4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
27 interface. Their functionality remains unchanged.
31 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
32 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
33 type is internally recognised so the workaround is no longer needed.
35 Functionality is still retained as it is, but will only work with
36 EVP_PKEYs with a legacy internal key.
40 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
41 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
45 * Changed all "STACK" functions to be macros instead of inline functions. Macro
46 parameters are still checked for type safety at compile time via helper
51 * Remove the RAND_DRBG API
53 The RAND_DRBG API did not fit well into the new provider concept as
54 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
55 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
56 and some of its API calls are rather low-level. This holds in particular
57 for the callback mechanism (RAND_DRBG_set_callbacks()).
59 Adding a compatibility layer to continue supporting the RAND_DRBG API as
60 a legacy API for a regular deprecation period turned out to come at the
61 price of complicating the new provider API unnecessarily. Since the
62 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
65 *Paul Dale and Matthias St. Pierre*
67 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
68 as well as actual hostnames.
72 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
73 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
74 conversely, silently ignore DTLS protocol version bounds when configuring
75 TLS-based contexts. The commands can be repeated to set bounds of both
76 types. The same applies with the corresponding "min_protocol" and
77 "max_protocol" command-line switches, in case some application uses both TLS
80 SSL_CTX instances that are created for a fixed protocol version (e.g.
81 TLSv1_server_method()) also silently ignore version bounds. Previously
82 attempts to apply bounds to these protocol versions would result in an
83 error. Now only the "version-flexible" SSL_CTX instances are subject to
84 limits in configuration files in command-line options.
88 * Deprecated the `ENGINE` API. Engines should be replaced with providers
93 * Reworked the recorded ERR codes to make better space for system errors.
94 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
95 given code is a system error (true) or an OpenSSL error (false).
99 * Reworked the test perl framework to better allow parallel testing.
101 *Nicola Tuveri and David von Oheimb*
103 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
104 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
108 * 'Configure' has been changed to figure out the configuration target if
109 none is given on the command line. Consequently, the 'config' script is
110 now only a mere wrapper. All documentation is changed to only mention
113 *Rich Salz and Richard Levitte*
115 * Added a library context `OSSL_LIB_CTX` that applications as well as
116 other libraries can use to form a separate context within which
117 libcrypto operations are performed.
119 There are two ways this can be used:
121 - Directly, by passing a library context to functions that take
122 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
124 - Indirectly, by creating a new library context and then assigning
125 it as the new default, with `OSSL_LIB_CTX_set0_default`.
127 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
128 apart from the functions directly related to `OSSL_LIB_CTX`, accept
129 NULL to indicate that the default library context should be used.
131 Library code that changes the default library context using
132 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
133 second call before returning to the caller.
135 _(Note: the library context was initially called `OPENSSL_CTX` and
136 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
140 * Handshake now fails if Extended Master Secret extension is dropped
145 * Dropped interactive mode from the 'openssl' program. From now on,
146 the `openssl` command without arguments is equivalent to `openssl
151 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
152 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
153 While the old function names have been retained for backward compatibility
154 they should not be used in new developments
155 because their return values are confusing: Unlike other `_cmp()` functions
156 they do not return 0 in case their arguments are equal.
160 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
161 EC_GROUP_get_field_type().
165 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
166 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
167 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
168 Applications should rely on the library automatically assigning a suitable
169 EC_METHOD internally upon EC_GROUP construction.
173 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
174 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
175 assigned internally without application intervention.
176 Users of EC_GROUP_new() should switch to a different suitable constructor.
180 * Add CAdES-BES signature verification support, mostly derived
181 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
183 *Filipe Raimundo da Silva*
185 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
189 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
190 functions are not widely used and now OpenSSL automatically perform this
191 conversion when needed.
195 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
196 EC_KEY_precompute_mult(). These functions are not widely used and
197 applications should instead switch to named curves which OpenSSL has
198 hardcoded lookup tables for.
202 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
203 should instead use the L<EC_POINT_mul(3)> function.
207 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
208 that are not applicable to the new provider model. Applications should
209 instead use EVP_default_properties_is_fips_enabled() and
210 EVP_default_properties_enable_fips().
214 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
215 is set, an unexpected EOF is ignored, it pretends a close notify was received
216 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
220 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
221 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
222 used and applications should instead use the
223 L<EC_POINT_set_affine_coordinates(3)> and
224 L<EC_POINT_get_affine_coordinates(3)> functions.
228 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
229 arrays to be more easily constructed via a series of utility functions.
230 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
231 the various push functions and finally convert to a passable OSSL_PARAM
232 array using OSSL_PARAM_BLD_to_param().
236 * The security strength of SHA1 and MD5 based signatures in TLS has been
237 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
238 working at the default security level of 1 and instead requires security
239 level 0. The security level can be changed either using the cipher string
240 with @SECLEVEL, or calling SSL_CTX_set_security_level().
244 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
245 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
246 internal keys, if they correspond to one of those built in types.
250 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
251 contain a provider side internal key.
255 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
256 They are old functions that we don't use, and that you could disable with
257 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
261 * Project text documents not yet having a proper file name extension
262 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
263 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
267 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
268 have been converted to Markdown with the goal to produce documents
269 which not only look pretty when viewed online in the browser, but
270 remain well readable inside a plain text editor.
272 To achieve this goal, a 'minimalistic' Markdown style has been applied
273 which avoids formatting elements that interfere too much with the
274 reading flow in the text file. For example, it
276 * avoids [ATX headings][] and uses [setext headings][] instead
277 (which works for `<h1>` and `<h2>` headings only).
278 * avoids [inline links][] and uses [reference links][] instead.
279 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
281 [ATX headings]: https://github.github.com/gfm/#atx-headings
282 [setext headings]: https://github.github.com/gfm/#setext-headings
283 [inline links]: https://github.github.com/gfm/#inline-link
284 [reference links]: https://github.github.com/gfm/#reference-link
285 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
286 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
288 *Matthias St. Pierre*
290 * The test suite is changed to preserve results of each test recipe.
291 A new directory test-runs/ with subdirectories named like the
292 test recipes are created in the build tree for this purpose.
296 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
297 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
298 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
300 *David von Oheimb, Martin Peylo*
302 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
303 The legacy OCSP-focused and only partly documented API is retained.
304 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
308 * Added `util/check-format.pl`, a tool for checking adherence to the
309 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
310 The checks performed are incomplete and yield some false positives.
311 Still the tool should be useful for detecting most typical glitches.
315 * BIO_do_connect and BIO_do_handshake have been extended:
316 If domain name resolution yields multiple IP addresses all of them are tried
317 after connect() failures.
321 * All of the low level RSA functions have been deprecated including:
323 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
324 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
325 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
326 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
327 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
328 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
329 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
330 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
331 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
332 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
333 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
334 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
335 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
336 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
337 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
338 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
339 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
340 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
341 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
342 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
343 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
344 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
345 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
346 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
347 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
348 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
349 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
350 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
352 Use of these low level functions has been informally discouraged for a long
353 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
354 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
355 L<EVP_PKEY_decrypt(3)>.
359 * X509 certificates signed using SHA1 are no longer allowed at security
361 In TLS/SSL the default security level is 1. It can be set either
362 using the cipher string with @SECLEVEL, or calling
363 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
364 a call to SSL_CTX_use_certificate() will fail if the security level is not
366 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
367 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
372 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
373 modified to use PKEY APIs. These commands are now in maintenance mode
374 and no new features will be added to them.
378 * The command line utility rsautl has been deprecated.
379 Instead use the pkeyutl program.
383 * The command line utilities genrsa and rsa have been modified to use PKEY
384 APIs. They now write PKCS#8 keys by default. These commands are now in
385 maintenance mode and no new features will be added to them.
389 * All of the low level DH functions have been deprecated including:
391 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
392 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
393 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
394 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
395 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
396 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
397 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
398 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
399 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
400 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
401 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
402 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
403 DH_meth_get_generate_params and DH_meth_set_generate_params.
405 Use of these low level functions has been informally discouraged for a long
406 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
407 and L<EVP_PKEY_derive(3)>.
411 * All of the low level DSA functions have been deprecated including:
413 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
414 DSA_get_default_method, DSA_set_method, DSA_get_method,
415 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
416 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
417 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
418 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
419 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
420 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
421 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
422 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
423 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
424 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
425 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
428 Use of these low level functions has been informally discouraged for a long
429 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
430 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
434 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
435 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
436 This means that applications don't have to look at the curve NID and
437 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
438 However, they still can, that EVP_PKEY_set_alias_type() call acts as
439 a no-op when the EVP_PKEY is already of the given type.
441 Parameter and key generation is also reworked to make it possible
442 to generate EVP_PKEY_SM2 parameters and keys without having to go
443 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
444 However, code that does the latter will still work as before.
448 * Deprecated low level ECDH and ECDSA functions. These include:
450 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
451 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
454 Use of these low level functions has been informally discouraged for a long
455 time. Instead applications should use the EVP_PKEY_derive(3),
456 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
460 * Deprecated the EC_KEY_METHOD functions. These include:
462 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
463 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
464 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
465 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
466 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
467 EC_KEY_METHOD_get_verify.
469 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
473 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
474 and EVP_PKEY_decrypt() instead.
475 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
476 and EVP_PKEY_encrypt() instead.
480 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
481 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
482 a new formulation to include all the things it can be used for,
483 as well as words of caution.
487 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
488 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
492 * All of the low level HMAC functions have been deprecated including:
494 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
495 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
498 Use of these low level functions has been informally discouraged for a long
499 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
500 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
501 and L<EVP_MAC_final(3)>.
505 * Over two thousand fixes were made to the documentation, including:
506 - Common options (such as -rand/-writerand, TLS version control, etc)
507 were refactored and point to newly-enhanced descriptions in openssl.pod.
508 - Added style conformance for all options (with help from Richard Levitte),
509 documented all reported missing options, added a CI build to check
510 that all options are documented and that no unimplemented options
512 - Documented some internals, such as all use of environment variables.
513 - Addressed all internal broken L<> references.
517 * All of the low level CMAC functions have been deprecated including:
519 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
520 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
522 Use of these low level functions has been informally discouraged for a long
523 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
524 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
525 and L<EVP_MAC_final(3)>.
529 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
530 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
533 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
534 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
535 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
536 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
537 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
538 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
539 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
540 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
541 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
542 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
544 Use of these low level functions has been informally discouraged
545 for a long time. Applications should use the EVP_DigestInit_ex(3),
546 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
550 * Corrected the documentation of the return values from the `EVP_DigestSign*`
551 set of functions. The documentation mentioned negative values for some
552 errors, but this was never the case, so the mention of negative values
555 Code that followed the documentation and thereby check with something
556 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
560 * All of the low level cipher functions have been deprecated including:
562 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
563 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
564 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
565 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
566 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
567 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
568 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
569 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
570 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
571 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
572 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
573 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
574 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
575 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
576 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
577 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
578 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
579 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
580 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
581 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
582 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
583 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
584 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
585 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
586 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
587 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
588 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
589 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
590 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
592 Use of these low level functions has been informally discouraged for
593 a long time. Applications should use the high level EVP APIs, e.g.
594 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
595 equivalently named decrypt functions instead.
597 *Matt Caswell and Paul Dale*
599 * Removed include/openssl/opensslconf.h.in and replaced it with
600 include/openssl/configuration.h.in, which differs in not including
601 <openssl/macros.h>. A short header include/openssl/opensslconf.h
602 was added to include both.
604 This allows internal hacks where one might need to modify the set
605 of configured macros, for example this if deprecated symbols are
606 still supposed to be available internally:
608 #include <openssl/configuration.h>
610 #undef OPENSSL_NO_DEPRECATED
611 #define OPENSSL_SUPPRESS_DEPRECATED
613 #include <openssl/macros.h>
615 This should not be used by applications that use the exported
616 symbols, as that will lead to linking errors.
620 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
621 used in exponentiation with 512-bit moduli. No EC algorithms are
622 affected. Analysis suggests that attacks against 2-prime RSA1024,
623 3-prime RSA1536, and DSA1024 as a result of this defect would be very
624 difficult to perform and are not believed likely. Attacks against DH512
625 are considered just feasible. However, for an attack the target would
626 have to re-use the DH512 private key, which is not recommended anyway.
627 Also applications directly using the low level API BN_mod_exp may be
628 affected if they use BN_FLG_CONSTTIME.
633 * Most memory-debug features have been deprecated, and the functionality
634 replaced with no-ops.
638 * Added documentation for the STACK API.
642 * Introduced a new method type and API, OSSL_ENCODER, to
643 represent generic encoders. An implementation is expected to
644 be able to encode an object associated with a given name (such
645 as an algorithm name for an asymmetric key) into forms given by
646 implementation properties.
648 Encoders are primarily used from inside libcrypto, through
649 calls to functions like EVP_PKEY_print_private(),
650 PEM_write_bio_PrivateKey() and similar.
652 Encoders are specified in such a way that they can be made to
653 directly handle the provider side portion of an object, if this
654 provider side part comes from the same provider as the encoder
655 itself, but can also be made to handle objects in parametrized
656 form (as an OSSL_PARAM array of data). This allows a provider to
657 offer generic encoders as a service for any other provider.
661 * Added a .pragma directive to the syntax of configuration files, to
662 allow varying behavior in a supported and predictable manner.
663 Currently added pragma:
667 This allows dollar signs to be a keyword character unless it's
668 followed by a opening brace or parenthesis. This is useful for
669 platforms where dollar signs are commonly used in names, such as
670 volume names and system directory names on VMS.
674 * Added functionality to create an EVP_PKEY from user data. This
675 is effectively the same as creating a RSA, DH or DSA object and
676 then assigning them to an EVP_PKEY, but directly using algorithm
677 agnostic EVP functions. A benefit is that this should be future
678 proof for public key algorithms to come.
682 * Change the interpretation of the '--api' configuration option to
683 mean that this is a desired API compatibility level with no
684 further meaning. The previous interpretation, that this would
685 also mean to remove all deprecated symbols up to and including
686 the given version, no requires that 'no-deprecated' is also used
687 in the configuration.
689 When building applications, the desired API compatibility level
690 can be set with the OPENSSL_API_COMPAT macro like before. For
691 API compatibility version below 3.0, the old style numerical
692 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
693 For version 3.0 and on, the value is expected to be the decimal
694 value calculated from the major and minor version like this:
696 MAJOR * 10000 + MINOR * 100
700 -DOPENSSL_API_COMPAT=30000 For 3.0
701 -DOPENSSL_API_COMPAT=30200 For 3.2
703 To hide declarations that are deprecated up to and including the
704 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
705 given when building the application as well.
709 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
710 access to certificate and CRL stores via URIs and OSSL_STORE
713 This adds the following functions:
715 - X509_LOOKUP_store()
716 - X509_STORE_load_file()
717 - X509_STORE_load_path()
718 - X509_STORE_load_store()
719 - SSL_add_store_cert_subjects_to_stack()
720 - SSL_CTX_set_default_verify_store()
721 - SSL_CTX_load_verify_file()
722 - SSL_CTX_load_verify_dir()
723 - SSL_CTX_load_verify_store()
727 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
728 The presence of this system service is determined at run-time.
732 * Added functionality to create an EVP_PKEY context based on data
733 for methods from providers. This takes an algorithm name and a
734 property query string and simply stores them, with the intent
735 that any operation that uses this context will use those strings
736 to fetch the needed methods implicitly, thereby making the port
737 of application written for pre-3.0 OpenSSL easier.
741 * The undocumented function NCONF_WIN32() has been deprecated; for
742 conversion details see the HISTORY section of doc/man5/config.pod
746 * Introduced the new functions EVP_DigestSignInit_ex() and
747 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
748 EVP_DigestVerifyUpdate() have been converted to functions. See the man
749 pages for further details.
753 * Over two thousand fixes were made to the documentation, including:
754 adding missing command flags, better style conformance, documentation
757 *Rich Salz, Richard Levitte*
759 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
760 X25519, X448, Ed25519 and Ed448.
764 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
769 * Deprecated the public definition of ERR_STATE as well as the function
770 ERR_get_state(). This is done in preparation of making ERR_STATE an
775 * Added ERR functionality to give callers access to the stored function
776 names that have replaced the older function code based functions.
778 New functions are ERR_get_error_func(), ERR_peek_error_func(),
779 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
780 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
781 and ERR_peek_last_error_all().
783 These functions have become deprecated: ERR_get_error_line_data(),
784 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
785 ERR_func_error_string().
789 * Extended testing to be verbose for failing tests only. The make variables
790 VERBOSE_FAILURE or VF can be used to enable this:
792 $ make VF=1 test # Unix
793 $ mms /macro=(VF=1) test ! OpenVMS
794 $ nmake VF=1 test # Windows
798 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
799 used even when parsing explicit parameters, when loading a encoded key
800 or calling `EC_GROUP_new_from_ecpkparameters()`/
801 `EC_GROUP_new_from_ecparameters()`.
802 This prevents bypass of security hardening and performance gains,
803 especially for curves with specialized EC_METHODs.
804 By default, if a key encoded with explicit parameters is loaded and later
805 encoded, the output is still encoded with explicit parameters, even if
806 internally a "named" EC_GROUP is used for computation.
810 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
811 this change, EC_GROUP_set_generator would accept order and/or cofactor as
812 NULL. After this change, only the cofactor parameter can be NULL. It also
813 does some minimal sanity checks on the passed order.
818 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
819 An attack is simple, if the first CMS_recipientInfo is valid but the
820 second CMS_recipientInfo is chosen ciphertext. If the second
821 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
822 encryption key will be replaced by garbage, and the message cannot be
823 decoded, but if the RSA decryption fails, the correct encryption key is
824 used and the recipient will not notice the attack.
825 As a work around for this potential attack the length of the decrypted
826 key must be equal to the cipher default key length, in case the
827 certifiate is not given and all recipientInfo are tried out.
828 The old behaviour can be re-enabled in the CMS code by setting the
829 CMS_DEBUG_DECRYPT flag.
833 * Early start up entropy quality from the DEVRANDOM seed source has been
834 improved for older Linux systems. The RAND subsystem will wait for
835 /dev/random to be producing output before seeding from /dev/urandom.
836 The seeded state is stored for future library initialisations using
837 a system global shared memory segment. The shared memory identifier
838 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
839 the desired value. The default identifier is 114.
843 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
844 when primes for RSA keys are computed.
845 Since we previously always generated primes == 2 (mod 3) for RSA keys,
846 the 2-prime and 3-prime RSA modules were easy to distinguish, since
847 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
848 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
849 This avoids possible fingerprinting of newly generated RSA modules.
853 * Correct the extended master secret constant on EBCDIC systems. Without this
854 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
855 negotiate EMS will fail. Unfortunately this also means that TLS connections
856 between EBCDIC systems with this fix, and EBCDIC systems without this
857 fix will fail if they negotiate EMS.
861 * Changed the library initialisation so that the config file is now loaded
862 by default. This was already the case for libssl. It now occurs for both
863 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
864 OPENSSL_init_crypto() to suppress automatic loading of a config file.
868 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
869 where the former acts as a replacement for ERR_put_error(), and the
870 latter replaces the combination ERR_put_error()+ERR_add_error_data().
871 ERR_raise_data() adds more flexibility by taking a format string and
872 an arbitrary number of arguments following it, to be processed with
877 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
878 to check if a named provider is loaded and available. When called, it
879 will also activate all fallback providers if such are still present.
883 * Enforce a minimum DH modulus size of 512 bits.
887 * Changed DH parameters to generate the order q subgroup instead of 2q.
888 Previously generated DH parameters are still accepted by DH_check
889 but DH_generate_key works around that by clearing bit 0 of the
890 private key for those. This avoids leaking bit 0 of the private key.
894 * Significantly reduce secure memory usage by the randomness pools.
898 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
903 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
904 algorithms. An implementation of a key exchange algorithm can be obtained
905 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
906 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
907 the older EVP_PKEY_derive_init() function. See the man pages for the new
908 functions for further details.
912 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
916 * Removed the function names from error messages and deprecated the
919 * Removed NextStep support and the macro OPENSSL_UNISTD
923 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
924 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
925 Also removed "export var as function" capability; we do not export
926 variables, only functions.
930 * RC5_32_set_key has been changed to return an int type, with 0 indicating
931 an error and 1 indicating success. In previous versions of OpenSSL this
932 was a void type. If a key was set longer than the maximum possible this
937 * Support SM2 signing and verification schemes with X509 certificate.
941 * Use SHA256 as the default digest for TS query in the ts app.
945 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
946 This checks that the salt length is at least 128 bits, the derived key
947 length is at least 112 bits, and that the iteration count is at least 1000.
948 For backwards compatibility these checks are disabled by default in the
949 default provider, but are enabled by default in the fips provider.
950 To enable or disable these checks use the control
951 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
955 * Default cipher lists/suites are now available via a function, the
956 #defines are deprecated.
960 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
961 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
962 for Windows Store apps easier. Also, the "no-uplink" option has been added.
966 * Join the directories crypto/x509 and crypto/x509v3
970 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
971 This changes the size when using the genpkey app when no size is given. It
972 fixes an omission in earlier changes that changed all RSA, DSA and DH
973 generation apps to use 2048 bits by default.
977 * Added command 'openssl kdf' that uses the EVP_KDF API.
981 * Added command 'openssl mac' that uses the EVP_MAC API.
985 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
986 as default directories. Also added the command 'openssl info'
987 for scripting purposes.
991 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
992 deprecated. These undocumented functions were never integrated into the EVP
993 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
994 Bi-directional IGE mode. These modes were never formally standardised and
995 usage of these functions is believed to be very small. In particular
996 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
997 is ever used. The security implications are believed to be minimal, but
998 this issue was never fixed for backwards compatibility reasons. New code
999 should not use these modes.
1003 * Add prediction resistance to the DRBG reseeding process.
1007 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1008 mandated by IEEE Std 1619-2018.
1012 * Added newline escaping functionality to a filename when using openssl dgst.
1013 This output format is to replicate the output format found in the `*sum`
1014 checksum programs. This aims to preserve backward compatibility.
1016 *Matt Eaton, Richard Levitte, and Paul Dale*
1018 * Removed the heartbeat message in DTLS feature, as it has very
1019 little usage and doesn't seem to fulfill a valuable purpose.
1020 The configuration option is now deprecated.
1024 * Changed the output of 'openssl {digestname} < file' to display the
1025 digest name in its output.
1029 * Added a new generic trace API which provides support for enabling
1030 instrumentation through trace output. This feature is mainly intended
1031 as an aid for developers and is disabled by default. To utilize it,
1032 OpenSSL needs to be configured with the `enable-trace` option.
1034 If the tracing API is enabled, the application can activate trace output
1035 by registering BIOs as trace channels for a number of tracing and debugging
1038 The 'openssl' application has been expanded to enable any of the types
1039 available via environment variables defined by the user, and serves as
1040 one possible example on how to use this functionality.
1042 *Richard Levitte & Matthias St. Pierre*
1044 * Added build tests for C++. These are generated files that only do one
1045 thing, to include one public OpenSSL head file each. This tests that
1046 the public header files can be usefully included in a C++ application.
1048 This test isn't enabled by default. It can be enabled with the option
1049 'enable-buildtest-c++'.
1053 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1057 * Add KMAC to EVP_MAC.
1061 * Added property based algorithm implementation selection framework to
1066 * Added SCA hardening for modular field inversion in EC_GROUP through
1067 a new dedicated field_inv() pointer in EC_METHOD.
1068 This also addresses a leakage affecting conversions from projective
1069 to affine coordinates.
1071 *Billy Bob Brumley, Nicola Tuveri*
1073 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1074 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1075 those algorithms that were already supported through the EVP_PKEY API
1076 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1077 and scrypt are now wrappers that call EVP_KDF.
1081 * Build devcrypto engine as a dynamic engine.
1083 *Eneas U de Queiroz*
1085 * Add keyed BLAKE2 to EVP_MAC.
1089 * Fix a bug in the computation of the endpoint-pair shared secret used
1090 by DTLS over SCTP. This breaks interoperability with older versions
1091 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1092 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1093 interoperability with such broken implementations. However, enabling
1094 this switch breaks interoperability with correct implementations.
1096 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1097 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1101 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1105 * Change the license to the Apache License v2.0.
1109 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1111 - Major releases (indicated by incrementing the MAJOR release number)
1112 may introduce incompatible API/ABI changes.
1113 - Minor releases (indicated by incrementing the MINOR release number)
1114 may introduce new features but retain API/ABI compatibility.
1115 - Patch releases (indicated by incrementing the PATCH number)
1116 are intended for bug fixes and other improvements of existing
1117 features only (like improving performance or adding documentation)
1118 and retain API/ABI compatibility.
1122 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1126 * Remove the 'dist' target and add a tarball building script. The
1127 'dist' target has fallen out of use, and it shouldn't be
1128 necessary to configure just to create a source distribution.
1132 * Recreate the OS390-Unix config target. It no longer relies on a
1133 special script like it did for OpenSSL pre-1.1.0.
1137 * Instead of having the source directories listed in Configure, add
1138 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1143 * Add GMAC to EVP_MAC.
1147 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1151 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1152 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1153 to facilitate the continued use of MACs through raw private keys in
1154 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1158 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1159 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1163 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1164 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1165 are retained for backwards compatibility.
1169 * AES-XTS mode now enforces that its two keys are different to mitigate
1170 the attacked described in "Efficient Instantiations of Tweakable
1171 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1172 Details of this attack can be obtained from:
1173 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1177 * Rename the object files, i.e. give them other names than in previous
1178 versions. Their names now include the name of the final product, as
1179 well as its type mnemonic (bin, lib, shlib).
1183 * Added new option for 'openssl list', '-objects', which will display the
1184 list of built in objects, i.e. OIDs with names.
1188 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1189 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1194 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1195 improves application performance by removing data copies and providing
1196 applications with zero-copy system calls such as sendfile and splice.
1200 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1201 option is set, openssl cleanses (zeroize) plaintext bytes from
1202 internal buffers after delivering them to the application. Note,
1203 the application is still responsible for cleansing other copies
1204 (e.g.: data received by SSL_read(3)).
1208 * `PKCS12_parse` now maintains the order of the parsed certificates
1209 when outputting them via `*ca` (rather than reversing it).
1216 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1220 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1222 * Certificates with explicit curve parameters are now disallowed in
1223 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1227 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1228 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1229 conversely, silently ignore DTLS protocol version bounds when configuring
1230 TLS-based contexts. The commands can be repeated to set bounds of both
1231 types. The same applies with the corresponding "min_protocol" and
1232 "max_protocol" command-line switches, in case some application uses both TLS
1235 SSL_CTX instances that are created for a fixed protocol version (e.g.
1236 TLSv1_server_method()) also silently ignore version bounds. Previously
1237 attempts to apply bounds to these protocol versions would result in an
1238 error. Now only the "version-flexible" SSL_CTX instances are subject to
1239 limits in configuration files in command-line options.
1243 * Handshake now fails if Extended Master Secret extension is dropped
1248 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1250 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1252 * Fixed segmentation fault in SSL_check_chain()
1253 Server or client applications that call the SSL_check_chain() function
1254 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1255 dereference as a result of incorrect handling of the
1256 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1257 or unrecognised signature algorithm is received from the peer. This could
1258 be exploited by a malicious peer in a Denial of Service attack.
1263 * Added AES consttime code for no-asm configurations
1264 an optional constant time support for AES was added
1265 when building openssl for no-asm.
1266 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1267 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1268 At this time this feature is by default disabled.
1269 It will be enabled by default in 3.0.
1273 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1275 * Revert the change of EOF detection while reading in libssl to avoid
1276 regressions in applications depending on the current way of reporting
1277 the EOF. As the existing method is not fully accurate the change to
1278 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1279 branch and will be present in the 3.0 release.
1283 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1284 when primes for RSA keys are computed.
1285 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1286 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1287 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1288 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1289 This avoids possible fingerprinting of newly generated RSA modules.
1293 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1295 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1296 while reading in libssl then we would report an error back to the
1297 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1298 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1299 therefore give a hint as to what went wrong.
1303 * Check that ed25519 and ed448 are allowed by the security level. Previously
1304 signature algorithms not using an MD were not being checked that they were
1305 allowed by the security level.
1309 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1310 was not quite right. The behaviour was not consistent between resumption
1311 and normal handshakes, and also not quite consistent with historical
1312 behaviour. The behaviour in various scenarios has been clarified and
1313 it has been updated to make it match historical behaviour as closely as
1318 * *[VMS only]* The header files that the VMS compilers include automatically,
1319 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1320 that the C++ compiler doesn't understand. This is a shortcoming in the
1321 compiler, but can be worked around with `__cplusplus` guards.
1323 C++ applications that use OpenSSL libraries must be compiled using the
1324 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1325 functions. Otherwise, only functions with symbols of less than 31
1326 characters can be used, as the linker will not be able to successfully
1327 resolve symbols with longer names.
1331 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1332 The presence of this system service is determined at run-time.
1336 * Added newline escaping functionality to a filename when using openssl dgst.
1337 This output format is to replicate the output format found in the `*sum`
1338 checksum programs. This aims to preserve backward compatibility.
1340 *Matt Eaton, Richard Levitte, and Paul Dale*
1342 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1347 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1349 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1350 number generator (RNG). This was intended to include protection in the
1351 event of a fork() system call in order to ensure that the parent and child
1352 processes did not share the same RNG state. However this protection was not
1353 being used in the default case.
1355 A partial mitigation for this issue is that the output from a high
1356 precision timer is mixed into the RNG state so the likelihood of a parent
1357 and child process sharing state is significantly reduced.
1359 If an application already calls OPENSSL_init_crypto() explicitly using
1360 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1363 *Matthias St. Pierre*
1365 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1366 used even when parsing explicit parameters, when loading a encoded key
1367 or calling `EC_GROUP_new_from_ecpkparameters()`/
1368 `EC_GROUP_new_from_ecparameters()`.
1369 This prevents bypass of security hardening and performance gains,
1370 especially for curves with specialized EC_METHODs.
1371 By default, if a key encoded with explicit parameters is loaded and later
1372 encoded, the output is still encoded with explicit parameters, even if
1373 internally a "named" EC_GROUP is used for computation.
1377 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1378 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1379 NULL. After this change, only the cofactor parameter can be NULL. It also
1380 does some minimal sanity checks on the passed order.
1385 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1386 An attack is simple, if the first CMS_recipientInfo is valid but the
1387 second CMS_recipientInfo is chosen ciphertext. If the second
1388 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1389 encryption key will be replaced by garbage, and the message cannot be
1390 decoded, but if the RSA decryption fails, the correct encryption key is
1391 used and the recipient will not notice the attack.
1392 As a work around for this potential attack the length of the decrypted
1393 key must be equal to the cipher default key length, in case the
1394 certifiate is not given and all recipientInfo are tried out.
1395 The old behaviour can be re-enabled in the CMS code by setting the
1396 CMS_DEBUG_DECRYPT flag.
1401 * Early start up entropy quality from the DEVRANDOM seed source has been
1402 improved for older Linux systems. The RAND subsystem will wait for
1403 /dev/random to be producing output before seeding from /dev/urandom.
1404 The seeded state is stored for future library initialisations using
1405 a system global shared memory segment. The shared memory identifier
1406 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1407 the desired value. The default identifier is 114.
1411 * Correct the extended master secret constant on EBCDIC systems. Without this
1412 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1413 negotiate EMS will fail. Unfortunately this also means that TLS connections
1414 between EBCDIC systems with this fix, and EBCDIC systems without this
1415 fix will fail if they negotiate EMS.
1419 * Use Windows installation paths in the mingw builds
1421 Mingw isn't a POSIX environment per se, which means that Windows
1422 paths should be used for installation.
1427 * Changed DH_check to accept parameters with order q and 2q subgroups.
1428 With order 2q subgroups the bit 0 of the private key is not secret
1429 but DH_generate_key works around that by clearing bit 0 of the
1430 private key for those. This avoids leaking bit 0 of the private key.
1434 * Significantly reduce secure memory usage by the randomness pools.
1438 * Revert the DEVRANDOM_WAIT feature for Linux systems
1440 The DEVRANDOM_WAIT feature added a select() call to wait for the
1441 /dev/random device to become readable before reading from the
1442 /dev/urandom device.
1444 It turned out that this change had negative side effects on
1445 performance which were not acceptable. After some discussion it
1446 was decided to revert this feature and leave it up to the OS
1447 resp. the platform maintainer to ensure a proper initialization
1448 during early boot time.
1450 *Matthias St. Pierre*
1452 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1454 * Add build tests for C++. These are generated files that only do one
1455 thing, to include one public OpenSSL head file each. This tests that
1456 the public header files can be usefully included in a C++ application.
1458 This test isn't enabled by default. It can be enabled with the option
1459 'enable-buildtest-c++'.
1463 * Enable SHA3 pre-hashing for ECDSA and DSA.
1467 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1468 This changes the size when using the genpkey app when no size is given. It
1469 fixes an omission in earlier changes that changed all RSA, DSA and DH
1470 generation apps to use 2048 bits by default.
1474 * Reorganize the manual pages to consistently have RETURN VALUES,
1475 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1476 util/fix-doc-nits accordingly.
1478 *Paul Yang, Joshua Lock*
1480 * Add the missing accessor EVP_PKEY_get0_engine()
1484 * Have apps like 's_client' and 's_server' output the signature scheme
1485 along with other cipher suite parameters when debugging.
1489 * Make OPENSSL_config() error agnostic again.
1493 * Do the error handling in RSA decryption constant time.
1497 * Prevent over long nonces in ChaCha20-Poly1305.
1499 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1500 for every encryption operation. RFC 7539 specifies that the nonce value
1501 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1502 and front pads the nonce with 0 bytes if it is less than 12
1503 bytes. However it also incorrectly allows a nonce to be set of up to 16
1504 bytes. In this case only the last 12 bytes are significant and any
1505 additional leading bytes are ignored.
1507 It is a requirement of using this cipher that nonce values are
1508 unique. Messages encrypted using a reused nonce value are susceptible to
1509 serious confidentiality and integrity attacks. If an application changes
1510 the default nonce length to be longer than 12 bytes and then makes a
1511 change to the leading bytes of the nonce expecting the new value to be a
1512 new unique nonce then such an application could inadvertently encrypt
1513 messages with a reused nonce.
1515 Additionally the ignored bytes in a long nonce are not covered by the
1516 integrity guarantee of this cipher. Any application that relies on the
1517 integrity of these ignored leading bytes of a long nonce may be further
1518 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1519 is safe because no such use sets such a long nonce value. However user
1520 applications that use this cipher directly and set a non-default nonce
1521 length to be longer than 12 bytes may be vulnerable.
1523 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1529 * Add DEVRANDOM_WAIT feature for Linux systems
1531 On older Linux systems where the getrandom() system call is not available,
1532 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1533 Contrary to getrandom(), the /dev/urandom device will not block during
1534 early boot when the kernel CSPRNG has not been seeded yet.
1536 To mitigate this known weakness, use select() to wait for /dev/random to
1537 become readable before reading from /dev/urandom.
1539 * Ensure that SM2 only uses SM3 as digest algorithm
1543 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1545 * Change the info callback signals for the start and end of a post-handshake
1546 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1547 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1548 confused by this and assume that a TLSv1.2 renegotiation has started. This
1549 can break KeyUpdate handling. Instead we no longer signal the start and end
1550 of a post handshake message exchange (although the messages themselves are
1551 still signalled). This could break some applications that were expecting
1552 the old signals. However without this KeyUpdate is not usable for many
1557 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1559 * Timing vulnerability in DSA signature generation
1561 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1562 timing side channel attack. An attacker could use variations in the signing
1563 algorithm to recover the private key.
1565 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1570 * Timing vulnerability in ECDSA signature generation
1572 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1573 timing side channel attack. An attacker could use variations in the signing
1574 algorithm to recover the private key.
1576 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1581 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1582 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1583 of two gigabytes and the error handling improved.
1585 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1586 categorized as a normal bug, not a security issue, because the DRBG reseeds
1587 automatically and is fully functional even without additional randomness
1588 provided by the application.
1590 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1592 * Add a new ClientHello callback. Provides a callback interface that gives
1593 the application the ability to adjust the nascent SSL object at the
1594 earliest stage of ClientHello processing, immediately after extensions have
1595 been collected but before they have been processed. In particular, this
1596 callback can adjust the supported TLS versions in response to the contents
1601 * Add SM2 base algorithm support.
1605 * s390x assembly pack: add (improved) hardware-support for the following
1606 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1607 aes-cfb/cfb8, aes-ecb.
1611 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1612 parameter is no longer accepted, as it leads to a corrupt table. NULL
1613 pem_str is reserved for alias entries only.
1617 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1618 step for prime curves. The new implementation is based on formulae from
1619 differential addition-and-doubling in homogeneous projective coordinates
1620 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1621 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1622 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1623 to work in projective coordinates.
1625 *Billy Bob Brumley, Nicola Tuveri*
1627 * Change generating and checking of primes so that the error rate of not
1628 being prime depends on the intended use based on the size of the input.
1629 For larger primes this will result in more rounds of Miller-Rabin.
1630 The maximal error rate for primes with more than 1080 bits is lowered
1633 *Kurt Roeckx, Annie Yousar*
1635 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1639 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1640 moving between systems, and to avoid confusion when a Windows build is
1641 done with mingw vs with MSVC. For POSIX installs, there's still a
1642 symlink or copy named 'tsget' to avoid that confusion as well.
1646 * Revert blinding in ECDSA sign and instead make problematic addition
1647 length-invariant. Switch even to fixed-length Montgomery multiplication.
1651 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1652 step for binary curves. The new implementation is based on formulae from
1653 differential addition-and-doubling in mixed Lopez-Dahab projective
1654 coordinates, modified to independently blind the operands.
1656 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1658 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1659 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1660 EC_METHODs to implement their own specialized "ladder step", to take
1661 advantage of more favorable coordinate systems or more efficient
1662 differential addition-and-doubling algorithms.
1664 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1666 * Modified the random device based seed sources to keep the relevant
1667 file descriptors open rather than reopening them on each access.
1668 This allows such sources to operate in a chroot() jail without
1669 the associated device nodes being available. This behaviour can be
1670 controlled using RAND_keep_random_devices_open().
1674 * Numerous side-channel attack mitigations have been applied. This may have
1675 performance impacts for some algorithms for the benefit of improved
1676 security. Specific changes are noted in this change log by their respective
1681 * AIX shared library support overhaul. Switch to AIX "natural" way of
1682 handling shared libraries, which means collecting shared objects of
1683 different versions and bitnesses in one common archive. This allows to
1684 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1685 doesn't affect the way 3rd party applications are linked, only how
1686 multi-version installation is managed.
1690 * Make ec_group_do_inverse_ord() more robust and available to other
1691 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1692 mitigations are applied to the fallback BN_mod_inverse().
1693 When using this function rather than BN_mod_inverse() directly, new
1694 EC cryptosystem implementations are then safer-by-default.
1698 * Add coordinate blinding for EC_POINT and implement projective
1699 coordinate blinding for generic prime curves as a countermeasure to
1700 chosen point SCA attacks.
1702 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1704 * Add blinding to ECDSA and DSA signatures to protect against side channel
1705 attacks discovered by Keegan Ryan (NCC Group).
1709 * Enforce checking in the pkeyutl command line app to ensure that the input
1710 length does not exceed the maximum supported digest length when performing
1711 a sign, verify or verifyrecover operation.
1715 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1716 I/O in combination with something like select() or poll() will hang. This
1717 can be turned off again using SSL_CTX_clear_mode().
1718 Many applications do not properly handle non-application data records, and
1719 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1720 around the problems in those applications, but can also break some.
1721 It's recommended to read the manpages about SSL_read(), SSL_write(),
1722 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1723 SSL_CTX_set_read_ahead() again.
1727 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1728 now allow empty (zero character) pass phrases.
1732 * Apply blinding to binary field modular inversion and remove patent
1733 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1737 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1738 binary and prime elliptic curves.
1742 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1743 constant time fixed point multiplication.
1747 * Revise elliptic curve scalar multiplication with timing attack
1748 defenses: ec_wNAF_mul redirects to a constant time implementation
1749 when computing fixed point and variable point multiplication (which
1750 in OpenSSL are mostly used with secret scalars in keygen, sign,
1751 ECDH derive operations).
1752 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1755 * Updated CONTRIBUTING
1759 * Updated DRBG / RAND to request nonce and additional low entropy
1760 randomness from the system.
1762 *Matthias St. Pierre*
1764 * Updated 'openssl rehash' to use OpenSSL consistent default.
1768 * Moved the load of the ssl_conf module to libcrypto, which helps
1769 loading engines that libssl uses before libssl is initialised.
1773 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1777 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1779 *Ingo Schwarze, Rich Salz*
1781 * Added output of accepting IP address and port for 'openssl s_server'
1785 * Added a new API for TLSv1.3 ciphersuites:
1786 SSL_CTX_set_ciphersuites()
1787 SSL_set_ciphersuites()
1791 * Memory allocation failures consistently add an error to the error
1796 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1797 in libcrypto when run as setuid/setgid.
1801 * Load any config file by default when libssl is used.
1805 * Added new public header file <openssl/rand_drbg.h> and documentation
1806 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1808 *Matthias St. Pierre*
1810 * QNX support removed (cannot find contributors to get their approval
1811 for the license change).
1815 * TLSv1.3 replay protection for early data has been implemented. See the
1816 SSL_read_early_data() man page for further details.
1820 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1821 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1822 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1823 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1824 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1825 configuration has been separated out. See the ciphers man page or the
1826 SSL_CTX_set_ciphersuites() man page for more information.
1830 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1831 in responder mode now supports the new "-multi" option, which
1832 spawns the specified number of child processes to handle OCSP
1833 requests. The "-timeout" option now also limits the OCSP
1834 responder's patience to wait to receive the full client request
1835 on a newly accepted connection. Child processes are respawned
1836 as needed, and the CA index file is automatically reloaded
1837 when changed. This makes it possible to run the "ocsp" responder
1838 as a long-running service, making the OpenSSL CA somewhat more
1839 feature-complete. In this mode, most diagnostic messages logged
1840 after entering the event loop are logged via syslog(3) rather than
1845 * Added support for X448 and Ed448. Heavily based on original work by
1850 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1851 objects loaded. This adds the functions OSSL_STORE_expect() and
1852 OSSL_STORE_find() as well as needed tools to construct searches and
1853 get the search data out of them.
1857 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1858 version of OpenSSL should review their configuration settings to ensure
1859 that they are still appropriate for TLSv1.3. For further information see:
1860 <https://wiki.openssl.org/index.php/TLS1.3>
1864 * Grand redesign of the OpenSSL random generator
1866 The default RAND method now utilizes an AES-CTR DRBG according to
1867 NIST standard SP 800-90Ar1. The new random generator is essentially
1868 a port of the default random generator from the OpenSSL FIPS 2.0
1869 object module. It is a hybrid deterministic random bit generator
1870 using an AES-CTR bit stream and which seeds and reseeds itself
1871 automatically using trusted system entropy sources.
1873 Some of its new features are:
1874 - Support for multiple DRBG instances with seed chaining.
1875 - The default RAND method makes use of a DRBG.
1876 - There is a public and private DRBG instance.
1877 - The DRBG instances are fork-safe.
1878 - Keep all global DRBG instances on the secure heap if it is enabled.
1879 - The public and private DRBG instance are per thread for lock free
1882 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1884 * Changed Configure so it only says what it does and doesn't dump
1885 so much data. Instead, ./configdata.pm should be used as a script
1886 to display all sorts of configuration data.
1890 * Added processing of "make variables" to Configure.
1894 * Added SHA512/224 and SHA512/256 algorithm support.
1898 * The last traces of Netware support, first removed in 1.1.0, have
1903 * Get rid of Makefile.shared, and in the process, make the processing
1904 of certain files (rc.obj, or the .def/.map/.opt files produced from
1905 the ordinal files) more visible and hopefully easier to trace and
1906 debug (or make silent).
1910 * Make it possible to have environment variable assignments as
1911 arguments to config / Configure.
1915 * Add multi-prime RSA (RFC 8017) support.
1919 * Add SM3 implemented according to GB/T 32905-2016
1920 *Jack Lloyd <jack.lloyd@ribose.com>,*
1921 *Ronald Tse <ronald.tse@ribose.com>,*
1922 *Erick Borsboom <erick.borsboom@ribose.com>*
1924 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1925 as documented in RFC6066.
1926 Based on a patch from Tomasz Moń
1928 *Filipe Raimundo da Silva*
1930 * Add SM4 implemented according to GB/T 32907-2016.
1931 *Jack Lloyd <jack.lloyd@ribose.com>,*
1932 *Ronald Tse <ronald.tse@ribose.com>,*
1933 *Erick Borsboom <erick.borsboom@ribose.com>*
1935 * Reimplement -newreq-nodes and ERR_error_string_n; the
1936 original author does not agree with the license change.
1940 * Add ARIA AEAD TLS support.
1944 * Some macro definitions to support VS6 have been removed. Visual
1945 Studio 6 has not worked since 1.1.0
1949 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1950 without clearing the errors.
1954 * Add "atfork" functions. If building on a system that without
1955 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1956 requirements. The RAND facility now uses/requires this.
1964 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1965 not possible to disable entirely. However, it's still possible to
1966 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1969 To disable, configure with 'no-ui-console'. 'no-ui' is still
1970 possible to use as an alias. Check at compile time with the
1971 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1972 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1976 * Add a STORE module, which implements a uniform and URI based reader of
1977 stores that can contain keys, certificates, CRLs and numerous other
1978 objects. The main API is loosely based on a few stdio functions,
1979 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1980 OSSL_STORE_error and OSSL_STORE_close.
1981 The implementation uses backends called "loaders" to implement arbitrary
1982 URI schemes. There is one built in "loader" for the 'file' scheme.
1986 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1987 then adjusted to work on FreeBSD 8.4 as well.
1988 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1989 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1993 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1994 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1995 error code calls like this:
1997 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1999 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2000 that can be encoded in C. For the foreseeable future, this will only
2003 *Richard Levitte and Tim Hudson*
2005 * Removed BSD cryptodev engine.
2009 * Add a build target 'build_all_generated', to build all generated files
2010 and only that. This can be used to prepare everything that requires
2011 things like perl for a system that lacks perl and then move everything
2012 to that system and do the rest of the build there.
2016 * In the UI interface, make it possible to duplicate the user data. This
2017 can be used by engines that need to retain the data for a longer time
2018 than just the call where this user data is passed.
2022 * Ignore the '-named_curve auto' value for compatibility of applications
2025 *Tomas Mraz <tmraz@fedoraproject.org>*
2027 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2028 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2029 alerts across multiple records (some of which could be empty). In practice
2030 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2031 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2032 support this at all. Supporting it adds significant complexity to the
2033 record layer, and its removal is unlikely to cause interoperability
2038 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2039 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2040 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2045 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2046 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2048 *Richard Levitte, Andy Polyakov*
2050 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2055 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2056 platform rather than 'mingw'.
2060 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2061 success if they are asked to add an object which already exists
2062 in the store. This change cascades to other functions which load
2063 certificates and CRLs.
2067 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2068 facilitate stack unwinding even from assembly subroutines.
2072 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2073 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2077 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2078 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2079 which is the minimum version we support.
2083 * Certificate time validation (X509_cmp_time) enforces stricter
2084 compliance with RFC 5280. Fractional seconds and timezone offsets
2085 are no longer allowed.
2089 * Add support for ARIA
2093 * s_client will now send the Server Name Indication (SNI) extension by
2094 default unless the new "-noservername" option is used. The server name is
2095 based on the host provided to the "-connect" option unless overridden by
2096 using "-servername".
2100 * Add support for SipHash
2104 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2105 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2106 prevent issues where no progress is being made and the peer continually
2107 sends unrecognised record types, using up resources processing them.
2111 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2112 using the algorithm defined in
2113 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2117 * Heartbeat support has been removed; the ABI is changed for now.
2119 *Richard Levitte, Rich Salz*
2121 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2125 * The RSA "null" method, which was partially supported to avoid patent
2126 issues, has been replaced to always returns NULL.
2133 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2135 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2136 used even when parsing explicit parameters, when loading a encoded key
2137 or calling `EC_GROUP_new_from_ecpkparameters()`/
2138 `EC_GROUP_new_from_ecparameters()`.
2139 This prevents bypass of security hardening and performance gains,
2140 especially for curves with specialized EC_METHODs.
2141 By default, if a key encoded with explicit parameters is loaded and later
2142 encoded, the output is still encoded with explicit parameters, even if
2143 internally a "named" EC_GROUP is used for computation.
2147 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2148 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2149 NULL. After this change, only the cofactor parameter can be NULL. It also
2150 does some minimal sanity checks on the passed order.
2155 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2156 An attack is simple, if the first CMS_recipientInfo is valid but the
2157 second CMS_recipientInfo is chosen ciphertext. If the second
2158 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2159 encryption key will be replaced by garbage, and the message cannot be
2160 decoded, but if the RSA decryption fails, the correct encryption key is
2161 used and the recipient will not notice the attack.
2162 As a work around for this potential attack the length of the decrypted
2163 key must be equal to the cipher default key length, in case the
2164 certifiate is not given and all recipientInfo are tried out.
2165 The old behaviour can be re-enabled in the CMS code by setting the
2166 CMS_DEBUG_DECRYPT flag.
2171 * Use Windows installation paths in the mingw builds
2173 Mingw isn't a POSIX environment per se, which means that Windows
2174 paths should be used for installation.
