipher-commands` is used.
17521 * Bugfix: s_client occasionally would sleep in select() when
17522 it should have checked SSL_pending() first.
17526 * New functions DSA_do_sign and DSA_do_verify to provide access to
17527 the raw DSA values prior to ASN.1 encoding.
17531 * Tweaks to Configure
17533 *Niels Poppe <niels@netbox.org>*
17535 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17540 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17544 * New config option to avoid instructions that are illegal on the 80386.
17545 The default code is faster, but requires at least a 486.
17549 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17550 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17551 same as SSL2_VERSION anyway.
17555 * New "-showcerts" option for s_client.
17559 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17560 application. Various cleanups and fixes.
17564 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17565 modify error routines to work internally. Add error codes and PBE init
17566 to library startup routines.
17570 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17571 packing functions to asn1 and evp. Changed function names and error
17572 codes along the way.
17576 * PKCS12 integration: and so it begins... First of several patches to
17577 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17578 objects to objects.h
17582 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17583 and display support for Thawte strong extranet extension.
17587 * Add LinuxPPC support.
17589 *Jeff Dubrule <igor@pobox.org>*
17591 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17592 bn_div_words in alpha.s.
17594 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17596 * Make sure the RSA OAEP test is skipped under -DRSAref because
17597 OAEP isn't supported when OpenSSL is built with RSAref.
17599 *Ulf Moeller <ulf@fitug.de>*
17601 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17602 so they no longer are missing under -DNOPROTO.
17604 *Soren S. Jorvang <soren@t.dk>*
17606 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17608 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17609 doesn't work when the session is reused. Coming soon!
17613 * Fix a security hole, that allows sessions to be reused in the wrong
17614 context thus bypassing client cert protection! All software that uses
17615 client certs and session caches in multiple contexts NEEDS PATCHING to
17616 allow session reuse! A fuller solution is in the works.
17618 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17620 * Some more source tree cleanups (removed obsolete files
17621 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17622 permission on "config" script to be executable) and a fix for the INSTALL
17625 *Ulf Moeller <ulf@fitug.de>*
17627 * Remove some legacy and erroneous uses of malloc, free instead of
17630 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17632 * Make rsa_oaep_test return non-zero on error.
17634 *Ulf Moeller <ulf@fitug.de>*
17636 * Add support for native Solaris shared libraries. Configure
17637 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17638 if someone would make that last step automatic.
17640 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17642 * ctx_size was not built with the right compiler during "make links". Fixed.
17646 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17647 except NULL ciphers". This means the default cipher list will no longer
17648 enable NULL ciphers. They need to be specifically enabled e.g. with
17649 the string "DEFAULT:eNULL".
17653 * Fix to RSA private encryption routines: if p < q then it would
17654 occasionally produce an invalid result. This will only happen with
17655 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17659 * Be less restrictive and allow also `perl util/perlpath.pl
17660 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17661 because this way one can also use an interpreter named `perl5` (which is
17662 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17663 installed as `perl`).
17665 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17667 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17669 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17671 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17672 advapi32.lib to Win32 build and change the pem test comparison
17673 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17674 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17675 and crypto/des/ede_cbcm_enc.c.
17679 * DES quad checksum was broken on big-endian architectures. Fixed.
17683 * Comment out two functions in bio.h that aren't implemented. Fix up the
17684 Win32 test batch file so it (might) work again. The Win32 test batch file
17685 is horrible: I feel ill....
17689 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17690 in e_os.h. Audit of header files to check ANSI and non ANSI
17691 sections: 10 functions were absent from non ANSI section and not exported
17692 from Windows DLLs. Fixed up libeay.num for new functions.
17696 * Make `openssl version` output lines consistent.
17698 *Ralf S. Engelschall*
17700 * Fix Win32 symbol export lists for BIO functions: Added
17701 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17702 to ms/libeay{16,32}.def.
17704 *Ralf S. Engelschall*
17706 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17707 fine under Unix and passes some trivial tests I've now added. But the
17708 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17709 added to make sure no one expects that this stuff really works in the
17710 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17711 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17714 *Ralf S. Engelschall*
17716 * Fix the generation of two part addresses in perl.
17718 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17720 * Add config entry for Linux on MIPS.