2179 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2181 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2182 This changes the size when using the genpkey app when no size is given. It
2183 fixes an omission in earlier changes that changed all RSA, DSA and DH
2184 generation apps to use 2048 bits by default.
2188 * Prevent over long nonces in ChaCha20-Poly1305.
2190 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2191 for every encryption operation. RFC 7539 specifies that the nonce value
2192 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2193 and front pads the nonce with 0 bytes if it is less than 12
2194 bytes. However it also incorrectly allows a nonce to be set of up to 16
2195 bytes. In this case only the last 12 bytes are significant and any
2196 additional leading bytes are ignored.
2198 It is a requirement of using this cipher that nonce values are
2199 unique. Messages encrypted using a reused nonce value are susceptible to
2200 serious confidentiality and integrity attacks. If an application changes
2201 the default nonce length to be longer than 12 bytes and then makes a
2202 change to the leading bytes of the nonce expecting the new value to be a
2203 new unique nonce then such an application could inadvertently encrypt
2204 messages with a reused nonce.
2206 Additionally the ignored bytes in a long nonce are not covered by the
2207 integrity guarantee of this cipher. Any application that relies on the
2208 integrity of these ignored leading bytes of a long nonce may be further
2209 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2210 is safe because no such use sets such a long nonce value. However user
2211 applications that use this cipher directly and set a non-default nonce
2212 length to be longer than 12 bytes may be vulnerable.
2214 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2220 * Added SCA hardening for modular field inversion in EC_GROUP through
2221 a new dedicated field_inv() pointer in EC_METHOD.
2222 This also addresses a leakage affecting conversions from projective
2223 to affine coordinates.
2225 *Billy Bob Brumley, Nicola Tuveri*
2227 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2228 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2232 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2236 * Remove the 'dist' target and add a tarball building script. The
2237 'dist' target has fallen out of use, and it shouldn't be
2238 necessary to configure just to create a source distribution.
2242 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2244 * Timing vulnerability in DSA signature generation
2246 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2247 timing side channel attack. An attacker could use variations in the signing
2248 algorithm to recover the private key.
2250 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2255 * Timing vulnerability in ECDSA signature generation
2257 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2258 timing side channel attack. An attacker could use variations in the signing
2259 algorithm to recover the private key.
2261 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2266 * Add coordinate blinding for EC_POINT and implement projective
2267 coordinate blinding for generic prime curves as a countermeasure to
2268 chosen point SCA attacks.
2270 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2272 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2274 * Client DoS due to large DH parameter
2276 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2277 malicious server can send a very large prime value to the client. This will
2278 cause the client to spend an unreasonably long period of time generating a
2279 key for this prime resulting in a hang until the client has finished. This
2280 could be exploited in a Denial Of Service attack.
2282 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2287 * Cache timing vulnerability in RSA Key Generation
2289 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2290 a cache timing side channel attack. An attacker with sufficient access to
2291 mount cache timing attacks during the RSA key generation process could
2292 recover the private key.
2294 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2295 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2300 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2301 parameter is no longer accepted, as it leads to a corrupt table. NULL
2302 pem_str is reserved for alias entries only.
2306 * Revert blinding in ECDSA sign and instead make problematic addition
2307 length-invariant. Switch even to fixed-length Montgomery multiplication.
2311 * Change generating and checking of primes so that the error rate of not
2312 being prime depends on the intended use based on the size of the input.
2313 For larger primes this will result in more rounds of Miller-Rabin.
2314 The maximal error rate for primes with more than 1080 bits is lowered
2317 *Kurt Roeckx, Annie Yousar*
2319 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2323 * Add blinding to ECDSA and DSA signatures to protect against side channel
2324 attacks discovered by Keegan Ryan (NCC Group).
2328 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2329 now allow empty (zero character) pass phrases.
2333 * Certificate time validation (X509_cmp_time) enforces stricter
2334 compliance with RFC 5280. Fractional seconds and timezone offsets
2335 are no longer allowed.
2339 * Fixed a text canonicalisation bug in CMS
2341 Where a CMS detached signature is used with text content the text goes
2342 through a canonicalisation process first prior to signing or verifying a
2343 signature. This process strips trailing space at the end of lines, converts
2344 line terminators to CRLF and removes additional trailing line terminators
2345 at the end of a file. A bug in the canonicalisation process meant that
2346 some characters, such as form-feed, were incorrectly treated as whitespace
2347 and removed. This is contrary to the specification (RFC5485). This fix
2348 could mean that detached text data signed with an earlier version of
2349 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2350 signed with a fixed OpenSSL may fail to verify with an earlier version of
2351 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2352 and use the "-binary" flag (for the "cms" command line application) or set
2353 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2357 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2359 * Constructed ASN.1 types with a recursive definition could exceed the stack
2361 Constructed ASN.1 types with a recursive definition (such as can be found
2362 in PKCS7) could eventually exceed the stack given malicious input with
2363 excessive recursion. This could result in a Denial Of Service attack. There
2364 are no such structures used within SSL/TLS that come from untrusted sources
2365 so this is considered safe.
2367 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2373 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2375 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2376 effectively reduced to only comparing the least significant bit of each
2377 byte. This allows an attacker to forge messages that would be considered as
2378 authenticated in an amount of tries lower than that guaranteed by the
2379 security claims of the scheme. The module can only be compiled by the
2380 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2382 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2388 * Add a build target 'build_all_generated', to build all generated files
2389 and only that. This can be used to prepare everything that requires
2390 things like perl for a system that lacks perl and then move everything
2391 to that system and do the rest of the build there.
2395 * Backport SSL_OP_NO_RENGOTIATION
2397 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2398 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2399 changes this is no longer possible in 1.1.0. Therefore the new
2400 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2401 1.1.0 to provide equivalent functionality.
2403 Note that if an application built against 1.1.0h headers (or above) is run
2404 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2405 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2409 * Removed the OS390-Unix config target. It relied on a script that doesn't
2414 * rsaz_1024_mul_avx2 overflow bug on x86_64
2416 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2417 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2418 Analysis suggests that attacks against RSA and DSA as a result of this
2419 defect would be very difficult to perform and are not believed likely.
2420 Attacks against DH1024 are considered just feasible, because most of the
2421 work necessary to deduce information about a private key may be performed
2422 offline. The amount of resources required for such an attack would be
2423 significant. However, for an attack on TLS to be meaningful, the server
2424 would have to share the DH1024 private key among multiple clients, which is
2425 no longer an option since CVE-2016-0701.
2427 This only affects processors that support the AVX2 but not ADX extensions
2428 like Intel Haswell (4th generation).
2430 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2431 was originally found via the OSS-Fuzz project.
2436 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2438 * bn_sqrx8x_internal carry bug on x86_64
2440 There is a carry propagating bug in the x86_64 Montgomery squaring
2441 procedure. No EC algorithms are affected. Analysis suggests that attacks
2442 against RSA and DSA as a result of this defect would be very difficult to
2443 perform and are not believed likely. Attacks against DH are considered just
2444 feasible (although very difficult) because most of the work necessary to
2445 deduce information about a private key may be performed offline. The amount
2446 of resources required for such an attack would be very significant and
2447 likely only accessible to a limited number of attackers. An attacker would
2448 additionally need online access to an unpatched system using the target
2449 private key in a scenario with persistent DH parameters and a private
2450 key that is shared between multiple clients.
2452 This only affects processors that support the BMI1, BMI2 and ADX extensions
2453 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2455 This issue was reported to OpenSSL by the OSS-Fuzz project.
2460 * Malformed X.509 IPAddressFamily could cause OOB read
2462 If an X.509 certificate has a malformed IPAddressFamily extension,
2463 OpenSSL could do a one-byte buffer overread. The most likely result
2464 would be an erroneous display of the certificate in text format.
2466 This issue was reported to OpenSSL by the OSS-Fuzz project.
2471 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2473 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2474 platform rather than 'mingw'.
2478 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2479 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2480 which is the minimum version we support.
2484 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2486 * Encrypt-Then-Mac renegotiation crash
2488 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2489 negotiated where it was not in the original handshake (or vice-versa) then
2490 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2491 and servers are affected.
2493 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2498 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2500 * Truncated packet could crash via OOB read
2502 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2503 cipher is being used, then a truncated packet can cause that host to
2504 perform an out-of-bounds read, usually resulting in a crash.
2506 This issue was reported to OpenSSL by Robert Święcki of Google.
2511 * Bad (EC)DHE parameters cause a client crash
2513 If a malicious server supplies bad parameters for a DHE or ECDHE key
2514 exchange then this can result in the client attempting to dereference a
2515 NULL pointer leading to a client crash. This could be exploited in a Denial
2518 This issue was reported to OpenSSL by Guido Vranken.
2523 * BN_mod_exp may produce incorrect results on x86_64
2525 There is a carry propagating bug in the x86_64 Montgomery squaring
2526 procedure. No EC algorithms are affected. Analysis suggests that attacks
2527 against RSA and DSA as a result of this defect would be very difficult to
2528 perform and are not believed likely. Attacks against DH are considered just
2529 feasible (although very difficult) because most of the work necessary to
2530 deduce information about a private key may be performed offline. The amount
2531 of resources required for such an attack would be very significant and
2532 likely only accessible to a limited number of attackers. An attacker would
2533 additionally need online access to an unpatched system using the target
2534 private key in a scenario with persistent DH parameters and a private
2535 key that is shared between multiple clients. For example this can occur by
2536 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2537 similar to CVE-2015-3193 but must be treated as a separate problem.
2539 This issue was reported to OpenSSL by the OSS-Fuzz project.
2544 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2546 * ChaCha20/Poly1305 heap-buffer-overflow
2548 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2549 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2550 crash. This issue is not considered to be exploitable beyond a DoS.
2552 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2557 * CMS Null dereference
2559 Applications parsing invalid CMS structures can crash with a NULL pointer
2560 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2561 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2562 structure callback if an attempt is made to free certain invalid encodings.
2563 Only CHOICE structures using a callback which do not handle NULL value are
2566 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2571 * Montgomery multiplication may produce incorrect results
2573 There is a carry propagating bug in the Broadwell-specific Montgomery
2574 multiplication procedure that handles input lengths divisible by, but
2575 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2576 and DH private keys are impossible. This is because the subroutine in
2577 question is not used in operations with the private key itself and an input
2578 of the attacker's direct choice. Otherwise the bug can manifest itself as
2579 transient authentication and key negotiation failures or reproducible
2580 erroneous outcome of public-key operations with specially crafted input.
2581 Among EC algorithms only Brainpool P-512 curves are affected and one
2582 presumably can attack ECDH key negotiation. Impact was not analyzed in
2583 detail, because pre-requisites for attack are considered unlikely. Namely
2584 multiple clients have to choose the curve in question and the server has to
2585 share the private key among them, neither of which is default behaviour.
2586 Even then only clients that chose the curve will be affected.
2588 This issue was publicly reported as transient failures and was not
2589 initially recognized as a security issue. Thanks to Richard Morgan for
2590 providing reproducible case.
2595 * Removed automatic addition of RPATH in shared libraries and executables,
2596 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2600 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2602 * Fix Use After Free for large message sizes
2604 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2605 message larger than approx 16k is received then the underlying buffer to
2606 store the incoming message is reallocated and moved. Unfortunately a
2607 dangling pointer to the old location is left which results in an attempt to
2608 write to the previously freed location. This is likely to result in a
2609 crash, however it could potentially lead to execution of arbitrary code.
2611 This issue only affects OpenSSL 1.1.0a.
2613 This issue was reported to OpenSSL by Robert Święcki.
2618 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2620 * OCSP Status Request extension unbounded memory growth
2622 A malicious client can send an excessively large OCSP Status Request
2623 extension. If that client continually requests renegotiation, sending a
2624 large OCSP Status Request extension each time, then there will be unbounded
2625 memory growth on the server. This will eventually lead to a Denial Of
2626 Service attack through memory exhaustion. Servers with a default
2627 configuration are vulnerable even if they do not support OCSP. Builds using
2628 the "no-ocsp" build time option are not affected.
2630 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2635 * SSL_peek() hang on empty record
2637 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2638 sends an empty record. This could be exploited by a malicious peer in a
2639 Denial Of Service attack.
2641 This issue was reported to OpenSSL by Alex Gaynor.
2646 * Excessive allocation of memory in tls_get_message_header() and
2647 dtls1_preprocess_fragment()
2649 A (D)TLS message includes 3 bytes for its length in the header for the
2650 message. This would allow for messages up to 16Mb in length. Messages of
2651 this length are excessive and OpenSSL includes a check to ensure that a
2652 peer is sending reasonably sized messages in order to avoid too much memory
2653 being consumed to service a connection. A flaw in the logic of version
2654 1.1.0 means that memory for the message is allocated too early, prior to
2655 the excessive message length check. Due to way memory is allocated in
2656 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2657 to service a connection. This could lead to a Denial of Service through
2658 memory exhaustion. However, the excessive message length check still takes
2659 place, and this would cause the connection to immediately fail. Assuming
2660 that the application calls SSL_free() on the failed connection in a timely
2661 manner then the 21Mb of allocated memory will then be immediately freed
2662 again. Therefore the excessive memory allocation will be transitory in
2663 nature. This then means that there is only a security impact if:
2665 1) The application does not call SSL_free() in a timely manner in the event
2666 that the connection fails
2668 2) The application is working in a constrained environment where there is
2669 very little free memory
2671 3) The attacker initiates multiple connection attempts such that there are
2672 multiple connections in a state where memory has been allocated for the
2673 connection; SSL_free() has not yet been called; and there is insufficient
2674 memory to service the multiple requests.
2676 Except in the instance of (1) above any Denial Of Service is likely to be
2677 transitory because as soon as the connection fails the memory is
2678 subsequently freed again in the SSL_free() call. However there is an
2679 increased risk during this period of application crashes due to the lack of
2680 memory - which would then mean a more serious Denial of Service.
2682 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2683 (CVE-2016-6307 and CVE-2016-6308)
2687 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2688 had to be removed. Primary reason is that vendor assembler can't
2689 assemble our modules with -KPIC flag. As result it, assembly
2690 support, was not even available as option. But its lack means
2691 lack of side-channel resistant code, which is incompatible with
2692 security by todays standards. Fortunately gcc is readily available
2693 prepackaged option, which we firmly point at...
2697 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2699 * Windows command-line tool supports UTF-8 opt-in option for arguments
2700 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2701 (to any value) allows Windows user to access PKCS#12 file generated
2702 with Windows CryptoAPI and protected with non-ASCII password, as well
2703 as files generated under UTF-8 locale on Linux also protected with
2708 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2709 have been disabled by default and removed from DEFAULT, just like RC4.
2710 See the RC4 item below to re-enable both.
2714 * The method for finding the storage location for the Windows RAND seed file
2715 has changed. First we check %RANDFILE%. If that is not set then we check
2716 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2717 all else fails we fall back to C:\.
2721 * The EVP_EncryptUpdate() function has had its return type changed from void
2722 to int. A return of 0 indicates and error while a return of 1 indicates
2727 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2728 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2729 off the constant time implementation for RSA, DSA and DH have been made
2730 no-ops and deprecated.
2734 * Windows RAND implementation was simplified to only get entropy by
2735 calling CryptGenRandom(). Various other RAND-related tickets
2738 *Joseph Wylie Yandle, Rich Salz*
2740 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2741 and `OPENSSL_LH_`, respectively. The old names are available
2742 with API compatibility. They new names are now completely documented.
2746 * Unify TYPE_up_ref(obj) methods signature.
2747 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2748 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2749 int (instead of void) like all others TYPE_up_ref() methods.
2750 So now these methods also check the return value of CRYPTO_atomic_add(),
2751 and the validity of object reference counter.
2753 *fdasilvayy@gmail.com*
2755 * With Windows Visual Studio builds, the .pdb files are installed
2756 alongside the installed libraries and executables. For a static
2757 library installation, ossl_static.pdb is the associate compiler
2758 generated .pdb file to be used when linking programs.
2762 * Remove openssl.spec. Packaging files belong with the packagers.
2766 * Automatic Darwin/OSX configuration has had a refresh, it will now
2767 recognise x86_64 architectures automatically. You can still decide
2768 to build for a different bitness with the environment variable
2769 KERNEL_BITS (can be 32 or 64), for example:
2771 KERNEL_BITS=32 ./config
2775 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2776 256 bit AES and HMAC with SHA256.
2780 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2784 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2788 * To enable users to have their own config files and build file templates,
2789 Configure looks in the directory indicated by the environment variable
2790 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2791 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2792 name and is used as is.
2796 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2797 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2798 X509_CERT_FILE_CTX was removed.
2802 * "shared" builds are now the default. To create only static libraries use
2803 the "no-shared" Configure option.
2807 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2808 All of these option have not worked for some while and are fundamental
2813 * Make various cleanup routines no-ops and mark them as deprecated. Most
2814 global cleanup functions are no longer required because they are handled
2815 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2816 Explicitly de-initing can cause problems (e.g. where a library that uses
2817 OpenSSL de-inits, but an application is still using it). The affected
2818 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2819 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2820 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2821 COMP_zlib_cleanup().
2825 * --strict-warnings no longer enables runtime debugging options
2826 such as REF_DEBUG. Instead, debug options are automatically
2827 enabled with '--debug' builds.
2829 *Andy Polyakov, Emilia Käsper*
2831 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2832 have been moved out of the public header files. New functions for managing
2833 these have been added.
2837 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2838 objects have been moved out of the public header files. New
2839 functions for managing these have been added.
2843 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2844 have been moved out of the public header files. New functions for managing
2845 these have been added.
2849 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2850 moved out of the public header files. New functions for managing these
2855 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2859 * Removed the mk1mf build scripts.
2863 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2864 it is always safe to #include a header now.
2868 * Removed the aged BC-32 config and all its supporting scripts
2872 * Removed support for Ultrix, Netware, and OS/2.
2876 * Add support for HKDF.
2878 *Alessandro Ghedini*
2880 * Add support for blake2b and blake2s
2884 * Added support for "pipelining". Ciphers that have the
2885 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2886 encryptions/decryptions simultaneously. There are currently no built-in
2887 ciphers with this property but the expectation is that engines will be able
2888 to offer it to significantly improve throughput. Support has been extended
2889 into libssl so that multiple records for a single connection can be
2890 processed in one go (for >=TLS 1.1).
2894 * Added the AFALG engine. This is an async capable engine which is able to
2895 offload work to the Linux kernel. In this initial version it only supports
2896 AES128-CBC. The kernel must be version 4.1.0 or greater.
2900 * OpenSSL now uses a new threading API. It is no longer necessary to
2901 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2902 are two supported threading models: pthreads and windows threads. It is
2903 also possible to configure OpenSSL at compile time for "no-threads". The
2904 old threading API should no longer be used. The functions have been
2905 replaced with "no-op" compatibility macros.
2907 *Alessandro Ghedini, Matt Caswell*
2909 * Modify behavior of ALPN to invoke callback after SNI/servername
2910 callback, such that updates to the SSL_CTX affect ALPN.
2914 * Add SSL_CIPHER queries for authentication and key-exchange.
2918 * Changes to the DEFAULT cipherlist:
2919 - Prefer (EC)DHE handshakes over plain RSA.
2920 - Prefer AEAD ciphers over legacy ciphers.
2921 - Prefer ECDSA over RSA when both certificates are available.
2922 - Prefer TLSv1.2 ciphers/PRF.
2923 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2928 * Change the ECC default curve list to be this, in order: x25519,
2929 secp256r1, secp521r1, secp384r1.
2933 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2934 disabled by default. They can be re-enabled using the
2935 enable-weak-ssl-ciphers option to Configure.
2939 * If the server has ALPN configured, but supports no protocols that the
2940 client advertises, send a fatal "no_application_protocol" alert.
2941 This behaviour is SHALL in RFC 7301, though it isn't universally
2942 implemented by other servers.
2946 * Add X25519 support.
2947 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2948 for public and private key encoding using the format documented in
2949 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2950 key generation and key derivation.
2952 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2957 * Deprecate SRP_VBASE_get_by_user.
2958 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2959 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
2960 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2961 seed, even if the seed is configured.
2963 Users should use SRP_VBASE_get1_by_user instead. Note that in
2964 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2965 also that even though configuring the SRP seed attempts to hide
2966 invalid usernames by continuing the handshake with fake
2967 credentials, this behaviour is not constant time and no strong
2968 guarantees are made that the handshake is indistinguishable from
2969 that of a valid user.
2973 * Configuration change; it's now possible to build dynamic engines
2974 without having to build shared libraries and vice versa. This
2975 only applies to the engines in engines/, those in crypto/engine/
2976 will always be built into libcrypto (i.e. "static").
2978 Building dynamic engines is enabled by default; to disable, use
2979 the configuration option "disable-dynamic-engine".
2981 The only requirements for building dynamic engines are the
2982 presence of the DSO module and building with position independent
2983 code, so they will also automatically be disabled if configuring
2984 with "disable-dso" or "disable-pic".
2986 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2987 are also taken away from openssl/opensslconf.h, as they are
2992 * Configuration change; if there is a known flag to compile
2993 position independent code, it will always be applied on the
2994 libcrypto and libssl object files, and never on the application
2995 object files. This means other libraries that use routines from
2996 libcrypto / libssl can be made into shared libraries regardless
2997 of how OpenSSL was configured.
2999 If this isn't desirable, the configuration options "disable-pic"
3000 or "no-pic" can be used to disable the use of PIC. This will
3001 also disable building shared libraries and dynamic engines.
3005 * Removed JPAKE code. It was experimental and has no wide use.
3009 * The INSTALL_PREFIX Makefile variable has been renamed to
3010 DESTDIR. That makes for less confusion on what this variable
3011 is for. Also, the configuration option --install_prefix is
3016 * Heartbeat for TLS has been removed and is disabled by default
3017 for DTLS; configure with enable-heartbeats. Code that uses the
3018 old #define's might need to be updated.
3020 *Emilia Käsper, Rich Salz*
3022 * Rename REF_CHECK to REF_DEBUG.
3026 * New "unified" build system
3028 The "unified" build system is aimed to be a common system for all
3029 platforms we support. With it comes new support for VMS.
3031 This system builds supports building in a different directory tree
3032 than the source tree. It produces one Makefile (for unix family
3033 or lookalikes), or one descrip.mms (for VMS).
3035 The source of information to make the Makefile / descrip.mms is
3036 small files called 'build.info', holding the necessary
3037 information for each directory with source to compile, and a
3038 template in Configurations, like unix-Makefile.tmpl or
3041 With this change, the library names were also renamed on Windows
3042 and on VMS. They now have names that are closer to the standard
3043 on Unix, and include the major version number, and in certain
3044 cases, the architecture they are built for. See "Notes on shared
3045 libraries" in INSTALL.
3047 We rely heavily on the perl module Text::Template.
3051 * Added support for auto-initialisation and de-initialisation of the library.
3052 OpenSSL no longer requires explicit init or deinit routines to be called,
3053 except in certain circumstances. See the OPENSSL_init_crypto() and
3054 OPENSSL_init_ssl() man pages for further information.
3058 * The arguments to the DTLSv1_listen function have changed. Specifically the
3059 "peer" argument is now expected to be a BIO_ADDR object.
3061 * Rewrite of BIO networking library. The BIO library lacked consistent
3062 support of IPv6, and adding it required some more extensive
3063 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3064 which hold all types of addresses and chains of address information.
3065 It also introduces a new API, with functions like BIO_socket,
3066 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3067 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3068 have been adapted accordingly.
3072 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3077 * CRIME protection: disable compression by default, even if OpenSSL is
3078 compiled with zlib enabled. Applications can still enable compression
3079 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3080 using the SSL_CONF library to configure compression.
3084 * The signature of the session callback configured with
3085 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3086 was explicitly marked as `const unsigned char*` instead of
3091 * Always DPURIFY. Remove the use of uninitialized memory in the
3092 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3096 * Removed many obsolete configuration items, including
3097 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3098 MD2_CHAR, MD2_INT, MD2_LONG
3100 IDEA_SHORT, IDEA_LONG
3101 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3103 *Rich Salz, with advice from Andy Polyakov*
3105 * Many BN internals have been moved to an internal header file.
3107 *Rich Salz with help from Andy Polyakov*
3109 * Configuration and writing out the results from it has changed.
3110 Files such as Makefile include/openssl/opensslconf.h and are now
3111 produced through general templates, such as Makefile.in and
3112 crypto/opensslconf.h.in and some help from the perl module
3115 Also, the center of configuration information is no longer
3116 Makefile. Instead, Configure produces a perl module in
3117 configdata.pm which holds most of the config data (in the hash
3118 table %config), the target data that comes from the target
3119 configuration in one of the `Configurations/*.conf` files (in
3124 * To clarify their intended purposes, the Configure options
3125 --prefix and --openssldir change their semantics, and become more
3126 straightforward and less interdependent.
3128 --prefix shall be used exclusively to give the location INSTALLTOP
3129 where programs, scripts, libraries, include files and manuals are
3130 going to be installed. The default is now /usr/local.
3132 --openssldir shall be used exclusively to give the default
3133 location OPENSSLDIR where certificates, private keys, CRLs are
3134 managed. This is also where the default openssl.cnf gets
3136 If the directory given with this option is a relative path, the
3137 values of both the --prefix value and the --openssldir value will
3138 be combined to become OPENSSLDIR.
3139 The default for --openssldir is INSTALLTOP/ssl.
3141 Anyone who uses --openssldir to specify where OpenSSL is to be
3142 installed MUST change to use --prefix instead.
3146 * The GOST engine was out of date and therefore it has been removed. An up
3147 to date GOST engine is now being maintained in an external repository.
3148 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3149 support for GOST ciphersuites (these are only activated if a GOST engine
3154 * EGD is no longer supported by default; use enable-egd when
3157 *Ben Kaduk and Rich Salz*
3159 * The distribution now has Makefile.in files, which are used to
3160 create Makefile's when Configure is run. *Configure must be run
3161 before trying to build now.*
3165 * The return value for SSL_CIPHER_description() for error conditions
3170 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3172 Obtaining and performing DNSSEC validation of TLSA records is
3173 the application's responsibility. The application provides
3174 the TLSA records of its choice to OpenSSL, and these are then
3175 used to authenticate the peer.
3177 The TLSA records need not even come from DNS. They can, for
3178 example, be used to implement local end-entity certificate or
3179 trust-anchor "pinning", where the "pin" data takes the form
3180 of TLSA records, which can augment or replace verification
3181 based on the usual WebPKI public certification authorities.
3185 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3186 continues to support deprecated interfaces in default builds.
3187 However, applications are strongly advised to compile their
3188 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3189 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3190 or the 1.1.0 releases.
3192 In environments in which all applications have been ported to
3193 not use any deprecated interfaces OpenSSL's Configure script
3194 should be used with the --api=1.1.0 option to entirely remove
3195 support for the deprecated features from the library and
3196 unconditionally disable them in the installed headers.
3197 Essentially the same effect can be achieved with the "no-deprecated"
3198 argument to Configure, except that this will always restrict
3199 the build to just the latest API, rather than a fixed API
3202 As applications are ported to future revisions of the API,
3203 they should update their compile-time OPENSSL_API_COMPAT define
3204 accordingly, but in most cases should be able to continue to
3205 compile with later releases.
3207 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3208 0x10000000L and 0x00908000L, respectively. However those
3209 versions did not support the OPENSSL_API_COMPAT feature, and
3210 so applications are not typically tested for explicit support
3211 of just the undeprecated features of either release.
3215 * Add support for setting the minimum and maximum supported protocol.
3216 It can bet set via the SSL_set_min_proto_version() and
3217 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3218 MaxProtocol. It's recommended to use the new APIs to disable
3219 protocols instead of disabling individual protocols using
3220 SSL_set_options() or SSL_CONF's Protocol. This change also
3221 removes support for disabling TLS 1.2 in the OpenSSL TLS
3222 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3226 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3230 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3231 and integrates ECDSA and ECDH functionality into EC. Implementations can
3232 now redirect key generation and no longer need to convert to or from
3235 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3236 include the ec.h header file instead.
3240 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3241 ciphers who are no longer supported and drops support the ephemeral RSA key
3242 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3246 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3247 opaque. For HMAC_CTX, the following constructors and destructors
3250 HMAC_CTX *HMAC_CTX_new(void);
3251 void HMAC_CTX_free(HMAC_CTX *ctx);
3253 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3254 destroy such methods has been added. See EVP_MD_meth_new(3) and
3255 EVP_CIPHER_meth_new(3) for documentation.
3258 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3259 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3260 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3261 an already created structure.
3262 2) For consistency with the majority of our object creators and
3263 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3264 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3265 for deprecated builds.
3269 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3270 cryptographic operations to be performed asynchronously as long as an
3271 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3272 further details. Libssl has also had this capability integrated with the
3273 introduction of the new mode SSL_MODE_ASYNC and associated error
3274 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3275 pages. This work was developed in partnership with Intel Corp.
3279 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3280 always enabled now. If you want to disable the support you should
3281 exclude it using the list of supported ciphers. This also means that the
3282 "-no_ecdhe" option has been removed from s_server.
3286 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3287 SSL_{CTX_}set1_curves() which can set a list.
3291 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3292 curve you want to support using SSL_{CTX_}set1_curves().
3296 * State machine rewrite. The state machine code has been significantly
3297 refactored in order to remove much duplication of code and solve issues
3298 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3299 further details). This change does have some associated API changes.
3300 Notably the SSL_state() function has been removed and replaced by
3301 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3302 SSL_set_state() has been removed altogether. The previous handshake states
3303 defined in ssl.h and ssl3.h have also been removed.
3307 * All instances of the string "ssleay" in the public API were replaced
3308 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3309 Some error codes related to internal RSA_eay API's were renamed.
3313 * The demo files in crypto/threads were moved to demo/threads.
3317 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3320 *Matt Caswell, Rich Salz*
3322 * New ASN.1 embed macro.
3324 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3325 structure is not allocated: it is part of the parent. That is instead of
3333 This reduces memory fragmentation and make it impossible to accidentally
3334 set a mandatory field to NULL.
3336 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3337 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3338 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3343 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3347 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3348 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3349 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3350 DES and RC4 ciphersuites.
3354 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3355 This changes the decoding behaviour for some invalid messages,
3356 though the change is mostly in the more lenient direction, and
3357 legacy behaviour is preserved as much as possible.
3361 * Fix no-stdio build.
3362 *David Woodhouse <David.Woodhouse@intel.com> and also*
3363 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3365 * New testing framework
3366 The testing framework has been largely rewritten and is now using
3367 perl and the perl modules Test::Harness and an extended variant of
3368 Test::More called OpenSSL::Test to do its work. All test scripts in
3369 test/ have been rewritten into test recipes, and all direct calls to
3370 executables in test/Makefile have become individual recipes using the
3371 simplified testing OpenSSL::Test::Simple.
3373 For documentation on our testing modules, do:
3375 perldoc test/testlib/OpenSSL/Test/Simple.pm
3376 perldoc test/testlib/OpenSSL/Test.pm
3380 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3381 are used; the latter aborts on memory leaks (usually checked on exit).
3382 Some undocumented "set malloc, etc., hooks" functions were removed
3383 and others were changed. All are now documented.
3387 * In DSA_generate_parameters_ex, if the provided seed is too short,
3390 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3392 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3393 from RFC4279, RFC4785, RFC5487, RFC5489.
3395 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3396 original RSA_PSK patch.
3400 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3401 era flag was never set throughout the codebase (only read). Also removed
3402 SSL3_FLAGS_POP_BUFFER which was only used if
3403 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3407 * Changed the default name options in the "ca", "crl", "req" and "x509"
3408 to be "oneline" instead of "compat".
3412 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3413 not aware of clients that still exhibit this bug, and the workaround
3414 hasn't been working properly for a while.
3418 * The return type of BIO_number_read() and BIO_number_written() as well as
3419 the corresponding num_read and num_write members in the BIO structure has
3420 changed from unsigned long to uint64_t. On platforms where an unsigned
3421 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3426 * Given the pervasive nature of TLS extensions it is inadvisable to run
3427 OpenSSL without support for them. It also means that maintaining
3428 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3429 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3433 * Removed support for the two export grade static DH ciphersuites
3434 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3435 were newly added (along with a number of other static DH ciphersuites) to
3436 1.0.2. However the two export ones have *never* worked since they were
3437 introduced. It seems strange in any case to be adding new export
3438 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3442 * Version negotiation has been rewritten. In particular SSLv23_method(),
3443 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3444 and turned into macros which simply call the new preferred function names
3445 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3446 should use the new names instead. Also as part of this change the ssl23.h
3447 header file has been removed.
3451 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3452 code and the associated standard is no longer considered fit-for-purpose.
3456 * RT2547 was closed. When generating a private key, try to make the
3457 output file readable only by the owner. This behavior change might
3458 be noticeable when interacting with other software.
3460 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3465 * Added HTTP GET support to the ocsp command.
3469 * Changed default digest for the dgst and enc commands from MD5 to
3474 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3478 * Added support for TLS extended master secret from
3479 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3480 initial patch which was a great help during development.
3484 * All libssl internal structures have been removed from the public header
3485 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3486 now redundant). Users should not attempt to access internal structures
3487 directly. Instead they should use the provided API functions.
3491 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3492 Access to deprecated functions can be re-enabled by running config with
3493 "enable-deprecated". In addition applications wishing to use deprecated
3494 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3495 will, by default, disable some transitive includes that previously existed
3496 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3500 * Added support for OCB mode. OpenSSL has been granted a patent license
3501 compatible with the OpenSSL license for use of OCB. Details are available
3502 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3503 for OCB can be removed by calling config with no-ocb.
3507 * SSLv2 support has been removed. It still supports receiving a SSLv2
3508 compatible client hello.
3512 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3513 done while fixing the error code for the key-too-small case.
3515 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3517 * CA.sh has been removed; use CA.pl instead.
3521 * Removed old DES API.
3525 * Remove various unsupported platforms:
3531 Sinix/ReliantUNIX RM400
3536 16-bit platforms such as WIN16
3540 * Clean up OPENSSL_NO_xxx #define's
3541 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3542 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3543 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3544 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3545 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3546 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3547 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3548 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3549 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3550 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3554 * Cleaned up dead code
3555 Remove all but one '#ifdef undef' which is to be looked at.
3559 * Clean up calling of xxx_free routines.
3560 Just like free(), fix most of the xxx_free routines to accept
3561 NULL. Remove the non-null checks from callers. Save much code.
3565 * Add secure heap for storage of private keys (when possible).
3566 Add BIO_s_secmem(), CBIGNUM, etc.
3567 Contributed by Akamai Technologies under our Corporate CLA.
3571 * Experimental support for a new, fast, unbiased prime candidate generator,
3572 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3574 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3576 * New output format NSS in the sess_id command line tool. This allows
3577 exporting the session id and the master key in NSS keylog format.
3579 *Martin Kaiser <martin@kaiser.cx>*
3581 * Harmonize version and its documentation. -f flag is used to display
3584 *mancha <mancha1@zoho.com>*
3586 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3587 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3589 *mancha <mancha1@zoho.com>*
3591 * Fix some double frees. These are not thought to be exploitable.
3593 *mancha <mancha1@zoho.com>*
3595 * A missing bounds check in the handling of the TLS heartbeat extension
3596 can be used to reveal up to 64k of memory to a connected client or
3599 Thanks for Neel Mehta of Google Security for discovering this bug and to
3600 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3601 preparing the fix ([CVE-2014-0160])
3603 *Adam Langley, Bodo Moeller*
3605 * Fix for the attack described in the paper "Recovering OpenSSL
3606 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3607 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3608 <http://eprint.iacr.org/2014/140>
3610 Thanks to Yuval Yarom and Naomi Benger for discovering this
3611 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3613 *Yuval Yarom and Naomi Benger*
3615 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3616 this fixes a limitation in previous versions of OpenSSL.
3620 * Experimental encrypt-then-mac support.
3622 Experimental support for encrypt then mac from
3623 draft-gutmann-tls-encrypt-then-mac-02.txt
3625 To enable it set the appropriate extension number (0x42 for the test
3626 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3628 For non-compliant peers (i.e. just about everything) this should have no
3631 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3635 * Add EVP support for key wrapping algorithms, to avoid problems with
3636 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3637 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3638 algorithms and include tests cases.
3642 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3647 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3648 MGF1 digest and OAEP label.
3652 * Make openssl verify return errors.
3654 *Chris Palmer <palmer@google.com> and Ben Laurie*
3656 * New function ASN1_TIME_diff to calculate the difference between two
3657 ASN1_TIME structures or one structure and the current time.
3661 * Update fips_test_suite to support multiple command line options. New
3662 test to induce all self test errors in sequence and check expected
3667 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3668 sign or verify all in one operation.
3672 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3673 test programs and fips_test_suite. Includes functionality to parse
3674 the minimal script output of fipsalgest.pl directly.
3678 * Add authorisation parameter to FIPS_module_mode_set().
3682 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3686 * Use separate DRBG fields for internal and external flags. New function
3687 FIPS_drbg_health_check() to perform on demand health checking. Add
3688 generation tests to fips_test_suite with reduced health check interval to
3689 demonstrate periodic health checking. Add "nodh" option to
3690 fips_test_suite to skip very slow DH test.
3694 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3699 * More extensive health check for DRBG checking many more failure modes.
3700 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3701 combination: call this in fips_test_suite.
3705 * Add support for canonical generation of DSA parameter 'g'. See
3708 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3709 POST to handle HMAC cases.
3713 * Add functions FIPS_module_version() and FIPS_module_version_text()
3714 to return numerical and string versions of the FIPS module number.
3718 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3719 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3720 outside the validated module in the FIPS capable OpenSSL.
3724 * Minor change to DRBG entropy callback semantics. In some cases
3725 there is no multiple of the block length between min_len and
3726 max_len. Allow the callback to return more than max_len bytes
3727 of entropy but discard any extra: it is the callback's responsibility
3728 to ensure that the extra data discarded does not impact the
3729 requested amount of entropy.
3733 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3734 information in FIPS186-3, SP800-57 and SP800-131A.
3738 * CCM support via EVP. Interface is very similar to GCM case except we
3739 must supply all data in one chunk (i.e. no update, final) and the
3740 message length must be supplied if AAD is used. Add algorithm test
3745 * Initial version of POST overhaul. Add POST callback to allow the status
3746 of POST to be monitored and/or failures induced. Modify fips_test_suite
3747 to use callback. Always run all selftests even if one fails.
3751 * XTS support including algorithm test driver in the fips_gcmtest program.
3752 Note: this does increase the maximum key length from 32 to 64 bytes but
3753 there should be no binary compatibility issues as existing applications
3754 will never use XTS mode.
3758 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3759 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3760 performs algorithm blocking for unapproved PRNG types. Also do not
3761 set PRNG type in FIPS_mode_set(): leave this to the application.
3762 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3763 the standard OpenSSL PRNG: set additional data to a date time vector.
3767 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3768 This shouldn't present any incompatibility problems because applications
3769 shouldn't be using these directly and any that are will need to rethink
3770 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3774 * Extensive self tests and health checking required by SP800-90 DRBG.
3775 Remove strength parameter from FIPS_drbg_instantiate and always
3776 instantiate at maximum supported strength.
3780 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3784 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3788 * New function DH_compute_key_padded() to compute a DH key and pad with
3789 leading zeroes if needed: this complies with SP800-56A et al.
3793 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3794 anything, incomplete, subject to change and largely untested at present.
3798 * Modify fipscanisteronly build option to only build the necessary object
3799 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3803 * Add experimental option FIPSSYMS to give all symbols in
3804 fipscanister.o and FIPS or fips prefix. This will avoid
3805 conflicts with future versions of OpenSSL. Add perl script
3806 util/fipsas.pl to preprocess assembly language source files
3807 and rename any affected symbols.
3811 * Add selftest checks and algorithm block of non-fips algorithms in
3812 FIPS mode. Remove DES2 from selftests.
3816 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3817 return internal method without any ENGINE dependencies. Add new
3818 tiny fips sign and verify functions.
3822 * New build option no-ec2m to disable characteristic 2 code.
3826 * New build option "fipscanisteronly". This only builds fipscanister.o
3827 and (currently) associated fips utilities. Uses the file Makefile.fips
3828 instead of Makefile.org as the prototype.
3832 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3833 Update fips_gcmtest to use IV generator.
3837 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3838 setting output buffer to NULL. The `*Final` function must be
3839 called although it will not retrieve any additional data. The tag
3840 can be set or retrieved with a ctrl. The IV length is by default 12
3841 bytes (96 bits) but can be set to an alternative value. If the IV
3842 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3847 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3848 underlying do_cipher function handles all cipher semantics itself
3849 including padding and finalisation. This is useful if (for example)
3850 an ENGINE cipher handles block padding itself. The behaviour of
3851 do_cipher is subtly changed if this flag is set: the return value
3852 is the number of characters written to the output buffer (zero is
3853 no longer an error code) or a negative error code. Also if the
3854 input buffer is NULL and length 0 finalisation should be performed.
3858 * If a candidate issuer certificate is already part of the constructed
3859 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3863 * Improve forward-security support: add functions
3865 void SSL_CTX_set_not_resumable_session_callback(
3866 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3867 void SSL_set_not_resumable_session_callback(
3868 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3870 for use by SSL/TLS servers; the callback function will be called whenever a
3871 new session is created, and gets to decide whether the session may be
3872 cached to make it resumable (return 0) or not (return 1). (As by the
3873 SSL/TLS protocol specifications, the session_id sent by the server will be
3874 empty to indicate that the session is not resumable; also, the server will
3875 not generate RFC 4507 (RFC 5077) session tickets.)
3877 A simple reasonable callback implementation is to return is_forward_secure.
3878 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3879 by the SSL/TLS server library, indicating whether it can provide forward
3882 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3884 * New -verify_name option in command line utilities to set verification
3889 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3890 Add CMAC pkey methods.
3894 * Experimental renegotiation in s_server -www mode. If the client
3895 browses /reneg connection is renegotiated. If /renegcert it is
3896 renegotiated requesting a certificate.
3900 * Add an "external" session cache for debugging purposes to s_server. This
3901 should help trace issues which normally are only apparent in deployed
3902 multi-process servers.
3906 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3907 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3908 BIO_set_cipher() and some obscure PEM functions were changed so they
3909 can now return an error. The RAND changes required a change to the
3910 RAND_METHOD structure.
3914 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3915 a gcc attribute to warn if the result of a function is ignored. This
3916 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3917 whose return value is often ignored.
3921 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3922 These allow SCTs (signed certificate timestamps) to be requested and
3923 validated when establishing a connection.
3925 *Rob Percival <robpercival@google.com>*
3930 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3932 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3933 used even when parsing explicit parameters, when loading a encoded key
3934 or calling `EC_GROUP_new_from_ecpkparameters()`/
3935 `EC_GROUP_new_from_ecparameters()`.
3936 This prevents bypass of security hardening and performance gains,
3937 especially for curves with specialized EC_METHODs.
3938 By default, if a key encoded with explicit parameters is loaded and later
3939 encoded, the output is still encoded with explicit parameters, even if
3940 internally a "named" EC_GROUP is used for computation.
3944 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3945 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3946 NULL. After this change, only the cofactor parameter can be NULL. It also
3947 does some minimal sanity checks on the passed order.
3952 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3953 An attack is simple, if the first CMS_recipientInfo is valid but the
3954 second CMS_recipientInfo is chosen ciphertext. If the second
3955 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3956 encryption key will be replaced by garbage, and the message cannot be
3957 decoded, but if the RSA decryption fails, the correct encryption key is
3958 used and the recipient will not notice the attack.
3959 As a work around for this potential attack the length of the decrypted
3960 key must be equal to the cipher default key length, in case the
3961 certifiate is not given and all recipientInfo are tried out.
3962 The old behaviour can be re-enabled in the CMS code by setting the
3963 CMS_DEBUG_DECRYPT flag.
3968 * Document issue with installation paths in diverse Windows builds
3970 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3971 binaries and run-time config file.
3976 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3978 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3979 This changes the size when using the genpkey app when no size is given. It
3980 fixes an omission in earlier changes that changed all RSA, DSA and DH
3981 generation apps to use 2048 bits by default.
3985 * Add FIPS support for Android Arm 64-bit
3987 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3988 Module in Version 2.0.10. For some reason, the corresponding target
3989 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3990 built with FIPS support on Android Arm 64-bit. This omission has been
3993 *Matthias St. Pierre*
3995 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3997 * 0-byte record padding oracle
3999 If an application encounters a fatal protocol error and then calls
4000 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4001 then OpenSSL can respond differently to the calling application if a 0 byte
4002 record is received with invalid padding compared to if a 0 byte record is
4003 received with an invalid MAC. If the application then behaves differently
4004 based on that in a way that is detectable to the remote peer, then this
4005 amounts to a padding oracle that could be used to decrypt data.
4007 In order for this to be exploitable "non-stitched" ciphersuites must be in
4008 use. Stitched ciphersuites are optimised implementations of certain
4009 commonly used ciphersuites. Also the application must call SSL_shutdown()
4010 twice even if a protocol error has occurred (applications should not do
4011 this but some do anyway).
4013 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4014 Aviram, with additional investigation by Steven Collison and Andrew
4015 Hourselt. It was reported to OpenSSL on 10th December 2018.
4020 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4024 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4026 * Microarchitecture timing vulnerability in ECC scalar multiplication
4028 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4029 shown to be vulnerable to a microarchitecture timing side channel attack.
4030 An attacker with sufficient access to mount local timing attacks during
4031 ECDSA signature generation could recover the private key.
4033 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4034 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4040 * Timing vulnerability in DSA signature generation
4042 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4043 timing side channel attack. An attacker could use variations in the signing
4044 algorithm to recover the private key.
4046 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4051 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4052 Module, accidentally introduced while backporting security fixes from the
4053 development branch and hindering the use of ECC in FIPS mode.
4057 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4059 * Client DoS due to large DH parameter
4061 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4062 malicious server can send a very large prime value to the client. This will
4063 cause the client to spend an unreasonably long period of time generating a
4064 key for this prime resulting in a hang until the client has finished. This
4065 could be exploited in a Denial Of Service attack.
4067 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4072 * Cache timing vulnerability in RSA Key Generation
4074 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4075 a cache timing side channel attack. An attacker with sufficient access to
4076 mount cache timing attacks during the RSA key generation process could
4077 recover the private key.
4079 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4080 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4085 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4086 parameter is no longer accepted, as it leads to a corrupt table. NULL
4087 pem_str is reserved for alias entries only.
4091 * Revert blinding in ECDSA sign and instead make problematic addition
4092 length-invariant. Switch even to fixed-length Montgomery multiplication.
4096 * Change generating and checking of primes so that the error rate of not
4097 being prime depends on the intended use based on the size of the input.
4098 For larger primes this will result in more rounds of Miller-Rabin.
4099 The maximal error rate for primes with more than 1080 bits is lowered
4102 *Kurt Roeckx, Annie Yousar*
4104 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4108 * Add blinding to ECDSA and DSA signatures to protect against side channel
4109 attacks discovered by Keegan Ryan (NCC Group).
4113 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4114 now allow empty (zero character) pass phrases.
4118 * Certificate time validation (X509_cmp_time) enforces stricter
4119 compliance with RFC 5280. Fractional seconds and timezone offsets
4120 are no longer allowed.
4124 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4126 * Constructed ASN.1 types with a recursive definition could exceed the stack
4128 Constructed ASN.1 types with a recursive definition (such as can be found
4129 in PKCS7) could eventually exceed the stack given malicious input with
4130 excessive recursion. This could result in a Denial Of Service attack. There
4131 are no such structures used within SSL/TLS that come from untrusted sources
4132 so this is considered safe.
4134 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4140 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4142 * Read/write after SSL object in error state
4144 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4145 mechanism. The intent was that if a fatal error occurred during a handshake
4146 then OpenSSL would move into the error state and would immediately fail if
4147 you attempted to continue the handshake. This works as designed for the
4148 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4149 SSL_connect()), however due to a bug it does not work correctly if
4150 SSL_read() or SSL_write() is called directly. In that scenario, if the
4151 handshake fails then a fatal error will be returned in the initial function
4152 call. If SSL_read()/SSL_write() is subsequently called by the application
4153 for the same SSL object then it will succeed and the data is passed without
4154 being decrypted/encrypted directly from the SSL/TLS record layer.
4156 In order to exploit this issue an application bug would have to be present
4157 that resulted in a call to SSL_read()/SSL_write() being issued after having
4158 already received a fatal error.
4160 This issue was reported to OpenSSL by David Benjamin (Google).
4165 * rsaz_1024_mul_avx2 overflow bug on x86_64
4167 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4168 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4169 Analysis suggests that attacks against RSA and DSA as a result of this
4170 defect would be very difficult to perform and are not believed likely.
4171 Attacks against DH1024 are considered just feasible, because most of the
4172 work necessary to deduce information about a private key may be performed
4173 offline. The amount of resources required for such an attack would be
4174 significant. However, for an attack on TLS to be meaningful, the server
4175 would have to share the DH1024 private key among multiple clients, which is
4176 no longer an option since CVE-2016-0701.
4178 This only affects processors that support the AVX2 but not ADX extensions
4179 like Intel Haswell (4th generation).
4181 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4182 was originally found via the OSS-Fuzz project.
4187 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4189 * bn_sqrx8x_internal carry bug on x86_64
4191 There is a carry propagating bug in the x86_64 Montgomery squaring
4192 procedure. No EC algorithms are affected. Analysis suggests that attacks
4193 against RSA and DSA as a result of this defect would be very difficult to
4194 perform and are not believed likely. Attacks against DH are considered just
4195 feasible (although very difficult) because most of the work necessary to
4196 deduce information about a private key may be performed offline. The amount
4197 of resources required for such an attack would be very significant and
4198 likely only accessible to a limited number of attackers. An attacker would
4199 additionally need online access to an unpatched system using the target
4200 private key in a scenario with persistent DH parameters and a private
4201 key that is shared between multiple clients.
4203 This only affects processors that support the BMI1, BMI2 and ADX extensions
4204 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4206 This issue was reported to OpenSSL by the OSS-Fuzz project.
4211 * Malformed X.509 IPAddressFamily could cause OOB read
4213 If an X.509 certificate has a malformed IPAddressFamily extension,
4214 OpenSSL could do a one-byte buffer overread. The most likely result
4215 would be an erroneous display of the certificate in text format.
4217 This issue was reported to OpenSSL by the OSS-Fuzz project.
4222 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4224 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4225 platform rather than 'mingw'.
4229 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4231 * Truncated packet could crash via OOB read
4233 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4234 cipher is being used, then a truncated packet can cause that host to
4235 perform an out-of-bounds read, usually resulting in a crash.
4237 This issue was reported to OpenSSL by Robert Święcki of Google.
4242 * BN_mod_exp may produce incorrect results on x86_64
4244 There is a carry propagating bug in the x86_64 Montgomery squaring
4245 procedure. No EC algorithms are affected. Analysis suggests that attacks
4246 against RSA and DSA as a result of this defect would be very difficult to
4247 perform and are not believed likely. Attacks against DH are considered just
4248 feasible (although very difficult) because most of the work necessary to
4249 deduce information about a private key may be performed offline. The amount
4250 of resources required for such an attack would be very significant and
4251 likely only accessible to a limited number of attackers. An attacker would
4252 additionally need online access to an unpatched system using the target
4253 private key in a scenario with persistent DH parameters and a private
4254 key that is shared between multiple clients. For example this can occur by
4255 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4256 similar to CVE-2015-3193 but must be treated as a separate problem.
4258 This issue was reported to OpenSSL by the OSS-Fuzz project.
4263 * Montgomery multiplication may produce incorrect results
4265 There is a carry propagating bug in the Broadwell-specific Montgomery
4266 multiplication procedure that handles input lengths divisible by, but
4267 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4268 and DH private keys are impossible. This is because the subroutine in
4269 question is not used in operations with the private key itself and an input
4270 of the attacker's direct choice. Otherwise the bug can manifest itself as
4271 transient authentication and key negotiation failures or reproducible
4272 erroneous outcome of public-key operations with specially crafted input.
4273 Among EC algorithms only Brainpool P-512 curves are affected and one
4274 presumably can attack ECDH key negotiation. Impact was not analyzed in
4275 detail, because pre-requisites for attack are considered unlikely. Namely
4276 multiple clients have to choose the curve in question and the server has to
4277 share the private key among them, neither of which is default behaviour.
4278 Even then only clients that chose the curve will be affected.
4280 This issue was publicly reported as transient failures and was not
4281 initially recognized as a security issue. Thanks to Richard Morgan for
4282 providing reproducible case.
4287 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4288 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4289 prevent issues where no progress is being made and the peer continually
4290 sends unrecognised record types, using up resources processing them.
4294 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4296 * Missing CRL sanity check
4298 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4299 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4300 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4302 This issue only affects the OpenSSL 1.0.2i
4307 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4309 * OCSP Status Request extension unbounded memory growth
4311 A malicious client can send an excessively large OCSP Status Request
4312 extension. If that client continually requests renegotiation, sending a
4313 large OCSP Status Request extension each time, then there will be unbounded
4314 memory growth on the server. This will eventually lead to a Denial Of
4315 Service attack through memory exhaustion. Servers with a default
4316 configuration are vulnerable even if they do not support OCSP. Builds using
4317 the "no-ocsp" build time option are not affected.
4319 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4324 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4327 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4333 * OOB write in MDC2_Update()
4335 An overflow can occur in MDC2_Update() either if called directly or
4336 through the EVP_DigestUpdate() function using MDC2. If an attacker
4337 is able to supply very large amounts of input data after a previous
4338 call to EVP_EncryptUpdate() with a partial block then a length check
4339 can overflow resulting in a heap corruption.
4341 The amount of data needed is comparable to SIZE_MAX which is impractical
4344 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4349 * Malformed SHA512 ticket DoS
4351 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4352 DoS attack where a malformed ticket will result in an OOB read which will
4355 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4356 a custom server callback and ticket lookup mechanism.
4358 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4363 * OOB write in BN_bn2dec()
4365 The function BN_bn2dec() does not check the return value of BN_div_word().
4366 This can cause an OOB write if an application uses this function with an
4367 overly large BIGNUM. This could be a problem if an overly large certificate
4368 or CRL is printed out from an untrusted source. TLS is not affected because
4369 record limits will reject an oversized certificate before it is parsed.
4371 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4376 * OOB read in TS_OBJ_print_bio()
4378 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4379 the total length the OID text representation would use and not the amount
4380 of data written. This will result in OOB reads when large OIDs are
4383 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4388 * Pointer arithmetic undefined behaviour
4390 Avoid some undefined pointer arithmetic
4392 A common idiom in the codebase is to check limits in the following manner:
4395 Where "p" points to some malloc'd data of SIZE bytes and
4398 "len" here could be from some externally supplied data (e.g. from a TLS
4401 The rules of C pointer arithmetic are such that "p + len" is only well
4402 defined where len <= SIZE. Therefore the above idiom is actually
4403 undefined behaviour.
4405 For example this could cause problems if some malloc implementation
4406 provides an address for "p" such that "p + len" actually overflows for
4407 values of len that are too big and therefore p + len < limit.
4409 This issue was reported to OpenSSL by Guido Vranken
4414 * Constant time flag not preserved in DSA signing
4416 Operations in the DSA signing algorithm should run in constant time in
4417 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4418 implementation means that a non-constant time codepath is followed for
4419 certain operations. This has been demonstrated through a cache-timing
4420 attack to be sufficient for an attacker to recover the private DSA key.
4422 This issue was reported by César Pereida (Aalto University), Billy Brumley
4423 (Tampere University of Technology), and Yuval Yarom (The University of
4424 Adelaide and NICTA).
4429 * DTLS buffered message DoS
4431 In a DTLS connection where handshake messages are delivered out-of-order
4432 those messages that OpenSSL is not yet ready to process will be buffered
4433 for later use. Under certain circumstances, a flaw in the logic means that
4434 those messages do not get removed from the buffer even though the handshake
4435 has been completed. An attacker could force up to approx. 15 messages to
4436 remain in the buffer when they are no longer required. These messages will
4437 be cleared when the DTLS connection is closed. The default maximum size for
4438 a message is 100k. Therefore the attacker could force an additional 1500k
4439 to be consumed per connection. By opening many simulataneous connections an
4440 attacker could cause a DoS attack through memory exhaustion.
4442 This issue was reported to OpenSSL by Quan Luo.
4447 * DTLS replay protection DoS
4449 A flaw in the DTLS replay attack protection mechanism means that records
4450 that arrive for future epochs update the replay protection "window" before
4451 the MAC for the record has been validated. This could be exploited by an
4452 attacker by sending a record for the next epoch (which does not have to
4453 decrypt or have a valid MAC), with a very large sequence number. This means
4454 that all subsequent legitimate packets are dropped causing a denial of
4455 service for a specific DTLS connection.
4457 This issue was reported to OpenSSL by the OCAP audit team.
4462 * Certificate message OOB reads
4464 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4465 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4466 theoretical DoS risk but this has not been observed in practice on common
4469 The messages affected are client certificate, client certificate request
4470 and server certificate. As a result the attack can only be performed
4471 against a client or a server which enables client authentication.
4473 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4478 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4480 * Prevent padding oracle in AES-NI CBC MAC check
4482 A MITM attacker can use a padding oracle attack to decrypt traffic
4483 when the connection uses an AES CBC cipher and the server support
4486 This issue was introduced as part of the fix for Lucky 13 padding
4487 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4488 constant time by making sure that always the same bytes are read and
4489 compared against either the MAC or padding bytes. But it no longer
4490 checked that there was enough data to have both the MAC and padding
4493 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4498 * Fix EVP_EncodeUpdate overflow
4500 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4501 Base64 encoding of binary data. If an attacker is able to supply very large
4502 amounts of input data then a length check can overflow resulting in a heap
4505 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4506 the `PEM_write_bio*` family of functions. These are mainly used within the
4507 OpenSSL command line applications, so any application which processes data
4508 from an untrusted source and outputs it as a PEM file should be considered
4509 vulnerable to this issue. User applications that call these APIs directly
4510 with large amounts of untrusted data may also be vulnerable.
4512 This issue was reported by Guido Vranken.
4517 * Fix EVP_EncryptUpdate overflow
4519 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4520 is able to supply very large amounts of input data after a previous call to
4521 EVP_EncryptUpdate() with a partial block then a length check can overflow
4522 resulting in a heap corruption. Following an analysis of all OpenSSL
4523 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4524 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4525 the first called function after an EVP_EncryptInit(), and therefore that
4526 specific call must be safe. The second form is where the length passed to
4527 EVP_EncryptUpdate() can be seen from the code to be some small value and
4528 therefore there is no possibility of an overflow. Since all instances are
4529 one of these two forms, it is believed that there can be no overflows in
4530 internal code due to this problem. It should be noted that
4531 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4532 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4533 of these calls have also been analysed too and it is believed there are no
4534 instances in internal usage where an overflow could occur.
4536 This issue was reported by Guido Vranken.
4541 * Prevent ASN.1 BIO excessive memory allocation
4543 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4544 a short invalid encoding can cause allocation of large amounts of memory
4545 potentially consuming excessive resources or exhausting memory.
4547 Any application parsing untrusted data through d2i BIO functions is
4548 affected. The memory based functions such as d2i_X509() are *not* affected.
4549 Since the memory based functions are used by the TLS library, TLS
4550 applications are not affected.
4552 This issue was reported by Brian Carpenter.
4559 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4560 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4561 in arbitrary stack data being returned in the buffer.
4563 This issue was reported by Guido Vranken.
4568 * Modify behavior of ALPN to invoke callback after SNI/servername
4569 callback, such that updates to the SSL_CTX affect ALPN.
4573 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4578 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4579 methods are enabled and ssl2 is disabled the methods return NULL.
4583 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4585 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4586 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4587 provide any "EXPORT" or "LOW" strength ciphers.
4591 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4592 is by default disabled at build-time. Builds that are not configured with
4593 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4594 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4595 will need to explicitly call either of:
4597 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4599 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4601 as appropriate. Even if either of those is used, or the application
4602 explicitly uses the version-specific SSLv2_method() or its client and
4603 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4604 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4605 ciphers, and SSLv2 56-bit DES are no longer available.
4610 * Fix a double-free in DSA code
4612 A double free bug was discovered when OpenSSL parses malformed DSA private
4613 keys and could lead to a DoS attack or memory corruption for applications
4614 that receive DSA private keys from untrusted sources. This scenario is
4617 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4623 * Disable SRP fake user seed to address a server memory leak.
4625 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4627 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4628 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4629 was changed to ignore the "fake user" SRP seed, even if the seed
4632 Users should use SRP_VBASE_get1_by_user instead. Note that in
4633 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4634 also that even though configuring the SRP seed attempts to hide
4635 invalid usernames by continuing the handshake with fake
4636 credentials, this behaviour is not constant time and no strong
4637 guarantees are made that the handshake is indistinguishable from
4638 that of a valid user.
4643 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4645 In the BN_hex2bn function the number of hex digits is calculated using an
4646 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4647 large values of `i` this can result in `bn_expand` not allocating any
4648 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4649 field as NULL leading to a subsequent NULL ptr deref. For very large values
4650 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4651 In this case memory is allocated to the internal BIGNUM data field, but it
4652 is insufficiently sized leading to heap corruption. A similar issue exists
4653 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4654 is ever called by user applications with very large untrusted hex/dec data.
4655 This is anticipated to be a rare occurrence.
4657 All OpenSSL internal usage of these functions use data that is not expected
4658 to be untrusted, e.g. config file data or application command line
4659 arguments. If user developed applications generate config file data based
4660 on untrusted data then it is possible that this could also lead to security
4661 consequences. This is also anticipated to be rare.
4663 This issue was reported to OpenSSL by Guido Vranken.
4668 * Fix memory issues in `BIO_*printf` functions
4670 The internal `fmtstr` function used in processing a "%s" format string in
4671 the `BIO_*printf` functions could overflow while calculating the length of a
4672 string and cause an OOB read when printing very long strings.
4674 Additionally the internal `doapr_outch` function can attempt to write to an
4675 OOB memory location (at an offset from the NULL pointer) in the event of a
4676 memory allocation failure. In 1.0.2 and below this could be caused where
4677 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4678 could be in processing a very long "%s" format string. Memory leaks can
4681 The first issue may mask the second issue dependent on compiler behaviour.
4682 These problems could enable attacks where large amounts of untrusted data
4683 is passed to the `BIO_*printf` functions. If applications use these functions
4684 in this way then they could be vulnerable. OpenSSL itself uses these
4685 functions when printing out human-readable dumps of ASN.1 data. Therefore
4686 applications that print this data could be vulnerable if the data is from
4687 untrusted sources. OpenSSL command line applications could also be
4688 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4689 as command line arguments.
4691 Libssl is not considered directly vulnerable. Additionally certificates etc
4692 received via remote connections via libssl are also unlikely to be able to
4693 trigger these issues because of message size limits enforced within libssl.
4695 This issue was reported to OpenSSL Guido Vranken.
4700 * Side channel attack on modular exponentiation
4702 A side-channel attack was found which makes use of cache-bank conflicts on
4703 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4704 of RSA keys. The ability to exploit this issue is limited as it relies on
4705 an attacker who has control of code in a thread running on the same
4706 hyper-threaded core as the victim thread which is performing decryptions.
4708 This issue was reported to OpenSSL by Yuval Yarom, The University of
4709 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4710 Nadia Heninger, University of Pennsylvania with more information at
4711 <http://cachebleed.info>.
4716 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4717 if no keysize is specified with default_bits. This fixes an
4718 omission in an earlier change that changed all RSA/DSA key generation
4719 apps to use 2048 bits by default.
4723 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4725 * DH small subgroups
4727 Historically OpenSSL only ever generated DH parameters based on "safe"
4728 primes. More recently (in version 1.0.2) support was provided for
4729 generating X9.42 style parameter files such as those required for RFC 5114
4730 support. The primes used in such files may not be "safe". Where an
4731 application is using DH configured with parameters based on primes that are
4732 not "safe" then an attacker could use this fact to find a peer's private
4733 DH exponent. This attack requires that the attacker complete multiple
4734 handshakes in which the peer uses the same private DH exponent. For example
4735 this could be used to discover a TLS server's private DH exponent if it's
4736 reusing the private DH exponent or it's using a static DH ciphersuite.
4738 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4739 TLS. It is not on by default. If the option is not set then the server
4740 reuses the same private DH exponent for the life of the server process and
4741 would be vulnerable to this attack. It is believed that many popular
4742 applications do set this option and would therefore not be at risk.
4744 The fix for this issue adds an additional check where a "q" parameter is
4745 available (as is the case in X9.42 based parameters). This detects the
4746 only known attack, and is the only possible defense for static DH
4747 ciphersuites. This could have some performance impact.
4749 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4750 default and cannot be disabled. This could have some performance impact.
4752 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4757 * SSLv2 doesn't block disabled ciphers
4759 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4760 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4761 been disabled, provided that the SSLv2 protocol was not also disabled via
4764 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4765 and Sebastian Schinzel.
4770 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4772 * BN_mod_exp may produce incorrect results on x86_64
4774 There is a carry propagating bug in the x86_64 Montgomery squaring
4775 procedure. No EC algorithms are affected. Analysis suggests that attacks
4776 against RSA and DSA as a result of this defect would be very difficult to
4777 perform and are not believed likely. Attacks against DH are considered just
4778 feasible (although very difficult) because most of the work necessary to
4779 deduce information about a private key may be performed offline. The amount
4780 of resources required for such an attack would be very significant and
4781 likely only accessible to a limited number of attackers. An attacker would
4782 additionally need online access to an unpatched system using the target
4783 private key in a scenario with persistent DH parameters and a private
4784 key that is shared between multiple clients. For example this can occur by
4785 default in OpenSSL DHE based SSL/TLS ciphersuites.
4787 This issue was reported to OpenSSL by Hanno Böck.
4792 * Certificate verify crash with missing PSS parameter
4794 The signature verification routines will crash with a NULL pointer
4795 dereference if presented with an ASN.1 signature using the RSA PSS
4796 algorithm and absent mask generation function parameter. Since these
4797 routines are used to verify certificate signature algorithms this can be
4798 used to crash any certificate verification operation and exploited in a
4799 DoS attack. Any application which performs certificate verification is
4800 vulnerable including OpenSSL clients and servers which enable client
4803 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4808 * X509_ATTRIBUTE memory leak
4810 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4811 memory. This structure is used by the PKCS#7 and CMS routines so any
4812 application which reads PKCS#7 or CMS data from untrusted sources is
4813 affected. SSL/TLS is not affected.
4815 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4821 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4822 This changes the decoding behaviour for some invalid messages,
4823 though the change is mostly in the more lenient direction, and
4824 legacy behaviour is preserved as much as possible.
4828 * In DSA_generate_parameters_ex, if the provided seed is too short,
4831 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4833 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4835 * Alternate chains certificate forgery
4837 During certificate verification, OpenSSL will attempt to find an
4838 alternative certificate chain if the first attempt to build such a chain
4839 fails. An error in the implementation of this logic can mean that an
4840 attacker could cause certain checks on untrusted certificates to be
4841 bypassed, such as the CA flag, enabling them to use a valid leaf
4842 certificate to act as a CA and "issue" an invalid certificate.
4844 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4849 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4851 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4852 incompatibility in the handling of HMAC. The previous ABI has now been
4857 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4859 * Malformed ECParameters causes infinite loop
4861 When processing an ECParameters structure OpenSSL enters an infinite loop
4862 if the curve specified is over a specially malformed binary polynomial
4865 This can be used to perform denial of service against any
4866 system which processes public keys, certificate requests or
4867 certificates. This includes TLS clients and TLS servers with
4868 client authentication enabled.
4870 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4875 * Exploitable out-of-bounds read in X509_cmp_time
4877 X509_cmp_time does not properly check the length of the ASN1_TIME
4878 string and can read a few bytes out of bounds. In addition,
4879 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4882 An attacker can use this to craft malformed certificates and CRLs of
4883 various sizes and potentially cause a segmentation fault, resulting in
4884 a DoS on applications that verify certificates or CRLs. TLS clients
4885 that verify CRLs are affected. TLS clients and servers with client
4886 authentication enabled may be affected if they use custom verification
4889 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4890 independently by Hanno Böck.
4895 * PKCS7 crash with missing EnvelopedContent
4897 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4898 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4899 with missing content and trigger a NULL pointer dereference on parsing.
4901 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4902 structures from untrusted sources are affected. OpenSSL clients and
4903 servers are not affected.
4905 This issue was reported to OpenSSL by Michal Zalewski (Google).
4910 * CMS verify infinite loop with unknown hash function
4912 When verifying a signedData message the CMS code can enter an infinite loop
4913 if presented with an unknown hash function OID. This can be used to perform
4914 denial of service against any system which verifies signedData messages using
4916 This issue was reported to OpenSSL by Johannes Bauer.
4921 * Race condition handling NewSessionTicket
4923 If a NewSessionTicket is received by a multi-threaded client when attempting to
4924 reuse a previous ticket then a race condition can occur potentially leading to
4925 a double free of the ticket data.
4930 * Only support 256-bit or stronger elliptic curves with the
4931 'ecdh_auto' setting (server) or by default (client). Of supported
4932 curves, prefer P-256 (both).
4936 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4938 * ClientHello sigalgs DoS fix
4940 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4941 invalid signature algorithms extension a NULL pointer dereference will
4942 occur. This can be exploited in a DoS attack against the server.
4944 This issue was was reported to OpenSSL by David Ramos of Stanford
4948 *Stephen Henson and Matt Caswell*
4950 * Multiblock corrupted pointer fix
4952 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4953 feature only applies on 64 bit x86 architecture platforms that support AES
4954 NI instructions. A defect in the implementation of "multiblock" can cause
4955 OpenSSL's internal write buffer to become incorrectly set to NULL when
4956 using non-blocking IO. Typically, when the user application is using a
4957 socket BIO for writing, this will only result in a failed connection.
4958 However if some other BIO is used then it is likely that a segmentation
4959 fault will be triggered, thus enabling a potential DoS attack.
4961 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4966 * Segmentation fault in DTLSv1_listen fix
4968 The DTLSv1_listen function is intended to be stateless and processes the
4969 initial ClientHello from many peers. It is common for user code to loop
4970 over the call to DTLSv1_listen until a valid ClientHello is received with
4971 an associated cookie. A defect in the implementation of DTLSv1_listen means
4972 that state is preserved in the SSL object from one invocation to the next
4973 that can lead to a segmentation fault. Errors processing the initial
4974 ClientHello can trigger this scenario. An example of such an error could be
4975 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4978 This issue was reported to OpenSSL by Per Allansson.
4983 * Segmentation fault in ASN1_TYPE_cmp fix
4985 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4986 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4987 certificate signature algorithm consistency this can be used to crash any
4988 certificate verification operation and exploited in a DoS attack. Any
4989 application which performs certificate verification is vulnerable including
4990 OpenSSL clients and servers which enable client authentication.
4995 * Segmentation fault for invalid PSS parameters fix
4997 The signature verification routines will crash with a NULL pointer
4998 dereference if presented with an ASN.1 signature using the RSA PSS
4999 algorithm and invalid parameters. Since these routines are used to verify
5000 certificate signature algorithms this can be used to crash any
5001 certificate verification operation and exploited in a DoS attack. Any
5002 application which performs certificate verification is vulnerable including
5003 OpenSSL clients and servers which enable client authentication.
5005 This issue was was reported to OpenSSL by Brian Carpenter.
5010 * ASN.1 structure reuse memory corruption fix
5012 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5013 memory corruption via an invalid write. Such reuse is and has been
5014 strongly discouraged and is believed to be rare.
5016 Applications that parse structures containing CHOICE or ANY DEFINED BY
5017 components may be affected. Certificate parsing (d2i_X509 and related
5018 functions) are however not affected. OpenSSL clients and servers are
5024 * PKCS7 NULL pointer dereferences fix
5026 The PKCS#7 parsing code does not handle missing outer ContentInfo
5027 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5028 missing content and trigger a NULL pointer dereference on parsing.
5030 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5031 otherwise parse PKCS#7 structures from untrusted sources are
5032 affected. OpenSSL clients and servers are not affected.
5034 This issue was reported to OpenSSL by Michal Zalewski (Google).
5039 * DoS via reachable assert in SSLv2 servers fix
5041 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5042 servers that both support SSLv2 and enable export cipher suites by sending
5043 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5045 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5046 (OpenSSL development team).
5051 * Empty CKE with client auth and DHE fix
5053 If client auth is used then a server can seg fault in the event of a DHE
5054 ciphersuite being selected and a zero length ClientKeyExchange message
5055 being sent by the client. This could be exploited in a DoS attack.
5060 * Handshake with unseeded PRNG fix
5062 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5063 with an unseeded PRNG. The conditions are:
5064 - The client is on a platform where the PRNG has not been seeded
5065 automatically, and the user has not seeded manually
5066 - A protocol specific client method version has been used (i.e. not
5067 SSL_client_methodv23)
5068 - A ciphersuite is used that does not require additional random data from
5069 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5071 If the handshake succeeds then the client random that has been used will
5072 have been generated from a PRNG with insufficient entropy and therefore the
5073 output may be predictable.
5075 For example using the following command with an unseeded openssl will
5076 succeed on an unpatched platform:
5078 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5083 * Use After Free following d2i_ECPrivatekey error fix
5085 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5086 could cause a use after free condition. This, in turn, could cause a double
5087 free in several private key parsing functions (such as d2i_PrivateKey
5088 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5089 for applications that receive EC private keys from untrusted
5090 sources. This scenario is considered rare.
5092 This issue was discovered by the BoringSSL project and fixed in their
5098 * X509_to_X509_REQ NULL pointer deref fix
5100 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5101 the certificate key is invalid. This function is rarely used in practice.
5103 This issue was discovered by Brian Carpenter.
5108 * Removed the export ciphers from the DEFAULT ciphers
5112 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5114 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5115 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5116 So far those who have to target multiple platforms would compromise
5117 and argue that binary targeting say ARMv5 would still execute on
5118 ARMv8. "Universal" build resolves this compromise by providing
5119 near-optimal performance even on newer platforms.
5123 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5124 (other platforms pending).
5126 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5128 * Add support for the SignedCertificateTimestampList certificate and
5129 OCSP response extensions from RFC6962.
5133 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5134 for corner cases. (Certain input points at infinity could lead to
5135 bogus results, with non-infinity inputs mapped to infinity too.)
5139 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5140 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5141 common cases are optimized and there still is room for further
5142 improvements. Vector Permutation AES for Altivec is also added.
5146 * Add support for little-endian ppc64 Linux target.
5148 *Marcelo Cerri (IBM)*
5150 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5151 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5152 are optimized and there still is room for further improvements.
5153 Both 32- and 64-bit modes are supported.
5155 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5157 * Improved ARMv7 NEON support.
5161 * Support for SPARC Architecture 2011 crypto extensions, first
5162 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5163 SHA256/512, MD5, GHASH and modular exponentiation.
5165 *Andy Polyakov, David Miller*
5167 * Accelerated modular exponentiation for Intel processors, a.k.a.
5170 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5172 * Support for new and upcoming Intel processors, including AVX2,
5173 BMI and SHA ISA extensions. This includes additional "stitched"
5174 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5177 This work was sponsored by Intel Corp.
5181 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5182 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5183 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5187 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5188 this fixes a limitation in previous versions of OpenSSL.
5192 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5193 MGF1 digest and OAEP label.
5197 * Add EVP support for key wrapping algorithms, to avoid problems with
5198 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5199 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5200 algorithms and include tests cases.
5204 * Add functions to allocate and set the fields of an ECDSA_METHOD
5207 *Douglas E. Engert, Steve Henson*
5209 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5210 difference in days and seconds between two tm or ASN1_TIME structures.
5214 * Add -rev test option to s_server to just reverse order of characters
5215 received by client and send back to server. Also prints an abbreviated
5216 summary of the connection parameters.
5220 * New option -brief for s_client and s_server to print out a brief summary
5221 of connection parameters.
5225 * Add callbacks for arbitrary TLS extensions.
5227 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5229 * New option -crl_download in several openssl utilities to download CRLs
5230 from CRLDP extension in certificates.
5234 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5238 * New function X509_CRL_diff to generate a delta CRL from the difference
5239 of two full CRLs. Add support to "crl" utility.
5243 * New functions to set lookup_crls function and to retrieve
5244 X509_STORE from X509_STORE_CTX.
5248 * Print out deprecated issuer and subject unique ID fields in
5253 * Extend OCSP I/O functions so they can be used for simple general purpose
5254 HTTP as well as OCSP. New wrapper function which can be used to download
5255 CRLs using the OCSP API.
5259 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5263 * `SSL_CONF*` functions. These provide a common framework for application
5264 configuration using configuration files or command lines.
5268 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5269 message callback and prints the results. Needs compile time option
5270 "enable-ssl-trace". New options to s_client and s_server to enable
5275 * New ctrl and macro to retrieve supported points extensions.
5276 Print out extension in s_server and s_client.
5280 * New functions to retrieve certificate signature and signature
5285 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5290 * New Suite B modes for TLS code. These use and enforce the requirements
5291 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5292 only use Suite B curves. The Suite B modes can be set by using the
5293 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5297 * New chain verification flags for Suite B levels of security. Check
5298 algorithms are acceptable when flags are set in X509_verify_cert.
5302 * Make tls1_check_chain return a set of flags indicating checks passed
5303 by a certificate chain. Add additional tests to handle client
5304 certificates: checks for matching certificate type and issuer name
5309 * If an attempt is made to use a signature algorithm not in the peer
5310 preference list abort the handshake. If client has no suitable
5311 signature algorithms in response to a certificate request do not
5312 use the certificate.
5316 * If server EC tmp key is not in client preference list abort handshake.
5320 * Add support for certificate stores in CERT structure. This makes it
5321 possible to have different stores per SSL structure or one store in
5322 the parent SSL_CTX. Include distinct stores for certificate chain
5323 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5324 to build and store a certificate chain in CERT structure: returning
5325 an error if the chain cannot be built: this will allow applications
5326 to test if a chain is correctly configured.
5328 Note: if the CERT based stores are not set then the parent SSL_CTX
5329 store is used to retain compatibility with existing behaviour.
5333 * New function ssl_set_client_disabled to set a ciphersuite disabled
5334 mask based on the current session, check mask when sending client
5335 hello and checking the requested ciphersuite.
5339 * New ctrls to retrieve and set certificate types in a certificate
5340 request message. Print out received values in s_client. If certificate
5341 types is not set with custom values set sensible values based on
5342 supported signature algorithms.
5346 * Support for distinct client and server supported signature algorithms.
5350 * Add certificate callback. If set this is called whenever a certificate
5351 is required by client or server. An application can decide which
5352 certificate chain to present based on arbitrary criteria: for example
5353 supported signature algorithms. Add very simple example to s_server.
5354 This fixes many of the problems and restrictions of the existing client
5355 certificate callback: for example you can now clear an existing
5356 certificate and specify the whole chain.
5360 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5361 the certificate can be used for (if anything). Set valid_flags field
5362 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5363 to have similar checks in it.
5365 Add new "cert_flags" field to CERT structure and include a "strict mode".
5366 This enforces some TLS certificate requirements (such as only permitting
5367 certificate signature algorithms contained in the supported algorithms
5368 extension) which some implementations ignore: this option should be used
5369 with caution as it could cause interoperability issues.
5373 * Update and tidy signature algorithm extension processing. Work out
5374 shared signature algorithms based on preferences and peer algorithms
5375 and print them out in s_client and s_server. Abort handshake if no
5376 shared signature algorithms.
5380 * Add new functions to allow customised supported signature algorithms
5381 for SSL and SSL_CTX structures. Add options to s_client and s_server
5386 * New function SSL_certs_clear() to delete all references to certificates
5387 from an SSL structure. Before this once a certificate had been added
5388 it couldn't be removed.
5392 * Integrate hostname, email address and IP address checking with certificate
5393 verification. New verify options supporting checking in openssl utility.
5397 * Fixes and wildcard matching support to hostname and email checking
5398 functions. Add manual page.
5400 *Florian Weimer (Red Hat Product Security Team)*
5402 * New functions to check a hostname email or IP address against a
5403 certificate. Add options x509 utility to print results of checks against
5408 * Fix OCSP checking.
5410 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5412 * Initial experimental support for explicitly trusted non-root CAs.
5413 OpenSSL still tries to build a complete chain to a root but if an
5414 intermediate CA has a trust setting included that is used. The first
5415 setting is used: whether to trust (e.g., -addtrust option to the x509
5420 * Add -trusted_first option which attempts to find certificates in the
5421 trusted store even if an untrusted chain is also supplied.
5425 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5426 platform support for Linux and Android.
5430 * Support for linux-x32, ILP32 environment in x86_64 framework.
5434 * Experimental multi-implementation support for FIPS capable OpenSSL.
5435 When in FIPS mode the approved implementations are used as normal,
5436 when not in FIPS mode the internal unapproved versions are used instead.
5437 This means that the FIPS capable OpenSSL isn't forced to use the
5438 (often lower performance) FIPS implementations outside FIPS mode.
5442 * Transparently support X9.42 DH parameters when calling
5443 PEM_read_bio_DHparameters. This means existing applications can handle
5444 the new parameter format automatically.
5448 * Initial experimental support for X9.42 DH parameter format: mainly
5449 to support use of 'q' parameter for RFC5114 parameters.
5453 * Add DH parameters from RFC5114 including test data to dhtest.
5457 * Support for automatic EC temporary key parameter selection. If enabled
5458 the most preferred EC parameters are automatically used instead of
5459 hardcoded fixed parameters. Now a server just has to call:
5460 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5461 support ECDH and use the most appropriate parameters.
5465 * Enhance and tidy EC curve and point format TLS extension code. Use
5466 static structures instead of allocation if default values are used.
5467 New ctrls to set curves we wish to support and to retrieve shared curves.
5468 Print out shared curves in s_server. New options to s_server and s_client
5469 to set list of supported curves.
5473 * New ctrls to retrieve supported signature algorithms and
5474 supported curve values as an array of NIDs. Extend openssl utility
5475 to print out received values.
5479 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5480 between NIDs and the more common NIST names such as "P-256". Enhance
5481 ecparam utility and ECC method to recognise the NIST names for curves.
5485 * Enhance SSL/TLS certificate chain handling to support different
5486 chains for each certificate instead of one chain in the parent SSL_CTX.
5490 * Support for fixed DH ciphersuite client authentication: where both
5491 server and client use DH certificates with common parameters.
5495 * Support for fixed DH ciphersuites: those requiring DH server
5500 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5502 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5503 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5504 X509_CINF_get_signature were reverted post internal team review.
5509 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5511 * OCSP Status Request extension unbounded memory growth
5513 A malicious client can send an excessively large OCSP Status Request
5514 extension. If that client continually requests renegotiation, sending a
5515 large OCSP Status Request extension each time, then there will be unbounded
5516 memory growth on the server. This will eventually lead to a Denial Of
5517 Service attack through memory exhaustion. Servers with a default
5518 configuration are vulnerable even if they do not support OCSP. Builds using
5519 the "no-ocsp" build time option are not affected.
5521 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5526 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5529 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5535 * OOB write in MDC2_Update()
5537 An overflow can occur in MDC2_Update() either if called directly or
5538 through the EVP_DigestUpdate() function using MDC2. If an attacker
5539 is able to supply very large amounts of input data after a previous
5540 call to EVP_EncryptUpdate() with a partial block then a length check
5541 can overflow resulting in a heap corruption.
5543 The amount of data needed is comparable to SIZE_MAX which is impractical
5546 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5551 * Malformed SHA512 ticket DoS
5553 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5554 DoS attack where a malformed ticket will result in an OOB read which will
5557 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5558 a custom server callback and ticket lookup mechanism.
5560 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5565 * OOB write in BN_bn2dec()
5567 The function BN_bn2dec() does not check the return value of BN_div_word().
5568 This can cause an OOB write if an application uses this function with an
5569 overly large BIGNUM. This could be a problem if an overly large certificate
5570 or CRL is printed out from an untrusted source. TLS is not affected because
5571 record limits will reject an oversized certificate before it is parsed.
5573 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5578 * OOB read in TS_OBJ_print_bio()
5580 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5581 the total length the OID text representation would use and not the amount
5582 of data written. This will result in OOB reads when large OIDs are
5585 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5590 * Pointer arithmetic undefined behaviour
5592 Avoid some undefined pointer arithmetic
5594 A common idiom in the codebase is to check limits in the following manner:
5597 Where "p" points to some malloc'd data of SIZE bytes and
5600 "len" here could be from some externally supplied data (e.g. from a TLS
5603 The rules of C pointer arithmetic are such that "p + len" is only well
5604 defined where len <= SIZE. Therefore the above idiom is actually
5605 undefined behaviour.
5607 For example this could cause problems if some malloc implementation
5608 provides an address for "p" such that "p + len" actually overflows for
5609 values of len that are too big and therefore p + len < limit.
5611 This issue was reported to OpenSSL by Guido Vranken
5616 * Constant time flag not preserved in DSA signing
5618 Operations in the DSA signing algorithm should run in constant time in
5619 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5620 implementation means that a non-constant time codepath is followed for
5621 certain operations. This has been demonstrated through a cache-timing
5622 attack to be sufficient for an attacker to recover the private DSA key.
5624 This issue was reported by César Pereida (Aalto University), Billy Brumley
5625 (Tampere University of Technology), and Yuval Yarom (The University of
5626 Adelaide and NICTA).
5631 * DTLS buffered message DoS
5633 In a DTLS connection where handshake messages are delivered out-of-order
5634 those messages that OpenSSL is not yet ready to process will be buffered
5635 for later use. Under certain circumstances, a flaw in the logic means that
5636 those messages do not get removed from the buffer even though the handshake
5637 has been completed. An attacker could force up to approx. 15 messages to
5638 remain in the buffer when they are no longer required. These messages will
5639 be cleared when the DTLS connection is closed. The default maximum size for
5640 a message is 100k. Therefore the attacker could force an additional 1500k
5641 to be consumed per connection. By opening many simulataneous connections an
5642 attacker could cause a DoS attack through memory exhaustion.
5644 This issue was reported to OpenSSL by Quan Luo.
5649 * DTLS replay protection DoS
5651 A flaw in the DTLS replay attack protection mechanism means that records
5652 that arrive for future epochs update the replay protection "window" before
5653 the MAC for the record has been validated. This could be exploited by an
5654 attacker by sending a record for the next epoch (which does not have to
5655 decrypt or have a valid MAC), with a very large sequence number. This means
5656 that all subsequent legitimate packets are dropped causing a denial of
5657 service for a specific DTLS connection.
5659 This issue was reported to OpenSSL by the OCAP audit team.
5664 * Certificate message OOB reads
5666 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5667 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5668 theoretical DoS risk but this has not been observed in practice on common
5671 The messages affected are client certificate, client certificate request
5672 and server certificate. As a result the attack can only be performed
5673 against a client or a server which enables client authentication.
5675 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5680 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5682 * Prevent padding oracle in AES-NI CBC MAC check
5684 A MITM attacker can use a padding oracle attack to decrypt traffic
5685 when the connection uses an AES CBC cipher and the server support
5688 This issue was introduced as part of the fix for Lucky 13 padding
5689 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5690 constant time by making sure that always the same bytes are read and
5691 compared against either the MAC or padding bytes. But it no longer
5692 checked that there was enough data to have both the MAC and padding
5695 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5700 * Fix EVP_EncodeUpdate overflow
5702 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5703 Base64 encoding of binary data. If an attacker is able to supply very large
5704 amounts of input data then a length check can overflow resulting in a heap
5707 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5708 the `PEM_write_bio*` family of functions. These are mainly used within the
5709 OpenSSL command line applications, so any application which processes data
5710 from an untrusted source and outputs it as a PEM file should be considered
5711 vulnerable to this issue. User applications that call these APIs directly
5712 with large amounts of untrusted data may also be vulnerable.
5714 This issue was reported by Guido Vranken.
5719 * Fix EVP_EncryptUpdate overflow
5721 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5722 is able to supply very large amounts of input data after a previous call to
5723 EVP_EncryptUpdate() with a partial block then a length check can overflow
5724 resulting in a heap corruption. Following an analysis of all OpenSSL
5725 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5726 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5727 the first called function after an EVP_EncryptInit(), and therefore that
5728 specific call must be safe. The second form is where the length passed to
5729 EVP_EncryptUpdate() can be seen from the code to be some small value and
5730 therefore there is no possibility of an overflow. Since all instances are
5731 one of these two forms, it is believed that there can be no overflows in
5732 internal code due to this problem. It should be noted that
5733 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5734 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5735 of these calls have also been analysed too and it is believed there are no
5736 instances in internal usage where an overflow could occur.
5738 This issue was reported by Guido Vranken.
5743 * Prevent ASN.1 BIO excessive memory allocation
5745 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5746 a short invalid encoding can casuse allocation of large amounts of memory
5747 potentially consuming excessive resources or exhausting memory.
5749 Any application parsing untrusted data through d2i BIO functions is
5750 affected. The memory based functions such as d2i_X509() are *not* affected.
5751 Since the memory based functions are used by the TLS library, TLS
5752 applications are not affected.
5754 This issue was reported by Brian Carpenter.
5761 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5762 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5763 in arbitrary stack data being returned in the buffer.
5765 This issue was reported by Guido Vranken.
5770 * Modify behavior of ALPN to invoke callback after SNI/servername
5771 callback, such that updates to the SSL_CTX affect ALPN.
5775 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5780 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5781 methods are enabled and ssl2 is disabled the methods return NULL.
5785 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5787 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5788 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5789 provide any "EXPORT" or "LOW" strength ciphers.
5793 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5794 is by default disabled at build-time. Builds that are not configured with
5795 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5796 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5797 will need to explicitly call either of:
5799 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5801 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5803 as appropriate. Even if either of those is used, or the application
5804 explicitly uses the version-specific SSLv2_method() or its client and
5805 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5806 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5807 ciphers, and SSLv2 56-bit DES are no longer available.
5812 * Fix a double-free in DSA code
5814 A double free bug was discovered when OpenSSL parses malformed DSA private
5815 keys and could lead to a DoS attack or memory corruption for applications
5816 that receive DSA private keys from untrusted sources. This scenario is
5819 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5825 * Disable SRP fake user seed to address a server memory leak.
5827 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5829 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5830 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5831 was changed to ignore the "fake user" SRP seed, even if the seed
5834 Users should use SRP_VBASE_get1_by_user instead. Note that in
5835 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5836 also that even though configuring the SRP seed attempts to hide
5837 invalid usernames by continuing the handshake with fake
5838 credentials, this behaviour is not constant time and no strong
5839 guarantees are made that the handshake is indistinguishable from
5840 that of a valid user.
5845 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5847 In the BN_hex2bn function the number of hex digits is calculated using an
5848 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5849 large values of `i` this can result in `bn_expand` not allocating any
5850 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5851 field as NULL leading to a subsequent NULL ptr deref. For very large values
5852 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5853 In this case memory is allocated to the internal BIGNUM data field, but it
5854 is insufficiently sized leading to heap corruption. A similar issue exists
5855 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5856 is ever called by user applications with very large untrusted hex/dec data.
5857 This is anticipated to be a rare occurrence.
5859 All OpenSSL internal usage of these functions use data that is not expected
5860 to be untrusted, e.g. config file data or application command line
5861 arguments. If user developed applications generate config file data based
5862 on untrusted data then it is possible that this could also lead to security
5863 consequences. This is also anticipated to be rare.
5865 This issue was reported to OpenSSL by Guido Vranken.
5870 * Fix memory issues in `BIO_*printf` functions
5872 The internal `fmtstr` function used in processing a "%s" format string in
5873 the `BIO_*printf` functions could overflow while calculating the length of a
5874 string and cause an OOB read when printing very long strings.
5876 Additionally the internal `doapr_outch` function can attempt to write to an
5877 OOB memory location (at an offset from the NULL pointer) in the event of a
5878 memory allocation failure. In 1.0.2 and below this could be caused where
5879 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5880 could be in processing a very long "%s" format string. Memory leaks can
5883 The first issue may mask the second issue dependent on compiler behaviour.
5884 These problems could enable attacks where large amounts of untrusted data
5885 is passed to the `BIO_*printf` functions. If applications use these functions
5886 in this way then they could be vulnerable. OpenSSL itself uses these
5887 functions when printing out human-readable dumps of ASN.1 data. Therefore
5888 applications that print this data could be vulnerable if the data is from
5889 untrusted sources. OpenSSL command line applications could also be
5890 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5891 as command line arguments.
5893 Libssl is not considered directly vulnerable. Additionally certificates etc
5894 received via remote connections via libssl are also unlikely to be able to
5895 trigger these issues because of message size limits enforced within libssl.
5897 This issue was reported to OpenSSL Guido Vranken.
5902 * Side channel attack on modular exponentiation
5904 A side-channel attack was found which makes use of cache-bank conflicts on
5905 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5906 of RSA keys. The ability to exploit this issue is limited as it relies on
5907 an attacker who has control of code in a thread running on the same
5908 hyper-threaded core as the victim thread which is performing decryptions.
5910 This issue was reported to OpenSSL by Yuval Yarom, The University of
5911 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5912 Nadia Heninger, University of Pennsylvania with more information at
5913 <http://cachebleed.info>.
5918 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5919 if no keysize is specified with default_bits. This fixes an
5920 omission in an earlier change that changed all RSA/DSA key generation
5921 apps to use 2048 bits by default.
5925 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5927 * Protection for DH small subgroup attacks
5929 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5930 switched on by default and cannot be disabled. This could have some
5935 * SSLv2 doesn't block disabled ciphers
5937 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5938 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5939 been disabled, provided that the SSLv2 protocol was not also disabled via
5942 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5943 and Sebastian Schinzel.
5948 * Reject DH handshakes with parameters shorter than 1024 bits.
5952 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5954 * Certificate verify crash with missing PSS parameter
5956 The signature verification routines will crash with a NULL pointer
5957 dereference if presented with an ASN.1 signature using the RSA PSS
5958 algorithm and absent mask generation function parameter. Since these
5959 routines are used to verify certificate signature algorithms this can be
5960 used to crash any certificate verification operation and exploited in a
5961 DoS attack. Any application which performs certificate verification is
5962 vulnerable including OpenSSL clients and servers which enable client
5965 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5970 * X509_ATTRIBUTE memory leak
5972 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5973 memory. This structure is used by the PKCS#7 and CMS routines so any
5974 application which reads PKCS#7 or CMS data from untrusted sources is
5975 affected. SSL/TLS is not affected.
5977 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5983 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5984 This changes the decoding behaviour for some invalid messages,
5985 though the change is mostly in the more lenient direction, and
5986 legacy behaviour is preserved as much as possible.
5990 * In DSA_generate_parameters_ex, if the provided seed is too short,
5991 use a random seed, as already documented.
5993 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5995 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5997 * Alternate chains certificate forgery
5999 During certificate verfification, OpenSSL will attempt to find an
6000 alternative certificate chain if the first attempt to build such a chain
6001 fails. An error in the implementation of this logic can mean that an
6002 attacker could cause certain checks on untrusted certificates to be
6003 bypassed, such as the CA flag, enabling them to use a valid leaf
6004 certificate to act as a CA and "issue" an invalid certificate.
6006 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6012 * Race condition handling PSK identify hint
6014 If PSK identity hints are received by a multi-threaded client then
6015 the values are wrongly updated in the parent SSL_CTX structure. This can
6016 result in a race condition potentially leading to a double free of the
6022 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6024 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6025 incompatibility in the handling of HMAC. The previous ABI has now been
6028 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6030 * Malformed ECParameters causes infinite loop
6032 When processing an ECParameters structure OpenSSL enters an infinite loop
6033 if the curve specified is over a specially malformed binary polynomial
6036 This can be used to perform denial of service against any
6037 system which processes public keys, certificate requests or
6038 certificates. This includes TLS clients and TLS servers with
6039 client authentication enabled.
6041 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6046 * Exploitable out-of-bounds read in X509_cmp_time
6048 X509_cmp_time does not properly check the length of the ASN1_TIME
6049 string and can read a few bytes out of bounds. In addition,
6050 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6053 An attacker can use this to craft malformed certificates and CRLs of
6054 various sizes and potentially cause a segmentation fault, resulting in
6055 a DoS on applications that verify certificates or CRLs. TLS clients
6056 that verify CRLs are affected. TLS clients and servers with client
6057 authentication enabled may be affected if they use custom verification
6060 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6061 independently by Hanno Böck.
6066 * PKCS7 crash with missing EnvelopedContent
6068 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6069 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6070 with missing content and trigger a NULL pointer dereference on parsing.
6072 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6073 structures from untrusted sources are affected. OpenSSL clients and
6074 servers are not affected.
6076 This issue was reported to OpenSSL by Michal Zalewski (Google).
6081 * CMS verify infinite loop with unknown hash function
6083 When verifying a signedData message the CMS code can enter an infinite loop
6084 if presented with an unknown hash function OID. This can be used to perform
6085 denial of service against any system which verifies signedData messages using
6087 This issue was reported to OpenSSL by Johannes Bauer.
6092 * Race condition handling NewSessionTicket
6094 If a NewSessionTicket is received by a multi-threaded client when attempting to
6095 reuse a previous ticket then a race condition can occur potentially leading to
6096 a double free of the ticket data.
6101 * Reject DH handshakes with parameters shorter than 768 bits.
6103 *Kurt Roeckx and Emilia Kasper*
6105 * dhparam: generate 2048-bit parameters by default.
6107 *Kurt Roeckx and Emilia Kasper*
6109 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6111 * Segmentation fault in ASN1_TYPE_cmp fix
6113 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6114 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6115 certificate signature algorithm consistency this can be used to crash any
6116 certificate verification operation and exploited in a DoS attack. Any
6117 application which performs certificate verification is vulnerable including
6118 OpenSSL clients and servers which enable client authentication.
6123 * ASN.1 structure reuse memory corruption fix
6125 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6126 memory corruption via an invalid write. Such reuse is and has been
6127 strongly discouraged and is believed to be rare.
6129 Applications that parse structures containing CHOICE or ANY DEFINED BY
6130 components may be affected. Certificate parsing (d2i_X509 and related
6131 functions) are however not affected. OpenSSL clients and servers are
6137 * PKCS7 NULL pointer dereferences fix
6139 The PKCS#7 parsing code does not handle missing outer ContentInfo
6140 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6141 missing content and trigger a NULL pointer dereference on parsing.
6143 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6144 otherwise parse PKCS#7 structures from untrusted sources are
6145 affected. OpenSSL clients and servers are not affected.
6147 This issue was reported to OpenSSL by Michal Zalewski (Google).
6152 * DoS via reachable assert in SSLv2 servers fix
6154 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6155 servers that both support SSLv2 and enable export cipher suites by sending
6156 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6158 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6159 (OpenSSL development team).