17722 *John Tobey <jtobey@channel1.com>*
17724 * Make links whenever Configure is run, unless we are on Windoze.
17728 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17729 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17734 * Add a useful kludge to allow package maintainers to specify compiler and
17735 other platforms details on the command line without having to patch the
17736 Configure script every time: One now can use
17737 `perl Configure <id>:<details>`,
17738 i.e. platform ids are allowed to have details appended
17739 to them (separated by colons). This is treated as there would be a static
17740 pre-configured entry in Configure's %table under key `<id>` with value
17741 `<details>` and `perl Configure <id>` is called. So, when you want to
17742 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17743 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17744 now, which overrides the FreeBSD-elf entry on-the-fly.
17746 *Ralf S. Engelschall*
17748 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17752 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17753 on the `perl Configure ...` command line. This way one can compile
17754 OpenSSL libraries with Position Independent Code (PIC) which is needed
17755 for linking it into DSOs.
17757 *Ralf S. Engelschall*
17759 * Remarkably, export ciphers were totally broken and no-one had noticed!
17764 * Cleaned up the LICENSE document: The official contact for any license
17765 questions now is the OpenSSL core team under openssl-core@openssl.org.
17766 And add a paragraph about the dual-license situation to make sure people
17767 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17768 to the OpenSSL toolkit.
17770 *Ralf S. Engelschall*
17772 * General source tree makefile cleanups: Made `making xxx in yyy...`
17773 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17774 Additionally cleaned up the `make links` target: Remove unnecessary
17775 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17776 to speed processing and no longer clutter the display with confusing
17777 stuff. Instead only the actually done links are displayed.
17779 *Ralf S. Engelschall*
17781 * Permit null encryption ciphersuites, used for authentication only. It used
17782 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17783 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17788 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17789 signed attributes when verifying signatures (this would break them),
17790 the detached data encoding was wrong and public keys obtained using
17791 X509_get_pubkey() weren't freed.
17795 * Add text documentation for the BUFFER functions. Also added a work around
17796 to a Win95 console bug. This was triggered by the password read stuff: the
17797 last character typed gets carried over to the next fread(). If you were
17798 generating a new cert request using 'req' for example then the last
17799 character of the passphrase would be CR which would then enter the first
17804 * Added the new 'Includes OpenSSL Cryptography Software' button as
17805 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17806 button and can be used by applications based on OpenSSL to show the
17807 relationship to the OpenSSL project.
17809 *Ralf S. Engelschall*
17811 * Remove confusing variables in function signatures in files
17812 ssl/ssl_lib.c and ssl/ssl.h.
17814 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17816 * Don't install bss_file.c under PREFIX/include/
17818 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17820 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17821 functions that return function pointers and has support for NT specific
17822 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17823 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17824 unsigned to signed types: this was killing the Win32 compile.
17828 * Add new certificate file to stack functions,
17829 SSL_add_dir_cert_subjects_to_stack() and
17830 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17831 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17832 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17833 This means that Apache-SSL and similar packages don't have to mess around
17834 to add as many CAs as they want to the preferred list.
17838 * Experiment with doxygen documentation. Currently only partially applied to
17840 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17841 openssl.doxy as the configuration file.
17845 * Get rid of remaining C++-style comments which strict C compilers hate.
17847 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17849 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17850 compiled in by default: it has problems with large keys.
17854 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17855 DH private keys and/or callback functions which directly correspond to
17856 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17857 is needed for applications which have to configure certificates on a
17858 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17860 For the RSA certificate situation is makes no difference, but
17861 for the DSA certificate situation this fixes the "no shared cipher"
17862 problem where the OpenSSL cipher selection procedure failed because the
17863 temporary keys were not overtaken from the context and the API provided
17864 no way to reconfigure them.
17865 The new functions now let applications reconfigure the stuff and they
17866 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17867 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17868 non-public-API function ssl_cert_instantiate() is used as a helper
17869 function and also to reduce code redundancy inside ssl_rsa.c.
17871 *Ralf S. Engelschall*
17873 * Move s_server -dcert and -dkey options out of the undocumented feature
17874 area because they are useful for the DSA situation and should be
17875 recognized by the users.
17877 *Ralf S. Engelschall*
17879 * Fix the cipher decision scheme for export ciphers: the export bits are
17880 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17881 SSL_EXP_MASK. So, the original variable has to be used instead of the
17882 already masked variable.