6164 * Use After Free following d2i_ECPrivatekey error fix
6166 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6167 could cause a use after free condition. This, in turn, could cause a double
6168 free in several private key parsing functions (such as d2i_PrivateKey
6169 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6170 for applications that receive EC private keys from untrusted
6171 sources. This scenario is considered rare.
6173 This issue was discovered by the BoringSSL project and fixed in their
6179 * X509_to_X509_REQ NULL pointer deref fix
6181 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6182 the certificate key is invalid. This function is rarely used in practice.
6184 This issue was discovered by Brian Carpenter.
6189 * Removed the export ciphers from the DEFAULT ciphers
6193 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6195 * Build fixes for the Windows and OpenVMS platforms
6197 *Matt Caswell and Richard Levitte*
6199 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6201 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6202 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6203 dereference. This could lead to a Denial Of Service attack. Thanks to
6204 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6209 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6210 dtls1_buffer_record function under certain conditions. In particular this
6211 could occur if an attacker sent repeated DTLS records with the same
6212 sequence number but for the next epoch. The memory leak could be exploited
6213 by an attacker in a Denial of Service attack through memory exhaustion.
6214 Thanks to Chris Mueller for reporting this issue.
6219 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6220 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6221 method would be set to NULL which could later result in a NULL pointer
6222 dereference. Thanks to Frank Schmirler for reporting this issue.
6227 * Abort handshake if server key exchange message is omitted for ephemeral
6230 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6231 reporting this issue.
6236 * Remove non-export ephemeral RSA code on client and server. This code
6237 violated the TLS standard by allowing the use of temporary RSA keys in
6238 non-export ciphersuites and could be used by a server to effectively
6239 downgrade the RSA key length used to a value smaller than the server
6240 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6241 INRIA or reporting this issue.
6246 * Fixed issue where DH client certificates are accepted without verification.
6247 An OpenSSL server will accept a DH certificate for client authentication
6248 without the certificate verify message. This effectively allows a client to
6249 authenticate without the use of a private key. This only affects servers
6250 which trust a client certificate authority which issues certificates
6251 containing DH keys: these are extremely rare and hardly ever encountered.
6252 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6258 * Ensure that the session ID context of an SSL is updated when its
6259 SSL_CTX is updated via SSL_set_SSL_CTX.
6261 The session ID context is typically set from the parent SSL_CTX,
6262 and can vary with the CTX.
6266 * Fix various certificate fingerprint issues.
6268 By using non-DER or invalid encodings outside the signed portion of a
6269 certificate the fingerprint can be changed without breaking the signature.
6270 Although no details of the signed portion of the certificate can be changed
6271 this can cause problems with some applications: e.g. those using the
6272 certificate fingerprint for blacklists.
6274 1. Reject signatures with non zero unused bits.
6276 If the BIT STRING containing the signature has non zero unused bits reject
6277 the signature. All current signature algorithms require zero unused bits.
6279 2. Check certificate algorithm consistency.
6281 Check the AlgorithmIdentifier inside TBS matches the one in the
6282 certificate signature. NB: this will result in signature failure
6283 errors for some broken certificates.
6285 Thanks to Konrad Kraszewski from Google for reporting this issue.
6287 3. Check DSA/ECDSA signatures use DER.
6289 Re-encode DSA/ECDSA signatures and compare with the original received
6290 signature. Return an error if there is a mismatch.
6292 This will reject various cases including garbage after signature
6293 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6294 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6295 (negative or with leading zeroes).
6297 Further analysis was conducted and fixes were developed by Stephen Henson
6298 of the OpenSSL core team.
6304 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6305 results on some platforms, including x86_64. This bug occurs at random
6306 with a very low probability, and is not known to be exploitable in any
6307 way, though its exact impact is difficult to determine. Thanks to Pieter
6308 Wuille (Blockstream) who reported this issue and also suggested an initial
6309 fix. Further analysis was conducted by the OpenSSL development team and
6310 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6311 the OpenSSL core team.
6316 * Do not resume sessions on the server if the negotiated protocol
6317 version does not match the session's version. Resuming with a different
6318 version, while not strictly forbidden by the RFC, is of questionable
6319 sanity and breaks all known clients.
6321 *David Benjamin, Emilia Käsper*
6323 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6324 early CCS messages during renegotiation. (Note that because
6325 renegotiation is encrypted, this early CCS was not exploitable.)
6329 * Tighten client-side session ticket handling during renegotiation:
6330 ensure that the client only accepts a session ticket if the server sends
6331 the extension anew in the ServerHello. Previously, a TLS client would
6332 reuse the old extension state and thus accept a session ticket if one was
6333 announced in the initial ServerHello.
6335 Similarly, ensure that the client requires a session ticket if one
6336 was advertised in the ServerHello. Previously, a TLS client would
6337 ignore a missing NewSessionTicket message.
6341 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6345 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6346 sends a carefully crafted handshake message, to cause OpenSSL to fail
6347 to free up to 64k of memory causing a memory leak. This could be
6348 exploited in a Denial Of Service attack. This issue affects OpenSSL
6349 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6350 whether SRTP is used or configured. Implementations of OpenSSL that
6351 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6353 The fix was developed by the OpenSSL team.
6358 * Session Ticket Memory Leak.
6360 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6361 integrity of that ticket is first verified. In the event of a session
6362 ticket integrity check failing, OpenSSL will fail to free memory
6363 causing a memory leak. By sending a large number of invalid session
6364 tickets an attacker could exploit this issue in a Denial Of Service
6370 * Build option no-ssl3 is incomplete.
6372 When OpenSSL is configured with "no-ssl3" as a build option, servers
6373 could accept and complete a SSL 3.0 handshake, and clients could be
6374 configured to send them.
6377 *Akamai and the OpenSSL team*
6379 * Add support for TLS_FALLBACK_SCSV.
6380 Client applications doing fallback retries should call
6381 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6384 *Adam Langley, Bodo Moeller*
6386 * Add additional DigestInfo checks.
6388 Re-encode DigestInto in DER and check against the original when
6389 verifying RSA signature: this will reject any improperly encoded
6390 DigestInfo structures.
6392 Note: this is a precautionary measure and no attacks are currently known.
6396 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6398 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6399 SRP code can be overrun an internal buffer. Add sanity check that
6400 g, A, B < N to SRP code.
6402 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6403 Group for discovering this issue.
6408 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6409 TLS 1.0 instead of higher protocol versions when the ClientHello message
6410 is badly fragmented. This allows a man-in-the-middle attacker to force a
6411 downgrade to TLS 1.0 even if both the server and the client support a
6412 higher protocol version, by modifying the client's TLS records.
6414 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6415 researching this issue.
6420 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6421 to a denial of service attack. A malicious server can crash the client
6422 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6423 ciphersuite and sending carefully crafted handshake messages.
6425 Thanks to Felix Gröbert (Google) for discovering and researching this
6431 * By sending carefully crafted DTLS packets an attacker could cause openssl
6432 to leak memory. This can be exploited through a Denial of Service attack.
6433 Thanks to Adam Langley for discovering and researching this issue.
6438 * An attacker can force openssl to consume large amounts of memory whilst
6439 processing DTLS handshake messages. This can be exploited through a
6440 Denial of Service attack.
6441 Thanks to Adam Langley for discovering and researching this issue.
6446 * An attacker can force an error condition which causes openssl to crash
6447 whilst processing DTLS packets due to memory being freed twice. This
6448 can be exploited through a Denial of Service attack.
6449 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6455 * If a multithreaded client connects to a malicious server using a resumed
6456 session and the server sends an ec point format extension it could write
6457 up to 255 bytes to freed memory.
6459 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6465 * A malicious server can crash an OpenSSL client with a null pointer
6466 dereference (read) by specifying an SRP ciphersuite even though it was not
6467 properly negotiated with the client. This can be exploited through a
6468 Denial of Service attack.
6470 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6471 discovering and researching this issue.
6476 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6477 X509_name_oneline, X509_name_print_ex et al. to leak some information
6478 from the stack. Applications may be affected if they echo pretty printing
6479 output to the attacker.
6481 Thanks to Ivan Fratric (Google) for discovering this issue.
6484 *Emilia Käsper, and Steve Henson*
6486 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6487 for corner cases. (Certain input points at infinity could lead to
6488 bogus results, with non-infinity inputs mapped to infinity too.)
6492 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6494 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6495 handshake can force the use of weak keying material in OpenSSL
6496 SSL/TLS clients and servers.
6498 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6499 researching this issue. ([CVE-2014-0224])
6501 *KIKUCHI Masashi, Steve Henson*
6503 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6504 OpenSSL DTLS client the code can be made to recurse eventually crashing
6507 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6510 *Imre Rad, Steve Henson*
6512 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6513 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6514 client or server. This is potentially exploitable to run arbitrary
6515 code on a vulnerable client or server.
6517 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6519 *Jüri Aedla, Steve Henson*
6521 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6522 are subject to a denial of service attack.
6524 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6525 this issue. ([CVE-2014-3470])
6527 *Felix Gröbert, Ivan Fratric, Steve Henson*
6529 * Harmonize version and its documentation. -f flag is used to display
6532 *mancha <mancha1@zoho.com>*
6534 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6535 in i2d_ECPrivateKey.
6537 *mancha <mancha1@zoho.com>*
6539 * Fix some double frees. These are not thought to be exploitable.
6541 *mancha <mancha1@zoho.com>*
6543 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6545 * A missing bounds check in the handling of the TLS heartbeat extension
6546 can be used to reveal up to 64k of memory to a connected client or
6549 Thanks for Neel Mehta of Google Security for discovering this bug and to
6550 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6551 preparing the fix ([CVE-2014-0160])
6553 *Adam Langley, Bodo Moeller*
6555 * Fix for the attack described in the paper "Recovering OpenSSL
6556 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6557 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6558 <http://eprint.iacr.org/2014/140>
6560 Thanks to Yuval Yarom and Naomi Benger for discovering this
6561 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6563 *Yuval Yarom and Naomi Benger*
6565 * TLS pad extension: draft-agl-tls-padding-03
6567 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6568 TLS client Hello record length value would otherwise be > 255 and
6569 less that 512 pad with a dummy extension containing zeroes so it
6570 is at least 512 bytes long.
6572 *Adam Langley, Steve Henson*
6574 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6576 * Fix for TLS record tampering bug. A carefully crafted invalid
6577 handshake could crash OpenSSL with a NULL pointer exception.
6578 Thanks to Anton Johansson for reporting this issues.
6581 * Keep original DTLS digest and encryption contexts in retransmission
6582 structures so we can use the previous session parameters if they need
6583 to be resent. ([CVE-2013-6450])
6587 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6588 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6589 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6590 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6591 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6592 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6594 *Rob Stradling, Adam Langley*
6596 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6598 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6599 supporting platforms or when small records were transferred.
6601 *Andy Polyakov, Steve Henson*
6603 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6605 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6607 This addresses the flaw in CBC record processing discovered by
6608 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6609 at: <http://www.isg.rhul.ac.uk/tls/>
6611 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6612 Security Group at Royal Holloway, University of London
6613 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6614 Emilia Käsper for the initial patch.
6617 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6619 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6620 ciphersuites which can be exploited in a denial of service attack.
6621 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6622 and detecting this bug and to Wolfgang Ettlinger
6623 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6628 * Return an error when checking OCSP signatures when key is NULL.
6629 This fixes a DoS attack. ([CVE-2013-0166])
6633 * Make openssl verify return errors.
6635 *Chris Palmer <palmer@google.com> and Ben Laurie*
6637 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6638 the right response is stapled. Also change SSL_get_certificate()
6639 so it returns the certificate actually sent.
6640 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6642 *Rob Stradling <rob.stradling@comodo.com>*
6644 * Fix possible deadlock when decoding public keys.
6648 * Don't use TLS 1.0 record version number in initial client hello
6653 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6655 * Sanity check record length before skipping explicit IV in TLS
6656 1.2, 1.1 and DTLS to fix DoS attack.
6658 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6659 fuzzing as a service testing platform.
6664 * Initialise tkeylen properly when encrypting CMS messages.
6665 Thanks to Solar Designer of Openwall for reporting this issue.
6669 * In FIPS mode don't try to use composite ciphers as they are not
6674 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6676 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6677 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6678 mean any application compiled against OpenSSL 1.0.0 headers setting
6679 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6680 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6681 0x10000000L Any application which was previously compiled against
6682 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6683 will need to be recompiled as a result. Letting be results in
6684 inability to disable specifically TLS 1.1 and in client context,
6685 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6689 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6690 disable just protocol X, but all protocols above X *if* there are
6691 protocols *below* X still enabled. In more practical terms it means
6692 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6693 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6694 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6699 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6701 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6702 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6703 in CRYPTO_realloc_clean.
6705 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6706 issue and to Adam Langley <agl@chromium.org> for fixing it.
6709 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6711 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6715 * Workarounds for some broken servers that "hang" if a client hello
6716 record length exceeds 255 bytes.
6718 1. Do not use record version number > TLS 1.0 in initial client
6719 hello: some (but not all) hanging servers will now work.
6720 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6721 the number of ciphers sent in the client hello. This should be
6722 set to an even number, such as 50, for example by passing:
6723 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6724 Most broken servers should now work.
6725 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6726 TLS 1.2 client support entirely.
6730 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6734 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6736 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6737 STRING form instead of a DigestInfo.
6741 * The format used for MDC2 RSA signatures is inconsistent between EVP
6742 and the RSA_sign/RSA_verify functions. This was made more apparent when
6743 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6744 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6745 the correct format in RSA_verify so both forms transparently work.
6749 * Some servers which support TLS 1.0 can choke if we initially indicate
6750 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6751 encrypted premaster secret. As a workaround use the maximum permitted
6752 client version in client hello, this should keep such servers happy
6753 and still work with previous versions of OpenSSL.
6757 * Add support for TLS/DTLS heartbeats.
6759 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6761 * Add support for SCTP.
6763 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6765 * Improved PRNG seeding for VOS.
6767 *Paul Green <Paul.Green@stratus.com>*
6769 * Extensive assembler packs updates, most notably:
6771 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6772 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6773 - x86_64: bit-sliced AES implementation;
6774 - ARM: NEON support, contemporary platforms optimizations;
6775 - s390x: z196 support;
6776 - `*`: GHASH and GF(2^m) multiplication implementations;
6780 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6781 (removal of unnecessary code)
6783 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6785 * Add TLS key material exporter from RFC 5705.
6789 * Add DTLS-SRTP negotiation from RFC 5764.
6793 * Add Next Protocol Negotiation,
6794 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6795 disabled with a no-npn flag to config or Configure. Code donated
6798 *Adam Langley <agl@google.com> and Ben Laurie*
6800 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6801 NIST-P256, NIST-P521, with constant-time single point multiplication on
6802 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6803 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6804 Code made available under Apache License version 2.0.
6806 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6807 line to include this in your build of OpenSSL, and run "make depend" (or
6808 "make update"). This enables the following EC_METHODs:
6810 EC_GFp_nistp224_method()
6811 EC_GFp_nistp256_method()
6812 EC_GFp_nistp521_method()
6814 EC_GROUP_new_by_curve_name() will automatically use these (while
6815 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6818 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6820 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6821 all platforms. Move ssize_t definition from e_os.h to the public
6822 header file e_os2.h as it now appears in public header file cms.h
6826 * New -sigopt option to the ca, req and x509 utilities. Additional
6827 signature parameters can be passed using this option and in
6832 * Add RSA PSS signing function. This will generate and set the
6833 appropriate AlgorithmIdentifiers for PSS based on those in the
6834 corresponding EVP_MD_CTX structure. No application support yet.
6838 * Support for companion algorithm specific ASN1 signing routines.
6839 New function ASN1_item_sign_ctx() signs a pre-initialised
6840 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6841 the appropriate parameters.
6845 * Add new algorithm specific ASN1 verification initialisation function
6846 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6847 handling will be the same no matter what EVP_PKEY_METHOD is used.
6848 Add a PSS handler to support verification of PSS signatures: checked
6849 against a number of sample certificates.
6853 * Add signature printing for PSS. Add PSS OIDs.
6855 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6857 * Add algorithm specific signature printing. An individual ASN1 method
6858 can now print out signatures instead of the standard hex dump.
6860 More complex signatures (e.g. PSS) can print out more meaningful
6861 information. Include DSA version that prints out the signature
6866 * Password based recipient info support for CMS library: implementing
6871 * Split password based encryption into PBES2 and PBKDF2 functions. This
6872 neatly separates the code into cipher and PBE sections and is required
6873 for some algorithms that split PBES2 into separate pieces (such as
6874 password based CMS).
6878 * Session-handling fixes:
6879 - Fix handling of connections that are resuming with a session ID,
6880 but also support Session Tickets.
6881 - Fix a bug that suppressed issuing of a new ticket if the client
6882 presented a ticket with an expired session.
6883 - Try to set the ticket lifetime hint to something reasonable.
6884 - Make tickets shorter by excluding irrelevant information.
6885 - On the client side, don't ignore renewed tickets.
6887 *Adam Langley, Bodo Moeller (Google)*
6889 * Fix PSK session representation.
6893 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6895 This work was sponsored by Intel.
6899 * Add GCM support to TLS library. Some custom code is needed to split
6900 the IV between the fixed (from PRF) and explicit (from TLS record)
6901 portions. This adds all GCM ciphersuites supported by RFC5288 and
6902 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6903 add a special AESGCM string for GCM only.
6907 * Expand range of ctrls for AES GCM. Permit setting invocation
6908 field on decrypt and retrieval of invocation field only on encrypt.
6912 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6913 As required by RFC5289 these ciphersuites cannot be used if for
6914 versions of TLS earlier than 1.2.
6918 * For FIPS capable OpenSSL interpret a NULL default public key method
6919 as unset and return the appropriate default but do *not* set the default.
6920 This means we can return the appropriate method in applications that
6921 switch between FIPS and non-FIPS modes.
6925 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6926 ENGINE is used then we cannot handle that in the FIPS module so we
6927 keep original code iff non-FIPS operations are allowed.
6931 * Add -attime option to openssl utilities.
6933 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6935 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6939 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6940 FIPS EC methods unconditionally for now.
6944 * New build option no-ec2m to disable characteristic 2 code.
6948 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6949 all cases can be covered as some introduce binary incompatibilities.
6953 * Redirect RSA operations to FIPS module including keygen,
6954 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6958 * Add similar low level API blocking to ciphers.
6962 * Low level digest APIs are not approved in FIPS mode: any attempt
6963 to use these will cause a fatal error. Applications that *really* want
6964 to use them can use the `private_*` version instead.
6968 * Redirect cipher operations to FIPS module for FIPS builds.
6972 * Redirect digest operations to FIPS module for FIPS builds.
6976 * Update build system to add "fips" flag which will link in fipscanister.o
6977 for static and shared library builds embedding a signature if needed.
6981 * Output TLS supported curves in preference order instead of numerical
6982 order. This is currently hardcoded for the highest order curves first.
6983 This should be configurable so applications can judge speed vs strength.
6987 * Add TLS v1.2 server support for client authentication.
6991 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6996 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6997 FIPS modules versions.
7001 * Add TLS v1.2 client side support for client authentication. Keep cache
7002 of handshake records longer as we don't know the hash algorithm to use
7003 until after the certificate request message is received.
7007 * Initial TLS v1.2 client support. Add a default signature algorithms
7008 extension including all the algorithms we support. Parse new signature
7009 format in client key exchange. Relax some ECC signing restrictions for
7010 TLS v1.2 as indicated in RFC5246.
7014 * Add server support for TLS v1.2 signature algorithms extension. Switch
7015 to new signature format when needed using client digest preference.
7016 All server ciphersuites should now work correctly in TLS v1.2. No client
7017 support yet and no support for client certificates.
7021 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7022 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7023 ciphersuites. At present only RSA key exchange ciphersuites work with
7024 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7025 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7026 and version checking.
7030 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7031 with this defined it will not be affected by any changes to ssl internal
7032 structures. Add several utility functions to allow openssl application
7033 to work with OPENSSL_NO_SSL_INTERN defined.
7037 * A long standing patch to add support for SRP from EdelWeb (Peter
7038 Sylvester and Christophe Renou) was integrated.
7039 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7040 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7043 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7047 * Permit abbreviated handshakes when renegotiating using the function
7048 SSL_renegotiate_abbreviated().
7050 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7052 * Add call to ENGINE_register_all_complete() to
7053 ENGINE_load_builtin_engines(), so some implementations get used
7054 automatically instead of needing explicit application support.
7058 * Add support for TLS key exporter as described in RFC5705.
7060 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7062 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7063 a few changes are required:
7065 Add SSL_OP_NO_TLSv1_1 flag.
7066 Add TLSv1_1 methods.
7067 Update version checking logic to handle version 1.1.
7068 Add explicit IV handling (ported from DTLS code).
7069 Add command line options to s_client/s_server.
7076 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7078 * X509_ATTRIBUTE memory leak
7080 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7081 memory. This structure is used by the PKCS#7 and CMS routines so any
7082 application which reads PKCS#7 or CMS data from untrusted sources is
7083 affected. SSL/TLS is not affected.
7085 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7091 * Race condition handling PSK identify hint
7093 If PSK identity hints are received by a multi-threaded client then
7094 the values are wrongly updated in the parent SSL_CTX structure. This can
7095 result in a race condition potentially leading to a double free of the
7101 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7103 * Malformed ECParameters causes infinite loop
7105 When processing an ECParameters structure OpenSSL enters an infinite loop
7106 if the curve specified is over a specially malformed binary polynomial
7109 This can be used to perform denial of service against any
7110 system which processes public keys, certificate requests or
7111 certificates. This includes TLS clients and TLS servers with
7112 client authentication enabled.
7114 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7119 * Exploitable out-of-bounds read in X509_cmp_time
7121 X509_cmp_time does not properly check the length of the ASN1_TIME
7122 string and can read a few bytes out of bounds. In addition,
7123 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7126 An attacker can use this to craft malformed certificates and CRLs of
7127 various sizes and potentially cause a segmentation fault, resulting in
7128 a DoS on applications that verify certificates or CRLs. TLS clients
7129 that verify CRLs are affected. TLS clients and servers with client
7130 authentication enabled may be affected if they use custom verification
7133 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7134 independently by Hanno Böck.
7139 * PKCS7 crash with missing EnvelopedContent
7141 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7142 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7143 with missing content and trigger a NULL pointer dereference on parsing.
7145 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7146 structures from untrusted sources are affected. OpenSSL clients and
7147 servers are not affected.
7149 This issue was reported to OpenSSL by Michal Zalewski (Google).
7154 * CMS verify infinite loop with unknown hash function
7156 When verifying a signedData message the CMS code can enter an infinite loop
7157 if presented with an unknown hash function OID. This can be used to perform
7158 denial of service against any system which verifies signedData messages using
7160 This issue was reported to OpenSSL by Johannes Bauer.
7165 * Race condition handling NewSessionTicket
7167 If a NewSessionTicket is received by a multi-threaded client when attempting to
7168 reuse a previous ticket then a race condition can occur potentially leading to
7169 a double free of the ticket data.
7174 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7176 * Segmentation fault in ASN1_TYPE_cmp fix
7178 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7179 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7180 certificate signature algorithm consistency this can be used to crash any
7181 certificate verification operation and exploited in a DoS attack. Any
7182 application which performs certificate verification is vulnerable including
7183 OpenSSL clients and servers which enable client authentication.
7188 * ASN.1 structure reuse memory corruption fix
7190 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7191 memory corruption via an invalid write. Such reuse is and has been
7192 strongly discouraged and is believed to be rare.
7194 Applications that parse structures containing CHOICE or ANY DEFINED BY
7195 components may be affected. Certificate parsing (d2i_X509 and related
7196 functions) are however not affected. OpenSSL clients and servers are
7202 * PKCS7 NULL pointer dereferences fix
7204 The PKCS#7 parsing code does not handle missing outer ContentInfo
7205 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7206 missing content and trigger a NULL pointer dereference on parsing.
7208 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7209 otherwise parse PKCS#7 structures from untrusted sources are
7210 affected. OpenSSL clients and servers are not affected.
7212 This issue was reported to OpenSSL by Michal Zalewski (Google).
7217 * DoS via reachable assert in SSLv2 servers fix
7219 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7220 servers that both support SSLv2 and enable export cipher suites by sending
7221 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7223 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7224 (OpenSSL development team).
7229 * Use After Free following d2i_ECPrivatekey error fix
7231 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7232 could cause a use after free condition. This, in turn, could cause a double
7233 free in several private key parsing functions (such as d2i_PrivateKey
7234 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7235 for applications that receive EC private keys from untrusted
7236 sources. This scenario is considered rare.
7238 This issue was discovered by the BoringSSL project and fixed in their
7244 * X509_to_X509_REQ NULL pointer deref fix
7246 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7247 the certificate key is invalid. This function is rarely used in practice.
7249 This issue was discovered by Brian Carpenter.
7254 * Removed the export ciphers from the DEFAULT ciphers
7258 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7260 * Build fixes for the Windows and OpenVMS platforms
7262 *Matt Caswell and Richard Levitte*
7264 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7266 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7267 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7268 dereference. This could lead to a Denial Of Service attack. Thanks to
7269 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7274 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7275 dtls1_buffer_record function under certain conditions. In particular this
7276 could occur if an attacker sent repeated DTLS records with the same
7277 sequence number but for the next epoch. The memory leak could be exploited
7278 by an attacker in a Denial of Service attack through memory exhaustion.
7279 Thanks to Chris Mueller for reporting this issue.
7284 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7285 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7286 method would be set to NULL which could later result in a NULL pointer
7287 dereference. Thanks to Frank Schmirler for reporting this issue.
7292 * Abort handshake if server key exchange message is omitted for ephemeral
7295 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7296 reporting this issue.
7301 * Remove non-export ephemeral RSA code on client and server. This code
7302 violated the TLS standard by allowing the use of temporary RSA keys in
7303 non-export ciphersuites and could be used by a server to effectively
7304 downgrade the RSA key length used to a value smaller than the server
7305 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7306 INRIA or reporting this issue.
7311 * Fixed issue where DH client certificates are accepted without verification.
7312 An OpenSSL server will accept a DH certificate for client authentication
7313 without the certificate verify message. This effectively allows a client to
7314 authenticate without the use of a private key. This only affects servers
7315 which trust a client certificate authority which issues certificates
7316 containing DH keys: these are extremely rare and hardly ever encountered.
7317 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7323 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7324 results on some platforms, including x86_64. This bug occurs at random
7325 with a very low probability, and is not known to be exploitable in any
7326 way, though its exact impact is difficult to determine. Thanks to Pieter
7327 Wuille (Blockstream) who reported this issue and also suggested an initial
7328 fix. Further analysis was conducted by the OpenSSL development team and
7329 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7330 the OpenSSL core team.
7335 * Fix various certificate fingerprint issues.
7337 By using non-DER or invalid encodings outside the signed portion of a
7338 certificate the fingerprint can be changed without breaking the signature.
7339 Although no details of the signed portion of the certificate can be changed
7340 this can cause problems with some applications: e.g. those using the
7341 certificate fingerprint for blacklists.
7343 1. Reject signatures with non zero unused bits.
7345 If the BIT STRING containing the signature has non zero unused bits reject
7346 the signature. All current signature algorithms require zero unused bits.
7348 2. Check certificate algorithm consistency.
7350 Check the AlgorithmIdentifier inside TBS matches the one in the
7351 certificate signature. NB: this will result in signature failure
7352 errors for some broken certificates.
7354 Thanks to Konrad Kraszewski from Google for reporting this issue.
7356 3. Check DSA/ECDSA signatures use DER.
7358 Reencode DSA/ECDSA signatures and compare with the original received
7359 signature. Return an error if there is a mismatch.
7361 This will reject various cases including garbage after signature
7362 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7363 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7364 (negative or with leading zeroes).
7366 Further analysis was conducted and fixes were developed by Stephen Henson
7367 of the OpenSSL core team.
7373 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7375 * Session Ticket Memory Leak.
7377 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7378 integrity of that ticket is first verified. In the event of a session
7379 ticket integrity check failing, OpenSSL will fail to free memory
7380 causing a memory leak. By sending a large number of invalid session
7381 tickets an attacker could exploit this issue in a Denial Of Service
7387 * Build option no-ssl3 is incomplete.
7389 When OpenSSL is configured with "no-ssl3" as a build option, servers
7390 could accept and complete a SSL 3.0 handshake, and clients could be
7391 configured to send them.
7394 *Akamai and the OpenSSL team*
7396 * Add support for TLS_FALLBACK_SCSV.
7397 Client applications doing fallback retries should call
7398 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7401 *Adam Langley, Bodo Moeller*
7403 * Add additional DigestInfo checks.
7405 Reencode DigestInto in DER and check against the original when
7406 verifying RSA signature: this will reject any improperly encoded
7407 DigestInfo structures.
7409 Note: this is a precautionary measure and no attacks are currently known.
7413 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7415 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7416 to a denial of service attack. A malicious server can crash the client
7417 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7418 ciphersuite and sending carefully crafted handshake messages.
7420 Thanks to Felix Gröbert (Google) for discovering and researching this
7426 * By sending carefully crafted DTLS packets an attacker could cause openssl
7427 to leak memory. This can be exploited through a Denial of Service attack.
7428 Thanks to Adam Langley for discovering and researching this issue.
7433 * An attacker can force openssl to consume large amounts of memory whilst
7434 processing DTLS handshake messages. This can be exploited through a
7435 Denial of Service attack.
7436 Thanks to Adam Langley for discovering and researching this issue.
7441 * An attacker can force an error condition which causes openssl to crash
7442 whilst processing DTLS packets due to memory being freed twice. This
7443 can be exploited through a Denial of Service attack.
7444 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7450 * If a multithreaded client connects to a malicious server using a resumed
7451 session and the server sends an ec point format extension it could write
7452 up to 255 bytes to freed memory.
7454 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7460 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7461 X509_name_oneline, X509_name_print_ex et al. to leak some information
7462 from the stack. Applications may be affected if they echo pretty printing
7463 output to the attacker.
7465 Thanks to Ivan Fratric (Google) for discovering this issue.
7468 *Emilia Käsper, and Steve Henson*
7470 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7471 for corner cases. (Certain input points at infinity could lead to
7472 bogus results, with non-infinity inputs mapped to infinity too.)
7476 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7478 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7479 handshake can force the use of weak keying material in OpenSSL
7480 SSL/TLS clients and servers.
7482 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7483 researching this issue. ([CVE-2014-0224])
7485 *KIKUCHI Masashi, Steve Henson*
7487 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7488 OpenSSL DTLS client the code can be made to recurse eventually crashing
7491 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7494 *Imre Rad, Steve Henson*
7496 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7497 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7498 client or server. This is potentially exploitable to run arbitrary
7499 code on a vulnerable client or server.
7501 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7503 *Jüri Aedla, Steve Henson*
7505 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7506 are subject to a denial of service attack.
7508 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7509 this issue. ([CVE-2014-3470])
7511 *Felix Gröbert, Ivan Fratric, Steve Henson*
7513 * Harmonize version and its documentation. -f flag is used to display
7516 *mancha <mancha1@zoho.com>*
7518 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7519 in i2d_ECPrivateKey.
7521 *mancha <mancha1@zoho.com>*
7523 * Fix some double frees. These are not thought to be exploitable.
7525 *mancha <mancha1@zoho.com>*
7527 * Fix for the attack described in the paper "Recovering OpenSSL
7528 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7529 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7530 <http://eprint.iacr.org/2014/140>
7532 Thanks to Yuval Yarom and Naomi Benger for discovering this
7533 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7535 *Yuval Yarom and Naomi Benger*
7537 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7539 * Keep original DTLS digest and encryption contexts in retransmission
7540 structures so we can use the previous session parameters if they need
7541 to be resent. ([CVE-2013-6450])
7545 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7546 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7547 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7548 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7549 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7550 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7552 *Rob Stradling, Adam Langley*
7554 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7556 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7558 This addresses the flaw in CBC record processing discovered by
7559 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7560 at: <http://www.isg.rhul.ac.uk/tls/>
7562 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7563 Security Group at Royal Holloway, University of London
7564 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7565 Emilia Käsper for the initial patch.
7568 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7570 * Return an error when checking OCSP signatures when key is NULL.
7571 This fixes a DoS attack. ([CVE-2013-0166])
7575 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7576 the right response is stapled. Also change SSL_get_certificate()
7577 so it returns the certificate actually sent.
7578 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7579 (This is a backport)
7581 *Rob Stradling <rob.stradling@comodo.com>*
7583 * Fix possible deadlock when decoding public keys.
7587 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7589 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7592 * Sanity check record length before skipping explicit IV in DTLS
7595 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7596 fuzzing as a service testing platform.
7601 * Initialise tkeylen properly when encrypting CMS messages.
7602 Thanks to Solar Designer of Openwall for reporting this issue.
7606 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7608 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7609 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7610 in CRYPTO_realloc_clean.
7612 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7613 issue and to Adam Langley <agl@chromium.org> for fixing it.
7616 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7618 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7620 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7621 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7622 content decryption and always return the same error. Note: this attack
7623 needs on average 2^20 messages so it only affects automated senders. The
7624 old behaviour can be re-enabled in the CMS code by setting the
7625 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7626 an MMA defence is not necessary.
7627 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7628 this issue. ([CVE-2012-0884])
7632 * Fix CVE-2011-4619: make sure we really are receiving a
7633 client hello before rejecting multiple SGC restarts. Thanks to
7634 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7638 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7640 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7641 Thanks to Antonio Martin, Enterprise Secure Access Research and
7642 Development, Cisco Systems, Inc. for discovering this bug and
7643 preparing a fix. ([CVE-2012-0050])
7647 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7649 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7650 of the Vaudenay padding oracle attack on CBC mode encryption
7651 which enables an efficient plaintext recovery attack against
7652 the OpenSSL implementation of DTLS. Their attack exploits timing
7653 differences arising during decryption processing. A research
7654 paper describing this attack can be found at:
7655 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7656 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7657 Security Group at Royal Holloway, University of London
7658 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7659 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7660 for preparing the fix. ([CVE-2011-4108])
7662 *Robin Seggelmann, Michael Tuexen*
7664 * Clear bytes used for block padding of SSL 3.0 records.
7667 *Adam Langley (Google)*
7669 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7670 Kadianakis <desnacked@gmail.com> for discovering this issue and
7671 Adam Langley for preparing the fix. ([CVE-2011-4619])
7673 *Adam Langley (Google)*
7675 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7677 *Andrey Kulikov <amdeich@gmail.com>*
7679 * Prevent malformed RFC3779 data triggering an assertion failure.
7680 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7681 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7683 *Rob Austein <sra@hactrn.net>*
7685 * Improved PRNG seeding for VOS.
7687 *Paul Green <Paul.Green@stratus.com>*
7689 * Fix ssl_ciph.c set-up race.
7691 *Adam Langley (Google)*
7693 * Fix spurious failures in ecdsatest.c.
7695 *Emilia Käsper (Google)*
7697 * Fix the BIO_f_buffer() implementation (which was mixing different
7698 interpretations of the `..._len` fields).
7700 *Adam Langley (Google)*
7702 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7703 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7704 threads won't reuse the same blinding coefficients.
7706 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7707 lock to call BN_BLINDING_invert_ex, and avoids one use of
7708 BN_BLINDING_update for each BN_BLINDING structure (previously,
7709 the last update always remained unused).
7711 *Emilia Käsper (Google)*
7713 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7715 *Bob Buckholz (Google)*
7717 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7719 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7720 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7722 *Kaspar Brand <ossl@velox.ch>*
7724 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7725 for multi-threaded use of ECDH. ([CVE-2011-3210])
7727 *Adam Langley (Google)*
7729 * Fix x509_name_ex_d2i memory leak on bad inputs.
7733 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7734 signature public key algorithm by using OID xref utilities instead.
7735 Before this you could only use some ECC ciphersuites with SHA1 only.
7739 * Add protection against ECDSA timing attacks as mentioned in the paper
7740 by Billy Bob Brumley and Nicola Tuveri, see:
7741 <http://eprint.iacr.org/2011/232.pdf>
7743 *Billy Bob Brumley and Nicola Tuveri*
7745 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7747 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7749 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7751 * Fix bug in string printing code: if *any* escaping is enabled we must
7752 escape the escape character (backslash) or the resulting string is
7757 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7759 * Disable code workaround for ancient and obsolete Netscape browsers
7760 and servers: an attacker can use it in a ciphersuite downgrade attack.
7761 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7765 * Fixed J-PAKE implementation error, originally discovered by
7766 Sebastien Martini, further info and confirmation from Stefan
7767 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7771 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7773 * Fix extension code to avoid race conditions which can result in a buffer
7774 overrun vulnerability: resumed sessions must not be modified as they can
7775 be shared by multiple threads. CVE-2010-3864
7779 * Fix WIN32 build system to correctly link an ENGINE directory into
7784 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7786 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7789 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7791 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7793 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7794 context. The operation can be customised via the ctrl mechanism in
7795 case ENGINEs want to include additional functionality.
7799 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7803 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7804 output hashes compatible with older versions of OpenSSL.
7806 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7808 * Fix compression algorithm handling: if resuming a session use the
7809 compression algorithm of the resumed session instead of determining
7810 it from client hello again. Don't allow server to change algorithm.
7814 * Add load_crls() function to apps tidying load_certs() too. Add option
7815 to verify utility to allow additional CRLs to be included.
7819 * Update OCSP request code to permit adding custom headers to the request:
7820 some responders need this.
7824 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7827 *Julia Lawall <julia@diku.dk>*
7829 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7830 needlessly dereferenced structures, used obsolete functions and
7831 didn't handle all updated verify codes correctly.
7835 * Disable MD2 in the default configuration.
7839 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7840 indicate the initial BIO being pushed or popped. This makes it possible
7841 to determine whether the BIO is the one explicitly called or as a result
7842 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7843 it handles reference counts correctly and doesn't zero out the I/O bio
7844 when it is not being explicitly popped. WARNING: applications which
7845 included workarounds for the old buggy behaviour will need to be modified
7846 or they could free up already freed BIOs.
7850 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7851 renaming to all platforms (within the 0.9.8 branch, this was
7852 done conditionally on Netware platforms to avoid a name clash).
7854 *Guenter <lists@gknw.net>*
7856 * Add ECDHE and PSK support to DTLS.
7858 *Michael Tuexen <tuexen@fh-muenster.de>*
7860 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7865 * Add "missing" function EVP_MD_flags() (without this the only way to
7866 retrieve a digest flags is by accessing the structure directly. Update
7867 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7868 or cipher is registered as in the "from" argument. Print out all
7869 registered digests in the dgst usage message instead of manually
7870 attempting to work them out.
7874 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7875 this allows the use of compression and extensions. Change default cipher
7876 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7877 by default unless an application cipher string requests it.
7881 * Alter match criteria in PKCS12_parse(). It used to try to use local
7882 key ids to find matching certificates and keys but some PKCS#12 files
7883 don't follow the (somewhat unwritten) rules and this strategy fails.
7884 Now just gather all certificates together and the first private key
7885 then look for the first certificate that matches the key.
7889 * Support use of registered digest and cipher names for dgst and cipher
7890 commands instead of having to add each one as a special case. So now
7897 openssl dgst -sha256 foo
7899 and this works for ENGINE based algorithms too.
7903 * Update Gost ENGINE to support parameter files.
7905 *Victor B. Wagner <vitus@cryptocom.ru>*
7907 * Support GeneralizedTime in ca utility.
7909 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7911 * Enhance the hash format used for certificate directory links. The new
7912 form uses the canonical encoding (meaning equivalent names will work
7913 even if they aren't identical) and uses SHA1 instead of MD5. This form
7914 is incompatible with the older format and as a result c_rehash should
7915 be used to rebuild symbolic links.
7919 * Make PKCS#8 the default write format for private keys, replacing the
7920 traditional format. This form is standardised, more secure and doesn't
7921 include an implicit MD5 dependency.
7925 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7926 committed to OpenSSL should pass this lot as a minimum.
7930 * Add session ticket override functionality for use by EAP-FAST.
7932 *Jouni Malinen <j@w1.fi>*
7934 * Modify HMAC functions to return a value. Since these can be implemented
7935 in an ENGINE errors can occur.
7939 * Type-checked OBJ_bsearch_ex.
7943 * Type-checked OBJ_bsearch. Also some constification necessitated
7944 by type-checking. Still to come: TXT_DB, bsearch(?),
7945 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7950 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7951 seconds to a tm structure directly, instead of going through OS
7952 specific date routines. This avoids any issues with OS routines such
7953 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7954 and X509_time_adj_ex() to cover the extended range. The existing
7955 X509_time_adj() is still usable and will no longer have any date issues.
7959 * Delta CRL support. New use deltas option which will attempt to locate
7960 and search any appropriate delta CRLs available.
7962 This work was sponsored by Google.
7966 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7967 code and add additional score elements. Validate alternate CRL paths
7968 as part of the CRL checking and indicate a new error "CRL path validation
7969 error" in this case. Applications wanting additional details can use
7970 the verify callback and check the new "parent" field. If this is not
7971 NULL CRL path validation is taking place. Existing applications won't
7972 see this because it requires extended CRL support which is off by
7975 This work was sponsored by Google.
7979 * Support for freshest CRL extension.
7981 This work was sponsored by Google.
7985 * Initial indirect CRL support. Currently only supported in the CRLs
7986 passed directly and not via lookup. Process certificate issuer
7987 CRL entry extension and lookup CRL entries by bother issuer name
7988 and serial number. Check and process CRL issuer entry in IDP extension.
7990 This work was sponsored by Google.
7994 * Add support for distinct certificate and CRL paths. The CRL issuer
7995 certificate is validated separately in this case. Only enabled if
7996 an extended CRL support flag is set: this flag will enable additional
7997 CRL functionality in future.
7999 This work was sponsored by Google.
8003 * Add support for policy mappings extension.
8005 This work was sponsored by Google.
8009 * Fixes to pathlength constraint, self issued certificate handling,
8010 policy processing to align with RFC3280 and PKITS tests.
8012 This work was sponsored by Google.
8016 * Support for name constraints certificate extension. DN, email, DNS
8017 and URI types are currently supported.
8019 This work was sponsored by Google.
8023 * To cater for systems that provide a pointer-based thread ID rather
8024 than numeric, deprecate the current numeric thread ID mechanism and
8025 replace it with a structure and associated callback type. This
8026 mechanism allows a numeric "hash" to be extracted from a thread ID in
8027 either case, and on platforms where pointers are larger than 'long',
8028 mixing is done to help ensure the numeric 'hash' is usable even if it
8029 can't be guaranteed unique. The default mechanism is to use "&errno"
8030 as a pointer-based thread ID to distinguish between threads.
8032 Applications that want to provide their own thread IDs should now use
8033 CRYPTO_THREADID_set_callback() to register a callback that will call
8034 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8036 Note that ERR_remove_state() is now deprecated, because it is tied
8037 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8038 to free the current thread's error state should be replaced by
8039 ERR_remove_thread_state(NULL).
8041 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8042 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8043 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8044 application was previously providing a numeric thread callback that
8045 was inappropriate for distinguishing threads, then uniqueness might
8046 have been obtained with &errno that happened immediately in the
8047 intermediate development versions of OpenSSL; this is no longer the
8048 case, the numeric thread callback will now override the automatic use
8051 *Geoff Thorpe, with help from Bodo Moeller*
8053 * Initial support for different CRL issuing certificates. This covers a
8054 simple case where the self issued certificates in the chain exist and
8055 the real CRL issuer is higher in the existing chain.
8057 This work was sponsored by Google.
8061 * Removed effectively defunct crypto/store from the build.
8065 * Revamp of STACK to provide stronger type-checking. Still to come:
8066 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8067 ASN1_STRING, CONF_VALUE.
8071 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8072 RAM on SSL connections. This option can save about 34k per idle SSL.
8076 * Revamp of LHASH to provide stronger type-checking. Still to come:
8077 STACK, TXT_DB, bsearch, qsort.
8081 * Initial support for Cryptographic Message Syntax (aka CMS) based
8082 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8083 support for data, signedData, compressedData, digestedData and
8084 encryptedData, envelopedData types included. Scripts to check against
8085 RFC4134 examples draft and interop and consistency checks of many
8086 content types and variants.
8090 * Add options to enc utility to support use of zlib compression BIO.
8094 * Extend mk1mf to support importing of options and assembly language
8095 files from Configure script, currently only included in VC-WIN32.
8096 The assembly language rules can now optionally generate the source
8097 files from the associated perl scripts.
8101 * Implement remaining functionality needed to support GOST ciphersuites.
8102 Interop testing has been performed using CryptoPro implementations.
8104 *Victor B. Wagner <vitus@cryptocom.ru>*
8106 * s390x assembler pack.
8110 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8115 * Implement Opaque PRF Input TLS extension as specified in
8116 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8117 official specification yet and no extension type assignment by
8118 IANA exists, this extension (for now) will have to be explicitly
8119 enabled when building OpenSSL by providing the extension number
8120 to use. For example, specify an option
8122 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8124 to the "config" or "Configure" script to enable the extension,
8125 assuming extension number 0x9527 (which is a completely arbitrary
8126 and unofficial assignment based on the MD5 hash of the Internet
8127 Draft). Note that by doing so, you potentially lose
8128 interoperability with other TLS implementations since these might
8129 be using the same extension number for other purposes.
8131 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8132 opaque PRF input value to use in the handshake. This will create
8133 an internal copy of the length-'len' string at 'src', and will
8134 return non-zero for success.
8136 To get more control and flexibility, provide a callback function
8139 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8140 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8144 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8147 Callback function 'cb' will be called in handshakes, and is
8148 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8149 Argument 'arg' is for application purposes (the value as given to
8150 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8151 be provided to the callback function). The callback function
8152 has to return non-zero to report success: usually 1 to use opaque
8153 PRF input just if possible, or 2 to enforce use of the opaque PRF
8154 input. In the latter case, the library will abort the handshake
8155 if opaque PRF input is not successfully negotiated.
8157 Arguments 'peerinput' and 'len' given to the callback function
8158 will always be NULL and 0 in the case of a client. A server will
8159 see the client's opaque PRF input through these variables if
8160 available (NULL and 0 otherwise). Note that if the server
8161 provides an opaque PRF input, the length must be the same as the
8162 length of the client's opaque PRF input.
8164 Note that the callback function will only be called when creating
8165 a new session (session resumption can resume whatever was
8166 previously negotiated), and will not be called in SSL 2.0
8167 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8168 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8169 for applications that need to enforce opaque PRF input.
8173 * Update ssl code to support digests other than SHA1+MD5 for handshake
8176 *Victor B. Wagner <vitus@cryptocom.ru>*
8178 * Add RFC4507 support to OpenSSL. This includes the corrections in
8179 RFC4507bis. The encrypted ticket format is an encrypted encoded
8180 SSL_SESSION structure, that way new session features are automatically
8183 If a client application caches session in an SSL_SESSION structure
8184 support is transparent because tickets are now stored in the encoded
8187 The SSL_CTX structure automatically generates keys for ticket
8188 protection in servers so again support should be possible
8189 with no application modification.
8191 If a client or server wishes to disable RFC4507 support then the option
8192 SSL_OP_NO_TICKET can be set.
8194 Add a TLS extension debugging callback to allow the contents of any client
8195 or server extensions to be examined.
8197 This work was sponsored by Google.
8201 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8202 OpenSSL should now compile cleanly on gcc 4.2
8204 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8206 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8207 support including streaming MAC support: this is required for GOST
8208 ciphersuite support.
8210 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8212 * Add option -stream to use PKCS#7 streaming in smime utility. New
8213 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8214 to output in BER and PEM format.
8218 * Experimental support for use of HMAC via EVP_PKEY interface. This
8219 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8220 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8221 ENGINE support for HMAC keys which are unextractable. New -mac and
8222 -macopt options to dgst utility.
8226 * New option -sigopt to dgst utility. Update dgst to use
8227 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8228 alternative signing parameters such as X9.31 or PSS in the dgst
8233 * Change ssl_cipher_apply_rule(), the internal function that does
8234 the work each time a ciphersuite string requests enabling
8235 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8236 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8237 the order of disabled ciphersuites such that those ciphersuites
8238 that most recently went from enabled to disabled not only stay
8239 in order with respect to each other, but also have higher priority
8240 than other disabled ciphersuites the next time ciphersuites are
8243 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8244 the same ciphersuites as with "HIGH" alone, but in a specific
8245 order where the PSK ciphersuites come first (since they are the
8246 most recently disabled ciphersuites when "HIGH" is parsed).
8248 Also, change ssl_create_cipher_list() (using this new
8249 functionality) such that between otherwise identical
8250 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8255 * Change ssl_create_cipher_list() so that it automatically
8256 arranges the ciphersuites in reasonable order before starting
8257 to process the rule string. Thus, the definition for "DEFAULT"
8258 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8259 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8260 This makes it much easier to arrive at a reasonable default order
8261 in applications for which anonymous ciphers are OK (meaning
8262 that you can't actually use DEFAULT).