17884 *Richard Levitte <levitte@stacken.kth.se>*
17886 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17888 *Richard Levitte <levitte@stacken.kth.se>*
17890 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17891 from `int` to `unsigned int` because it is a length and initialized by
17892 EVP_DigestFinal() which expects an `unsigned int *`.
17894 *Richard Levitte <levitte@stacken.kth.se>*
17896 * Don't hard-code path to Perl interpreter on shebang line of Configure
17897 script. Instead use the usual Shell->Perl transition trick.
17899 *Ralf S. Engelschall*
17901 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17902 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17903 -noout -modulus` as it's already the case for `openssl rsa -noout
17904 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17905 currently the public key is printed (a decision which was already done by
17906 `openssl dsa -modulus` in the past) which serves a similar purpose.
17907 Additionally the NO_RSA no longer completely removes the whole -modulus
17908 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17911 *Ralf S. Engelschall*
17913 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17914 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17916 *Arne Ansper <arne@ats.cyber.ee>*
17918 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17919 to be added. Now both 'req' and 'ca' can use new objects defined in the
17924 * Add cool BIO that does syslog (or event log on NT).
17926 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17928 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17929 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17930 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17931 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17935 * Add preliminary config info for new extension code.
17939 * Make RSA_NO_PADDING really use no padding.
17941 *Ulf Moeller <ulf@fitug.de>*
17943 * Generate errors when private/public key check is done.
17947 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17948 for some CRL extensions and new objects added.
17952 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17953 key usage extension and fuller support for authority key id.
17957 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17958 padding method for RSA, which is recommended for new applications in PKCS
17959 #1 v2.0 (RFC 2437, October 1998).
17960 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17961 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17962 against Bleichbacher's attack on RSA.
17963 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17966 * Updates to the new SSL compression code
17968 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17970 * Fix so that the version number in the master secret, when passed
17971 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17972 (because the server will not accept higher), that the version number
17973 is 0x03,0x01, not 0x03,0x00
17975 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17977 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17978 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17979 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17983 * Support for RAW extensions where an arbitrary extension can be
17984 created by including its DER encoding. See apps/openssl.cnf for
17989 * Make sure latest Perl versions don't interpret some generated C array
17990 code as Perl array code in the crypto/err/err_genc.pl script.
17992 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17994 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17995 not many people have the assembler. Various Win32 compilation fixes and
17996 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17997 build instructions.
18001 * Modify configure script 'Configure' to automatically create crypto/date.h
18002 file under Win32 and also build pem.h from pem.org. New script
18003 util/mkfiles.pl to create the MINFO file on environments that can't do a
18004 'make files': perl util/mkfiles.pl >MINFO should work.
18008 * Major rework of DES function declarations, in the pursuit of correctness
18009 and purity. As a result, many evil casts evaporated, and some weirdness,
18010 too. You may find this causes warnings in your code. Zapping your evil
18011 casts will probably fix them. Mostly.
18015 * Fix for a typo in asn1.h. Bug fix to object creation script
18016 obj_dat.pl. It considered a zero in an object definition to mean
18017 "end of object": none of the objects in objects.h have any zeros
18018 so it wasn't spotted.
18020 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18022 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18023 Masking (CBCM). In the absence of test vectors, the best I have been able
18024 to do is check that the decrypt undoes the encrypt, so far. Send me test
18025 vectors if you have them.
18029 * Correct calculation of key length for export ciphers (too much space was
18030 allocated for null ciphers). This has not been tested!
18034 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18035 message is now correct (it understands "crypto" and "ssl" on its
18036 command line). There is also now an "update" option. This will update
18037 the util/ssleay.num and util/libeay.num files with any new functions.
18039 perl util/mkdef.pl crypto ssl update
18040 it will update them.
18044 * Overhauled the Perl interface:
18045 - ported BN stuff to OpenSSL's different BN library
18046 - made the perl/ source tree CVS-aware
18047 - renamed the package from SSLeay to OpenSSL (the files still contain
18048 their history because I've copied them in the repository)
18049 - removed obsolete files (the test scripts will be replaced
18050 by better Test::Harness variants in the future)
18052 *Ralf S. Engelschall*
18054 * First cut for a very conservative source tree cleanup:
18055 1. merge various obsolete readme texts into doc/ssleay.txt
18056 where we collect the old documents and readme texts.