8264 *Bodo Moeller; suggested by Victor Duchovni*
8266 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8267 processing) into multiple integers instead of setting
8268 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8269 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8270 (These masks as well as the individual bit definitions are hidden
8271 away into the non-exported interface ssl/ssl_locl.h, so this
8272 change to the definition of the SSL_CIPHER structure shouldn't
8273 affect applications.) This give us more bits for each of these
8274 categories, so there is no longer a need to coagulate AES128 and
8275 AES256 into a single algorithm bit, and to coagulate Camellia128
8276 and Camellia256 into a single algorithm bit, which has led to all
8279 Thus, among other things, the kludge introduced in 0.9.7m and
8280 0.9.8e for masking out AES256 independently of AES128 or masking
8281 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8283 With the change, we also introduce new ciphersuite aliases that
8284 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8289 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8290 Use the leftmost N bytes of the signature input if the input is
8291 larger than the prime q (with N being the size in bytes of q).
8295 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8296 it yet and it is largely untested.
8300 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8304 * Initial incomplete changes to avoid need for function casts in OpenSSL
8305 some compilers (gcc 4.2 and later) reject their use. Safestack is
8306 reimplemented. Update ASN1 to avoid use of legacy functions.
8310 * Win32/64 targets are linked with Winsock2.
8314 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8315 to external functions. This can be used to increase CRL handling
8316 efficiency especially when CRLs are very large by (for example) storing
8317 the CRL revoked certificates in a database.
8321 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8322 new CRLs added to a directory can be used. New command line option
8323 -verify_return_error to s_client and s_server. This causes real errors
8324 to be returned by the verify callback instead of carrying on no matter
8325 what. This reflects the way a "real world" verify callback would behave.
8329 * GOST engine, supporting several GOST algorithms and public key formats.
8330 Kindly donated by Cryptocom.
8334 * Partial support for Issuing Distribution Point CRL extension. CRLs
8335 partitioned by DP are handled but no indirect CRL or reason partitioning
8336 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8337 selected via a scoring technique which handles IDP and AKID in CRLs.
8341 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8342 will ultimately be used for all verify operations: this will remove the
8343 X509_STORE dependency on certificate verification and allow alternative
8344 lookup methods. X509_STORE based implementations of these two callbacks.
8348 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8349 Modify get_crl() to find a valid (unexpired) CRL if possible.
8353 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8354 this would be called X509_CRL_cmp() but that name is already used by
8355 a function that just compares CRL issuer names. Cache several CRL
8356 extensions in X509_CRL structure and cache CRLDP in X509.
8360 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8361 this maps equivalent X509_NAME structures into a consistent structure.
8362 Name comparison can then be performed rapidly using memcmp().
8366 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8371 * Allow digests to supply their own micalg string for S/MIME type using
8372 the ctrl EVP_MD_CTRL_MICALG.
8376 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8377 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8378 ctrl. It can then customise the structure before and/or after signing
8383 * New function OBJ_add_sigid() to allow application defined signature OIDs
8384 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8385 to free up any added signature OIDs.
8389 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8390 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8391 digest and cipher tables. New options added to openssl utility:
8392 list-message-digest-algorithms and list-cipher-algorithms.
8396 * Change the array representation of binary polynomials: the list
8397 of degrees of non-zero coefficients is now terminated with -1.
8398 Previously it was terminated with 0, which was also part of the
8399 value; thus, the array representation was not applicable to
8400 polynomials where t^0 has coefficient zero. This change makes
8401 the array representation useful in a more general context.
8405 * Various modifications and fixes to SSL/TLS cipher string
8406 handling. For ECC, the code now distinguishes between fixed ECDH
8407 with RSA certificates on the one hand and with ECDSA certificates
8408 on the other hand, since these are separate ciphersuites. The
8409 unused code for Fortezza ciphersuites has been removed.
8411 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8412 (not "ECDHE"). For consistency with the code for DH
8413 certificates, use of ECDH certificates is now considered ECDH
8414 authentication, not RSA or ECDSA authentication (the latter is
8415 merely the CA's signing algorithm and not actively used in the
8418 The temporary ciphersuite alias "ECCdraft" is no longer
8419 available, and ECC ciphersuites are no longer excluded from "ALL"
8420 and "DEFAULT". The following aliases now exist for RFC 4492
8421 ciphersuites, most of these by analogy with the DH case:
8423 kECDHr - ECDH cert, signed with RSA
8424 kECDHe - ECDH cert, signed with ECDSA
8425 kECDH - ECDH cert (signed with either RSA or ECDSA)
8426 kEECDH - ephemeral ECDH
8427 ECDH - ECDH cert or ephemeral ECDH
8433 AECDH - anonymous ECDH
8434 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8438 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8439 Use correct micalg parameters depending on digest(s) in signed message.
8443 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8444 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8448 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8449 an engine to register a method. Add ENGINE lookups for methods and
8450 functional reference processing.
8454 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8455 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8460 * New -resign option to smime utility. This adds one or more signers
8461 to an existing PKCS#7 signedData structure. Also -md option to use an
8462 alternative message digest algorithm for signing.
8466 * Tidy up PKCS#7 routines and add new functions to make it easier to
8467 create PKCS7 structures containing multiple signers. Update smime
8468 application to support multiple signers.
8472 * New -macalg option to pkcs12 utility to allow setting of an alternative
8477 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8478 Reorganize PBE internals to lookup from a static table using NIDs,
8479 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8480 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8481 PRF which will be automatically used with PBES2.
8485 * Replace the algorithm specific calls to generate keys in "req" with the
8490 * Update PKCS#7 enveloped data routines to use new API. This is now
8491 supported by any public key method supporting the encrypt operation. A
8492 ctrl is added to allow the public key algorithm to examine or modify
8493 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8498 * Add a ctrl to asn1 method to allow a public key algorithm to express
8499 a default digest type to use. In most cases this will be SHA1 but some
8500 algorithms (such as GOST) need to specify an alternative digest. The
8501 return value indicates how strong the preference is 1 means optional and
8502 2 is mandatory (that is it is the only supported type). Modify
8503 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8504 use the default md. Update openssl utilities to use the default digest
8505 type for signing if it is not explicitly indicated.
8509 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8510 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8511 signing method from the key type. This effectively removes the link
8512 between digests and public key types.
8516 * Add an OID cross reference table and utility functions. Its purpose is to
8517 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8518 rsaEncryption. This will allow some of the algorithm specific hackery
8519 needed to use the correct OID to be removed.
8523 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8524 structures for PKCS7_sign(). They are now set up by the relevant public
8529 * Add provisional EC pkey method with support for ECDSA and ECDH.
8533 * Add support for key derivation (agreement) in the API, DH method and
8538 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8539 public and private key formats. As a side effect these add additional
8540 command line functionality not previously available: DSA signatures can be
8541 generated and verified using pkeyutl and DH key support and generation in
8548 *Oliver Tappe <zooey@hirschkaefer.de>*
8550 * New make target "install_html_docs" installs HTML renditions of the
8553 *Oliver Tappe <zooey@hirschkaefer.de>*
8555 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8556 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8557 support key and parameter generation and add initial key generation
8558 functionality for RSA.
8562 * Add functions for main EVP_PKEY_method operations. The undocumented
8563 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8564 `EVP_PKEY_{encrypt,decrypt}_old`.
8568 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8569 key API, doesn't do much yet.
8573 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8574 public key algorithms. New option to openssl utility:
8575 "list-public-key-algorithms" to print out info.
8579 * Implement the Supported Elliptic Curves Extension for
8580 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8584 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8585 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8589 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8590 utilities such as rsa, dsa, dsaparam etc except they process any key
8595 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8596 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8597 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8602 * Initial support for pluggable public key ASN1.
8603 De-spaghettify the public key ASN1 handling. Move public and private
8604 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8605 algorithm specific handling to a single module within the relevant
8606 algorithm directory. Add functions to allow (near) opaque processing
8607 of public and private key structures.
8611 * Implement the Supported Point Formats Extension for
8612 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8616 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8617 for the psk identity [hint] and the psk callback functions to the
8618 SSL_SESSION, SSL and SSL_CTX structure.
8621 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8625 SSL_CTX_use_psk_identity_hint
8626 SSL_get_psk_identity_hint
8627 SSL_get_psk_identity
8628 SSL_use_psk_identity_hint
8630 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8632 * Add RFC 3161 compliant time stamp request creation, response generation
8633 and response verification functionality.
8635 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8637 * Add initial support for TLS extensions, specifically for the server_name
8638 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8639 have new members for a host name. The SSL data structure has an
8640 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8641 stored in that context to allow for session resumption, even after the
8642 SSL has been switched to a new SSL_CTX in reaction to a client's
8643 server_name extension.
8645 New functions (subject to change):
8647 SSL_get_servername()
8648 SSL_get_servername_type()
8651 New CTRL codes and macros (subject to change):
8653 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8654 - SSL_CTX_set_tlsext_servername_callback()
8655 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8656 - SSL_CTX_set_tlsext_servername_arg()
8657 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8659 openssl s_client has a new '-servername ...' option.
8661 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8662 '-key2 ...', '-servername_fatal' (subject to change). This allows
8663 testing the HostName extension for a specific single host name ('-cert'
8664 and '-key' remain fallbacks for handshakes without HostName
8665 negotiation). If the unrecognized_name alert has to be sent, this by
8666 default is a warning; it becomes fatal with the '-servername_fatal'
8669 *Peter Sylvester, Remy Allais, Christophe Renou*
8671 * Whirlpool hash implementation is added.
8675 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8676 bn(64,32). Because of instruction set limitations it doesn't have
8677 any negative impact on performance. This was done mostly in order
8678 to make it possible to share assembler modules, such as bn_mul_mont
8679 implementations, between 32- and 64-bit builds without hassle.
8683 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8684 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8689 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8690 dedicated Montgomery multiplication procedure, is introduced.
8691 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8692 "64-bit" performance on certain 32-bit targets.
8696 * New option SSL_OP_NO_COMP to disable use of compression selectively
8697 in SSL structures. New SSL ctrl to set maximum send fragment size.
8698 Save memory by setting the I/O buffer sizes dynamically instead of
8699 using the maximum available value.
8703 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8704 in addition to the text details.
8708 * Very, very preliminary EXPERIMENTAL support for printing of general
8709 ASN1 structures. This currently produces rather ugly output and doesn't
8710 handle several customised structures at all.
8714 * Integrated support for PVK file format and some related formats such
8715 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8716 these in the 'rsa' and 'dsa' utilities.
8720 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8724 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8725 place for the (very old) "NETSCAPE" format certificates which are now
8726 handled using new ASN1 code equivalents.
8730 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8731 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8732 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8736 * Modify CRL distribution points extension code to print out previously
8737 unsupported fields. Enhance extension setting code to allow setting of
8742 * Add print and set support for Issuing Distribution Point CRL extension.
8746 * Change 'Configure' script to enable Camellia by default.
8753 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8755 * When rejecting SSL/TLS records due to an incorrect version number, never
8756 update s->server with a new major version number. As of
8757 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8758 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8759 the previous behavior could result in a read attempt at NULL when
8760 receiving specific incorrect SSL/TLS records once record payload
8761 protection is active. ([CVE-2010-0740])
8763 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8765 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8766 could be crashed if the relevant tables were not present (e.g. chrooted).
8768 *Tomas Hoger <thoger@redhat.com>*
8770 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8772 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8774 *Martin Olsson, Neel Mehta*
8776 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8777 accommodate for stack sorting, always a write lock!).
8781 * On some versions of WIN32 Heap32Next is very slow. This can cause
8782 excessive delays in the RAND_poll(): over a minute. As a workaround
8783 include a time check in the inner Heap32Next loop too.
8787 * The code that handled flushing of data in SSL/TLS originally used the
8788 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8789 the problem outlined in PR#1949. The fix suggested there however can
8790 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8791 of Apache). So instead simplify the code to flush unconditionally.
8792 This should be fine since flushing with no data to flush is a no op.
8796 * Handle TLS versions 2.0 and later properly and correctly use the
8797 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8798 off ancient servers have a habit of sticking around for a while...
8802 * Modify compression code so it frees up structures without using the
8803 ex_data callbacks. This works around a problem where some applications
8804 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8805 restarting) then use compression (e.g. SSL with compression) later.
8806 This results in significant per-connection memory leaks and
8807 has caused some security issues including CVE-2008-1678 and
8812 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8813 change when encrypting or decrypting.
8817 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8818 connect and renegotiate with servers which do not support RI.
8819 Until RI is more widely deployed this option is enabled by default.
8823 * Add "missing" ssl ctrls to clear options and mode.
8827 * If client attempts to renegotiate and doesn't support RI respond with
8828 a no_renegotiation alert as required by RFC5746. Some renegotiating
8829 TLS clients will continue a connection gracefully when they receive
8830 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8831 waiting for a server hello which it will never receive. Now we treat a
8832 received no_renegotiation alert as a fatal error. This is because
8833 applications requesting a renegotiation might well expect it to succeed
8834 and would have no code in place to handle the server denying it so the
8835 only safe thing to do is to terminate the connection.
8839 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8840 peer supports secure renegotiation and 0 otherwise. Print out peer
8841 renegotiation support in s_client/s_server.
8845 * Replace the highly broken and deprecated SPKAC certification method with
8846 the updated NID creation version. This should correctly handle UTF8.
8850 * Implement RFC5746. Re-enable renegotiation but require the extension
8851 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8852 turns out to be a bad idea. It has been replaced by
8853 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8854 SSL_CTX_set_options(). This is really not recommended unless you
8855 know what you are doing.
8857 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8859 * Fixes to stateless session resumption handling. Use initial_ctx when
8860 issuing and attempting to decrypt tickets in case it has changed during
8861 servername handling. Use a non-zero length session ID when attempting
8862 stateless session resumption: this makes it possible to determine if
8863 a resumption has occurred immediately after receiving server hello
8864 (several places in OpenSSL subtly assume this) instead of later in
8869 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8870 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8871 fixes for a few places where the return code is not checked
8874 *Julia Lawall <julia@diku.dk>*
8876 * Add --strict-warnings option to Configure script to include devteam
8877 warnings in other configurations.
8881 * Add support for --libdir option and LIBDIR variable in makefiles. This
8882 makes it possible to install openssl libraries in locations which
8883 have names other than "lib", for example "/usr/lib64" which some
8886 *Steve Henson, based on patch from Jeremy Utley*
8888 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8889 X690 8.9.12 and can produce some misleading textual output of OIDs.
8891 *Steve Henson, reported by Dan Kaminsky*
8893 * Delete MD2 from algorithm tables. This follows the recommendation in
8894 several standards that it is not used in new applications due to
8895 several cryptographic weaknesses. For binary compatibility reasons
8896 the MD2 API is still compiled in by default.
8900 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8905 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8906 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8909 *Guenter <lists@gknw.net>*
8911 * Fix the server certificate chain building code to use X509_verify_cert(),
8912 it used to have an ad-hoc builder which was unable to cope with anything
8913 other than a simple chain.
8915 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8917 * Don't check self signed certificate signatures in X509_verify_cert()
8918 by default (a flag can override this): it just wastes time without
8919 adding any security. As a useful side effect self signed root CAs
8920 with non-FIPS digests are now usable in FIPS mode.
8924 * In dtls1_process_out_of_seq_message() the check if the current message
8925 is already buffered was missing. For every new message was memory
8926 allocated, allowing an attacker to perform an denial of service attack
8927 with sending out of seq handshake messages until there is no memory
8928 left. Additionally every future message was buffered, even if the
8929 sequence number made no sense and would be part of another handshake.
8930 So only messages with sequence numbers less than 10 in advance will be
8931 buffered. ([CVE-2009-1378])
8933 *Robin Seggelmann, discovered by Daniel Mentz*
8935 * Records are buffered if they arrive with a future epoch to be
8936 processed after finishing the corresponding handshake. There is
8937 currently no limitation to this buffer allowing an attacker to perform
8938 a DOS attack with sending records with future epochs until there is no
8939 memory left. This patch adds the pqueue_size() function to determine
8940 the size of a buffer and limits the record buffer to 100 entries.
8943 *Robin Seggelmann, discovered by Daniel Mentz*
8945 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8946 parent structure is freed. ([CVE-2009-1379])
8950 * Handle non-blocking I/O properly in SSL_shutdown() call.
8952 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8954 * Add `2.5.4.*` OIDs
8956 *Ilya O. <vrghost@gmail.com>*
8958 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8960 * Disable renegotiation completely - this fixes a severe security
8961 problem ([CVE-2009-3555]) at the cost of breaking all
8962 renegotiation. Renegotiation can be re-enabled by setting
8963 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8964 run-time. This is really not recommended unless you know what
8969 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8971 * Don't set val to NULL when freeing up structures, it is freed up by
8972 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8973 zeroing past the valid field. ([CVE-2009-0789])
8975 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8977 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8978 checked correctly. This would allow some invalid signed attributes to
8979 appear to verify correctly. ([CVE-2009-0591])
8981 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8983 * Reject UniversalString and BMPString types with invalid lengths. This
8984 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8985 a legal length. ([CVE-2009-0590])
8989 * Set S/MIME signing as the default purpose rather than setting it
8990 unconditionally. This allows applications to override it at the store
8995 * Permit restricted recursion of ASN1 strings. This is needed in practice
8996 to handle some structures.
9000 * Improve efficiency of mem_gets: don't search whole buffer each time
9003 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9005 * New -hex option for openssl rand.
9009 * Print out UTF8String and NumericString when parsing ASN1.
9013 * Support NumericString type for name components.
9017 * Allow CC in the environment to override the automatically chosen
9018 compiler. Note that nothing is done to ensure flags work with the
9023 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9025 * Properly check EVP_VerifyFinal() and similar return values
9028 *Ben Laurie, Bodo Moeller, Google Security Team*
9030 * Enable TLS extensions by default.
9034 * Allow the CHIL engine to be loaded, whether the application is
9035 multithreaded or not. (This does not release the developer from the
9036 obligation to set up the dynamic locking callbacks.)
9038 *Sander Temme <sander@temme.net>*
9040 * Use correct exit code if there is an error in dgst command.
9042 *Steve Henson; problem pointed out by Roland Dirlewanger*
9044 * Tweak Configure so that you need to say "experimental-jpake" to enable
9045 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9049 * Add experimental JPAKE support, including demo authentication in
9050 s_client and s_server.
9054 * Set the comparison function in v3_addr_canonize().
9056 *Rob Austein <sra@hactrn.net>*
9058 * Add support for XMPP STARTTLS in s_client.
9060 *Philip Paeps <philip@freebsd.org>*
9062 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9063 to ensure that even with this option, only ciphersuites in the
9064 server's preference list will be accepted. (Note that the option
9065 applies only when resuming a session, so the earlier behavior was
9066 just about the algorithm choice for symmetric cryptography.)
9070 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9072 * Fix NULL pointer dereference if a DTLS server received
9073 ChangeCipherSpec as first record ([CVE-2009-1386]).
9077 * Fix a state transition in s3_srvr.c and d1_srvr.c
9078 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9082 * The fix in 0.9.8c that supposedly got rid of unsafe
9083 double-checked locking was incomplete for RSA blinding,
9084 addressing just one layer of what turns out to have been
9085 doubly unsafe triple-checked locking.
9087 So now fix this for real by retiring the MONT_HELPER macro
9088 in crypto/rsa/rsa_eay.c.
9090 *Bodo Moeller; problem pointed out by Marius Schilder*
9092 * Various precautionary measures:
9094 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9096 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9097 (NB: This would require knowledge of the secret session ticket key
9098 to exploit, in which case you'd be SOL either way.)
9100 - Change bn_nist.c so that it will properly handle input BIGNUMs
9101 outside the expected range.
9103 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9106 *Neel Mehta, Bodo Moeller*
9108 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9109 the load fails. Useful for distros.
9111 *Ben Laurie and the FreeBSD team*
9113 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9117 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9121 * Expand ENGINE to support engine supplied SSL client certificate functions.
9123 This work was sponsored by Logica.
9127 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9128 keystores. Support for SSL/TLS client authentication too.
9129 Not compiled unless enable-capieng specified to Configure.
9131 This work was sponsored by Logica.
9135 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9136 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9137 attribute creation routines such as certificate requests and PKCS#12
9142 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9144 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9145 handshake which could lead to a client crash as found using the
9146 Codenomicon TLS test suite ([CVE-2008-1672])
9148 *Steve Henson, Mark Cox*
9150 * Fix double free in TLS server name extensions which could lead to
9151 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9155 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9157 Clear the error queue to ensure that error entries left from
9158 older function calls do not interfere with the correct operation.
9160 *Lutz Jaenicke, Erik de Castro Lopo*
9162 * Remove root CA certificates of commercial CAs:
9164 The OpenSSL project does not recommend any specific CA and does not
9165 have any policy with respect to including or excluding any CA.
9166 Therefore it does not make any sense to ship an arbitrary selection
9167 of root CA certificates with the OpenSSL software.
9171 * RSA OAEP patches to fix two separate invalid memory reads.
9172 The first one involves inputs when 'lzero' is greater than
9173 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9174 before the beginning of from). The second one involves inputs where
9175 the 'db' section contains nothing but zeroes (there is a one-byte
9176 invalid read after the end of 'db').
9178 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9180 * Partial backport from 0.9.9-dev:
9182 Introduce bn_mul_mont (dedicated Montgomery multiplication
9183 procedure) as a candidate for BIGNUM assembler implementation.
9184 While 0.9.9-dev uses assembler for various architectures, only
9185 x86_64 is available by default here in the 0.9.8 branch, and
9186 32-bit x86 is available through a compile-time setting.
9188 To try the 32-bit x86 assembler implementation, use Configure
9189 option "enable-montasm" (which exists only for this backport).
9191 As "enable-montasm" for 32-bit x86 disclaims code stability
9192 anyway, in this constellation we activate additional code
9193 backported from 0.9.9-dev for further performance improvements,
9194 namely BN_from_montgomery_word. (To enable this otherwise,
9195 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9197 *Andy Polyakov (backport partially by Bodo Moeller)*
9199 * Add TLS session ticket callback. This allows an application to set
9200 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9201 values. This is useful for key rollover for example where several key
9202 sets may exist with different names.
9206 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9207 This was broken until now in 0.9.8 releases, such that the only way
9208 a registered ENGINE could be used (assuming it initialises
9209 successfully on the host) was to explicitly set it as the default
9210 for the relevant algorithms. This is in contradiction with 0.9.7
9211 behaviour and the documentation. With this fix, when an ENGINE is
9212 registered into a given algorithm's table of implementations, the
9213 'uptodate' flag is reset so that auto-discovery will be used next
9214 time a new context for that algorithm attempts to select an
9217 *Ian Lister (tweaked by Geoff Thorpe)*
9219 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9220 implementation in the following ways:
9222 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9225 Lack of BER streaming support means one pass streaming processing is
9226 only supported if data is detached: setting the streaming flag is
9227 ignored for embedded content.
9229 CMS support is disabled by default and must be explicitly enabled
9230 with the enable-cms configuration option.
9234 * Update the GMP engine glue to do direct copies between BIGNUM and
9235 mpz_t when openssl and GMP use the same limb size. Otherwise the
9236 existing "conversion via a text string export" trick is still used.
9238 *Paul Sheer <paulsheer@gmail.com>*
9240 * Zlib compression BIO. This is a filter BIO which compressed and
9241 uncompresses any data passed through it.
9245 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9246 RFC3394 compatible AES key wrapping.
9250 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9251 sets string data without copying. X509_ALGOR_set0() and
9252 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9253 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9254 from an X509_ATTRIBUTE structure optionally checking it occurs only
9255 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9260 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9261 to get the expected BN_FLG_CONSTTIME behavior.
9263 *Bodo Moeller (Google)*
9267 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9268 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9269 - added some more tests to do_tests.pl
9270 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9271 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9272 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9273 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9274 - various changes to netware.pl to enable gcc-cross builds on Win32
9276 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9277 - various changes to fix missing prototype warnings
9278 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9279 - added AES, WHIRLPOOL and CPUID assembler code to build files
9280 - added missing AES assembler make rules to mk1mf.pl
9281 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9283 *Guenter Knauf <eflash@gmx.net>*
9285 * Implement certificate status request TLS extension defined in RFC3546.
9286 A client can set the appropriate parameters and receive the encoded
9287 OCSP response via a callback. A server can query the supplied parameters
9288 and set the encoded OCSP response in the callback. Add simplified examples
9289 to s_client and s_server.
9293 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9296 + Binary incompatibility of ssl_ctx_st structure
9297 + DTLS interoperation with non-compliant servers
9298 + Don't call get_session_cb() without proposed session
9299 + Fix ia64 assembler code
9301 *Andy Polyakov, Steve Henson*
9303 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9305 * DTLS Handshake overhaul. There were longstanding issues with
9306 OpenSSL DTLS implementation, which were making it impossible for
9307 RFC 4347 compliant client to communicate with OpenSSL server.
9308 Unfortunately just fixing these incompatibilities would "cut off"
9309 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9310 server keeps tolerating non RFC compliant syntax. The opposite is
9311 not true, 0.9.8f client can not communicate with earlier server.
9312 This update even addresses CVE-2007-4995.
9316 * Changes to avoid need for function casts in OpenSSL: some compilers
9317 (gcc 4.2 and later) reject their use.
9318 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9321 * Add RFC4507 support to OpenSSL. This includes the corrections in
9322 RFC4507bis. The encrypted ticket format is an encrypted encoded
9323 SSL_SESSION structure, that way new session features are automatically
9326 If a client application caches session in an SSL_SESSION structure
9327 support is transparent because tickets are now stored in the encoded
9330 The SSL_CTX structure automatically generates keys for ticket
9331 protection in servers so again support should be possible
9332 with no application modification.
9334 If a client or server wishes to disable RFC4507 support then the option
9335 SSL_OP_NO_TICKET can be set.
9337 Add a TLS extension debugging callback to allow the contents of any client
9338 or server extensions to be examined.
9340 This work was sponsored by Google.
9344 * Add initial support for TLS extensions, specifically for the server_name
9345 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9346 have new members for a host name. The SSL data structure has an
9347 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9348 stored in that context to allow for session resumption, even after the
9349 SSL has been switched to a new SSL_CTX in reaction to a client's
9350 server_name extension.
9352 New functions (subject to change):
9354 SSL_get_servername()
9355 SSL_get_servername_type()
9358 New CTRL codes and macros (subject to change):
9360 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9361 - SSL_CTX_set_tlsext_servername_callback()
9362 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9363 - SSL_CTX_set_tlsext_servername_arg()
9364 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9366 openssl s_client has a new '-servername ...' option.
9368 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9369 '-key2 ...', '-servername_fatal' (subject to change). This allows
9370 testing the HostName extension for a specific single host name ('-cert'
9371 and '-key' remain fallbacks for handshakes without HostName
9372 negotiation). If the unrecognized_name alert has to be sent, this by
9373 default is a warning; it becomes fatal with the '-servername_fatal'
9376 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9378 * Add AES and SSE2 assembly language support to VC++ build.
9382 * Mitigate attack on final subtraction in Montgomery reduction.
9386 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9387 (which previously caused an internal error).
9391 * Squeeze another 10% out of IGE mode when in != out.
9395 * AES IGE mode speedup.
9397 *Dean Gaudet (Google)*
9399 * Add the Korean symmetric 128-bit cipher SEED (see
9400 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9401 add SEED ciphersuites from RFC 4162:
9403 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9404 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9405 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9406 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9408 To minimize changes between patchlevels in the OpenSSL 0.9.8
9409 series, SEED remains excluded from compilation unless OpenSSL
9410 is configured with 'enable-seed'.
9412 *KISA, Bodo Moeller*
9414 * Mitigate branch prediction attacks, which can be practical if a
9415 single processor is shared, allowing a spy process to extract
9416 information. For detailed background information, see
9417 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9418 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9419 and Necessary Software Countermeasures"). The core of the change
9420 are new versions BN_div_no_branch() and
9421 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9422 respectively, which are slower, but avoid the security-relevant
9423 conditional branches. These are automatically called by BN_div()
9424 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9425 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9426 remove a conditional branch.
9428 BN_FLG_CONSTTIME is the new name for the previous
9429 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9430 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9431 in the exponent causes BN_mod_exp_mont() to use the alternative
9432 implementation in BN_mod_exp_mont_consttime().) The old name
9433 remains as a deprecated alias.
9435 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9436 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9437 constant-time implementations for more than just exponentiation.
9438 Here too the old name is kept as a deprecated alias.
9440 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9441 the BN_BLINDING structure gets an independent copy of the
9442 modulus. This means that the previous `BIGNUM *m` argument to
9443 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9444 essentially becomes `const BIGNUM *m`, although we can't actually
9445 change this in the header file before 0.9.9. It allows
9446 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9447 enable BN_FLG_CONSTTIME.
9449 *Matthew D Wood (Intel Corp)*
9451 * In the SSL/TLS server implementation, be strict about session ID
9452 context matching (which matters if an application uses a single
9453 external cache for different purposes). Previously,
9454 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9455 set. This did ensure strict client verification, but meant that,
9456 with applications using a single external cache for quite
9457 different requirements, clients could circumvent ciphersuite
9458 restrictions for a given session ID context by starting a session
9459 in a different context.
9463 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9464 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9465 authentication-only ciphersuites.
9469 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9470 not complete and could lead to a possible single byte overflow
9471 ([CVE-2007-5135]) [Ben Laurie]
9473 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9475 * Since AES128 and AES256 (and similarly Camellia128 and
9476 Camellia256) share a single mask bit in the logic of
9477 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9478 kludge to work properly if AES128 is available and AES256 isn't
9479 (or if Camellia128 is available and Camellia256 isn't).
9483 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9484 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9485 When a point or a seed is encoded in a BIT STRING, we need to
9486 prevent the removal of trailing zero bits to get the proper DER
9487 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9488 of a NamedBitList, for which trailing 0 bits need to be removed.)
9492 * Have SSL/TLS server implementation tolerate "mismatched" record
9493 protocol version while receiving ClientHello even if the
9494 ClientHello is fragmented. (The server can't insist on the
9495 particular protocol version it has chosen before the ServerHello
9496 message has informed the client about his choice.)
9500 * Add RFC 3779 support.
9502 *Rob Austein for ARIN, Ben Laurie*
9504 * Load error codes if they are not already present instead of using a
9505 static variable. This allows them to be cleanly unloaded and reloaded.
9506 Improve header file function name parsing.
9510 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9511 or CAPABILITY handshake as required by RFCs.
9515 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9517 * Introduce limits to prevent malicious keys being able to
9518 cause a denial of service. ([CVE-2006-2940])
9520 *Steve Henson, Bodo Moeller*
9522 * Fix ASN.1 parsing of certain invalid structures that can result
9523 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9525 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9526 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9528 * Fix SSL client code which could crash if connecting to a
9529 malicious SSLv2 server. ([CVE-2006-4343])
9531 *Tavis Ormandy and Will Drewry, Google Security Team*
9533 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9534 match only those. Before that, "AES256-SHA" would be interpreted
9535 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9536 the same strength classification in 0.9.7h) as we currently only
9537 have a single AES bit in the ciphersuite description bitmap.
9538 That change, however, also applied to ciphersuite strings such as
9539 "RC4-MD5" that intentionally matched multiple ciphersuites --
9540 namely, SSL 2.0 ciphersuites in addition to the more common ones
9541 from SSL 3.0/TLS 1.0.
9543 So we change the selection algorithm again: Naming an explicit
9544 ciphersuite selects this one ciphersuite, and any other similar
9545 ciphersuite (same bitmap) from *other* protocol versions.
9546 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9547 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9549 Since SSL 2.0 does not have any ciphersuites for which the
9550 128/256 bit distinction would be relevant, this works for now.
9551 The proper fix will be to use different bits for AES128 and
9552 AES256, which would have avoided the problems from the beginning;
9553 however, bits are scarce, so we can only do this in a new release
9554 (not just a patchlevel) when we can change the SSL_CIPHER
9555 definition to split the single 'unsigned long mask' bitmap into
9556 multiple values to extend the available space.
9560 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9562 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9563 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9565 * Add AES IGE and biIGE modes.
9569 * Change the Unix randomness entropy gathering to use poll() when
9570 possible instead of select(), since the latter has some
9571 undesirable limitations.
9573 *Darryl Miles via Richard Levitte and Bodo Moeller*
9575 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9576 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9577 cannot be implicitly activated as part of, e.g., the "AES" alias.
9578 However, please upgrade to OpenSSL 0.9.9[-dev] for
9579 non-experimental use of the ECC ciphersuites to get TLS extension
9580 support, which is required for curve and point format negotiation
9581 to avoid potential handshake problems.
9585 * Disable rogue ciphersuites:
9587 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9588 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9589 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9591 The latter two were purportedly from
9592 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9595 Also deactivate the remaining ciphersuites from
9596 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9597 unofficial, and the ID has long expired.
9601 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9602 dual-core machines) and other potential thread-safety issues.
9606 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9607 versions), which is now available for royalty-free use
9608 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9609 Also, add Camellia TLS ciphersuites from RFC 4132.
9611 To minimize changes between patchlevels in the OpenSSL 0.9.8
9612 series, Camellia remains excluded from compilation unless OpenSSL
9613 is configured with 'enable-camellia'.
9617 * Disable the padding bug check when compression is in use. The padding
9618 bug check assumes the first packet is of even length, this is not
9619 necessarily true if compression is enabled and can result in false
9620 positives causing handshake failure. The actual bug test is ancient
9621 code so it is hoped that implementations will either have fixed it by
9622 now or any which still have the bug do not support compression.
9626 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9628 * When applying a cipher rule check to see if string match is an explicit
9629 cipher suite and only match that one cipher suite if it is.
9633 * Link in manifests for VC++ if needed.
9635 *Austin Ziegler <halostatue@gmail.com>*
9637 * Update support for ECC-based TLS ciphersuites according to
9638 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9639 TLS extensions, which are supported starting with the 0.9.9
9640 branch, not in the OpenSSL 0.9.8 branch).
9644 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9645 opaque EVP_CIPHER_CTX handling.
9649 * Fixes and enhancements to zlib compression code. We now only use
9650 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9651 to conform with the standards mentioned here:
9652 <http://www.zlib.net/DLL_FAQ.txt>
9653 Static zlib linking now works on Windows and the new --with-zlib-include
9654 --with-zlib-lib options to Configure can be used to supply the location
9655 of the headers and library. Gracefully handle case where zlib library
9660 * Several fixes and enhancements to the OID generation code. The old code
9661 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9662 handle numbers larger than ULONG_MAX, truncated printing and had a
9663 non standard OBJ_obj2txt() behaviour.
9667 * Add support for building of engines under engine/ as shared libraries
9668 under VC++ build system.
9672 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9673 Hopefully, we will not see any false combination of paths any more.
9677 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9679 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9680 (part of SSL_OP_ALL). This option used to disable the
9681 countermeasure against man-in-the-middle protocol-version
9682 rollback in the SSL 2.0 server implementation, which is a bad
9683 idea. ([CVE-2005-2969])
9685 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9686 for Information Security, National Institute of Advanced Industrial
9687 Science and Technology [AIST], Japan)*
9689 * Add two function to clear and return the verify parameter flags.
9693 * Keep cipherlists sorted in the source instead of sorting them at
9694 runtime, thus removing the need for a lock.
9698 * Avoid some small subgroup attacks in Diffie-Hellman.
9700 *Nick Mathewson and Ben Laurie*
9702 * Add functions for well-known primes.
9706 * Extended Windows CE support.
9708 *Satoshi Nakamura and Andy Polyakov*
9710 * Initialize SSL_METHOD structures at compile time instead of during
9711 runtime, thus removing the need for a lock.
9715 * Make PKCS7_decrypt() work even if no certificate is supplied by
9716 attempting to decrypt each encrypted key in turn. Add support to
9721 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9723 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9726 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9730 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9731 key into the same file any more.
9735 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9739 * Add -utf8 command line and config file option to 'ca'.
9741 *Stefan <stf@udoma.org*
9743 * Removed the macro des_crypt(), as it seems to conflict with some
9744 libraries. Use DES_crypt().
9748 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9749 involves renaming the source and generated shared-libs for
9750 both. The engines will accept the corrected or legacy ids
9751 ('ncipher' and '4758_cca' respectively) when binding. NB,
9752 this only applies when building 'shared'.
9754 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9756 * Add attribute functions to EVP_PKEY structure. Modify
9757 PKCS12_create() to recognize a CSP name attribute and
9758 use it. Make -CSP option work again in pkcs12 utility.
9762 * Add new functionality to the bn blinding code:
9763 - automatic re-creation of the BN_BLINDING parameters after
9764 a fixed number of uses (currently 32)
9765 - add new function for parameter creation
9766 - introduce flags to control the update behaviour of the
9767 BN_BLINDING parameters
9768 - hide BN_BLINDING structure
9769 Add a second BN_BLINDING slot to the RSA structure to improve
9770 performance when a single RSA object is shared among several
9775 * Add support for DTLS.
9777 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9779 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9780 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9784 * Remove buggy and incomplete DH cert support from
9785 ssl/ssl_rsa.c and ssl/s3_both.c
9789 * Use SHA-1 instead of MD5 as the default digest algorithm for
9790 the apps/openssl applications.
9794 * Compile clean with "-Wall -Wmissing-prototypes
9795 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9796 DEBUG_SAFESTACK must also be set.
9800 * Change ./Configure so that certain algorithms can be disabled by default.
9801 The new counterpiece to "no-xxx" is "enable-xxx".
9803 The patented RC5 and MDC2 algorithms will now be disabled unless
9804 "enable-rc5" and "enable-mdc2", respectively, are specified.
9806 (IDEA remains enabled despite being patented. This is because IDEA
9807 is frequently required for interoperability, and there is no license
9808 fee for non-commercial use. As before, "no-idea" can be used to
9809 avoid this algorithm.)
9813 * Add processing of proxy certificates (see RFC 3820). This work was
9814 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9815 EGEE (Enabling Grids for E-science in Europe).
9819 * RC4 performance overhaul on modern architectures/implementations, such
9820 as Intel P4, IA-64 and AMD64.
9824 * New utility extract-section.pl. This can be used specify an alternative
9825 section number in a pod file instead of having to treat each file as
9826 a separate case in Makefile. This can be done by adding two lines to the
9829 =for comment openssl_section:XXX
9831 The blank line is mandatory.
9835 * New arguments -certform, -keyform and -pass for s_client and s_server
9836 to allow alternative format key and certificate files and passphrase
9841 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9842 update associated structures and add various utility functions.
9844 Add new policy related verify parameters, include policy checking in
9845 standard verify code. Enhance 'smime' application with extra parameters
9846 to support policy checking and print out.
9850 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9851 Nehemiah processors. These extensions support AES encryption in hardware
9852 as well as RNG (though RNG support is currently disabled).
9854 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9856 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9860 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9862 *Andy Polyakov and a number of other people*
9864 * Improved PowerPC platform support. Most notably BIGNUM assembler
9865 implementation contributed by IBM.
9867 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9869 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9870 exponent rather than 'unsigned long'. There is a corresponding change to
9871 the new 'rsa_keygen' element of the RSA_METHOD structure.
9873 *Jelte Jansen, Geoff Thorpe*
9875 * Functionality for creating the initial serial number file is now
9876 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9878 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9879 number file to 1, which is bound to cause problems. To avoid
9880 the problems while respecting compatibility between different 0.9.7
9881 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9882 CA.pl for serial number initialization. With the new release 0.9.8,
9883 we can fix the problem directly in the 'ca' utility.)
9887 * Reduced header interdependencies by declaring more opaque objects in
9888 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9889 give fewer recursive includes, which could break lazy source code - so
9890 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9891 developers should define this symbol when building and using openssl to
9892 ensure they track the recommended behaviour, interfaces, [etc], but
9893 backwards-compatible behaviour prevails when this isn't defined.
9897 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9901 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9902 This will generate a random key of the appropriate length based on the
9903 cipher context. The EVP_CIPHER can provide its own random key generation
9904 routine to support keys of a specific form. This is used in the des and
9905 3des routines to generate a key of the correct parity. Update S/MIME
9906 code to use new functions and hence generate correct parity DES keys.
9907 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9908 valid (weak or incorrect parity).
9912 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9913 as looking them up. This is useful when the verified structure may contain
9914 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9915 present unless the new PKCS7_NO_CRL flag is asserted.
9919 * Extend ASN1 oid configuration module. It now additionally accepts the
9922 shortName = some long name, 1.2.3.4
9926 * Reimplemented the BN_CTX implementation. There is now no more static
9927 limitation on the number of variables it can handle nor the depth of the
9928 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9929 information can now expand as required, and rather than having a single
9930 static array of bignums, BN_CTX now uses a linked-list of such arrays
9931 allowing it to expand on demand whilst maintaining the usefulness of
9932 BN_CTX's "bundling".
9936 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9937 to allow all RSA operations to function using a single BN_CTX.
9941 * Preliminary support for certificate policy evaluation and checking. This
9942 is initially intended to pass the tests outlined in "Conformance Testing
9943 of Relying Party Client Certificate Path Processing Logic" v1.07.
9947 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9948 remained unused and not that useful. A variety of other little bignum
9949 tweaks and fixes have also been made continuing on from the audit (see
9954 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9955 associated ASN1, EVP and SSL functions and old ASN1 macros.
9959 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9960 and this should never fail. So the return value from the use of
9961 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9962 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9966 * BN_CTX_get() should return zero-valued bignums, providing the same
9967 initialised value as BN_new().
9969 *Geoff Thorpe, suggested by Ulf Möller*
9971 * Support for inhibitAnyPolicy certificate extension.
9975 * An audit of the BIGNUM code is underway, for which debugging code is
9976 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9977 is considered valid when processing BIGNUMs, and causes execution to
9978 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9979 further steps are taken to deliberately pollute unused data in BIGNUM
9980 structures to try and expose faulty code further on. For now, openssl will
9981 (in its default mode of operation) continue to tolerate the inconsistent
9982 forms that it has tolerated in the past, but authors and packagers should
9983 consider trying openssl and their own applications when compiled with
9984 these debugging symbols defined. It will help highlight potential bugs in
9985 their own code, and will improve the test coverage for OpenSSL itself. At
9986 some point, these tighter rules will become openssl's default to improve
9987 maintainability, though the assert()s and other overheads will remain only
9988 in debugging configurations. See bn.h for more details.
9990 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9992 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9993 that can only be obtained through BN_CTX_new() (which implicitly
9994 initialises it). The presence of this function only made it possible
9995 to overwrite an existing structure (and cause memory leaks).
9999 * Because of the callback-based approach for implementing LHASH as a
10000 template type, lh_insert() adds opaque objects to hash-tables and
10001 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10002 to clean up those corresponding objects before destroying the hash table
10003 (and losing the object pointers). So some over-zealous constifications in
10004 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10005 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10006 prototyped to have "const" restrictions on the object pointers they are
10007 given (and so aren't required to cast them away any more).
10011 * The tmdiff.h API was so ugly and minimal that our own timing utility
10012 (speed) prefers to use its own implementation. The two implementations
10013 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10014 its object type properly exposed (MS_TM) instead of casting to/from
10015 `char *`. This may still change yet if someone realises MS_TM and
10017 aren't necessarily the greatest nomenclatures - but this is what was used
10018 internally to the implementation so I've used that for now.
10022 * Ensure that deprecated functions do not get compiled when
10023 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10024 the self-tests were still using deprecated key-generation functions so
10025 these have been updated also.
10029 * Reorganise PKCS#7 code to separate the digest location functionality
10030 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10031 New function PKCS7_set_digest() to set the digest type for PKCS#7
10032 digestedData type. Add additional code to correctly generate the
10033 digestedData type and add support for this type in PKCS7 initialization
10038 * New function PKCS7_set0_type_other() this initializes a PKCS7
10039 structure of type "other".
10043 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10044 sure the loop does correctly stop and breaking ("division by zero")
10045 modulus operations are not performed. The (pre-generated) prime
10046 table crypto/bn/bn_prime.h was already correct, but it could not be
10047 re-generated on some platforms because of the "division by zero"
10048 situation in the script.
10050 *Ralf S. Engelschall*
10052 * Update support for ECC-based TLS ciphersuites according to
10053 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10054 SHA-1 now is only used for "small" curves (where the
10055 representation of a field element takes up to 24 bytes); for
10056 larger curves, the field element resulting from ECDH is directly
10057 used as premaster secret.
10059 *Douglas Stebila (Sun Microsystems Laboratories)*
10061 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10062 curve secp160r1 to the tests.
10064 *Douglas Stebila (Sun Microsystems Laboratories)*
10066 * Add the possibility to load symbols globally with DSO.
10068 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10070 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10071 control of the error stack.
10075 * Add support for STORE in ENGINE.
10079 * Add the STORE type. The intention is to provide a common interface
10080 to certificate and key stores, be they simple file-based stores, or
10081 HSM-type store, or LDAP stores, or...
10082 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10086 * Add a generic structure called OPENSSL_ITEM. This can be used to
10087 pass a list of arguments to any function as well as provide a way
10088 for a function to pass data back to the caller.
10092 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10093 works like BUF_strdup() but can be used to duplicate a portion of
10094 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10099 * Add the function sk_find_ex() which works like sk_find(), but will
10100 return an index to an element even if an exact match couldn't be
10101 found. The index is guaranteed to point at the element where the
10102 searched-for key would be inserted to preserve sorting order.
10106 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10107 takes an extra flags argument for optional functionality. Currently,
10108 the following flags are defined:
10110 OBJ_BSEARCH_VALUE_ON_NOMATCH
10111 This one gets OBJ_bsearch_ex() to return a pointer to the first
10112 element where the comparing function returns a negative or zero
10115 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10116 This one gets OBJ_bsearch_ex() to return a pointer to the first
10117 element where the comparing function returns zero. This is useful
10118 if there are more than one element where the comparing function
10123 * Make it possible to create self-signed certificates with 'openssl ca'
10124 in such a way that the self-signed certificate becomes part of the
10125 CA database and uses the same mechanisms for serial number generation
10126 as all other certificate signing. The new flag '-selfsign' enables
10127 this functionality. Adapt CA.sh and CA.pl.in.
10131 * Add functionality to check the public key of a certificate request
10132 against a given private. This is useful to check that a certificate
10133 request can be signed by that key (self-signing).
10137 * Make it possible to have multiple active certificates with the same
10138 subject in the CA index file. This is done only if the keyword
10139 'unique_subject' is set to 'no' in the main CA section (default
10140 if 'CA_default') of the configuration file. The value is saved
10141 with the database itself in a separate index attribute file,
10142 named like the index file with '.attr' appended to the name.
10146 * Generate multi-valued AVAs using '+' notation in config files for
10151 * Support for nameConstraints certificate extension.
10155 * Support for policyConstraints certificate extension.
10159 * Support for policyMappings certificate extension.
10163 * Make sure the default DSA_METHOD implementation only uses its
10164 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10165 and change its own handlers to be NULL so as to remove unnecessary
10166 indirection. This lets alternative implementations fallback to the
10167 default implementation more easily.
10171 * Support for directoryName in GeneralName related extensions
10176 * Make it possible to link applications using Makefile.shared.
10177 Make that possible even when linking against static libraries!
10181 * Support for single pass processing for S/MIME signing. This now
10182 means that S/MIME signing can be done from a pipe, in addition
10183 cleartext signing (multipart/signed type) is effectively streaming
10184 and the signed data does not need to be all held in memory.
10186 This is done with a new flag PKCS7_STREAM. When this flag is set
10187 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10188 is done after the data is output (and digests calculated) in
10189 SMIME_write_PKCS7().
10193 * Add full support for -rpath/-R, both in shared libraries and
10194 applications, at least on the platforms where it's known how
10199 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10200 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10201 will now compute a table of multiples of the generator that
10202 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10203 faster (notably in the case of a single point multiplication,
10204 scalar * generator).
10206 *Nils Larsch, Bodo Moeller*
10208 * IPv6 support for certificate extensions. The various extensions
10209 which use the IP:a.b.c.d can now take IPv6 addresses using the
10210 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10215 * Added an ENGINE that implements RSA by performing private key
10216 exponentiations with the GMP library. The conversions to and from
10217 GMP's mpz_t format aren't optimised nor are any montgomery forms
10218 cached, and on x86 it appears OpenSSL's own performance has caught up.
10219 However there are likely to be other architectures where GMP could
10220 provide a boost. This ENGINE is not built in by default, but it can be
10221 specified at Configure time and should be accompanied by the necessary
10222 linker additions, eg;
10223 ./config -DOPENSSL_USE_GMP -lgmp
10227 * "openssl engine" will not display ENGINE/DSO load failure errors when
10228 testing availability of engines with "-t" - the old behaviour is
10229 produced by increasing the feature's verbosity with "-tt".
10233 * ECDSA routines: under certain error conditions uninitialized BN objects
10234 could be freed. Solution: make sure initialization is performed early
10235 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10240 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10241 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10242 software implementations. For DSA and DH, parameter generation can
10243 also be overridden by providing the appropriate method callbacks.
10247 * Change the "progress" mechanism used in key-generation and
10248 primality testing to functions that take a new BN_GENCB pointer in
10249 place of callback/argument pairs. The new API functions have `_ex`
10250 postfixes and the older functions are reimplemented as wrappers for
10251 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10252 declarations of the old functions to help (graceful) attempts to
10253 migrate to the new functions. Also, the new key-generation API
10254 functions operate on a caller-supplied key-structure and return
10255 success/failure rather than returning a key or NULL - this is to
10256 help make "keygen" another member function of RSA_METHOD etc.
10258 Example for using the new callback interface:
10260 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10261 void *my_arg = ...;
10264 BN_GENCB_set(&my_cb, my_callback, my_arg);
10266 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10267 /* For the meaning of a, b in calls to my_callback(), see the
10268 * documentation of the function that calls the callback.
10269 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10270 * my_callback should return 1 if it wants BN_is_prime_ex()
10271 * to continue, or 0 to stop.
10276 * Change the ZLIB compression method to be stateful, and make it
10277 available to TLS with the number defined in
10278 draft-ietf-tls-compression-04.txt.
10282 * Add the ASN.1 structures and functions for CertificatePair, which
10283 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10285 CertificatePair ::= SEQUENCE {
10286 forward [0] Certificate OPTIONAL,
10287 reverse [1] Certificate OPTIONAL,
10288 -- at least one of the pair shall be present -- }
10290 Also implement the PEM functions to read and write certificate
10291 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10293 This needed to be defined, mostly for the sake of the LDAP
10294 attribute crossCertificatePair, but may prove useful elsewhere as
10299 * Make it possible to inhibit symlinking of shared libraries in
10300 Makefile.shared, for Cygwin's sake.
10304 * Extend the BIGNUM API by creating a function
10305 void BN_set_negative(BIGNUM *a, int neg);
10306 and a macro that behave like
10307 int BN_is_negative(const BIGNUM *a);
10309 to avoid the need to access 'a->neg' directly in applications.
10313 * Implement fast modular reduction for pseudo-Mersenne primes
10314 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10315 EC_GROUP_new_curve_GFp() will now automatically use this
10318 *Nils Larsch <nla@trustcenter.de>*
10320 * Add new lock type (CRYPTO_LOCK_BN).
10324 * Change the ENGINE framework to automatically load engines
10325 dynamically from specific directories unless they could be
10326 found to already be built in or loaded. Move all the
10327 current engines except for the cryptodev one to a new
10328 directory engines/.
10329 The engines in engines/ are built as shared libraries if
10330 the "shared" options was given to ./Configure or ./config.
10331 Otherwise, they are inserted in libcrypto.a.
10332 /usr/local/ssl/engines is the default directory for dynamic
10333 engines, but that can be overridden at configure time through
10334 the usual use of --prefix and/or --openssldir, and at run
10335 time with the environment variable OPENSSL_ENGINES.
10337 *Geoff Thorpe and Richard Levitte*
10339 * Add Makefile.shared, a helper makefile to build shared
10340 libraries. Adapt Makefile.org.
10344 * Add version info to Win32 DLLs.
10346 *Peter 'Luna' Runestig" <peter@runestig.com>*
10348 * Add new 'medium level' PKCS#12 API. Certificates and keys
10349 can be added using this API to created arbitrary PKCS#12
10350 files while avoiding the low level API.
10352 New options to PKCS12_create(), key or cert can be NULL and
10353 will then be omitted from the output file. The encryption
10354 algorithm NIDs can be set to -1 for no encryption, the mac
10355 iteration count can be set to 0 to omit the mac.
10357 Enhance pkcs12 utility by making the -nokeys and -nocerts
10358 options work when creating a PKCS#12 file. New option -nomac
10359 to omit the mac, NONE can be set for an encryption algorithm.
10360 New code is modified to use the enhanced PKCS12_create()
10361 instead of the low level API.
10365 * Extend ASN1 encoder to support indefinite length constructed
10366 encoding. This can output sequences tags and octet strings in
10367 this form. Modify pk7_asn1.c to support indefinite length
10368 encoding. This is experimental and needs additional code to
10369 be useful, such as an ASN1 bio and some enhanced streaming
10372 Extend template encode functionality so that tagging is passed
10373 down to the template encoder.
10377 * Let 'openssl req' fail if an argument to '-newkey' is not
10378 recognized instead of using RSA as a default.
10382 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10383 As these are not official, they are not included in "ALL";
10384 the "ECCdraft" ciphersuite group alias can be used to select them.
10386 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10388 * Add ECDH engine support.
10390 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10392 * Add ECDH in new directory crypto/ecdh/.
10394 *Douglas Stebila (Sun Microsystems Laboratories)*
10396 * Let BN_rand_range() abort with an error after 100 iterations
10397 without success (which indicates a broken PRNG).
10401 * Change BN_mod_sqrt() so that it verifies that the input value
10402 is really the square of the return value. (Previously,
10403 BN_mod_sqrt would show GIGO behaviour.)
10407 * Add named elliptic curves over binary fields from X9.62, SECG,
10408 and WAP/WTLS; add OIDs that were still missing.
10410 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10412 * Extend the EC library for elliptic curves over binary fields
10413 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10416 EC_GF2m_simple_method
10420 EC_GROUP_new_curve_GF2m
10421 EC_GROUP_set_curve_GF2m
10422 EC_GROUP_get_curve_GF2m
10423 EC_POINT_set_affine_coordinates_GF2m
10424 EC_POINT_get_affine_coordinates_GF2m
10425 EC_POINT_set_compressed_coordinates_GF2m
10427 Point compression for binary fields is disabled by default for
10428 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10431 As binary polynomials are represented as BIGNUMs, various members
10432 of the EC_GROUP and EC_POINT data structures can be shared
10433 between the implementations for prime fields and binary fields;
10434 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10435 are essentially identical to their `..._GFp` counterparts.
10436 (For simplicity, the `..._GFp` prefix has been dropped from
10437 various internal method names.)
10439 An internal 'field_div' method (similar to 'field_mul' and
10440 'field_sqr') has been added; this is used only for binary fields.
10442 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10444 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10445 through methods ('mul', 'precompute_mult').
10447 The generic implementations (now internally called 'ec_wNAF_mul'
10448 and 'ec_wNAF_precomputed_mult') remain the default if these
10449 methods are undefined.
10451 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10453 * New function EC_GROUP_get_degree, which is defined through
10454 EC_METHOD. For curves over prime fields, this returns the bit
10455 length of the modulus.
10457 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10459 * New functions EC_GROUP_dup, EC_POINT_dup.
10460 (These simply call ..._new and ..._copy).
10462 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10464 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10465 Polynomials are represented as BIGNUMs (where the sign bit is not
10466 used) in the following functions [macros]:
10469 BN_GF2m_sub [= BN_GF2m_add]
10470 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10471 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10472 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10474 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10475 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10476 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10477 BN_GF2m_cmp [= BN_ucmp]
10479 (Note that only the 'mod' functions are actually for fields GF(2^m).
10480 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10482 For some functions, an the irreducible polynomial defining a
10483 field can be given as an 'unsigned int[]' with strictly
10484 decreasing elements giving the indices of those bits that are set;
10485 i.e., p[] represents the polynomial
10486 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10488 p[0] > p[1] > ... > p[k] = 0.
10489 This applies to the following functions:
10492 BN_GF2m_mod_mul_arr
10493 BN_GF2m_mod_sqr_arr
10494 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10495 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10496 BN_GF2m_mod_exp_arr
10497 BN_GF2m_mod_sqrt_arr
10498 BN_GF2m_mod_solve_quad_arr
10502 Conversion can be performed by the following functions:
10507 bntest.c has additional tests for binary polynomial arithmetic.
10509 Two implementations for BN_GF2m_mod_div() are available.
10510 The default algorithm simply uses BN_GF2m_mod_inv() and
10511 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10512 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10513 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10515 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10517 * Add new error code 'ERR_R_DISABLED' that can be used when some
10518 functionality is disabled at compile-time.
10520 *Douglas Stebila <douglas.stebila@sun.com>*
10522 * Change default behaviour of 'openssl asn1parse' so that more
10523 information is visible when viewing, e.g., a certificate:
10525 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10526 mode the content of non-printable OCTET STRINGs is output in a
10527 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10528 avoid the appearance of a printable string.
10530 *Nils Larsch <nla@trustcenter.de>*
10532 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10534 EC_GROUP_set_asn1_flag()
10535 EC_GROUP_get_asn1_flag()
10536 EC_GROUP_set_point_conversion_form()
10537 EC_GROUP_get_point_conversion_form()
10538 These control ASN1 encoding details:
10539 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10540 has been set to OPENSSL_EC_NAMED_CURVE.
10541 - Points are encoded in uncompressed form by default; options for
10542 asn1_for are as for point2oct, namely
10543 POINT_CONVERSION_COMPRESSED
10544 POINT_CONVERSION_UNCOMPRESSED
10545 POINT_CONVERSION_HYBRID
10547 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10549 EC_GROUP_set_seed()
10550 EC_GROUP_get0_seed()
10551 EC_GROUP_get_seed_len()
10552 This is used only for ASN1 purposes (so far).
10554 *Nils Larsch <nla@trustcenter.de>*
10556 * Add 'field_type' member to EC_METHOD, which holds the NID
10557 of the appropriate field type OID. The new function
10558 EC_METHOD_get_field_type() returns this value.
10560 *Nils Larsch <nla@trustcenter.de>*
10563 EC_POINT_point2bn()
10564 EC_POINT_bn2point()
10565 EC_POINT_point2hex()
10566 EC_POINT_hex2point()
10567 providing useful interfaces to EC_POINT_point2oct() and
10568 EC_POINT_oct2point().
10570 *Nils Larsch <nla@trustcenter.de>*
10572 * Change internals of the EC library so that the functions
10573 EC_GROUP_set_generator()
10574 EC_GROUP_get_generator()
10575 EC_GROUP_get_order()
10576 EC_GROUP_get_cofactor()
10577 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10578 to methods, which would lead to unnecessary code duplication when
10579 adding different types of curves.
10581 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10583 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10584 arithmetic, and such that modified wNAFs are generated
10585 (which avoid length expansion in many cases).
10589 * Add a function EC_GROUP_check_discriminant() (defined via
10590 EC_METHOD) that verifies that the curve discriminant is non-zero.
10592 Add a function EC_GROUP_check() that makes some sanity tests
10593 on a EC_GROUP, its generator and order. This includes
10594 EC_GROUP_check_discriminant().
10596 *Nils Larsch <nla@trustcenter.de>*
10598 * Add ECDSA in new directory crypto/ecdsa/.
10600 Add applications 'openssl ecparam' and 'openssl ecdsa'
10601 (these are based on 'openssl dsaparam' and 'openssl dsa').
10603 ECDSA support is also included in various other files across the
10604 library. Most notably,
10605 - 'openssl req' now has a '-newkey ecdsa:file' option;
10606 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10607 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10608 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10609 them suitable for ECDSA where domain parameters must be
10610 extracted before the specific public key;
10611 - ECDSA engine support has been added.
10613 *Nils Larsch <nla@trustcenter.de>*
10615 * Include some named elliptic curves, and add OIDs from X9.62,
10616 SECG, and WAP/WTLS. Each curve can be obtained from the new
10618 EC_GROUP_new_by_curve_name(),
10619 and the list of available named curves can be obtained with
10620 EC_get_builtin_curves().
10621 Also add a 'curve_name' member to EC_GROUP objects, which can be
10623 EC_GROUP_set_curve_name()
10624 EC_GROUP_get_curve_name()
10626 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10628 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10629 was actually never needed) and in BN_mul(). The removal in BN_mul()
10630 required a small change in bn_mul_part_recursive() and the addition
10631 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10632 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10633 bn_sub_words() and bn_add_words() except they take arrays with
10638 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10640 * Cleanse PEM buffers before freeing them since they may contain
10643 *Benjamin Bennett <ben@psc.edu>*
10645 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10646 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10647 authentication-only ciphersuites.
10651 * Since AES128 and AES256 share a single mask bit in the logic of
10652 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10653 kludge to work properly if AES128 is available and AES256 isn't.
10657 * Expand security boundary to match 1.1.1 module.
10661 * Remove redundant features: hash file source, editing of test vectors
10662 modify fipsld to use external fips_premain.c signature.
10666 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10667 run algorithm test programs.
10671 * Make algorithm test programs more tolerant of whitespace.
10675 * Have SSL/TLS server implementation tolerate "mismatched" record
10676 protocol version while receiving ClientHello even if the
10677 ClientHello is fragmented. (The server can't insist on the
10678 particular protocol version it has chosen before the ServerHello
10679 message has informed the client about his choice.)
10683 * Load error codes if they are not already present instead of using a
10684 static variable. This allows them to be cleanly unloaded and reloaded.
10688 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10690 * Introduce limits to prevent malicious keys being able to
10691 cause a denial of service. ([CVE-2006-2940])
10693 *Steve Henson, Bodo Moeller*
10695 * Fix ASN.1 parsing of certain invalid structures that can result
10696 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10698 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10699 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10701 * Fix SSL client code which could crash if connecting to a
10702 malicious SSLv2 server. ([CVE-2006-4343])
10704 *Tavis Ormandy and Will Drewry, Google Security Team*
10706 * Change ciphersuite string processing so that an explicit
10707 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10708 will no longer include "AES128-SHA"), and any other similar
10709 ciphersuite (same bitmap) from *other* protocol versions (so that
10710 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10711 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10712 changes from 0.9.8b and 0.9.8d.
10716 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10718 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10719 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10721 * Change the Unix randomness entropy gathering to use poll() when
10722 possible instead of select(), since the latter has some
10723 undesirable limitations.
10725 *Darryl Miles via Richard Levitte and Bodo Moeller*
10727 * Disable rogue ciphersuites:
10729 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10730 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10731 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10733 The latter two were purportedly from
10734 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10737 Also deactivate the remaining ciphersuites from
10738 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10739 unofficial, and the ID has long expired.
10743 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10744 dual-core machines) and other potential thread-safety issues.
10748 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10750 * Adapt fipsld and the build system to link against the validated FIPS
10751 module in FIPS mode.
10755 * Fixes for VC++ 2005 build under Windows.
10759 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10760 from a Windows bash shell such as MSYS. It is autodetected from the
10761 "config" script when run from a VC++ environment. Modify standard VC++
10762 build to use fipscanister.o from the GNU make build.
10766 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10768 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10769 The value now differs depending on if you build for FIPS or not.
10770 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10771 safely run with a non-FIPSed libcrypto, as it may crash because of
10772 the difference induced by this change.
10776 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10778 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10779 (part of SSL_OP_ALL). This option used to disable the
10780 countermeasure against man-in-the-middle protocol-version
10781 rollback in the SSL 2.0 server implementation, which is a bad
10782 idea. ([CVE-2005-2969])
10784 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10785 for Information Security, National Institute of Advanced Industrial
10786 Science and Technology [AIST, Japan)]*
10788 * Minimal support for X9.31 signatures and PSS padding modes. This is
10789 mainly for FIPS compliance and not fully integrated at this stage.
10793 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10794 the exponentiation using a fixed-length exponent. (Otherwise,
10795 the information leaked through timing could expose the secret key
10796 after many signatures; cf. Bleichenbacher's attack on DSA with
10801 * Make a new fixed-window mod_exp implementation the default for
10802 RSA, DSA, and DH private-key operations so that the sequence of
10803 squares and multiplies and the memory access pattern are
10804 independent of the particular secret key. This will mitigate
10805 cache-timing and potential related attacks.
10807 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10808 and this is automatically used by BN_mod_exp_mont() if the new flag
10809 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10810 will use this BN flag for private exponents unless the flag
10811 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10812 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10814 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10816 * Change the client implementation for SSLv23_method() and
10817 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10818 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10819 (Previously, the SSL 2.0 backwards compatible Client Hello
10820 message format would be used even with SSL_OP_NO_SSLv2.)
10824 * Add support for smime-type MIME parameter in S/MIME messages which some
10829 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10830 a threadsafe manner. Modify rsa code to use new function and add calls
10831 to dsa and dh code (which had race conditions before).
10835 * Include the fixed error library code in the C error file definitions
10836 instead of fixing them up at runtime. This keeps the error code
10837 structures constant.
10841 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10843 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10846 * Fixes for newer kerberos headers. NB: the casts are needed because
10847 the 'length' field is signed on one version and unsigned on another
10848 with no (?) obvious way to tell the difference, without these VC++
10849 complains. Also the "definition" of FAR (blank) is no longer included
10850 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10851 some needed definitions.
10855 * Undo Cygwin change.
10859 * Added support for proxy certificates according to RFC 3820.
10860 Because they may be a security thread to unaware applications,
10861 they must be explicitly allowed in run-time. See
10862 docs/HOWTO/proxy_certificates.txt for further information.
10866 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10868 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10869 server and client random values. Previously
10870 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10871 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10873 This change has negligible security impact because:
10875 1. Server and client random values still have 24 bytes of pseudo random
10878 2. Server and client random values are sent in the clear in the initial
10881 3. The master secret is derived using the premaster secret (48 bytes in
10882 size for static RSA ciphersuites) as well as client server and random
10885 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10888 *Stephen Henson, reported by UK NISCC*
10890 * Use Windows randomness collection on Cygwin.
10894 * Fix hang in EGD/PRNGD query when communication socket is closed
10895 prematurely by EGD/PRNGD.
10897 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10899 * Prompt for pass phrases when appropriate for PKCS12 input format.
10903 * Back-port of selected performance improvements from development
10904 branch, as well as improved support for PowerPC platforms.
10908 * Add lots of checks for memory allocation failure, error codes to indicate
10909 failure and freeing up memory if a failure occurs.
10911 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10913 * Add new -passin argument to dgst.
10917 * Perform some character comparisons of different types in X509_NAME_cmp:
10918 this is needed for some certificates that re-encode DNs into UTF8Strings
10919 (in violation of RFC3280) and can't or won't issue name rollover
10924 * Make an explicit check during certificate validation to see that
10925 the CA setting in each certificate on the chain is correct. As a
10926 side effect always do the following basic checks on extensions,
10927 not just when there's an associated purpose to the check:
10929 - if there is an unhandled critical extension (unless the user
10930 has chosen to ignore this fault)
10931 - if the path length has been exceeded (if one is set at all)
10932 - that certain extensions fit the associated purpose (if one has
10937 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10939 * Avoid a race condition when CRLs are checked in a multi threaded
10940 environment. This would happen due to the reordering of the revoked
10941 entries during signature checking and serial number lookup. Now the
10942 encoding is cached and the serial number sort performed under a lock.
10943 Add new STACK function sk_is_sorted().
10947 * Add Delta CRL to the extension code.
10951 * Various fixes to s3_pkt.c so alerts are sent properly.
10953 *David Holmes <d.holmes@f5.com>*
10955 * Reduce the chances of duplicate issuer name and serial numbers (in
10956 violation of RFC3280) using the OpenSSL certificate creation utilities.
10957 This is done by creating a random 64 bit value for the initial serial
10958 number when a serial number file is created or when a self signed
10959 certificate is created using 'openssl req -x509'. The initial serial
10960 number file is created using 'openssl x509 -next_serial' in CA.pl
10961 rather than being initialized to 1.
10965 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10967 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10968 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
10970 *Joe Orton, Steve Henson*
10972 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10975 *Joe Orton, Steve Henson*
10977 * Make it possible to have multiple active certificates with the same
10978 subject in the CA index file. This is done only if the keyword
10979 'unique_subject' is set to 'no' in the main CA section (default
10980 if 'CA_default') of the configuration file. The value is saved
10981 with the database itself in a separate index attribute file,
10982 named like the index file with '.attr' appended to the name.
10986 * X509 verify fixes. Disable broken certificate workarounds when
10987 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10988 keyUsage extension present. Don't accept CRLs with unhandled critical
10989 extensions: since verify currently doesn't process CRL extensions this
10990 rejects a CRL with *any* critical extensions. Add new verify error codes
10995 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10996 A clarification of RFC2560 will require the use of OCTET STRINGs and
10997 some implementations cannot handle the current raw format. Since OpenSSL
10998 copies and compares OCSP nonces as opaque blobs without any attempt at
10999 parsing them this should not create any compatibility issues.
11003 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11004 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11005 this HMAC (and other) operations are several times slower than OpenSSL
11010 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11012 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11014 * Use the correct content when signing type "other".
11018 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11020 * Fix various bugs revealed by running the NISCC test suite:
11022 Stop out of bounds reads in the ASN1 code when presented with
11023 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11025 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11027 If verify callback ignores invalid public key errors don't try to check
11028 certificate signature with the NULL public key.
11032 * New -ignore_err option in ocsp application to stop the server
11033 exiting on the first error in a request.
11037 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11038 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11043 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11044 extra data after the compression methods not only for TLS 1.0
11045 but also for SSL 3.0 (as required by the specification).
11047 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11049 * Change X509_certificate_type() to mark the key as exported/exportable
11050 when it's 512 *bits* long, not 512 bytes.
11054 * Change AES_cbc_encrypt() so it outputs exact multiple of
11055 blocks during encryption.
11059 * Various fixes to base64 BIO and non blocking I/O. On write
11060 flushes were not handled properly if the BIO retried. On read
11061 data was not being buffered properly and had various logic bugs.
11062 This also affects blocking I/O when the data being decoded is a
11067 * Various S/MIME bugfixes and compatibility changes:
11068 output correct application/pkcs7 MIME type if
11069 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11070 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11071 of files as .eml work). Correctly handle very long lines in MIME
11076 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11078 * Countermeasure against the Klima-Pokorny-Rosa extension of
11079 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11080 a protocol version number mismatch like a decryption error
11081 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11085 * Turn on RSA blinding by default in the default implementation
11086 to avoid a timing attack. Applications that don't want it can call
11087 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11088 They would be ill-advised to do so in most cases.
11090 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11092 * Change RSA blinding code so that it works when the PRNG is not
11093 seeded (in this case, the secret RSA exponent is abused as
11094 an unpredictable seed -- if it is not unpredictable, there
11095 is no point in blinding anyway). Make RSA blinding thread-safe
11096 by remembering the creator's thread ID in rsa->blinding and
11097 having all other threads use local one-time blinding factors
11098 (this requires more computation than sharing rsa->blinding, but
11099 avoids excessive locking; and if an RSA object is not shared
11100 between threads, blinding will still be very fast).
11104 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11105 ENGINE as defaults for all supported algorithms irrespective of
11106 the 'flags' parameter. 'flags' is now honoured, so applications
11107 should make sure they are passing it correctly.
11111 * Target "mingw" now allows native Windows code to be generated in
11112 the Cygwin environment as well as with the MinGW compiler.
11116 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11118 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11119 via timing by performing a MAC computation even if incorrect
11120 block cipher padding has been found. This is a countermeasure
11121 against active attacks where the attacker has to distinguish
11122 between bad padding and a MAC verification error. ([CVE-2003-0078])
11124 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11125 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11126 Martin Vuagnoux (EPFL, Ilion)*
11128 * Make the no-err option work as intended. The intention with no-err
11129 is not to have the whole error stack handling routines removed from
11130 libcrypto, it's only intended to remove all the function name and
11131 reason texts, thereby removing some of the footprint that may not
11132 be interesting if those errors aren't displayed anyway.
11134 NOTE: it's still possible for any application or module to have its
11135 own set of error texts inserted. The routines are there, just not
11136 used by default when no-err is given.
11140 * Add support for FreeBSD on IA64.
11142 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11144 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11145 Kerberos function mit_des_cbc_cksum(). Before this change,
11146 the value returned by DES_cbc_cksum() was like the one from
11147 mit_des_cbc_cksum(), except the bytes were swapped.
11149 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11151 * Allow an application to disable the automatic SSL chain building.
11152 Before this a rather primitive chain build was always performed in
11153 ssl3_output_cert_chain(): an application had no way to send the
11154 correct chain if the automatic operation produced an incorrect result.
11156 Now the chain builder is disabled if either:
11158 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11160 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11162 The reasoning behind this is that an application would not want the
11163 auto chain building to take place if extra chain certificates are
11164 present and it might also want a means of sending no additional
11165 certificates (for example the chain has two certificates and the
11170 * Add the possibility to build without the ENGINE framework.
11172 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11174 * Under Win32 gmtime() can return NULL: check return value in
11175 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11179 * DSA routines: under certain error conditions uninitialized BN objects
11180 could be freed. Solution: make sure initialization is performed early
11181 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11182 Nils Larsch <nla@trustcenter.de> via PR#459)
11186 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11187 checked on reconnect on the client side, therefore session resumption
11188 could still fail with a "ssl session id is different" error. This
11189 behaviour is masked when SSL_OP_ALL is used due to
11190 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11191 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11192 followup to PR #377.
11196 * IA-32 assembler support enhancements: unified ELF targets, support
11197 for SCO/Caldera platforms, fix for Cygwin shared build.
11201 * Add support for FreeBSD on sparc64. As a consequence, support for
11202 FreeBSD on non-x86 processors is separate from x86 processors on
11203 the config script, much like the NetBSD support.
11205 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11207 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11209 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11212 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11213 code (06) was taken as the first octet of the session ID and the last
11214 octet was ignored consequently. As a result SSLv2 client side session
11215 caching could not have worked due to the session ID mismatch between
11217 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11222 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11223 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11228 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11229 seems that in spite of existing for more than a year, many application
11230 author have done nothing to provide the necessary callbacks, which
11231 means that this particular engine will not work properly anywhere.
11232 This is a very unfortunate situation which forces us, in the name
11233 of usability, to give the hw_ncipher.c a static lock, which is part
11235 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11236 appear in 0.9.8 or later. We EXPECT application authors to have
11237 dealt properly with this when 0.9.8 is released (unless we actually
11238 make such changes in the libcrypto locking code that changes will
11239 have to be made anyway).
11243 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11244 octets have been read, EOF or an error occurs. Without this change
11245 some truncated ASN1 structures will not produce an error.
11249 * Disable Heimdal support, since it hasn't been fully implemented.
11250 Still give the possibility to force the use of Heimdal, but with
11251 warnings and a request that patches get sent to openssl-dev.
11255 * Add the VC-CE target, introduce the WINCE sysname, and add
11256 INSTALL.WCE and appropriate conditionals to make it build.
11258 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11260 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11261 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11262 edit numbers of the version.
11264 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11266 * Introduce safe string copy and catenation functions
11267 (BUF_strlcpy() and BUF_strlcat()).
11269 *Ben Laurie (CHATS) and Richard Levitte*
11271 * Avoid using fixed-size buffers for one-line DNs.
11273 *Ben Laurie (CHATS)*
11275 * Add BUF_MEM_grow_clean() to avoid information leakage when
11276 resizing buffers containing secrets, and use where appropriate.
11278 *Ben Laurie (CHATS)*
11280 * Avoid using fixed size buffers for configuration file location.
11282 *Ben Laurie (CHATS)*
11284 * Avoid filename truncation for various CA files.
11286 *Ben Laurie (CHATS)*
11288 * Use sizeof in preference to magic numbers.
11290 *Ben Laurie (CHATS)*
11292 * Avoid filename truncation in cert requests.
11294 *Ben Laurie (CHATS)*
11296 * Add assertions to check for (supposedly impossible) buffer
11299 *Ben Laurie (CHATS)*
11301 * Don't cache truncated DNS entries in the local cache (this could
11302 potentially lead to a spoofing attack).
11304 *Ben Laurie (CHATS)*
11306 * Fix various buffers to be large enough for hex/decimal
11307 representations in a platform independent manner.
11309 *Ben Laurie (CHATS)*
11311 * Add CRYPTO_realloc_clean() to avoid information leakage when
11312 resizing buffers containing secrets, and use where appropriate.
11314 *Ben Laurie (CHATS)*
11316 * Add BIO_indent() to avoid much slightly worrying code to do
11319 *Ben Laurie (CHATS)*
11321 * Convert sprintf()/BIO_puts() to BIO_printf().
11323 *Ben Laurie (CHATS)*
11325 * buffer_gets() could terminate with the buffer only half
11328 *Ben Laurie (CHATS)*
11330 * Add assertions to prevent user-supplied crypto functions from
11331 overflowing internal buffers by having large block sizes, etc.
11333 *Ben Laurie (CHATS)*
11335 * New OPENSSL_assert() macro (similar to assert(), but enabled
11338 *Ben Laurie (CHATS)*
11340 * Eliminate unused copy of key in RC4.
11342 *Ben Laurie (CHATS)*
11344 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11346 *Ben Laurie (CHATS)*
11348 * Fix off-by-one error in EGD path.
11350 *Ben Laurie (CHATS)*
11352 * If RANDFILE path is too long, ignore instead of truncating.
11354 *Ben Laurie (CHATS)*
11356 * Eliminate unused and incorrectly sized X.509 structure
11359 *Ben Laurie (CHATS)*
11361 * Eliminate unused and dangerous function knumber().
11363 *Ben Laurie (CHATS)*
11365 * Eliminate unused and dangerous structure, KSSL_ERR.
11367 *Ben Laurie (CHATS)*
11369 * Protect against overlong session ID context length in an encoded
11370 session object. Since these are local, this does not appear to be
11373 *Ben Laurie (CHATS)*
11375 * Change from security patch (see 0.9.6e below) that did not affect
11376 the 0.9.6 release series:
11378 Remote buffer overflow in SSL3 protocol - an attacker could
11379 supply an oversized master key in Kerberos-enabled versions.
11382 *Ben Laurie (CHATS)*
11384 * Change the SSL kerb5 codes to match RFC 2712.
11388 * Make -nameopt work fully for req and add -reqopt switch.
11390 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11392 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11394 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11396 * Make sure tests can be performed even if the corresponding algorithms
11397 have been removed entirely. This was also the last step to make
11398 OpenSSL compilable with DJGPP under all reasonable conditions.
11400 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11402 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11403 to allow version independent disabling of normally unselected ciphers,
11404 which may be activated as a side-effect of selecting a single cipher.
11406 (E.g., cipher list string "RSA" enables ciphersuites that are left
11407 out of "ALL" because they do not provide symmetric encryption.
11408 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11410 *Lutz Jaenicke, Bodo Moeller*
11412 * Add appropriate support for separate platform-dependent build
11413 directories. The recommended way to make a platform-dependent
11414 build directory is the following (tested on Linux), maybe with
11417 # Place yourself outside of the OpenSSL source tree. In
11418 # this example, the environment variable OPENSSL_SOURCE
11419 # is assumed to contain the absolute OpenSSL source directory.
11420 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11421 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11422 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11423 mkdir -p `dirname $F`
11424 ln -s $OPENSSL_SOURCE/$F $F
11427 To be absolutely sure not to disturb the source tree, a "make clean"
11428 is a good thing. If it isn't successful, don't worry about it,
11429 it probably means the source directory is very clean.
11433 * Make sure any ENGINE control commands make local copies of string
11434 pointers passed to them whenever necessary. Otherwise it is possible
11435 the caller may have overwritten (or deallocated) the original string
11436 data when a later ENGINE operation tries to use the stored values.
11438 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11440 * Improve diagnostics in file reading and command-line digests.
11442 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11444 * Add AES modes CFB and OFB to the object database. Correct an
11445 error in AES-CFB decryption.
11449 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11450 allows existing EVP_CIPHER_CTX structures to be reused after
11451 calling `EVP_*Final()`. This behaviour is used by encryption
11452 BIOs and some applications. This has the side effect that
11453 applications must explicitly clean up cipher contexts with
11454 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11458 * Check the values of dna and dnb in bn_mul_recursive before calling
11459 bn_mul_comba (a non zero value means the a or b arrays do not contain
11460 n2 elements) and fallback to bn_mul_normal if either is not zero.
11464 * Fix escaping of non-ASCII characters when using the -subj option
11465 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11469 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11470 form for "surname", serialNumber has no short form.
11471 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11472 therefore remove "mail" short name for "internet 7".
11473 The OID for unique identifiers in X509 certificates is
11474 x500UniqueIdentifier, not uniqueIdentifier.
11475 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11479 * Add an "init" command to the ENGINE config module and auto initialize
11480 ENGINEs. Without any "init" command the ENGINE will be initialized
11481 after all ctrl commands have been executed on it. If init=1 the
11482 ENGINE is initialized at that point (ctrls before that point are run
11483 on the uninitialized ENGINE and after on the initialized one). If
11484 init=0 then the ENGINE will not be initialized at all.
11488 * Fix the 'app_verify_callback' interface so that the user-defined
11489 argument is actually passed to the callback: In the
11490 SSL_CTX_set_cert_verify_callback() prototype, the callback
11491 declaration has been changed from
11494 int (*cb)(X509_STORE_CTX *,void *);
11495 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11496 i=s->ctx->app_verify_callback(&ctx)
11497 has been changed into
11498 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11500 To update applications using SSL_CTX_set_cert_verify_callback(),
11501 a dummy argument can be added to their callback functions.
11503 *D. K. Smetters <smetters@parc.xerox.com>*
11505 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11507 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11509 * Add and OPENSSL_LOAD_CONF define which will cause
11510 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11511 This allows older applications to transparently support certain
11512 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11513 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11514 load the config file and OPENSSL_add_all_algorithms_conf() which will
11515 always load it have also been added.
11519 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11520 Adjust NIDs and EVP layer.
11522 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11524 * Config modules support in openssl utility.
11526 Most commands now load modules from the config file,
11527 though in a few (such as version) this isn't done
11528 because it couldn't be used for anything.
11530 In the case of ca and req the config file used is
11531 the same as the utility itself: that is the -config
11532 command line option can be used to specify an
11537 * Move default behaviour from OPENSSL_config(). If appname is NULL
11538 use "openssl_conf" if filename is NULL use default openssl config file.
11542 * Add an argument to OPENSSL_config() to allow the use of an alternative
11543 config section name. Add a new flag to tolerate a missing config file
11544 and move code to CONF_modules_load_file().
11548 * Support for crypto accelerator cards from Accelerated Encryption
11549 Processing, www.aep.ie. (Use engine 'aep')
11550 The support was copied from 0.9.6c [engine] and adapted/corrected
11551 to work with the new engine framework.
11553 *AEP Inc. and Richard Levitte*
11555 * Support for SureWare crypto accelerator cards from Baltimore
11556 Technologies. (Use engine 'sureware')
11557 The support was copied from 0.9.6c [engine] and adapted
11558 to work with the new engine framework.
11562 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11563 make the newer ENGINE framework commands for the CHIL engine work.
11565 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11567 * Make it possible to produce shared libraries on ReliantUNIX.
11569 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11571 * Add the configuration target debug-linux-ppro.
11572 Make 'openssl rsa' use the general key loading routines
11573 implemented in apps.c, and make those routines able to
11574 handle the key format FORMAT_NETSCAPE and the variant
11577 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11579 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11581 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11583 * Add -keyform to rsautl, and document -engine.
11585 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11587 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11588 BIO_R_NO_SUCH_FILE error code rather than the generic
11589 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11593 * Add new functions
11594 ERR_peek_last_error
11595 ERR_peek_last_error_line
11596 ERR_peek_last_error_line_data.
11597 These are similar to
11599 ERR_peek_error_line
11600 ERR_peek_error_line_data,
11601 but report on the latest error recorded rather than the first one
11602 still in the error queue.
11604 *Ben Laurie, Bodo Moeller*
11606 * default_algorithms option in ENGINE config module. This allows things
11608 default_algorithms = ALL
11609 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11613 * Preliminary ENGINE config module.
11617 * New experimental application configuration code.
11621 * Change the AES code to follow the same name structure as all other
11622 symmetric ciphers, and behave the same way. Move everything to
11623 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11625 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11627 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11629 *Ben Laurie and Theo de Raadt*
11631 * Add option to output public keys in req command.
11633 *Massimiliano Pala madwolf@openca.org*
11635 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11636 (up to about 10% better than before for P-192 and P-224).
11640 * New functions/macros
11642 SSL_CTX_set_msg_callback(ctx, cb)
11643 SSL_CTX_set_msg_callback_arg(ctx, arg)
11644 SSL_set_msg_callback(ssl, cb)
11645 SSL_set_msg_callback_arg(ssl, arg)
11647 to request calling a callback function
11649 void cb(int write_p, int version, int content_type,
11650 const void *buf, size_t len, SSL *ssl, void *arg)
11652 whenever a protocol message has been completely received
11653 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11654 protocol version according to which the SSL library interprets
11655 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11656 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11657 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11658 specification (change_cipher_spec(20), alert(21), handshake(22)).
11659 'buf' and 'len' point to the actual message, 'ssl' to the
11660 SSL object, and 'arg' is the application-defined value set by
11661 SSL[_CTX]_set_msg_callback_arg().
11663 'openssl s_client' and 'openssl s_server' have new '-msg' options
11664 to enable a callback that displays all protocol messages.
11668 * Change the shared library support so shared libraries are built as
11669 soon as the corresponding static library is finished, and thereby get
11670 openssl and the test programs linked against the shared library.
11671 This still only happens when the keyword "shard" has been given to
11672 the configuration scripts.
11674 NOTE: shared library support is still an experimental thing, and
11675 backward binary compatibility is still not guaranteed.
11677 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11679 * Add support for Subject Information Access extension.
11681 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11683 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11684 additional bytes when new memory had to be allocated, not just
11685 when reusing an existing buffer.
11689 * New command line and configuration option 'utf8' for the req command.
11690 This allows field values to be specified as UTF8 strings.
11694 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11695 runs for the former and machine-readable output for the latter.
11699 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11700 of the e-mail address in the DN (i.e., it will go into a certificate
11701 extension only). The new configuration file option 'email_in_dn = no'
11702 has the same effect.
11704 *Massimiliano Pala madwolf@openca.org*
11706 * Change all functions with names starting with `des_` to be starting
11707 with `DES_` instead. Add wrappers that are compatible with libdes,
11708 but are named `_ossl_old_des_*`. Finally, add macros that map the
11709 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11710 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11711 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11714 Since we provide two compatibility mappings, the user needs to
11715 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11716 compatibility is desired. The default (i.e., when that macro
11717 isn't defined) is OpenSSL 0.9.6c compatibility.
11719 There are also macros that enable and disable the support of old
11720 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11721 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11722 are defined, the default will apply: to support the old des routines.
11724 In either case, one must include openssl/des.h to get the correct
11725 definitions. Do not try to just include openssl/des_old.h, that
11728 NOTE: This is a major break of an old API into a new one. Software
11729 authors are encouraged to switch to the `DES_` style functions. Some
11730 time in the future, des_old.h and the libdes compatibility functions
11731 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11732 default), and then completely removed.
11736 * Test for certificates which contain unsupported critical extensions.
11737 If such a certificate is found during a verify operation it is
11738 rejected by default: this behaviour can be overridden by either
11739 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11740 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11741 X509_supported_extension() has also been added which returns 1 if a
11742 particular extension is supported.
11746 * Modify the behaviour of EVP cipher functions in similar way to digests
11747 to retain compatibility with existing code.
11751 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11752 compatibility with existing code. In particular the 'ctx' parameter does
11753 not have to be to be initialized before the call to EVP_DigestInit() and
11754 it is tidied up after a call to EVP_DigestFinal(). New function
11755 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11756 EVP_MD_CTX_copy() changed to not require the destination to be
11757 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11758 requires the destination to be valid.
11760 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11761 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11765 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11766 so that complete 'Handshake' protocol structures are kept in memory
11767 instead of overwriting 'msg_type' and 'length' with 'body' data.
11771 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11773 *Massimo Santin via Richard Levitte*
11775 * Major restructuring to the underlying ENGINE code. This includes
11776 reduction of linker bloat, separation of pure "ENGINE" manipulation
11777 (initialisation, etc) from functionality dealing with implementations
11778 of specific crypto interfaces. This change also introduces integrated
11779 support for symmetric ciphers and digest implementations - so ENGINEs
11780 can now accelerate these by providing EVP_CIPHER and EVP_MD
11781 implementations of their own. This is detailed in
11782 [crypto/engine/README.md](crypto/engine/README.md)
11783 as it couldn't be adequately described here. However, there are a few
11784 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11785 were changed in the original introduction of ENGINE code have now
11786 reverted back - the hooking from this code to ENGINE is now a good
11787 deal more passive and at run-time, operations deal directly with
11788 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11789 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11790 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11791 they were not being used by the framework as there is no concept of a
11792 BIGNUM_METHOD and they could not be generalised to the new
11793 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11794 ENGINE_cpy() has been removed as it cannot be consistently defined in
11799 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11803 * Change mkdef.pl to sort symbols that get the same entry number,
11804 and make sure the automatically generated functions `ERR_load_*`
11805 become part of libeay.num as well.
11809 * New function SSL_renegotiate_pending(). This returns true once
11810 renegotiation has been requested (either SSL_renegotiate() call
11811 or HelloRequest/ClientHello received from the peer) and becomes
11812 false once a handshake has been completed.
11813 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11814 sends a HelloRequest, but does not ensure that a handshake takes
11815 place. SSL_renegotiate_pending() is useful for checking if the
11816 client has followed the request.)
11820 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11821 By default, clients may request session resumption even during
11822 renegotiation (if session ID contexts permit); with this option,
11823 session resumption is possible only in the first handshake.
11825 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11826 more bits available for options that should not be part of
11827 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11831 * Add some demos for certificate and certificate request creation.
11835 * Make maximum certificate chain size accepted from the peer application
11836 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11837 "Douglas E. Engert" <deengert@anl.gov>.
11841 * Add support for shared libraries for Unixware-7
11842 (Boyd Lynn Gerber <gerberb@zenez.com>).
11846 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11847 be done prior to destruction. Use this to unload error strings from
11848 ENGINEs that load their own error strings. NB: This adds two new API
11849 functions to "get" and "set" this destroy handler in an ENGINE.
11853 * Alter all existing ENGINE implementations (except "openssl" and
11854 "openbsd") to dynamically instantiate their own error strings. This
11855 makes them more flexible to be built both as statically-linked ENGINEs
11856 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11857 Also, add stub code to each that makes building them as self-contained
11858 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11862 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11863 implementations into applications that are completely implemented in
11864 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11865 commands that can be used to configure what shared-library to load and
11866 to control aspects of the way it is handled. Also, made an update to
11867 the [README-Engine.md](README-Engine.md) file
11868 that brings its information up-to-date and
11869 provides some information and instructions on the "dynamic" ENGINE
11870 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11874 * Make it possible to unload ranges of ERR strings with a new
11875 "ERR_unload_strings" function.
11879 * Add a copy() function to EVP_MD.
11883 * Make EVP_MD routines take a context pointer instead of just the
11884 md_data void pointer.
11888 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11889 that the digest can only process a single chunk of data
11890 (typically because it is provided by a piece of
11891 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11892 is only going to provide a single chunk of data, and hence the
11893 framework needn't accumulate the data for oneshot drivers.
11897 * As with "ERR", make it possible to replace the underlying "ex_data"
11898 functions. This change also alters the storage and management of global
11899 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11900 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11901 index counters. The API functions that use this state have been changed
11902 to take a "class_index" rather than pointers to the class's local STACK
11903 and counter, and there is now an API function to dynamically create new
11904 classes. This centralisation allows us to (a) plug a lot of the
11905 thread-safety problems that existed, and (b) makes it possible to clean
11906 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11907 such data would previously have always leaked in application code and
11908 workarounds were in place to make the memory debugging turn a blind eye
11909 to it. Application code that doesn't use this new function will still
11910 leak as before, but their memory debugging output will announce it now
11911 rather than letting it slide.
11913 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11914 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11915 has a return value to indicate success or failure.
11919 * Make it possible to replace the underlying "ERR" functions such that the
11920 global state (2 LHASH tables and 2 locks) is only used by the "default"
11921 implementation. This change also adds two functions to "get" and "set"
11922 the implementation prior to it being automatically set the first time
11923 any other ERR function takes place. Ie. an application can call "get",
11924 pass the return value to a module it has just loaded, and that module
11925 can call its own "set" function using that value. This means the
11926 module's "ERR" operations will use (and modify) the error state in the
11927 application and not in its own statically linked copy of OpenSSL code.
11931 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11932 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11933 the operation, and provides a more encapsulated way for external code
11934 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11935 to use these functions rather than manually incrementing the counts.
11937 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11941 * Add EVP test program.
11945 * Add symmetric cipher support to ENGINE. Expect the API to change!
11949 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11950 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11951 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11952 These allow a CRL to be built without having to access X509_CRL fields
11953 directly. Modify 'ca' application to use new functions.
11957 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11958 bug workarounds. Rollback attack detection is a security feature.
11959 The problem will only arise on OpenSSL servers when TLSv1 is not
11960 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11961 Software authors not wanting to support TLSv1 will have special reasons
11962 for their choice and can explicitly enable this option.
11964 *Bodo Moeller, Lutz Jaenicke*
11966 * Rationalise EVP so it can be extended: don't include a union of
11967 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11968 (similar to those existing for EVP_CIPHER_CTX).
11973 EVP_MD_CTX_init(&md); /* new function call */
11974 EVP_DigestInit(&md, EVP_sha1());
11975 EVP_DigestUpdate(&md, in, len);
11976 EVP_DigestFinal(&md, out, NULL);
11977 EVP_MD_CTX_cleanup(&md); /* new function call */
11981 * Make DES key schedule conform to the usual scheme, as well as
11982 correcting its structure. This means that calls to DES functions
11983 now have to pass a pointer to a des_key_schedule instead of a
11984 plain des_key_schedule (which was actually always a pointer
11987 des_key_schedule ks;
11989 des_set_key_checked(..., &ks);
11990 des_ncbc_encrypt(..., &ks, ...);
11992 (Note that a later change renames 'des_...' into 'DES_...'.)
11996 * Initial reduction of linker bloat: the use of some functions, such as
11997 PEM causes large amounts of unused functions to be linked in due to
11998 poor organisation. For example pem_all.c contains every PEM function
11999 which has a knock on effect of linking in large amounts of (unused)
12000 ASN1 code. Grouping together similar functions and splitting unrelated
12001 functions prevents this.
12005 * Cleanup of EVP macros.
12009 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12010 correct `_ecb suffix`.
12014 * Add initial OCSP responder support to ocsp application. The
12015 revocation information is handled using the text based index
12016 use by the ca application. The responder can either handle
12017 requests generated internally, supplied in files (for example
12018 via a CGI script) or using an internal minimal server.
12022 * Add configuration choices to get zlib compression for TLS.
12026 * Changes to Kerberos SSL for RFC 2712 compliance:
12027 1. Implemented real KerberosWrapper, instead of just using
12028 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12029 2. Implemented optional authenticator field of KerberosWrapper.
12031 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12032 and authenticator structs; see crypto/krb5/.
12034 Generalized Kerberos calls to support multiple Kerberos libraries.
12035 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12036 via Richard Levitte*
12038 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12039 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12040 values for each of the key sizes rather than having just
12041 parameters (and 'speed' generating keys each time).
12045 * Speed up EVP routines.
12048 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12049 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12050 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12051 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12053 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12054 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12055 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12058 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12060 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12064 * Added the OS2-EMX target.
12066 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12068 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12069 to support NCONF routines in extension code. New function CONF_set_nconf()
12070 to allow functions which take an NCONF to also handle the old LHASH
12071 structure: this means that the old CONF compatible routines can be
12072 retained (in particular wrt extensions) without having to duplicate the
12073 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12077 * Enhance the general user interface with mechanisms for inner control
12078 and with possibilities to have yes/no kind of prompts.
12082 * Change all calls to low level digest routines in the library and
12083 applications to use EVP. Add missing calls to HMAC_cleanup() and
12084 don't assume HMAC_CTX can be copied using memcpy().
12086 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12088 * Add the possibility to control engines through control names but with
12089 arbitrary arguments instead of just a string.
12090 Change the key loaders to take a UI_METHOD instead of a callback
12091 function pointer. NOTE: this breaks binary compatibility with earlier
12092 versions of OpenSSL [engine].
12093 Adapt the nCipher code for these new conditions and add a card insertion
12098 * Enhance the general user interface with mechanisms to better support
12099 dialog box interfaces, application-defined prompts, the possibility
12100 to use defaults (for example default passwords from somewhere else)
12101 and interrupts/cancellations.
12105 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12106 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12110 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12111 tidy up some unnecessarily weird code in 'sk_new()').
12113 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12115 * Change the key loading routines for ENGINEs to use the same kind
12116 callback (pem_password_cb) as all other routines that need this
12121 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12122 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12123 than this minimum value is recommended.
12127 * New random seeder for OpenVMS, using the system process statistics
12128 that are easily reachable.
12132 * Windows apparently can't transparently handle global
12133 variables defined in DLLs. Initialisations such as:
12135 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12137 won't compile. This is used by the any applications that need to
12138 declare their own ASN1 modules. This was fixed by adding the option
12139 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12140 needed for static libraries under Win32.
12144 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12145 setting of purpose and trust fields. New X509_STORE trust and
12146 purpose functions and tidy up setting in other SSL functions.
12150 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12151 structure. These are inherited by X509_STORE_CTX when it is
12152 initialised. This allows various defaults to be set in the
12153 X509_STORE structure (such as flags for CRL checking and custom
12154 purpose or trust settings) for functions which only use X509_STORE_CTX
12155 internally such as S/MIME.
12157 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12158 trust settings if they are not set in X509_STORE. This allows X509_STORE
12159 purposes and trust (in S/MIME for example) to override any set by default.
12161 Add command line options for CRL checking to smime, s_client and s_server
12166 * Initial CRL based revocation checking. If the CRL checking flag(s)
12167 are set then the CRL is looked up in the X509_STORE structure and
12168 its validity and signature checked, then if the certificate is found
12169 in the CRL the verify fails with a revoked error.
12171 Various new CRL related callbacks added to X509_STORE_CTX structure.
12173 Command line options added to 'verify' application to support this.
12175 This needs some additional work, such as being able to handle multiple
12176 CRLs with different times, extension based lookup (rather than just
12177 by subject name) and ultimately more complete V2 CRL extension
12182 * Add a general user interface API (crypto/ui/). This is designed
12183 to replace things like des_read_password and friends (backward
12184 compatibility functions using this new API are provided).
12185 The purpose is to remove prompting functions from the DES code
12186 section as well as provide for prompting through dialog boxes in
12187 a window system and the like.
12191 * Add "ex_data" support to ENGINE so implementations can add state at a
12192 per-structure level rather than having to store it globally.