18057 2. remove the first part of files where I'm already sure that we no
18058 longer need them because of three reasons: either they are just temporary
18059 files which were left by Eric or they are preserved original files where
18060 I've verified that the diff is also available in the CVS via "cvs diff
18061 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18062 the crypto/md/ stuff).
18064 *Ralf S. Engelschall*
18066 * More extension code. Incomplete support for subject and issuer alt
18067 name, issuer and authority key id. Change the i2v function parameters
18068 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18069 what that's for :-) Fix to ASN1 macro which messed up
18070 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18074 * Preliminary support for ENUMERATED type. This is largely copied from the
18079 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18081 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18083 * Make sure `make rehash` target really finds the `openssl` program.
18085 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18087 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18088 like to hear about it if this slows down other processors.
18092 * Add CygWin32 platform information to Configure script.
18094 *Alan Batie <batie@aahz.jf.intel.com>*
18096 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18098 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18100 * New program nseq to manipulate netscape certificate sequences
18104 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18109 * Fixes to BN code. Previously the default was to define BN_RECURSION
18110 but the BN code had some problems that would cause failures when
18111 doing certificate verification and some other functions.
18113 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18115 * Add ASN1 and PEM code to support netscape certificate sequences.
18119 * Add ASN1 and PEM code to support netscape certificate sequences.
18123 * Add several PKIX and private extended key usage OIDs.
18127 * Modify the 'ca' program to handle the new extension code. Modify
18128 openssl.cnf for new extension format, add comments.
18132 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18133 and add a sample to openssl.cnf so req -x509 now adds appropriate
18138 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18139 error code, add initial support to X509_print() and x509 application.
18143 * Takes a deep breath and start adding X509 V3 extension support code. Add
18144 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18145 stuff is currently isolated and isn't even compiled yet.
18149 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18150 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18151 Removed the versions check from X509 routines when loading extensions:
18152 this allows certain broken certificates that don't set the version
18153 properly to be processed.
18157 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18158 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18159 can still be regenerated with "make depend".
18163 * Spelling mistake in C version of CAST-128.
18165 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18167 * Changes to the error generation code. The perl script err-code.pl
18168 now reads in the old error codes and retains the old numbers, only
18169 adding new ones if necessary. It also only changes the .err files if new
18170 codes are added. The makefiles have been modified to only insert errors
18171 when needed (to avoid needlessly modifying header files). This is done
18172 by only inserting errors if the .err file is newer than the auto generated
18173 C file. To rebuild all the error codes from scratch (the old behaviour)
18174 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18175 or delete all the .err files.
18179 * CAST-128 was incorrectly implemented for short keys. The C version has
18180 been fixed, but is untested. The assembler versions are also fixed, but
18181 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18182 to regenerate it if needed.
18183 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18184 Hagino <itojun@kame.net>*
18186 * File was opened incorrectly in randfile.c.
18188 *Ulf Möller <ulf@fitug.de>*
18190 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18191 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18192 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18193 al: it's just almost always a UTCTime. Note this patch adds new error
18194 codes so do a "make errors" if there are problems.
18198 * Correct Linux 1 recognition in config.
18200 *Ulf Möller <ulf@fitug.de>*
18202 * Remove pointless MD5 hash when using DSA keys in ca.
18204 *Anonymous <nobody@replay.com>*
18206 * Generate an error if given an empty string as a cert directory. Also
18207 generate an error if handed NULL (previously returned 0 to indicate an
18208 error, but didn't set one).
18210 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18212 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18216 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18217 parameters. This was causing a warning which killed off the Win32 compile.
18221 * Remove C++ style comments from crypto/bn/bn_local.h.
18223 *Neil Costigan <neil.costigan@celocom.com>*
18225 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18226 based on a text string, looking up short and long names and finally
18227 "dot" format. The "dot" format stuff didn't work. Added new function
18228 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18229 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18230 OID is not part of the table.
18234 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18235 X509_LOOKUP_by_alias().
18239 * Sort openssl functions by name.
18243 * Get the gendsa program working (hopefully) and add it to app list. Remove
18244 encryption from sample DSA keys (in case anyone is interested the password
18249 * Make *all* `*_free` functions accept a NULL pointer.
18251 *Frans Heymans <fheymans@isaserver.be>*
18253 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18256 *Anonymous <nobody@replay.com>*
18258 * s_server should send the CAfile as acceptable CAs, not its own cert.
18260 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18262 * Don't blow it for numeric -newkey arguments to apps/req.
18264 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18266 * Temp key "for export" tests were wrong in s3_srvr.c.
18268 *Anonymous <nobody@replay.com>*
18270 * Add prototype for temp key callback functions
18271 SSL_CTX_set_tmp_{rsa,dh}_callback().
18275 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18276 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18280 * X509_name_add_entry() freed the wrong thing after an error.
18282 *Arne Ansper <arne@ats.cyber.ee>*
18284 * rsa_eay.c would attempt to free a NULL context.
18286 *Arne Ansper <arne@ats.cyber.ee>*
18288 * BIO_s_socket() had a broken should_retry() on Windoze.
18290 *Arne Ansper <arne@ats.cyber.ee>*
18292 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18294 *Arne Ansper <arne@ats.cyber.ee>*
18296 * Make sure the already existing X509_STORE->depth variable is initialized
18297 in X509_STORE_new(), but document the fact that this variable is still
18298 unused in the certificate verification process.
18300 *Ralf S. Engelschall*
18302 * Fix the various library and apps files to free up pkeys obtained from
18303 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18307 * Fix reference counting in X509_PUBKEY_get(). This makes
18308 demos/maurice/example2.c work, amongst others, probably.
18310 *Steve Henson and Ben Laurie*
18312 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18313 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18314 are no longer created. This way we have a single and consistent command
18315 line interface `openssl <command>`, similar to `cvs <command>`.
18317 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18319 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18320 BIT STRING wrapper always have zero unused bits.
18324 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18328 * Make the top-level INSTALL documentation easier to understand.
18332 * Makefiles updated to exit if an error occurs in a sub-directory
18333 make (including if user presses ^C) [Paul Sutton]
18335 * Make Montgomery context stuff explicit in RSA data structure.
18339 * Fix build order of pem and err to allow for generated pem.h.
18343 * Fix renumbering bug in X509_NAME_delete_entry().
18347 * Enhanced the err-ins.pl script so it makes the error library number
18348 global and can add a library name. This is needed for external ASN1 and
18349 other error libraries.
18353 * Fixed sk_insert which never worked properly.
18357 * Fix ASN1 macros so they can handle indefinite length constructed
18358 EXPLICIT tags. Some non standard certificates use these: they can now
18363 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18364 into a single doc/ssleay.txt bundle. This way the information is still
18365 preserved but no longer messes up this directory. Now it's new room for
18366 the new set of documentation files.
18368 *Ralf S. Engelschall*
18370 * SETs were incorrectly DER encoded. This was a major pain, because they
18371 shared code with SEQUENCEs, which aren't coded the same. This means that
18372 almost everything to do with SETs or SEQUENCEs has either changed name or
18373 number of arguments.
18375 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18377 * Fix test data to work with the above.
18381 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18382 was already fixed by Eric for 0.9.1 it seems.
18384 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18386 * Autodetect FreeBSD3.
18390 * Fix various bugs in Configure. This affects the following platforms:
18394 unixware-2.0-pentium
18399 * Eliminate generated files from CVS. Reorder tests to regenerate files
18400 before they are needed.
18404 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18408 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18410 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18411 changed SSLeay to OpenSSL in version strings.
18413 *Ralf S. Engelschall*
18415 * Some fixups to the top-level documents.
18419 * Fixed the nasty bug where rsaref.h was not found under compile-time
18420 because the symlink to include/ was missing.
18422 *Ralf S. Engelschall*
18424 * Incorporated the popular no-RSA/DSA-only patches
18425 which allow to compile a RSA-free SSLeay.
18427 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18429 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18430 when "ssleay" is still not found.
18432 *Ralf S. Engelschall*
18434 * Added more platforms to Configure: Cray T3E, HPUX 11,
18436 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18438 * Updated the README file.
18440 *Ralf S. Engelschall*
18442 * Added various .cvsignore files in the CVS repository subdirs
18443 to make a "cvs update" really silent.