12196 * Make it possible for ENGINE structures to be copied when retrieved by
12197 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12198 This causes the "original" ENGINE structure to act like a template,
12199 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12200 operational state can be localised to each ENGINE structure, despite the
12201 fact they all share the same "methods". New ENGINE structures returned in
12202 this case have no functional references and the return value is the single
12203 structural reference. This matches the single structural reference returned
12204 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12209 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12210 needs to match any other type at all we need to manually clear the
12215 * Changes to the "openssl engine" utility to include;
12216 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12217 about an ENGINE's available control commands.
12218 - executing control commands from command line arguments using the
12219 '-pre' and '-post' switches. '-post' is only used if '-t' is
12220 specified and the ENGINE is successfully initialised. The syntax for
12221 the individual commands are colon-separated, for example;
12222 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12226 * New dynamic control command support for ENGINEs. ENGINEs can now
12227 declare their own commands (numbers), names (strings), descriptions,
12228 and input types for run-time discovery by calling applications. A
12229 subset of these commands are implicitly classed as "executable"
12230 depending on their input type, and only these can be invoked through
12231 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12232 can be based on user input, config files, etc). The distinction is
12233 that "executable" commands cannot return anything other than a boolean
12234 result and can only support numeric or string input, whereas some
12235 discoverable commands may only be for direct use through
12236 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12237 pointers, or other custom uses. The "executable" commands are to
12238 support parameterisations of ENGINE behaviour that can be
12239 unambiguously defined by ENGINEs and used consistently across any
12240 OpenSSL-based application. Commands have been added to all the
12241 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12242 control over shared-library paths without source code alterations.
12246 * Changed all ENGINE implementations to dynamically allocate their
12247 ENGINEs rather than declaring them statically. Apart from this being
12248 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12249 this also allows the implementations to compile without using the
12250 internal engine_int.h header.
12254 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12255 'const' value. Any code that should be able to modify a RAND_METHOD
12256 should already have non-const pointers to it (ie. they should only
12257 modify their own ones).
12261 * Made a variety of little tweaks to the ENGINE code.
12262 - "atalla" and "ubsec" string definitions were moved from header files
12263 to C code. "nuron" string definitions were placed in variables
12264 rather than hard-coded - allowing parameterisation of these values
12265 later on via ctrl() commands.
12266 - Removed unused "#if 0"'d code.
12267 - Fixed engine list iteration code so it uses ENGINE_free() to release
12268 structural references.
12269 - Constified the RAND_METHOD element of ENGINE structures.
12270 - Constified various get/set functions as appropriate and added
12271 missing functions (including a catch-all ENGINE_cpy that duplicates
12272 all ENGINE values onto a new ENGINE except reference counts/state).
12273 - Removed NULL parameter checks in get/set functions. Setting a method
12274 or function to NULL is a way of cancelling out a previously set
12275 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12276 and doesn't justify the extra error symbols and code.
12277 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12278 flags from engine_int.h to engine.h.
12279 - Changed prototypes for ENGINE handler functions (init(), finish(),
12280 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12284 * Implement binary inversion algorithm for BN_mod_inverse in addition
12285 to the algorithm using long division. The binary algorithm can be
12286 used only if the modulus is odd. On 32-bit systems, it is faster
12287 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12288 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12289 up to 450 bits. In 64-bit environments, the binary algorithm
12290 appears to be advantageous for much longer moduli; here we use it
12291 for moduli up to 2048 bits.
12295 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12296 could not support the combine flag in choice fields.
12300 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12301 extensions from a certificate request to the certificate.
12305 * Allow multiple 'certopt' and 'nameopt' options to be separated
12306 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12307 file: this allows the display of the certificate about to be
12308 signed to be customised, to allow certain fields to be included
12309 or excluded and extension details. The old system didn't display
12310 multicharacter strings properly, omitted fields not in the policy
12311 and couldn't display additional details such as extensions.
12315 * Function EC_POINTs_mul for multiple scalar multiplication
12316 of an arbitrary number of elliptic curve points
12317 \sum scalars[i]*points[i],
12318 optionally including the generator defined for the EC_GROUP:
12319 scalar*generator + \sum scalars[i]*points[i].
12321 EC_POINT_mul is a simple wrapper function for the typical case
12322 that the point list has just one item (besides the optional
12327 * First EC_METHODs for curves over GF(p):
12329 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12330 operations and provides various method functions that can also
12331 operate with faster implementations of modular arithmetic.
12333 EC_GFp_mont_method() reuses most functions that are part of
12334 EC_GFp_simple_method, but uses Montgomery arithmetic.
12336 *Bodo Moeller; point addition and point doubling
12337 implementation directly derived from source code provided by
12338 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12340 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12341 crypto/ec/ec_lib.c):
12343 Curves are EC_GROUP objects (with an optional group generator)
12344 based on EC_METHODs that are built into the library.
12346 Points are EC_POINT objects based on EC_GROUP objects.
12348 Most of the framework would be able to handle curves over arbitrary
12349 finite fields, but as there are no obvious types for fields other
12350 than GF(p), some functions are limited to that for now.
12354 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12355 that the file contains a complete HTTP response.
12359 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12360 change the def and num file printf format specifier from "%-40sXXX"
12361 to "%-39s XXX". The latter will always guarantee a space after the
12362 field while the former will cause them to run together if the field
12363 is 40 of more characters long.
12367 * Constify the cipher and digest 'method' functions and structures
12368 and modify related functions to take constant EVP_MD and EVP_CIPHER
12373 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12374 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12378 * Modify `EVP_Digest*()` routines so they now return values. Although the
12379 internal software routines can never fail additional hardware versions
12384 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12386 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12387 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12390 ERR_R_NESTED_ASN1_ERROR
12392 ERR_R_MISSING_ASN1_EOS
12393 were 4 .. 9, conflicting with
12394 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12396 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12397 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12399 Add new error code 'ERR_R_INTERNAL_ERROR'.
12403 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12408 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12409 sets the subject name for a new request or supersedes the
12410 subject name in a given request. Formats that can be parsed are
12411 'CN=Some Name, OU=myOU, C=IT'
12413 'CN=Some Name/OU=myOU/C=IT'.
12415 Add options '-batch' and '-verbose' to 'openssl req'.
12417 *Massimiliano Pala <madwolf@hackmasters.net>*
12419 * Introduce the possibility to access global variables through
12420 functions on platform were that's the best way to handle exporting
12421 global variables in shared libraries. To enable this functionality,
12422 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12423 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12424 is normally done by Configure or something similar).
12426 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12427 in the source file (foo.c) like this:
12429 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12430 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12432 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12433 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12435 OPENSSL_DECLARE_GLOBAL(int,foo);
12436 #define foo OPENSSL_GLOBAL_REF(foo)
12437 OPENSSL_DECLARE_GLOBAL(double,bar);
12438 #define bar OPENSSL_GLOBAL_REF(bar)
12440 The #defines are very important, and therefore so is including the
12441 header file everywhere where the defined globals are used.
12443 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12444 of ASN.1 items, but that structure is a bit different.
12446 The largest change is in util/mkdef.pl which has been enhanced with
12447 better and easier to understand logic to choose which symbols should
12448 go into the Windows .def files as well as a number of fixes and code
12449 cleanup (among others, algorithm keywords are now sorted
12450 lexicographically to avoid constant rewrites).
12454 * In BN_div() keep a copy of the sign of 'num' before writing the
12455 result to 'rm' because if rm==num the value will be overwritten
12456 and produce the wrong result if 'num' is negative: this caused
12457 problems with BN_mod() and BN_nnmod().
12461 * Function OCSP_request_verify(). This checks the signature on an
12462 OCSP request and verifies the signer certificate. The signer
12463 certificate is just checked for a generic purpose and OCSP request
12468 * Add OCSP_check_validity() function to check the validity of OCSP
12469 responses. OCSP responses are prepared in real time and may only
12470 be a few seconds old. Simply checking that the current time lies
12471 between thisUpdate and nextUpdate max reject otherwise valid responses
12472 caused by either OCSP responder or client clock inaccuracy. Instead
12473 we allow thisUpdate and nextUpdate to fall within a certain period of
12474 the current time. The age of the response can also optionally be
12475 checked. Two new options -validity_period and -status_age added to
12480 * If signature or public key algorithm is unrecognized print out its
12481 OID rather that just UNKNOWN.
12485 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12486 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12487 ID to be generated from the issuer certificate alone which can then be
12488 passed to OCSP_id_issuer_cmp().
12492 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12493 ASN1 modules to export functions returning ASN1_ITEM pointers
12494 instead of the ASN1_ITEM structures themselves. This adds several
12495 new macros which allow the underlying ASN1 function/structure to
12496 be accessed transparently. As a result code should not use ASN1_ITEM
12497 references directly (such as &X509_it) but instead use the relevant
12498 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12499 use of the new ASN1 code on platforms where exporting structures
12500 is problematical (for example in shared libraries) but exporting
12501 functions returning pointers to structures is not.
12505 * Add support for overriding the generation of SSL/TLS session IDs.
12506 These callbacks can be registered either in an SSL_CTX or per SSL.
12507 The purpose of this is to allow applications to control, if they wish,
12508 the arbitrary values chosen for use as session IDs, particularly as it
12509 can be useful for session caching in multiple-server environments. A
12510 command-line switch for testing this (and any client code that wishes
12511 to use such a feature) has been added to "s_server".
12513 *Geoff Thorpe, Lutz Jaenicke*
12515 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12516 of the form `#if defined(...) || defined(...) || ...` and
12517 `#if !defined(...) && !defined(...) && ...`. This also avoids
12518 the growing number of special cases it was previously handling.
12522 * Make all configuration macros available for application by making
12523 sure they are available in opensslconf.h, by giving them names starting
12524 with `OPENSSL_` to avoid conflicts with other packages and by making
12525 sure e_os2.h will cover all platform-specific cases together with
12527 Additionally, it is now possible to define configuration/platform-
12528 specific names (called "system identities"). In the C code, these
12529 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12530 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12531 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12536 * New option -set_serial to 'req' and 'x509' this allows the serial
12537 number to use to be specified on the command line. Previously self
12538 signed certificates were hard coded with serial number 0 and the
12539 CA options of 'x509' had to use a serial number in a file which was
12544 * New options to 'ca' utility to support V2 CRL entry extensions.
12545 Currently CRL reason, invalidity date and hold instruction are
12546 supported. Add new CRL extensions to V3 code and some new objects.
12550 * New function EVP_CIPHER_CTX_set_padding() this is used to
12551 disable standard block padding (aka PKCS#5 padding) in the EVP
12552 API, which was previously mandatory. This means that the data is
12553 not padded in any way and so the total length much be a multiple
12554 of the block size, otherwise an error occurs.
12558 * Initial (incomplete) OCSP SSL support.
12562 * New function OCSP_parse_url(). This splits up a URL into its host,
12563 port and path components: primarily to parse OCSP URLs. New -url
12564 option to ocsp utility.
12568 * New nonce behavior. The return value of OCSP_check_nonce() now
12569 reflects the various checks performed. Applications can decide
12570 whether to tolerate certain situations such as an absent nonce
12571 in a response when one was present in a request: the ocsp application
12572 just prints out a warning. New function OCSP_add1_basic_nonce()
12573 this is to allow responders to include a nonce in a response even if
12574 the request is nonce-less.
12578 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12579 skipped when using openssl x509 multiple times on a single input file,
12580 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12584 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12585 set string type: to handle setting ASN1_TIME structures. Fix ca
12586 utility to correctly initialize revocation date of CRLs.
12590 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12591 the clients preferred ciphersuites and rather use its own preferences.
12592 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12593 Internet Explorer by ensuring unchanged hash method during stepup.
12594 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12598 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12599 to aes and add a new 'exist' option to print out symbols that don't
12604 * Additional options to ocsp utility to allow flags to be set and
12605 additional certificates supplied.
12609 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12610 OCSP client a number of certificate to only verify the response
12615 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12616 handle the new API. Currently only ECB, CBC modes supported. Add new
12619 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12620 Encryption Standard (AES) Ciphersuites for Transport Layer
12621 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12622 not enabled by default and were not part of the "ALL" ciphersuite
12623 alias because they were not yet official; they could be
12624 explicitly requested by specifying the "AESdraft" ciphersuite
12625 group alias. In the final release of OpenSSL 0.9.7, the group
12626 alias is called "AES" and is part of "ALL".)
12628 *Ben Laurie, Steve Henson, Bodo Moeller*
12630 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12631 request to response.
12635 * Functions for OCSP responders. OCSP_request_onereq_count(),
12636 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12637 extract information from a certificate request. OCSP_response_create()
12638 creates a response and optionally adds a basic response structure.
12639 OCSP_basic_add1_status() adds a complete single response to a basic
12640 response and returns the OCSP_SINGLERESP structure just added (to allow
12641 extensions to be included for example). OCSP_basic_add1_cert() adds a
12642 certificate to a basic response and OCSP_basic_sign() signs a basic
12643 response with various flags. New helper functions ASN1_TIME_check()
12644 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12645 (converts ASN1_TIME to GeneralizedTime).
12649 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12650 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12651 structure from a certificate. X509_pubkey_digest() digests the public_key
12652 contents: this is used in various key identifiers.
12656 * Make sk_sort() tolerate a NULL argument.
12658 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12660 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12661 passed by the function are trusted implicitly. If any of them signed the
12662 response then it is assumed to be valid and is not verified.
12666 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12667 to data. This was previously part of the PKCS7 ASN1 code. This
12668 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12669 *Steve Henson, reported by Kenneth R. Robinette
12670 <support@securenetterm.com>*
12672 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12673 routines: without these tracing memory leaks is very painful.
12674 Fix leaks in PKCS12 and PKCS7 routines.
12678 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12679 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12680 effectively meant GeneralizedTime would never be used. Now it
12681 is initialised to -1 but X509_time_adj() now has to check the value
12682 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12683 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12684 *Steve Henson, reported by Kenneth R. Robinette
12685 <support@securenetterm.com>*
12687 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12688 result in a zero length in the ASN1_INTEGER structure which was
12689 not consistent with the structure when d2i_ASN1_INTEGER() was used
12690 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12691 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12692 where it did not print out a minus for negative ASN1_INTEGER.
12696 * Add summary printout to ocsp utility. The various functions which
12697 convert status values to strings have been renamed to:
12698 OCSP_response_status_str(), OCSP_cert_status_str() and
12699 OCSP_crl_reason_str() and are no longer static. New options
12700 to verify nonce values and to disable verification. OCSP response
12701 printout format cleaned up.
12705 * Add additional OCSP certificate checks. These are those specified
12706 in RFC2560. This consists of two separate checks: the CA of the
12707 certificate being checked must either be the OCSP signer certificate
12708 or the issuer of the OCSP signer certificate. In the latter case the
12709 OCSP signer certificate must contain the OCSP signing extended key
12710 usage. This check is performed by attempting to match the OCSP
12711 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12712 in the OCSP_CERTID structures of the response.
12716 * Initial OCSP certificate verification added to OCSP_basic_verify()
12717 and related routines. This uses the standard OpenSSL certificate
12718 verify routines to perform initial checks (just CA validity) and
12719 to obtain the certificate chain. Then additional checks will be
12720 performed on the chain. Currently the root CA is checked to see
12721 if it is explicitly trusted for OCSP signing. This is used to set
12722 a root CA as a global signing root: that is any certificate that
12723 chains to that CA is an acceptable OCSP signing certificate.
12727 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12728 extensions from a separate configuration file.
12729 As when reading extensions from the main configuration file,
12730 the '-extensions ...' option may be used for specifying the
12733 *Massimiliano Pala <madwolf@comune.modena.it>*
12735 * New OCSP utility. Allows OCSP requests to be generated or
12736 read. The request can be sent to a responder and the output
12737 parsed, outputted or printed in text form. Not complete yet:
12738 still needs to check the OCSP response validity.
12742 * New subcommands for 'openssl ca':
12743 `openssl ca -status <serial>` prints the status of the cert with
12744 the given serial number (according to the index file).
12745 `openssl ca -updatedb` updates the expiry status of certificates
12748 *Massimiliano Pala <madwolf@comune.modena.it>*
12750 * New '-newreq-nodes' command option to CA.pl. This is like
12751 '-newreq', but calls 'openssl req' with the '-nodes' option
12752 so that the resulting key is not encrypted.
12754 *Damien Miller <djm@mindrot.org>*
12756 * New configuration for the GNU Hurd.
12758 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12760 * Initial code to implement OCSP basic response verify. This
12761 is currently incomplete. Currently just finds the signer's
12762 certificate and verifies the signature on the response.
12766 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12767 value of OPENSSLDIR. This is available via the new '-d' option
12768 to 'openssl version', and is also included in 'openssl version -a'.
12772 * Allowing defining memory allocation callbacks that will be given
12773 file name and line number information in additional arguments
12774 (a `const char*` and an int). The basic functionality remains, as
12775 well as the original possibility to just replace malloc(),
12776 realloc() and free() by functions that do not know about these
12777 additional arguments. To register and find out the current
12778 settings for extended allocation functions, the following
12779 functions are provided:
12781 CRYPTO_set_mem_ex_functions
12782 CRYPTO_set_locked_mem_ex_functions
12783 CRYPTO_get_mem_ex_functions
12784 CRYPTO_get_locked_mem_ex_functions
12786 These work the same way as CRYPTO_set_mem_functions and friends.
12787 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12788 extended allocation function is enabled.
12789 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12790 a conventional allocation function is enabled.
12792 *Richard Levitte, Bodo Moeller*
12794 * Finish off removing the remaining LHASH function pointer casts.
12795 There should no longer be any prototype-casting required when using
12796 the LHASH abstraction, and any casts that remain are "bugs". See
12797 the callback types and macros at the head of lhash.h for details
12798 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12802 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12803 If /dev/[u]random devices are not available or do not return enough
12804 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12806 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12807 /etc/entropy will be queried once each in this sequence, querying stops
12808 when enough entropy was collected without querying more sockets.
12812 * Change the Unix RAND_poll() variant to be able to poll several
12813 random devices, as specified by DEVRANDOM, until a sufficient amount
12814 of data has been collected. We spend at most 10 ms on each file
12815 (select timeout) and read in non-blocking mode. DEVRANDOM now
12816 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12817 (previously it was just the string "/dev/urandom"), so on typical
12818 platforms the 10 ms delay will never occur.
12819 Also separate out the Unix variant to its own file, rand_unix.c.
12820 For VMS, there's a currently-empty rand_vms.c.
12824 * Move OCSP client related routines to ocsp_cl.c. These
12825 provide utility functions which an application needing
12826 to issue a request to an OCSP responder and analyse the
12827 response will typically need: as opposed to those which an
12828 OCSP responder itself would need which will be added later.
12830 OCSP_request_sign() signs an OCSP request with an API similar
12831 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12832 response. OCSP_response_get1_basic() extracts basic response
12833 from response. OCSP_resp_find_status(): finds and extracts status
12834 information from an OCSP_CERTID structure (which will be created
12835 when the request structure is built). These are built from lower
12836 level functions which work on OCSP_SINGLERESP structures but
12837 won't normally be used unless the application wishes to examine
12838 extensions in the OCSP response for example.
12840 Replace nonce routines with a pair of functions.
12841 OCSP_request_add1_nonce() adds a nonce value and optionally
12842 generates a random value. OCSP_check_nonce() checks the
12843 validity of the nonce in an OCSP response.
12847 * Change function OCSP_request_add() to OCSP_request_add0_id().
12848 This doesn't copy the supplied OCSP_CERTID and avoids the
12849 need to free up the newly created id. Change return type
12850 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12851 This can then be used to add extensions to the request.
12852 Deleted OCSP_request_new(), since most of its functionality
12853 is now in OCSP_REQUEST_new() (and the case insensitive name
12854 clash) apart from the ability to set the request name which
12855 will be added elsewhere.
12859 * Update OCSP API. Remove obsolete extensions argument from
12860 various functions. Extensions are now handled using the new
12861 OCSP extension code. New simple OCSP HTTP function which
12862 can be used to send requests and parse the response.
12866 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12867 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12868 uses the special reorder version of SET OF to sort the attributes
12869 and reorder them to match the encoded order. This resolves a long
12870 standing problem: a verify on a PKCS7 structure just after signing
12871 it used to fail because the attribute order did not match the
12872 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12873 it uses the received order. This is necessary to tolerate some broken
12874 software that does not order SET OF. This is handled by encoding
12875 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12876 to produce the required SET OF.
12880 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12881 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12882 files to get correct declarations of the ASN.1 item variables.
12886 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12887 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12888 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12889 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12890 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12891 ASN1_ITEM and no wrapper functions.
12895 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12896 replace the old function pointer based I/O routines. Change most of
12897 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12901 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12902 lines, recognize more "algorithms" that can be deselected, and make
12903 it complain about algorithm deselection that isn't recognised.
12907 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12908 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12909 to use new functions. Add NO_ASN1_OLD which can be set to remove
12910 some old style ASN1 functions: this can be used to determine if old
12911 code will still work when these eventually go away.
12915 * New extension functions for OCSP structures, these follow the
12916 same conventions as certificates and CRLs.
12920 * New function X509V3_add1_i2d(). This automatically encodes and
12921 adds an extension. Its behaviour can be customised with various
12922 flags to append, replace or delete. Various wrappers added for
12923 certificates and CRLs.
12927 * Fix to avoid calling the underlying ASN1 print routine when
12928 an extension cannot be parsed. Correct a typo in the
12929 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12933 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12934 entries for variables.
12938 * Add functionality to apps/openssl.c for detecting locking
12939 problems: As the program is single-threaded, all we have
12940 to do is register a locking callback using an array for
12941 storing which locks are currently held by the program.
12945 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12946 SSL_get_ex_data_X509_STORE_idx(), which is used in
12947 ssl_verify_cert_chain() and thus can be called at any time
12948 during TLS/SSL handshakes so that thread-safety is essential.
12949 Unfortunately, the ex_data design is not at all suited
12950 for multi-threaded use, so it probably should be abolished.
12954 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12956 *Broadcom, tweaked and integrated by Geoff Thorpe*
12958 * Move common extension printing code to new function
12959 X509V3_print_extensions(). Reorganise OCSP print routines and
12960 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12964 * New function X509_signature_print() to remove duplication in some
12969 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12970 set (this was treated exactly the same as SET OF previously). This
12971 is used to reorder the STACK representing the structure to match the
12972 encoding. This will be used to get round a problem where a PKCS7
12973 structure which was signed could not be verified because the STACK
12974 order did not reflect the encoded order.
12978 * Reimplement the OCSP ASN1 module using the new code.
12982 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12983 for its ASN1 operations. The old style function pointers still exist
12984 for now but they will eventually go away.
12988 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12989 completely replaces the old ASN1 functionality with a table driven
12990 encoder and decoder which interprets an ASN1_ITEM structure describing
12991 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12992 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12993 has also been converted to the new form.
12997 * Change BN_mod_exp_recp so that negative moduli are tolerated
12998 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12999 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13000 for negative moduli.
13004 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13005 of not touching the result's sign bit.
13009 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13014 * Changed the LHASH code to use prototypes for callbacks, and created
13015 macros to declare and implement thin (optionally static) functions
13016 that provide type-safety and avoid function pointer casting for the
13017 type-specific callbacks.
13021 * Added Kerberos Cipher Suites to be used with TLS, as written in
13023 *Veers Staats <staatsvr@asc.hpc.mil>,
13024 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13026 * Reformat the FAQ so the different questions and answers can be divided
13027 in sections depending on the subject.
13031 * Have the zlib compression code load ZLIB.DLL dynamically under
13036 * New function BN_mod_sqrt for computing square roots modulo a prime
13037 (using the probabilistic Tonelli-Shanks algorithm unless
13038 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13039 be handled deterministically).
13041 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13043 * Make BN_mod_inverse faster by explicitly handling small quotients
13044 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13045 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13049 * New function BN_kronecker.
13053 * Fix BN_gcd so that it works on negative inputs; the result is
13054 positive unless both parameters are zero.
13055 Previously something reasonably close to an infinite loop was
13056 possible because numbers could be growing instead of shrinking
13057 in the implementation of Euclid's algorithm.
13061 * Fix BN_is_word() and BN_is_one() macros to take into account the
13062 sign of the number in question.
13064 Fix BN_is_word(a,w) to work correctly for w == 0.
13066 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13067 because its test if the absolute value of 'a' equals 'w'.
13068 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13069 it exists mostly for use in the implementations of BN_is_zero(),
13070 BN_is_one(), and BN_is_word().
13074 * New function BN_swap.
13078 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13079 the exponentiation functions are more likely to produce reasonable
13080 results on negative inputs.
13084 * Change BN_mod_mul so that the result is always non-negative.
13085 Previously, it could be negative if one of the factors was negative;
13086 I don't think anyone really wanted that behaviour.
13090 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13091 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13092 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13093 and add new functions:
13102 BN_mod_lshift1_quick
13104 BN_mod_lshift_quick
13106 These functions always generate non-negative results.
13108 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13109 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13111 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13112 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13113 be reduced modulo `m`.
13115 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13118 The following entry accidentally appeared in the CHANGES file
13119 distributed with OpenSSL 0.9.7. The modifications described in
13120 it do *not* apply to OpenSSL 0.9.7.
13122 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13123 was actually never needed) and in BN_mul(). The removal in BN_mul()
13124 required a small change in bn_mul_part_recursive() and the addition
13125 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13126 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13127 bn_sub_words() and bn_add_words() except they take arrays with
13133 * In 'openssl passwd', verify passwords read from the terminal
13134 unless the '-salt' option is used (which usually means that
13135 verification would just waste user's time since the resulting
13136 hash is going to be compared with some given password hash)
13137 or the new '-noverify' option is used.
13139 This is an incompatible change, but it does not affect
13140 non-interactive use of 'openssl passwd' (passwords on the command
13141 line, '-stdin' option, '-in ...' option) and thus should not
13142 cause any problems.
13146 * Remove all references to RSAref, since there's no more need for it.
13150 * Make DSO load along a path given through an environment variable
13151 (SHLIB_PATH) with shl_load().
13155 * Constify the ENGINE code as a result of BIGNUM constification.
13156 Also constify the RSA code and most things related to it. In a
13157 few places, most notable in the depth of the ASN.1 code, ugly
13158 casts back to non-const were required (to be solved at a later
13163 * Make it so the openssl application has all engines loaded by default.
13167 * Constify the BIGNUM routines a little more.
13171 * Add the following functions:
13173 ENGINE_load_cswift()
13175 ENGINE_load_atalla()
13176 ENGINE_load_nuron()
13177 ENGINE_load_builtin_engines()
13179 That way, an application can itself choose if external engines that
13180 are built-in in OpenSSL shall ever be used or not. The benefit is
13181 that applications won't have to be linked with libdl or other dso
13182 libraries unless it's really needed.
13184 Changed 'openssl engine' to load all engines on demand.
13185 Changed the engine header files to avoid the duplication of some
13186 declarations (they differed!).
13190 * 'openssl engine' can now list capabilities.
13194 * Better error reporting in 'openssl engine'.
13198 * Never call load_dh_param(NULL) in s_server.
13202 * Add engine application. It can currently list engines by name and
13203 identity, and test if they are actually available.
13207 * Improve RPM specification file by forcing symbolic linking and making
13208 sure the installed documentation is also owned by root.root.
13210 *Damien Miller <djm@mindrot.org>*
13212 * Give the OpenSSL applications more possibilities to make use of
13213 keys (public as well as private) handled by engines.
13217 * Add OCSP code that comes from CertCo.
13221 * Add VMS support for the Rijndael code.
13225 * Added untested support for Nuron crypto accelerator.
13229 * Add support for external cryptographic devices. This code was
13230 previously distributed separately as the "engine" branch.
13232 *Geoff Thorpe, Richard Levitte*
13234 * Rework the filename-translation in the DSO code. It is now possible to
13235 have far greater control over how a "name" is turned into a filename
13236 depending on the operating environment and any oddities about the
13237 different shared library filenames on each system.
13241 * Support threads on FreeBSD-elf in Configure.
13245 * Fix for SHA1 assembly problem with MASM: it produces
13246 warnings about corrupt line number information when assembling
13247 with debugging information. This is caused by the overlapping
13250 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13253 NCONF_get_number() has no error checking at all. As a replacement,
13254 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13255 promoted strongly. The old NCONF_get_number is kept around for
13256 binary backward compatibility.
13257 Make it possible for methods to load from something other than a BIO,
13258 by providing a function pointer that is given a name instead of a BIO.
13259 For example, this could be used to load configuration data from an
13264 * Fix for non blocking accept BIOs. Added new I/O special reason
13265 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13266 with non blocking I/O was not possible because no retry code was
13267 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13272 * Added the beginnings of Rijndael support.
13276 * Fix for bug in DirectoryString mask setting. Add support for
13277 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13278 to allow certificate printing to more controllable, additional
13279 'certopt' option to 'x509' to allow new printing options to be
13284 * Clean old EAY MD5 hack from e_os.h.
13288 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13290 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13291 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13293 *Joe Orton, Steve Henson*
13295 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13297 * Fix additional bug revealed by the NISCC test suite:
13299 Stop bug triggering large recursion when presented with
13300 certain ASN.1 tags ([CVE-2003-0851])
13304 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13306 * Fix various bugs revealed by running the NISCC test suite:
13308 Stop out of bounds reads in the ASN1 code when presented with
13309 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13311 If verify callback ignores invalid public key errors don't try to check
13312 certificate signature with the NULL public key.
13316 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13317 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13322 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13323 extra data after the compression methods not only for TLS 1.0
13324 but also for SSL 3.0 (as required by the specification).
13326 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13328 * Change X509_certificate_type() to mark the key as exported/exportable
13329 when it's 512 *bits* long, not 512 bytes.
13333 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13335 * Countermeasure against the Klima-Pokorny-Rosa extension of
13336 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13337 a protocol version number mismatch like a decryption error
13338 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13342 * Turn on RSA blinding by default in the default implementation
13343 to avoid a timing attack. Applications that don't want it can call
13344 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13345 They would be ill-advised to do so in most cases.
13347 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13349 * Change RSA blinding code so that it works when the PRNG is not
13350 seeded (in this case, the secret RSA exponent is abused as
13351 an unpredictable seed -- if it is not unpredictable, there
13352 is no point in blinding anyway). Make RSA blinding thread-safe
13353 by remembering the creator's thread ID in rsa->blinding and
13354 having all other threads use local one-time blinding factors
13355 (this requires more computation than sharing rsa->blinding, but
13356 avoids excessive locking; and if an RSA object is not shared
13357 between threads, blinding will still be very fast).
13361 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13363 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13364 via timing by performing a MAC computation even if incorrect
13365 block cipher padding has been found. This is a countermeasure
13366 against active attacks where the attacker has to distinguish
13367 between bad padding and a MAC verification error. ([CVE-2003-0078])
13369 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13370 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13371 Martin Vuagnoux (EPFL, Ilion)*
13373 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13375 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13376 memory from its contents. This is done with a counter that will
13377 place alternating values in each byte. This can be used to solve
13378 two issues: 1) the removal of calls to memset() by highly optimizing
13379 compilers, and 2) cleansing with other values than 0, since those can
13380 be read through on certain media, for example a swap space on disk.
13384 * Bugfix: client side session caching did not work with external caching,
13385 because the session->cipher setting was not restored when reloading
13386 from the external cache. This problem was masked, when
13387 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13388 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13392 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13393 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13395 *Zeev Lieber <zeev-l@yahoo.com>*
13397 * Undo an undocumented change introduced in 0.9.6e which caused
13398 repeated calls to OpenSSL_add_all_ciphers() and
13399 OpenSSL_add_all_digests() to be ignored, even after calling
13404 * Change the default configuration reader to deal with last line not
13405 being properly terminated.
13409 * Change X509_NAME_cmp() so it applies the special rules on handling
13410 DN values that are of type PrintableString, as well as RDNs of type
13411 emailAddress where the value has the type ia5String.
13413 *stefank@valicert.com via Richard Levitte*
13415 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13416 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13417 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13418 the bitwise-OR of the two for use by the majority of applications
13419 wanting this behaviour, and update the docs. The documented
13420 behaviour and actual behaviour were inconsistent and had been
13421 changing anyway, so this is more a bug-fix than a behavioural
13424 *Geoff Thorpe, diagnosed by Nadav Har'El*
13426 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13427 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13431 * Fix initialization code race conditions in
13432 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13433 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13434 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13435 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13436 ssl2_get_cipher_by_char(),
13437 ssl3_get_cipher_by_char().
13439 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13441 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13442 the cached sessions are flushed, as the remove_cb() might use ex_data
13443 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13444 (see [openssl.org #212]).
13446 *Geoff Thorpe, Lutz Jaenicke*
13448 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13449 length, instead of the encoding length to d2i_ASN1_OBJECT.
13453 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13455 * [In 0.9.6g-engine release:]
13456 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13458 *Lynn Gazis <lgazis@rainbow.com>*
13460 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13462 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13463 and get fix the header length calculation.
13464 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13465 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13467 * Use proper error handling instead of 'assertions' in buffer
13468 overflow checks added in 0.9.6e. This prevents DoS (the
13469 assertions could call abort()).
13471 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13473 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13475 * Add various sanity checks to asn1_get_length() to reject
13476 the ASN1 length bytes if they exceed sizeof(long), will appear
13477 negative or the content length exceeds the length of the
13480 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13482 * Fix cipher selection routines: ciphers without encryption had no flags
13483 for the cipher strength set and where therefore not handled correctly
13484 by the selection routines (PR #130).
13488 * Fix EVP_dsa_sha macro.
13493 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13494 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13495 that was added in OpenSSL 0.9.6d.
13497 As the countermeasure turned out to be incompatible with some
13498 broken SSL implementations, the new option is part of SSL_OP_ALL.
13499 SSL_OP_ALL is usually employed when compatibility with weird SSL
13500 implementations is desired (e.g. '-bugs' option to 's_client' and
13501 's_server'), so the new option is automatically set in many
13506 * Changes in security patch:
13508 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13509 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13510 Air Force Materiel Command, USAF, under agreement number
13513 * Add various sanity checks to asn1_get_length() to reject
13514 the ASN1 length bytes if they exceed sizeof(long), will appear
13515 negative or the content length exceeds the length of the
13516 supplied buffer. ([CVE-2002-0659])
13518 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13520 * Assertions for various potential buffer overflows, not known to
13521 happen in practice.
13523 *Ben Laurie (CHATS)*
13525 * Various temporary buffers to hold ASCII versions of integers were
13526 too small for 64 bit platforms. ([CVE-2002-0655])
13527 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13529 * Remote buffer overflow in SSL3 protocol - an attacker could
13530 supply an oversized session ID to a client. ([CVE-2002-0656])
13532 *Ben Laurie (CHATS)*
13534 * Remote buffer overflow in SSL2 protocol - an attacker could
13535 supply an oversized client master key. ([CVE-2002-0656])
13537 *Ben Laurie (CHATS)*
13539 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13541 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13542 encoded as NULL) with id-dsa-with-sha1.
13544 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13546 * Check various `X509_...()` return values in apps/req.c.
13548 *Nils Larsch <nla@trustcenter.de>*
13550 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13551 an end-of-file condition would erroneously be flagged, when the CRLF
13552 was just at the end of a processed block. The bug was discovered when
13553 processing data through a buffering memory BIO handing the data to a
13554 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13555 <ptsekov@syntrex.com> and Nedelcho Stanev.
13559 * Implement a countermeasure against a vulnerability recently found
13560 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13561 before application data chunks to avoid the use of known IVs
13562 with data potentially chosen by the attacker.
13566 * Fix length checks in ssl3_get_client_hello().
13570 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13571 to prevent ssl3_read_internal() from incorrectly assuming that
13572 ssl3_read_bytes() found application data while handshake
13573 processing was enabled when in fact s->s3->in_read_app_data was
13574 merely automatically cleared during the initial handshake.
13576 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13578 * Fix object definitions for Private and Enterprise: they were not
13579 recognized in their shortname (=lowercase) representation. Extend
13580 obj_dat.pl to issue an error when using undefined keywords instead
13581 of silently ignoring the problem (Svenning Sorensen
13582 <sss@sss.dnsalias.net>).
13586 * Fix DH_generate_parameters() so that it works for 'non-standard'
13587 generators, i.e. generators other than 2 and 5. (Previously, the
13588 code did not properly initialise the 'add' and 'rem' values to
13589 BN_generate_prime().)
13591 In the new general case, we do not insist that 'generator' is
13592 actually a primitive root: This requirement is rather pointless;
13593 a generator of the order-q subgroup is just as good, if not
13598 * Map new X509 verification errors to alerts. Discovered and submitted by
13599 Tom Wu <tom@arcot.com>.
13603 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13604 returning non-zero before the data has been completely received
13605 when using non-blocking I/O.
13607 *Bodo Moeller; problem pointed out by John Hughes*
13609 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13611 *Ben Laurie, Lutz Jaenicke*
13613 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13614 Yoram Zahavi <YoramZ@gilian.com>).
13618 * Add information about CygWin 1.3 and on, and preserve proper
13619 configuration for the versions before that.
13621 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13623 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13624 check whether we deal with a copy of a session and do not delete from
13625 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13626 <izhar@checkpoint.com>.
13630 * Do not store session data into the internal session cache, if it
13631 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13632 flag is set). Proposed by Aslam <aslam@funk.com>.
13636 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13641 * [In 0.9.6d-engine release:]
13642 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13644 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13646 * Add the configuration target linux-s390x.
13648 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13650 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13651 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13652 variable as an indication that a ClientHello message has been
13653 received. As the flag value will be lost between multiple
13654 invocations of ssl3_accept when using non-blocking I/O, the
13655 function may not be aware that a handshake has actually taken
13656 place, thus preventing a new session from being added to the
13659 To avoid this problem, we now set s->new_session to 2 instead of
13660 using a local variable.
13662 *Lutz Jaenicke, Bodo Moeller*
13664 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13665 if the SSL_R_LENGTH_MISMATCH error is detected.
13667 *Geoff Thorpe, Bodo Moeller*
13669 * New 'shared_ldflag' column in Configure platform table.
13673 * Fix EVP_CIPHER_mode macro.
13675 *"Dan S. Camper" <dan@bti.net>*
13677 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13678 type, we must throw them away by setting rr->length to 0.
13680 *D P Chang <dpc@qualys.com>*
13682 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13684 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13685 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13686 worked incorrectly for those cases where range = `10..._2` and
13687 `3*range` is two bits longer than range.)
13691 * Only add signing time to PKCS7 structures if it is not already
13696 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13697 OBJ_ld_ce should be OBJ_id_ce.
13698 Also some ip-pda OIDs in crypto/objects/objects.txt were
13699 incorrect (cf. RFC 3039).
13701 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13703 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13704 returns early because it has nothing to do.
13706 *Andy Schneider <andy.schneider@bjss.co.uk>*
13708 * [In 0.9.6c-engine release:]
13709 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13711 *Andy Schneider <andy.schneider@bjss.co.uk>*
13713 * [In 0.9.6c-engine release:]
13714 Add support for Cryptographic Appliance's keyserver technology.
13715 (Use engine 'keyclient')
13717 *Cryptographic Appliances and Geoff Thorpe*
13719 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13720 is called via tools/c89.sh because arguments have to be
13721 rearranged (all '-L' options must appear before the first object
13724 *Richard Shapiro <rshapiro@abinitio.com>*
13726 * [In 0.9.6c-engine release:]
13727 Add support for Broadcom crypto accelerator cards, backported
13730 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13732 * [In 0.9.6c-engine release:]
13733 Add support for SureWare crypto accelerator cards from
13734 Baltimore Technologies. (Use engine 'sureware')
13736 *Baltimore Technologies and Mark Cox*
13738 * [In 0.9.6c-engine release:]
13739 Add support for crypto accelerator cards from Accelerated
13740 Encryption Processing, www.aep.ie. (Use engine 'aep')
13742 *AEP Inc. and Mark Cox*
13744 * Add a configuration entry for gcc on UnixWare.
13746 *Gary Benson <gbenson@redhat.com>*
13748 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13749 messages are stored in a single piece (fixed-length part and
13750 variable-length part combined) and fix various bugs found on the way.
13754 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13755 instead. BIO_gethostbyname() does not know what timeouts are
13756 appropriate, so entries would stay in cache even when they have
13758 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13760 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13761 faced with a pathologically small ClientHello fragment that does
13762 not contain client_version: Instead of aborting with an error,
13763 simply choose the highest available protocol version (i.e.,
13764 TLS 1.0 unless it is disabled). In practice, ClientHello
13765 messages are never sent like this, but this change gives us
13766 strictly correct behaviour at least for TLS.
13770 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13771 never resets s->method to s->ctx->method when called from within
13772 one of the SSL handshake functions.
13774 *Bodo Moeller; problem pointed out by Niko Baric*
13776 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13777 (sent using the client's version number) if client_version is
13778 smaller than the protocol version in use. Also change
13779 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13780 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13781 the client will at least see that alert.
13785 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13790 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13791 client receives HelloRequest while in a handshake.
13793 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13795 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13796 should end in 'break', not 'goto end' which circumvents various
13797 cleanups done in state SSL_ST_OK. But session related stuff
13798 must be disabled for SSL_ST_OK in the case that we just sent a
13801 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13802 before just sending a HelloRequest.
13804 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13806 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13807 reveal whether illegal block cipher padding was found or a MAC
13808 verification error occurred. (Neither SSLerr() codes nor alerts
13809 are directly visible to potential attackers, but the information
13810 may leak via logfiles.)
13812 Similar changes are not required for the SSL 2.0 implementation
13813 because the number of padding bytes is sent in clear for SSL 2.0,
13814 and the extra bytes are just ignored. However ssl/s2_pkt.c
13815 failed to verify that the purported number of padding bytes is in
13820 * Add OpenUNIX-8 support including shared libraries
13821 (Boyd Lynn Gerber <gerberb@zenez.com>).
13825 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13826 'wristwatch attack' using huge encoding parameters (cf.
13827 James H. Manger's CRYPTO 2001 paper). Note that the
13828 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13829 encoding parameters and hence was not vulnerable.
13833 * BN_sqr() bug fix.
13835 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13837 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13838 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13839 followed by modular reduction.
13841 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13843 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13844 equivalent based on BN_pseudo_rand() instead of BN_rand().
13848 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13849 This function was broken, as the check for a new client hello message
13850 to handle SGC did not allow these large messages.
13851 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13855 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13859 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13860 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13864 * Rework the configuration and shared library support for Tru64 Unix.
13865 The configuration part makes use of modern compiler features and
13866 still retains old compiler behavior for those that run older versions
13867 of the OS. The shared library support part includes a variant that
13868 uses the RPATH feature, and is available through the special
13869 configuration target "alpha-cc-rpath", which will never be selected
13872 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13874 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13875 with the same message size as in ssl3_get_certificate_request().
13876 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13877 messages might inadvertently be reject as too long.
13879 *Petr Lampa <lampa@fee.vutbr.cz>*
13881 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13885 * Modified SSL library such that the verify_callback that has been set
13886 specifically for an SSL object with SSL_set_verify() is actually being
13887 used. Before the change, a verify_callback set with this function was
13888 ignored and the verify_callback() set in the SSL_CTX at the time of
13889 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13890 to allow the necessary settings.
13894 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13895 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13896 done automatically (in contradiction to the requirements of the C
13897 standard). This made problems when used from OpenSSH.
13901 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13902 dh->length and always used
13904 BN_rand_range(priv_key, dh->p).
13906 BN_rand_range() is not necessary for Diffie-Hellman, and this
13907 specific range makes Diffie-Hellman unnecessarily inefficient if
13908 dh->length (recommended exponent length) is much smaller than the
13909 length of dh->p. We could use BN_rand_range() if the order of
13910 the subgroup was stored in the DH structure, but we only have
13915 BN_rand(priv_key, l, ...)
13917 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13924 RSA_eay_public_encrypt
13925 RSA_eay_private_decrypt
13926 RSA_eay_private_encrypt (signing)
13927 RSA_eay_public_decrypt (signature verification)
13929 (default implementations for RSA_public_encrypt,
13930 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13931 always reject numbers >= n.
13935 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13936 to synchronize access to 'locking_thread'. This is necessary on
13937 systems where access to 'locking_thread' (an 'unsigned long'
13938 variable) is not atomic.
13942 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13943 *before* setting the 'crypto_lock_rand' flag. The previous code had
13944 a race condition if 0 is a valid thread ID.
13946 *Travis Vitek <vitek@roguewave.com>*
13948 * Add support for shared libraries under Irix.
13950 *Albert Chin-A-Young <china@thewrittenword.com>*
13952 * Add configuration option to build on Linux on both big-endian and
13953 little-endian MIPS.
13955 *Ralf Baechle <ralf@uni-koblenz.de>*
13957 * Add the possibility to create shared libraries on HP-UX.
13961 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13963 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13964 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13965 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13966 PRNG state recovery was possible based on the output of
13967 one PRNG request appropriately sized to gain knowledge on
13968 'md' followed by enough consecutive 1-byte PRNG requests
13969 to traverse all of 'state'.
13971 1. When updating 'md_local' (the current thread's copy of 'md')
13972 during PRNG output generation, hash all of the previous
13973 'md_local' value, not just the half used for PRNG output.
13975 2. Make the number of bytes from 'state' included into the hash
13976 independent from the number of PRNG bytes requested.
13978 The first measure alone would be sufficient to avoid
13979 Markku-Juhani's attack. (Actually it had never occurred
13980 to me that the half of 'md_local' used for chaining was the
13981 half from which PRNG output bytes were taken -- I had always
13982 assumed that the secret half would be used.) The second
13983 measure makes sure that additional data from 'state' is never
13984 mixed into 'md_local' in small portions; this heuristically
13985 further strengthens the PRNG.
13989 * Fix crypto/bn/asm/mips3.s.
13993 * When only the key is given to "enc", the IV is undefined. Print out
13994 an error message in this case.
13998 * Handle special case when X509_NAME is empty in X509 printing routines.
14002 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14003 positive and less than q.
14007 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14008 used: it isn't thread safe and the add_lock_callback should handle
14011 *Paul Rose <Paul.Rose@bridge.com>*
14013 * Verify that incoming data obeys the block size in
14014 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14020 *Ulf Möller, Bodo Möller*
14022 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14023 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14024 when fixing the server behaviour for backwards-compatible 'client
14025 hello' messages. (Note that the attack is impractical against
14026 SSL 3.0 and TLS 1.0 anyway because length and version checking
14027 means that the probability of guessing a valid ciphertext is
14028 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14031 Before 0.9.5, the countermeasure (hide the error by generating a
14032 random 'decryption result') did not work properly because
14033 ERR_clear_error() was missing, meaning that SSL_get_error() would
14034 detect the supposedly ignored error.
14036 Both problems are now fixed.
14040 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14041 (previously it was 1024).
14045 * Fix for compatibility mode trust settings: ignore trust settings
14046 unless some valid trust or reject settings are present.
14050 * Fix for blowfish EVP: its a variable length cipher.
14054 * Fix various bugs related to DSA S/MIME verification. Handle missing
14055 parameters in DSA public key structures and return an error in the
14056 DSA routines if parameters are absent.
14060 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14061 in the current directory if neither $RANDFILE nor $HOME was set.
14062 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14063 caused some confusion to Windows users who haven't defined $HOME.
14064 Thus RAND_file_name() is changed again: e_os.h can define a
14065 DEFAULT_HOME, which will be used if $HOME is not set.
14066 For Windows, we use "C:"; on other platforms, we still require
14067 environment variables.
14069 * Move 'if (!initialized) RAND_poll()' into regions protected by
14070 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14071 having multiple threads call RAND_poll() concurrently.
14075 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14076 combination of a flag and a thread ID variable.
14077 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14078 flag), *other* threads can enter ssleay_add_bytes without obeying
14079 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14080 that they do not hold after the first thread unsets add_do_not_lock).
14084 * Change bctest again: '-x' expressions are not available in all
14085 versions of 'test'.
14089 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14091 * Fix a couple of memory leaks in PKCS7_dataDecode()
14093 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14095 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14096 the default extension for executables, if any. Also, make the perl
14097 scripts that use symlink() to test if it really exists and use "cp"
14098 if it doesn't. All this made OpenSSL compilable and installable in
14103 * Fix for asn1_GetSequence() for indefinite length constructed data.
14104 If SEQUENCE is length is indefinite just set c->slen to the total
14105 amount of data available.
14107 *Steve Henson, reported by shige@FreeBSD.org*
14109 *This change does not apply to 0.9.7.*
14111 * Change bctest to avoid here-documents inside command substitution
14112 (workaround for FreeBSD /bin/sh bug).
14113 For compatibility with Ultrix, avoid shell functions (introduced
14114 in the bctest version that searches along $PATH).
14118 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14119 with des_encrypt() defined on some operating systems, like Solaris
14124 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14125 On the Importance of Eliminating Errors in Cryptographic
14126 Computations, J. Cryptology 14 (2001) 2, 101-119,
14127 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14131 * MIPS assembler BIGNUM division bug fix.
14135 * Disabled incorrect Alpha assembler code.
14139 * Fix PKCS#7 decode routines so they correctly update the length
14140 after reading an EOC for the EXPLICIT tag.
14144 *This change does not apply to 0.9.7.*
14146 * Fix bug in PKCS#12 key generation routines. This was triggered
14147 if a 3DES key was generated with a 0 initial byte. Include
14148 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14149 (but broken) behaviour.