18445 *Ralf S. Engelschall*
18447 * Recompiled the error-definition header files and added
18448 missing symbols to the Win32 linker tables.
18450 *Ralf S. Engelschall*
18452 * Cleaned up the top-level documents;
18453 o new files: CHANGES and LICENSE
18454 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18455 o merged COPYRIGHT into LICENSE
18456 o removed obsolete TODO file
18457 o renamed MICROSOFT to INSTALL.W32
18459 *Ralf S. Engelschall*
18461 * Removed dummy files from the 0.9.1b source tree:
18462 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18463 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18464 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18465 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18466 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18468 *Ralf S. Engelschall*
18470 * Added various platform portability fixes.
18474 * The Genesis of the OpenSSL rpject:
18475 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18476 Young and Tim J. Hudson created while they were working for C2Net until
18479 *The OpenSSL Project*
18481 ### Changes between 0.9.0b and 0.9.1b [not released]
18483 * Updated a few CA certificates under certs/
18487 * Changed some BIGNUM api stuff.
18491 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18492 DGUX x86, Linux Alpha, etc.
18496 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18497 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18502 * Add -strparse option to asn1pars program which parses nested
18505 *Dr Stephen Henson <shenson@bigfoot.com>*
18507 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18511 * DSA fix for "ca" program.
18515 * Added "-genkey" option to "dsaparam" program.
18519 * Added RIPE MD160 (rmd160) message digest.
18523 * Added -a (all) option to "ssleay version" command.
18527 * Added PLATFORM define which is the id given to Configure.
18531 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18535 * Extended the ASN.1 parser routines.
18539 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18543 * Added a BN_CTX to the BN library.
18547 * Fixed the weak key values in DES library
18551 * Changed API in EVP library for cipher aliases.
18555 * Added support for RC2/64bit cipher.
18559 * Converted the lhash library to the crypto/mem.c functions.
18563 * Added more recognized ASN.1 object ids.
18567 * Added more RSA padding checks for SSL/TLS.
18571 * Added BIO proxy/filter functionality.
18575 * Added extra_certs to SSL_CTX which can be used
18576 send extra CA certificates to the client in the CA cert chain sending
18577 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18581 * Now Fortezza is denied in the authentication phase because
18582 this is key exchange mechanism is not supported by SSLeay at all.
18586 * Additional PKCS1 checks.
18590 * Support the string "TLSv1" for all TLS v1 ciphers.
18594 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18595 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18599 * Fixed a few memory leaks.
18603 * Fixed various code and comment typos.
18607 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18608 bytes sent in the client random.
18610 *Edward Bishop <ebishop@spyglass.com>*
18614 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18615 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18616 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18617 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18618 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18619 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18620 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18621 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18622 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18623 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18624 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18625 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18626 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18627 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18628 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18629 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18630 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18631 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18632 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18633 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18634 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18635 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18636 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18637 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18638 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18639 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18640 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18641 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18642 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18643 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18644 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18645 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18646 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18647 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18648 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18649 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18650 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18651 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18652 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18653 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18654 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18655 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18656 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18657 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18658 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18659 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18660 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18661 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18662 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18663 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18664 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18665 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18666 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18667 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18668 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18669 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18670 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18671 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18672 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18673 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18674 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18675 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18676 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18677 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18678 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18679 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18680 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18681 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18682 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18683 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18684 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18685 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18686 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18687 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18688 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18689 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18690 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18691 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18692 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18693 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18694 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18695 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18696 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18697 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18698 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18699 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18700 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18701 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18702 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18703 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18704 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18705 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18706 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18707 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18708 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18709 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18710 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18711 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18712 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18713 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18714 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18715 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18716 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18717 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18718 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18719 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18720 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18721 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18722 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18723 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18724 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18725 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18726 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18727 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18728 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18729 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18730 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18731 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18732 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18733 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18734 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18735 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18736 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18737 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18738 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18739 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18740 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18741 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18742 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18743 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18744 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18745 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18746 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18747 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18748 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18749 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18750 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18751 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18752 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18753 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18754 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18755 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18756 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18757 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18758 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18759 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18760 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18761 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18762 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18763 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18764 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18765 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18766 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18767 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18768 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18769 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18770 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18771 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18772 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18773 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18774 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655