14153 * Enhance bctest to search for a working bc along $PATH and print
14156 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14158 * Fix memory leaks in err.c: free err_data string if necessary;
14159 don't write to the wrong index in ERR_set_error_data.
14163 * Implement ssl23_peek (analogous to ssl23_read), which previously
14168 * Replace rdtsc with `_emit` statements for VC++ version 5.
14170 *Jeremy Cooper <jeremy@baymoo.org>*
14172 * Make it possible to reuse SSLv2 sessions.
14176 * In copy_email() check for >= 0 as a return value for
14177 X509_NAME_get_index_by_NID() since 0 is a valid index.
14179 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14181 * Avoid coredump with unsupported or invalid public keys by checking if
14182 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14183 PKCS7_verify() fails with non detached data.
14187 * Don't use getenv in library functions when run as setuid/setgid.
14188 New function OPENSSL_issetugid().
14192 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14193 due to incorrect handling of multi-threading:
14195 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14197 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14199 3. Count how many times MemCheck_off() has been called so that
14200 nested use can be treated correctly. This also avoids
14201 inband-signalling in the previous code (which relied on the
14202 assumption that thread ID 0 is impossible).
14206 * Add "-rand" option also to s_client and s_server.
14210 * Fix CPU detection on Irix 6.x.
14211 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14212 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14214 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14219 *This change does not apply to 0.9.7.*
14221 * Use the cached encoding of an X509_NAME structure rather than
14222 copying it. This is apparently the reason for the libsafe "errors"
14223 but the code is actually correct.
14227 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14228 Bleichenbacher's DSA attack.
14229 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14230 to be set and top=0 forces the highest bit to be set; top=-1 is new
14231 and leaves the highest bit random.
14233 *Ulf Moeller, Bodo Moeller*
14235 * In the `NCONF_...`-based implementations for `CONF_...` queries
14236 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14237 a temporary CONF structure with the data component set to NULL
14238 (which gives segmentation faults in lh_retrieve).
14239 Instead, use NULL for the CONF pointer in CONF_get_string and
14240 CONF_get_number (which may use environment variables) and directly
14241 return NULL from CONF_get_section.
14245 * Fix potential buffer overrun for EBCDIC.
14249 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14250 keyUsage if basicConstraints absent for a CA.
14254 * Make SMIME_write_PKCS7() write mail header values with a format that
14255 is more generally accepted (no spaces before the semicolon), since
14256 some programs can't parse those values properly otherwise. Also make
14257 sure BIO's that break lines after each write do not create invalid
14262 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14263 macros previously used would not encode an empty SEQUENCE OF
14264 and break the signature.
14268 *This change does not apply to 0.9.7.*
14270 * Zero the premaster secret after deriving the master secret in
14275 * Add some EVP_add_digest_alias registrations (as found in
14276 OpenSSL_add_all_digests()) to SSL_library_init()
14277 aka OpenSSL_add_ssl_algorithms(). This provides improved
14278 compatibility with peers using X.509 certificates
14279 with unconventional AlgorithmIdentifier OIDs.
14283 * Fix for Irix with NO_ASM.
14285 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14287 * ./config script fixes.
14289 *Ulf Moeller, Richard Levitte*
14291 * Fix 'openssl passwd -1'.
14295 * Change PKCS12_key_gen_asc() so it can cope with non null
14296 terminated strings whose length is passed in the passlen
14297 parameter, for example from PEM callbacks. This was done
14298 by adding an extra length parameter to asc2uni().
14300 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14302 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14303 call failed, free the DSA structure.
14307 * Fix to uni2asc() to cope with zero length Unicode strings.
14308 These are present in some PKCS#12 files.
14312 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14313 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14314 when writing a 32767 byte record.
14316 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14318 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14319 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14321 (RSA objects have a reference count access to which is protected
14322 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14323 so they are meant to be shared between threads.)
14324 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14325 "Reddie, Steven" <Steven.Reddie@ca.com>*
14327 * Fix a deadlock in CRYPTO_mem_leaks().
14331 * Use better test patterns in bntest.
14335 * rand_win.c fix for Borland C.
14339 * BN_rshift bugfix for n == 0.
14343 * Add a 'bctest' script that checks for some known 'bc' bugs
14344 so that 'make test' does not abort just because 'bc' is broken.
14348 * Store verify_result within SSL_SESSION also for client side to
14349 avoid potential security hole. (Re-used sessions on the client side
14350 always resulted in verify_result==X509_V_OK, not using the original
14351 result of the server certificate verification.)
14355 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14356 SSL3_RT_APPLICATION_DATA, return 0.
14357 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14362 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14363 releases, have been re-implemented by renaming the previous
14364 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14365 and ssl3_read_internal, respectively, and adding 'peek' parameters
14366 to them. The new ssl[23]_{read,peek} functions are calls to
14367 ssl[23]_read_internal with the 'peek' flag set appropriately.
14368 A 'peek' parameter has also been added to ssl3_read_bytes, which
14369 does the actual work for ssl3_read_internal.
14373 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14374 the method-specific "init()" handler. Also clean up ex_data after
14375 calling the method-specific "finish()" handler. Previously, this was
14376 happening the other way round.
14380 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14381 The previous value, 12, was not always sufficient for BN_mod_exp().
14385 * Make sure that shared libraries get the internal name engine with
14386 the full version number and not just 0. This should mark the
14387 shared libraries as not backward compatible. Of course, this should
14388 be changed again when we can guarantee backward binary compatibility.
14392 * Fix typo in get_cert_by_subject() in by_dir.c
14394 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14396 * Rework the system to generate shared libraries:
14398 - Make note of the expected extension for the shared libraries and
14399 if there is a need for symbolic links from for example libcrypto.so.0
14400 to libcrypto.so.0.9.7. There is extended info in Configure for
14403 - Make as few rebuilds of the shared libraries as possible.
14405 - Still avoid linking the OpenSSL programs with the shared libraries.
14407 - When installing, install the shared libraries separately from the
14412 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14414 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14415 and not in SSL_clear because the latter is also used by the
14416 accept/connect functions; previously, the settings made by
14417 SSL_set_read_ahead would be lost during the handshake.
14419 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14421 * Correct util/mkdef.pl to be selective about disabled algorithms.
14422 Previously, it would create entries for disabled algorithms no
14427 * Added several new manual pages for SSL_* function.
14431 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14433 * In ssl23_get_client_hello, generate an error message when faced
14434 with an initial SSL 3.0/TLS record that is too small to contain the
14435 first two bytes of the ClientHello message, i.e. client_version.
14436 (Note that this is a pathologic case that probably has never happened
14437 in real life.) The previous approach was to use the version number
14438 from the record header as a substitute; but our protocol choice
14439 should not depend on that one because it is not authenticated
14440 by the Finished messages.
14444 * More robust randomness gathering functions for Windows.
14446 *Jeffrey Altman <jaltman@columbia.edu>*
14448 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14449 not set then we don't setup the error code for issuer check errors
14450 to avoid possibly overwriting other errors which the callback does
14451 handle. If an application does set the flag then we assume it knows
14452 what it is doing and can handle the new informational codes
14457 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14458 a general "ANY" type, as such it should be able to decode anything
14459 including tagged types. However it didn't check the class so it would
14460 wrongly interpret tagged types in the same way as their universal
14461 counterpart and unknown types were just rejected. Changed so that the
14462 tagged and unknown types are handled in the same way as a SEQUENCE:
14463 that is the encoding is stored intact. There is also a new type
14464 "V_ASN1_OTHER" which is used when the class is not universal, in this
14465 case we have no idea what the actual type is so we just lump them all
14470 * On VMS, stdout may very well lead to a file that is written to
14471 in a record-oriented fashion. That means that every write() will
14472 write a separate record, which will be read separately by the
14473 programs trying to read from it. This can be very confusing.
14475 The solution is to put a BIO filter in the way that will buffer
14476 text until a linefeed is reached, and then write everything a
14477 line at a time, so every record written will be an actual line,
14478 not chunks of lines and not (usually doesn't happen, but I've
14479 seen it once) several lines in one record. BIO_f_linebuffer() is
14482 Currently, it's a VMS-only method, because that's where it has
14483 been tested well enough.
14487 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14488 it can return incorrect results.
14489 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14490 but it was in 0.9.6-beta[12].)
14494 * Disable the check for content being present when verifying detached
14495 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14496 include zero length content when signing messages.
14500 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14501 BIO_ctrl (for BIO pairs).
14505 * Add DSO method for VMS.
14509 * Bug fix: Montgomery multiplication could produce results with the
14514 * Add RPM specification openssl.spec and modify it to build three
14515 packages. The default package contains applications, application
14516 documentation and run-time libraries. The devel package contains
14517 include files, static libraries and function documentation. The
14518 doc package contains the contents of the doc directory. The original
14519 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14523 * Add a large number of documentation files for many SSL routines.
14525 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14527 * Add a configuration entry for Sony News 4.
14529 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14531 * Don't set the two most significant bits to one when generating a
14532 random number < q in the DSA library.
14536 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14537 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14538 the underlying transport is blocking) if a handshake took place.
14539 (The default behaviour is needed by applications such as s_client
14540 and s_server that use select() to determine when to use SSL_read;
14541 but for applications that know in advance when to expect data, it
14542 just makes things more complicated.)
14546 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14551 * Add a few more EBCDIC conditionals that make `req` and `x509`
14552 work better on such systems.
14554 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14556 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14557 Update PKCS12_parse() so it copies the friendlyName and the
14558 keyid to the certificates aux info.
14562 * Fix bug in PKCS7_verify() which caused an infinite loop
14563 if there was more than one signature.
14565 *Sven Uszpelkat <su@celocom.de>*
14567 * Major change in util/mkdef.pl to include extra information
14568 about each symbol, as well as presenting variables as well
14569 as functions. This change means that there's n more need
14570 to rebuild the .num files when some algorithms are excluded.
14574 * Allow the verify time to be set by an application,
14575 rather than always using the current time.
14579 * Phase 2 verify code reorganisation. The certificate
14580 verify code now looks up an issuer certificate by a
14581 number of criteria: subject name, authority key id
14582 and key usage. It also verifies self signed certificates
14583 by the same criteria. The main comparison function is
14584 X509_check_issued() which performs these checks.
14586 Lot of changes were necessary in order to support this
14587 without completely rewriting the lookup code.
14589 Authority and subject key identifier are now cached.
14591 The LHASH 'certs' is X509_STORE has now been replaced
14592 by a STACK_OF(X509_OBJECT). This is mainly because an
14593 LHASH can't store or retrieve multiple objects with
14594 the same hash value.
14596 As a result various functions (which were all internal
14597 use only) have changed to handle the new X509_STORE
14598 structure. This will break anything that messed round
14599 with X509_STORE internally.
14601 The functions X509_STORE_add_cert() now checks for an
14602 exact match, rather than just subject name.
14604 The X509_STORE API doesn't directly support the retrieval
14605 of multiple certificates matching a given criteria, however
14606 this can be worked round by performing a lookup first
14607 (which will fill the cache with candidate certificates)
14608 and then examining the cache for matches. This is probably
14609 the best we can do without throwing out X509_LOOKUP
14610 entirely (maybe later...).
14612 The X509_VERIFY_CTX structure has been enhanced considerably.
14614 All certificate lookup operations now go via a get_issuer()
14615 callback. Although this currently uses an X509_STORE it
14616 can be replaced by custom lookups. This is a simple way
14617 to bypass the X509_STORE hackery necessary to make this
14618 work and makes it possible to use more efficient techniques
14619 in future. A very simple version which uses a simple
14620 STACK for its trusted certificate store is also provided
14621 using X509_STORE_CTX_trusted_stack().
14623 The verify_cb() and verify() callbacks now have equivalents
14624 in the X509_STORE_CTX structure.
14626 X509_STORE_CTX also has a 'flags' field which can be used
14627 to customise the verify behaviour.
14631 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14632 excludes S/MIME capabilities.
14636 * When a certificate request is read in keep a copy of the
14637 original encoding of the signed data and use it when outputting
14638 again. Signatures then use the original encoding rather than
14639 a decoded, encoded version which may cause problems if the
14640 request is improperly encoded.
14644 * For consistency with other BIO_puts implementations, call
14645 buffer_write(b, ...) directly in buffer_puts instead of calling
14648 In BIO_puts, increment b->num_write as in BIO_write.
14650 *Peter.Sylvester@EdelWeb.fr*
14652 * Fix BN_mul_word for the case where the word is 0. (We have to use
14653 BN_zero, we may not return a BIGNUM with an array consisting of
14654 words set to zero.)
14658 * Avoid calling abort() from within the library when problems are
14659 detected, except if preprocessor symbols have been defined
14660 (such as REF_CHECK, BN_DEBUG etc.).
14664 * New openssl application 'rsautl'. This utility can be
14665 used for low level RSA operations. DER public key
14666 BIO/fp routines also added.
14670 * New Configure entry and patches for compiling on QNX 4.
14672 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14674 * A demo state-machine implementation was sponsored by
14675 Nuron (<http://www.nuron.com/>) and is now available in
14676 demos/state_machine.
14680 * New options added to the 'dgst' utility for signature
14681 generation and verification.
14685 * Unrecognized PKCS#7 content types are now handled via a
14686 catch all ASN1_TYPE structure. This allows unsupported
14687 types to be stored as a "blob" and an application can
14688 encode and decode it manually.
14692 * Fix various signed/unsigned issues to make a_strex.c
14693 compile under VC++.
14695 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14697 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14698 length if passed a buffer. ASN1_INTEGER_to_BN failed
14699 if passed a NULL BN and its argument was negative.
14701 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14703 * Modification to PKCS#7 encoding routines to output definite
14704 length encoding. Since currently the whole structures are in
14705 memory there's not real point in using indefinite length
14706 constructed encoding. However if OpenSSL is compiled with
14707 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14711 * Added BIO_vprintf() and BIO_vsnprintf().
14715 * Added more prefixes to parse for in the strings written
14716 through a logging bio, to cover all the levels that are available
14717 through syslog. The prefixes are now:
14719 PANIC, EMERG, EMR => LOG_EMERG
14720 ALERT, ALR => LOG_ALERT
14721 CRIT, CRI => LOG_CRIT
14722 ERROR, ERR => LOG_ERR
14723 WARNING, WARN, WAR => LOG_WARNING
14724 NOTICE, NOTE, NOT => LOG_NOTICE
14725 INFO, INF => LOG_INFO
14726 DEBUG, DBG => LOG_DEBUG
14728 and as before, if none of those prefixes are present at the
14729 beginning of the string, LOG_ERR is chosen.
14731 On Win32, the `LOG_*` levels are mapped according to this:
14733 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14734 LOG_WARNING => EVENTLOG_WARNING_TYPE
14735 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14739 * Made it possible to reconfigure with just the configuration
14740 argument "reconf" or "reconfigure". The command line arguments
14741 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14742 and are retrieved from there when reconfiguring.
14748 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14750 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14754 * The obj_dat.pl script was messing up the sorting of object
14755 names. The reason was that it compared the quoted version
14756 of strings as a result "OCSP" > "OCSP Signing" because
14757 " > SPACE. Changed script to store unquoted versions of
14758 names and add quotes on output. It was also omitting some
14759 names from the lookup table if they were given a default
14760 value (that is if SN is missing it is given the same
14761 value as LN and vice versa), these are now added on the
14762 grounds that if an object has a name we should be able to
14763 look it up. Finally added warning output when duplicate
14764 short or long names are found.
14768 * Changes needed for Tandem NSK.
14770 *Scott Uroff <scott@xypro.com>*
14772 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14773 RSA_padding_check_SSLv23(), special padding was never detected
14774 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14775 version rollback attacks was not effective.
14777 In s23_clnt.c, don't use special rollback-attack detection padding
14778 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14779 client; similarly, in s23_srvr.c, don't do the rollback check if
14780 SSL 2.0 is the only protocol enabled in the server.
14784 * Make it possible to get hexdumps of unprintable data with 'openssl
14785 asn1parse'. By implication, the functions ASN1_parse_dump() and
14786 BIO_dump_indent() are added.
14790 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14791 these print out strings and name structures based on various
14792 flags including RFC2253 support and proper handling of
14793 multibyte characters. Added options to the 'x509' utility
14794 to allow the various flags to be set.
14798 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14799 Also change the functions X509_cmp_current_time() and
14800 X509_gmtime_adj() work with an ASN1_TIME structure,
14801 this will enable certificates using GeneralizedTime in validity
14802 dates to be checked.
14806 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14807 negative public key encodings) on by default,
14808 NO_NEG_PUBKEY_BUG can be set to disable it.
14812 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14813 content octets. An i2c_ASN1_OBJECT is unnecessary because
14814 the encoding can be trivially obtained from the structure.
14818 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14819 not read locks (`CRYPTO_r_[un]lock`).
14823 * A first attempt at creating official support for shared
14824 libraries through configuration. I've kept it so the
14825 default is static libraries only, and the OpenSSL programs
14826 are always statically linked for now, but there are
14827 preparations for dynamic linking in place.
14828 This has been tested on Linux and Tru64.
14832 * Randomness polling function for Win9x, as described in:
14833 Peter Gutmann, Software Generation of Practically Strong
14838 * Fix so PRNG is seeded in req if using an already existing
14843 * New options to smime application. -inform and -outform
14844 allow alternative formats for the S/MIME message including
14845 PEM and DER. The -content option allows the content to be
14846 specified separately. This should allow things like Netscape
14847 form signing output easier to verify.
14851 * Fix the ASN1 encoding of tags using the 'long form'.
14855 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14856 STRING types. These convert content octets to and from the
14857 underlying type. The actual tag and length octets are
14858 already assumed to have been read in and checked. These
14859 are needed because all other string types have virtually
14860 identical handling apart from the tag. By having versions
14861 of the ASN1 functions that just operate on content octets
14862 IMPLICIT tagging can be handled properly. It also allows
14863 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14864 and ASN1_INTEGER are identical apart from the tag.
14868 * Change the handling of OID objects as follows:
14870 - New object identifiers are inserted in objects.txt, following
14871 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14872 - objects.pl is used to process obj_mac.num and create a new
14874 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14877 This is currently kind of a hack, and the perl code in objects.pl
14878 isn't very elegant, but it works as I intended. The simplest way
14879 to check that it worked correctly is to look in obj_dat.h and
14880 check the array nid_objs and make sure the objects haven't moved
14881 around (this is important!). Additions are OK, as well as
14882 consistent name changes.
14886 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14890 * Addition of the command line parameter '-rand file' to 'openssl req'.
14891 The given file adds to whatever has already been seeded into the
14892 random pool through the RANDFILE configuration file option or
14893 environment variable, or the default random state file.
14897 * mkstack.pl now sorts each macro group into lexical order.
14898 Previously the output order depended on the order the files
14899 appeared in the directory, resulting in needless rewriting
14904 * Patches to make OpenSSL compile under Win32 again. Mostly
14905 work arounds for the VC++ problem that it treats func() as
14906 func(void). Also stripped out the parts of mkdef.pl that
14907 added extra typesafe functions: these no longer exist.
14911 * Reorganisation of the stack code. The macros are now all
14912 collected in safestack.h . Each macro is defined in terms of
14913 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14914 DEBUG_SAFESTACK is now handled in terms of function casts,
14915 this has the advantage of retaining type safety without the
14916 use of additional functions. If DEBUG_SAFESTACK is not defined
14917 then the non typesafe macros are used instead. Also modified the
14918 mkstack.pl script to handle the new form. Needs testing to see
14919 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14920 the default if no major problems. Similar behaviour for ASN1_SET_OF
14921 and PKCS12_STACK_OF.
14925 * When some versions of IIS use the 'NET' form of private key the
14926 key derivation algorithm is different. Normally MD5(password) is
14927 used as a 128 bit RC4 key. In the modified case
14928 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14929 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14930 as the old Netscape_RSA functions except they have an additional
14931 'sgckey' parameter which uses the modified algorithm. Also added
14932 an -sgckey command line option to the rsa utility. Thanks to
14933 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14934 algorithm to openssl-dev.
14938 * The evp_local.h macros were using 'c.##kname' which resulted in
14939 invalid expansion on some systems (SCO 5.0.5 for example).
14940 Corrected to 'c.kname'.
14942 *Phillip Porch <root@theporch.com>*
14944 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14945 a STACK of email addresses from a certificate or request, these look
14946 in the subject name and the subject alternative name extensions and
14947 omit any duplicate addresses.
14951 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14952 This makes DSA verification about 2 % faster.
14956 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14957 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14958 plus overhead for 1024 bit moduli).
14959 This makes exponentiations about 0.5 % faster for 1024 bit
14960 exponents (as measured by "openssl speed rsa2048").
14964 * Rename memory handling macros to avoid conflicts with other
14966 Malloc => OPENSSL_malloc
14967 Malloc_locked => OPENSSL_malloc_locked
14968 Realloc => OPENSSL_realloc
14969 Free => OPENSSL_free
14973 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14974 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14978 * CygWin32 support.
14980 *John Jarvie <jjarvie@newsguy.com>*
14982 * The type-safe stack code has been rejigged. It is now only compiled
14983 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14984 by default all type-specific stack functions are "#define"d back to
14985 standard stack functions. This results in more streamlined output
14986 but retains the type-safety checking possibilities of the original
14991 * The STACK code has been cleaned up, and certain type declarations
14992 that didn't make a lot of sense have been brought in line. This has
14993 also involved a cleanup of sorts in safestack.h to more correctly
14994 map type-safe stack functions onto their plain stack counterparts.
14995 This work has also resulted in a variety of "const"ifications of
14996 lots of the code, especially `_cmp` operations which should normally
14997 be prototyped with "const" parameters anyway.
15001 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15002 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15003 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15004 where all of 'md' is used each time the PRNG is used, but 'state'
15005 is used only indexed by a cyclic counter. As entropy may not be
15006 well distributed from the beginning, 'md' is important as a
15007 chaining variable. However, the output function chains only half
15008 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15009 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15010 in all of 'state' being rewritten, with the new values depending
15011 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15015 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15016 the handshake is continued after ssl_verify_cert_chain();
15017 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15018 can lead to 'unexplainable' connection aborts later.
15020 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15022 * Major EVP API cipher revision.
15023 Add hooks for extra EVP features. This allows various cipher
15024 parameters to be set in the EVP interface. Support added for variable
15025 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15026 setting of RC2 and RC5 parameters.
15028 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15031 Remove lots of duplicated code from the EVP library. For example *every*
15032 cipher init() function handles the 'iv' in the same way according to the
15033 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15034 for CFB and OFB modes they zero ctx->num.
15036 New functionality allows removal of S/MIME code RC2 hack.
15038 Most of the routines have the same form and so can be declared in terms
15041 By shifting this to the top level EVP_CipherInit() it can be removed from
15042 all individual ciphers. If the cipher wants to handle IVs or keys
15043 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15046 Change lots of functions like EVP_EncryptUpdate() to now return a
15047 value: although software versions of the algorithms cannot fail
15048 any installed hardware versions can.
15052 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15053 this option is set, tolerate broken clients that send the negotiated
15054 protocol version number instead of the requested protocol version
15059 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15060 i.e. non-zero for export ciphersuites, zero otherwise.
15061 Previous versions had this flag inverted, inconsistent with
15062 rsa_tmp_cb (..._TMP_RSA_CB).
15064 *Bodo Moeller; problem reported by Amit Chopra*
15066 * Add missing DSA library text string. Work around for some IIS
15067 key files with invalid SEQUENCE encoding.
15071 * Add a document (doc/standards.txt) that list all kinds of standards
15072 and so on that are implemented in OpenSSL.
15076 * Enhance c_rehash script. Old version would mishandle certificates
15077 with the same subject name hash and wouldn't handle CRLs at all.
15078 Added -fingerprint option to crl utility, to support new c_rehash
15083 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15087 * Fix for SSL server purpose checking. Server checking was
15088 rejecting certificates which had extended key usage present
15089 but no ssl client purpose.
15091 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15093 * Make PKCS#12 code work with no password. The PKCS#12 spec
15094 is a little unclear about how a blank password is handled.
15095 Since the password in encoded as a BMPString with terminating
15096 double NULL a zero length password would end up as just the
15097 double NULL. However no password at all is different and is
15098 handled differently in the PKCS#12 key generation code. NS
15099 treats a blank password as zero length. MSIE treats it as no
15100 password on export: but it will try both on import. We now do
15101 the same: PKCS12_parse() tries zero length and no password if
15102 the password is set to "" or NULL (NULL is now a valid password:
15103 it wasn't before) as does the pkcs12 application.
15107 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15108 perror when PEM_read_bio_X509_REQ fails, the error message must
15109 be obtained from the error queue.
15113 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15114 it in ERR_remove_state if appropriate, and change ERR_get_state
15115 accordingly to avoid race conditions (this is necessary because
15116 thread_hash is no longer constant once set).
15120 * Bugfix for linux-elf makefile.one.
15124 * RSA_get_default_method() will now cause a default
15125 RSA_METHOD to be chosen if one doesn't exist already.
15126 Previously this was only set during a call to RSA_new()
15127 or RSA_new_method(NULL) meaning it was possible for
15128 RSA_get_default_method() to return NULL.
15132 * Added native name translation to the existing DSO code
15133 that will convert (if the flag to do so is set) filenames
15134 that are sufficiently small and have no path information
15135 into a canonical native form. Eg. "blah" converted to
15136 "libblah.so" or "blah.dll" etc.
15140 * New function ERR_error_string_n(e, buf, len) which is like
15141 ERR_error_string(e, buf), but writes at most 'len' bytes
15142 including the 0 terminator. For ERR_error_string_n, 'buf'
15145 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15147 * CONF library reworked to become more general. A new CONF
15148 configuration file reader "class" is implemented as well as a
15149 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15150 old `CONF_*` functions are still there, but are reimplemented to
15151 work in terms of the new functions. Also, a set of functions
15152 to handle the internal storage of the configuration data is
15153 provided to make it easier to write new configuration file
15154 reader "classes" (I can definitely see something reading a
15155 configuration file in XML format, for example), called `_CONF_*`,
15156 or "the configuration storage API"...
15158 The new configuration file reading functions are:
15160 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15161 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15163 NCONF_default, NCONF_WIN32
15165 NCONF_dump_fp, NCONF_dump_bio
15167 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15168 NCONF_new creates a new CONF object. This works in the same way
15169 as other interfaces in OpenSSL, like the BIO interface.
15170 `NCONF_dump_*` dump the internal storage of the configuration file,
15171 which is useful for debugging. All other functions take the same
15172 arguments as the old `CONF_*` functions with the exception of the
15173 first that must be a `CONF *` instead of a `LHASH *`.
15175 To make it easier to use the new classes with the old `CONF_*` functions,
15176 the function CONF_set_default_method is provided.
15180 * Add '-tls1' option to 'openssl ciphers', which was already
15181 mentioned in the documentation but had not been implemented.
15182 (This option is not yet really useful because even the additional
15183 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15187 * Initial DSO code added into libcrypto for letting OpenSSL (and
15188 OpenSSL-based applications) load shared libraries and bind to
15189 them in a portable way.
15191 *Geoff Thorpe, with contributions from Richard Levitte*
15193 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15195 * Make sure _lrotl and _lrotr are only used with MSVC.
15197 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15198 (the default implementation of RAND_status).
15200 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15201 to '-clrext' (= clear extensions), as intended and documented.
15202 *Bodo Moeller; inconsistency pointed out by Michael Attili
15203 <attili@amaxo.com>*
15205 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15206 was larger than the MD block size.
15208 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15210 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15211 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15212 using the passed key: if the passed key was a private key the result
15213 of X509_print(), for example, would be to print out all the private key
15218 * des_quad_cksum() byte order bug fix.
15219 *Ulf Möller, using the problem description in krb4-0.9.7, where
15220 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15222 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15225 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15227 * For easily testing in shell scripts whether some command
15228 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15229 returns with exit code 0 iff no command of the given name is available.
15230 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15231 the output goes to stdout and nothing is printed to stderr.
15232 Additional arguments are always ignored.
15234 Since for each cipher there is a command of the same name,
15235 the 'no-cipher' compilation switches can be tested this way.
15237 ('openssl no-XXX' is not able to detect pseudo-commands such
15238 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15242 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15246 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15247 is set; it will be thrown away anyway because each handshake creates
15249 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15250 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15251 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15252 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15256 * New s_client option -ign_eof: EOF at stdin is ignored, and
15257 'Q' and 'R' lose their special meanings (quit/renegotiate).
15258 This is part of what -quiet does; unlike -quiet, -ign_eof
15259 does not suppress any output.
15263 * Add compatibility options to the purpose and trust code. The
15264 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15265 accepts a certificate or CA, this was the previous behaviour,
15266 with all the associated security issues.
15268 X509_TRUST_COMPAT is the old trust behaviour: only and
15269 automatically trust self signed roots in certificate store. A
15270 new trust setting X509_TRUST_DEFAULT is used to specify that
15271 a purpose has no associated trust setting and it should instead
15272 use the value in the default purpose.
15276 * Fix the PKCS#8 DSA private key code so it decodes keys again
15277 and fix a memory leak.
15281 * In util/mkerr.pl (which implements 'make errors'), preserve
15282 reason strings from the previous version of the .c file, as
15283 the default to have only downcase letters (and digits) in
15284 automatically generated reasons codes is not always appropriate.
15288 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15289 using strerror. Previously, ERR_reason_error_string() returned
15290 library names as reason strings for SYSerr; but SYSerr is a special
15291 case where small numbers are errno values, not library numbers.
15295 * Add '-dsaparam' option to 'openssl dhparam' application. This
15296 converts DSA parameters into DH parameters. (When creating parameters,
15297 DSA_generate_parameters is used.)
15301 * Include 'length' (recommended exponent length) in C code generated
15302 by 'openssl dhparam -C'.
15306 * The second argument to set_label in perlasm was already being used
15307 so couldn't be used as a "file scope" flag. Moved to third argument
15312 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15313 instead of RAND_bytes for encryption IVs and salts.
15317 * Include RAND_status() into RAND_METHOD instead of implementing
15318 it only for md_rand.c Otherwise replacing the PRNG by calling
15319 RAND_set_rand_method would be impossible.
15323 * Don't let DSA_generate_key() enter an infinite loop if the random
15324 number generation fails.
15328 * New 'rand' application for creating pseudo-random output.
15332 * Added configuration support for Linux/IA64
15334 *Rolf Haberrecker <rolf@suse.de>*
15336 * Assembler module support for Mingw32.
15340 * Shared library support for HPUX (in shlib/).
15342 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15344 * Shared library support for Solaris gcc.
15346 *Lutz Behnke <behnke@trustcenter.de>*
15348 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15350 * PKCS7_encrypt() was adding text MIME headers twice because they
15351 were added manually and by SMIME_crlf_copy().
15355 * In bntest.c don't call BN_rand with zero bits argument.
15357 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15359 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15360 case was implemented. This caused BN_div_recp() to fail occasionally.
15364 * Add an optional second argument to the set_label() in the perl
15365 assembly language builder. If this argument exists and is set
15366 to 1 it signals that the assembler should use a symbol whose
15367 scope is the entire file, not just the current function. This
15368 is needed with MASM which uses the format label:: for this scope.
15370 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15372 * Change the ASN1 types so they are typedefs by default. Before
15373 almost all types were #define'd to ASN1_STRING which was causing
15374 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15379 * Change names of new functions to the new get1/get0 naming
15380 convention: After 'get1', the caller owns a reference count
15381 and has to call `..._free`; 'get0' returns a pointer to some
15382 data structure without incrementing reference counters.
15383 (Some of the existing 'get' functions increment a reference
15384 counter, some don't.)
15385 Similarly, 'set1' and 'add1' functions increase reference
15386 counters or duplicate objects.
15390 * Allow for the possibility of temp RSA key generation failure:
15391 the code used to assume it always worked and crashed on failure.
15395 * Fix potential buffer overrun problem in BIO_printf().
15396 *Ulf Möller, using public domain code by Patrick Powell; problem
15397 pointed out by David Sacerdote <das33@cornell.edu>*
15399 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15400 RAND_egd() and RAND_status(). In the command line application,
15401 the EGD socket can be specified like a seed file using RANDFILE
15406 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15407 Some CAs (e.g. Verisign) distribute certificates in this form.
15411 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15412 list to exclude them. This means that no special compilation option
15413 is needed to use anonymous DH: it just needs to be included in the
15418 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15419 EVP_MD_type. The old functionality is available in a new macro called
15420 EVP_MD_md(). Change code that uses it and update docs.
15424 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15425 where the `void *` argument is replaced by a function pointer argument.
15426 Previously `void *` was abused to point to functions, which works on
15427 many platforms, but is not correct. As these functions are usually
15428 called by macros defined in OpenSSL header files, most source code
15429 should work without changes.
15433 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15434 sections with information on -D... compiler switches used for
15435 compiling the library so that applications can see them. To enable
15436 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15437 must be defined. E.g.,
15438 #define OPENSSL_ALGORITHM_DEFINES
15439 #include <openssl/opensslconf.h>
15440 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15442 *Richard Levitte, Ulf and Bodo Möller*
15444 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15449 * Change the 'other' type in certificate aux info to a STACK_OF
15450 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15451 the required ASN1 format: arbitrary types determined by an OID.
15455 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15456 argument to 'req'. This is not because the function is newer or
15457 better than others it just uses the work 'NEW' in the certificate
15458 request header lines. Some software needs this.
15462 * Reorganise password command line arguments: now passwords can be
15463 obtained from various sources. Delete the PEM_cb function and make
15464 it the default behaviour: i.e. if the callback is NULL and the
15465 usrdata argument is not NULL interpret it as a null terminated pass
15466 phrase. If usrdata and the callback are NULL then the pass phrase
15467 is prompted for as usual.
15471 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15472 the support is automatically enabled. The resulting binaries will
15473 autodetect the card and use it if present.
15475 *Ben Laurie and Compaq Inc.*
15477 * Work around for Netscape hang bug. This sends certificate request
15478 and server done in one record. Since this is perfectly legal in the
15479 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15480 the bugs/SSLv3 entry for more info.
15484 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15488 * Add -rand argument to smime and pkcs12 applications and read/write
15493 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15497 * Add command line password options to the remaining applications.
15501 * Bug fix for BN_div_recp() for numerators with an even number of
15506 * More tests in bntest.c, and changed test_bn output.
15510 * ./config recognizes MacOS X now.
15514 * Bug fix for BN_div() when the first words of num and divisor are
15515 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15519 * Add support for various broken PKCS#8 formats, and command line
15520 options to produce them.
15524 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15525 get temporary BIGNUMs from a BN_CTX.
15529 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15534 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15535 include a #define from the old name to the new. The original intent
15536 was that statically linked binaries could for example just call
15537 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15538 link with digests. This never worked because SSLeay_add_all_digests()
15539 and SSLeay_add_all_ciphers() were in the same source file so calling
15540 one would link with the other. They are now in separate source files.
15544 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15548 * Use a less unusual form of the Miller-Rabin primality test (it used
15549 a binary algorithm for exponentiation integrated into the Miller-Rabin
15550 loop, our standard modexp algorithms are faster).
15554 * Support for the EBCDIC character set completed.
15556 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15558 * Source code cleanups: use const where appropriate, eliminate casts,
15559 use `void *` instead of `char *` in lhash.
15563 * Bugfix: ssl3_send_server_key_exchange was not restartable
15564 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15565 this the server could overwrite ephemeral keys that the client
15570 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15571 using 50 iterations of the Rabin-Miller test.
15573 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15574 iterations of the Rabin-Miller test as required by the appendix
15575 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15576 As BN_is_prime_fasttest includes trial division, DSA parameter
15577 generation becomes much faster.
15579 This implies a change for the callback functions in DSA_is_prime
15580 and DSA_generate_parameters: The callback function is called once
15581 for each positive witness in the Rabin-Miller test, not just
15582 occasionally in the inner loop; and the parameters to the
15583 callback function now provide an iteration count for the outer
15584 loop rather than for the current invocation of the inner loop.
15585 DSA_generate_parameters additionally can call the callback
15586 function with an 'iteration count' of -1, meaning that a
15587 candidate has passed the trial division test (when q is generated
15588 from an application-provided seed, trial division is skipped).
15592 * New function BN_is_prime_fasttest that optionally does trial
15593 division before starting the Rabin-Miller test and has
15594 an additional BN_CTX * argument (whereas BN_is_prime always
15595 has to allocate at least one BN_CTX).
15596 'callback(1, -1, cb_arg)' is called when a number has passed the
15597 trial division stage.
15601 * Fix for bug in CRL encoding. The validity dates weren't being handled
15606 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15610 * New function BN_pseudo_rand().
15614 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15615 bignum version of BN_from_montgomery() with the working code from
15616 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15621 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15622 made it impossible to use the same SSL_SESSION data structure in
15623 SSL2 clients in multiple threads.
15627 * The return value of RAND_load_file() no longer counts bytes obtained
15628 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15629 to seed the PRNG (previously an explicit byte count was required).
15631 *Ulf Möller, Bodo Möller*
15633 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15634 used `char *` instead of `void *` and had casts all over the place.
15638 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15642 * Retain source code compatibility for BN_prime_checks macro:
15643 BN_is_prime(..., BN_prime_checks, ...) now uses
15644 BN_prime_checks_for_size to determine the appropriate number of
15645 Rabin-Miller iterations.
15649 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15650 DH_CHECK_P_NOT_SAFE_PRIME.
15651 (Check if this is true? OpenPGP calls them "strong".)
15655 * Merge the functionality of "dh" and "gendh" programs into a new program
15656 "dhparam". The old programs are retained for now but will handle DH keys
15657 (instead of parameters) in future.
15661 * Make the ciphers, s_server and s_client programs check the return values
15662 when a new cipher list is set.
15666 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15667 ciphers. Before when the 56bit ciphers were enabled the sorting was
15670 The syntax for the cipher sorting has been extended to support sorting by
15671 cipher-strength (using the strength_bits hard coded in the tables).
15672 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15674 Fix a bug in the cipher-command parser: when supplying a cipher command
15675 string with an "undefined" symbol (neither command nor alphanumeric
15676 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15677 an error is flagged.
15679 Due to the strength-sorting extension, the code of the
15680 ssl_create_cipher_list() function was completely rearranged. I hope that
15681 the readability was also increased :-)
15683 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15685 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15686 for the first serial number and places 2 in the serial number file. This
15687 avoids problems when the root CA is created with serial number zero and
15688 the first user certificate has the same issuer name and serial number
15693 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15694 the new code. Add documentation for this stuff.
15698 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15699 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15700 structures and behave in an analogous way to the X509v3 functions:
15701 they shouldn't be called directly but wrapper functions should be used
15704 So we also now have some wrapper functions that call the X509at functions
15705 when passed certificate requests. (TO DO: similar things can be done with
15706 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15707 things. Some of these need some d2i or i2d and print functionality
15708 because they handle more complex structures.)
15712 * Add missing #ifndefs that caused missing symbols when building libssl
15713 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15714 NO_RSA in `ssl/s2*.c`.
15716 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15718 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15719 has a return value which indicates the quality of the random data
15720 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15721 error queue. New function RAND_pseudo_bytes() generates output that is
15722 guaranteed to be unique but not unpredictable. RAND_add is like
15723 RAND_seed, but takes an extra argument for an entropy estimate
15724 (RAND_seed always assumes full entropy).
15728 * Do more iterations of Rabin-Miller probable prime test (specifically,
15729 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15730 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15731 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15732 false-positive rate of at most 2^-80 for random input.
15736 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15740 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15741 in the 0.9.5 release), this returns the chain
15742 from an X509_CTX structure with a dup of the stack and all
15743 the X509 reference counts upped: so the stack will exist
15744 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15747 Also make SSL_SESSION_print() print out the verify return
15752 * Add manpage for the pkcs12 command. Also change the default
15753 behaviour so MAC iteration counts are used unless the new
15754 -nomaciter option is used. This improves file security and
15755 only older versions of MSIE (4.0 for example) need it.
15759 * Honor the no-xxx Configure options when creating .DEF files.
15763 * Add PKCS#10 attributes to field table: challengePassword,
15764 unstructuredName and unstructuredAddress. These are taken from
15765 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15766 international characters are used.
15768 More changes to X509_ATTRIBUTE code: allow the setting of types
15769 based on strings. Remove the 'loc' parameter when adding
15770 attributes because these will be a SET OF encoding which is sorted
15775 * Initial changes to the 'req' utility to allow request generation
15776 automation. This will allow an application to just generate a template
15777 file containing all the field values and have req construct the
15780 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15781 used all over the place including certificate requests and PKCS#7
15782 structures. They are currently handled manually where necessary with
15783 some primitive wrappers for PKCS#7. The new functions behave in a
15784 manner analogous to the X509 extension functions: they allow
15785 attributes to be looked up by NID and added.
15787 Later something similar to the X509V3 code would be desirable to
15788 automatically handle the encoding, decoding and printing of the
15789 more complex types. The string types like challengePassword can
15790 be handled by the string table functions.
15792 Also modified the multi byte string table handling. Now there is
15793 a 'global mask' which masks out certain types. The table itself
15794 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15795 is useful when for example there is only one permissible type
15796 (as in countryName) and using the mask might result in no valid
15801 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15802 SSL_get_peer_finished to allow applications to obtain the latest
15803 Finished messages sent to the peer or expected from the peer,
15804 respectively. (SSL_get_peer_finished is usually the Finished message
15805 actually received from the peer, otherwise the protocol will be aborted.)
15807 As the Finished message are message digests of the complete handshake
15808 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15809 be used for external authentication procedures when the authentication
15810 provided by SSL/TLS is not desired or is not enough.
15814 * Enhanced support for Alpha Linux is added. Now ./config checks if
15815 the host supports BWX extension and if Compaq C is present on the
15816 $PATH. Just exploiting of the BWX extension results in 20-30%
15817 performance kick for some algorithms, e.g. DES and RC4 to mention
15818 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15823 * Add support for MS "fast SGC". This is arguably a violation of the
15824 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15825 weak crypto and after checking the certificate is SGC a second one
15826 with strong crypto. MS SGC stops the first handshake after receiving
15827 the server certificate message and sends a second client hello. Since
15828 a server will typically do all the time consuming operations before
15829 expecting any further messages from the client (server key exchange
15830 is the most expensive) there is little difference between the two.
15832 To get OpenSSL to support MS SGC we have to permit a second client
15833 hello message after we have sent server done. In addition we have to
15834 reset the MAC if we do get this second client hello.
15838 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15839 if a DER encoded private key is RSA or DSA traditional format. Changed
15840 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15841 format DER encoded private key. Newer code should use PKCS#8 format which
15842 has the key type encoded in the ASN1 structure. Added DER private key
15843 support to pkcs8 application.
15847 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15848 ciphersuites has been selected (as required by the SSL 3/TLS 1
15849 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15850 is set, we interpret this as a request to violate the specification
15851 (the worst that can happen is a handshake failure, and 'correct'
15852 behaviour would result in a handshake failure anyway).
15856 * In SSL_CTX_add_session, take into account that there might be multiple
15857 SSL_SESSION structures with the same session ID (e.g. when two threads
15858 concurrently obtain them from an external cache).
15859 The internal cache can handle only one SSL_SESSION with a given ID,
15860 so if there's a conflict, we now throw out the old one to achieve
15865 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15866 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15867 some routines that use cipher OIDs: some ciphers do not have OIDs
15868 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15873 * Simplify the trust setting structure and code. Now we just have
15874 two sequences of OIDs for trusted and rejected settings. These will
15875 typically have values the same as the extended key usage extension
15876 and any application specific purposes.
15878 The trust checking code now has a default behaviour: it will just
15879 check for an object with the same NID as the passed id. Functions can
15880 be provided to override either the default behaviour or the behaviour
15881 for a given id. SSL client, server and email already have functions
15882 in place for compatibility: they check the NID and also return "trusted"
15883 if the certificate is self signed.
15887 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15888 traditional format into an EVP_PKEY structure.
15892 * Add a password callback function PEM_cb() which either prompts for
15893 a password if usr_data is NULL or otherwise assumes it is a null
15894 terminated password. Allow passwords to be passed on command line
15895 environment or config files in a few more utilities.
15899 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15900 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15901 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15902 Update documentation.
15906 * Support for ASN1 "NULL" type. This could be handled before by using
15907 ASN1_TYPE but there wasn't any function that would try to read a NULL
15908 and produce an error if it couldn't. For compatibility we also have
15909 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15910 don't allocate anything because they don't need to.
15914 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15917 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15919 * Rebuild of the memory allocation routines used by OpenSSL code and
15920 possibly others as well. The purpose is to make an interface that
15921 provide hooks so anyone can build a separate set of allocation and
15922 deallocation routines to be used by OpenSSL, for example memory
15923 pool implementations, or something else, which was previously hard
15924 since Malloc(), Realloc() and Free() were defined as macros having
15925 the values malloc, realloc and free, respectively (except for Win32
15926 compilations). The same is provided for memory debugging code.
15927 OpenSSL already comes with functionality to find memory leaks, but
15928 this gives people a chance to debug other memory problems.
15930 With these changes, a new set of functions and macros have appeared:
15932 CRYPTO_set_mem_debug_functions() [F]
15933 CRYPTO_get_mem_debug_functions() [F]
15934 CRYPTO_dbg_set_options() [F]
15935 CRYPTO_dbg_get_options() [F]
15936 CRYPTO_malloc_debug_init() [M]
15938 The memory debug functions are NULL by default, unless the library
15939 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15940 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15941 gives the standard debugging functions that come with OpenSSL) or
15942 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15943 provided by the library user) must be used. When the standard
15944 debugging functions are used, CRYPTO_dbg_set_options can be used to
15945 request additional information:
15946 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15947 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15949 Also, things like CRYPTO_set_mem_functions will always give the
15950 expected result (the new set of functions is used for allocation
15951 and deallocation) at all times, regardless of platform and compiler
15954 To finish it up, some functions that were never use in any other
15955 way than through macros have a new API and new semantic:
15957 CRYPTO_dbg_malloc()
15958 CRYPTO_dbg_realloc()
15961 All macros of value have retained their old syntax.
15963 *Richard Levitte and Bodo Moeller*
15965 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15966 ordering of SMIMECapabilities wasn't in "strength order" and there
15967 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15972 * Some ASN1 types with illegal zero length encoding (INTEGER,
15973 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15975 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15977 * Merge in my S/MIME library for OpenSSL. This provides a simple
15978 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15979 functionality to handle multipart/signed properly) and a utility
15980 called 'smime' to call all this stuff. This is based on code I
15981 originally wrote for Celo who have kindly allowed it to be
15982 included in OpenSSL.
15986 * Add variants des_set_key_checked and des_set_key_unchecked of
15987 des_set_key (aka des_key_sched). Global variable des_check_key
15988 decides which of these is called by des_set_key; this way
15989 des_check_key behaves as it always did, but applications and
15990 the library itself, which was buggy for des_check_key == 1,
15991 have a cleaner way to pick the version they need.
15995 * New function PKCS12_newpass() which changes the password of a
16000 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16001 dynamic mix. In both cases the ids can be used as an index into the
16002 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16003 functions so they accept a list of the field values and the
16004 application doesn't need to directly manipulate the X509_TRUST
16009 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16014 * Modify the way the V3 extension code looks up extensions. This now
16015 works in a similar way to the object code: we have some "standard"
16016 extensions in a static table which is searched with OBJ_bsearch()
16017 and the application can add dynamic ones if needed. The file
16018 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16019 updated whenever a new extension is added to the core code and kept
16020 in ext_nid order. There is a simple program 'tabtest.c' which checks
16021 this. New extensions are not added too often so this file can readily
16022 be maintained manually.
16024 There are two big advantages in doing things this way. The extensions
16025 can be looked up immediately and no longer need to be "added" using
16026 X509V3_add_standard_extensions(): this function now does nothing.
16027 Side note: I get *lots* of email saying the extension code doesn't
16028 work because people forget to call this function.
16029 Also no dynamic allocation is done unless new extensions are added:
16030 so if we don't add custom extensions there is no need to call
16031 X509V3_EXT_cleanup().
16035 * Modify enc utility's salting as follows: make salting the default. Add a
16036 magic header, so unsalted files fail gracefully instead of just decrypting
16037 to garbage. This is because not salting is a big security hole, so people
16038 should be discouraged from doing it.
16042 * Fixes and enhancements to the 'x509' utility. It allowed a message
16043 digest to be passed on the command line but it only used this
16044 parameter when signing a certificate. Modified so all relevant
16045 operations are affected by the digest parameter including the
16046 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16047 DSA key was used because it didn't fix the digest.
16051 * Initial certificate chain verify code. Currently tests the untrusted
16052 certificates for consistency with the verify purpose (which is set
16053 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16055 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16056 this is because it will reject chains with invalid extensions whereas
16057 every previous version of OpenSSL and SSLeay made no checks at all.
16059 Trust code: checks the root CA for the relevant trust settings. Trust
16060 settings have an initial value consistent with the verify purpose: e.g.
16061 if the verify purpose is for SSL client use it expects the CA to be
16062 trusted for SSL client use. However the default value can be changed to
16063 permit custom trust settings: one example of this would be to only trust
16064 certificates from a specific "secure" set of CAs.
16066 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16067 which should be used for version portability: especially since the
16068 verify structure is likely to change more often now.
16070 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16071 to set them. If not set then assume SSL clients will verify SSL servers
16074 Two new options to the verify program: -untrusted allows a set of
16075 untrusted certificates to be passed in and -purpose which sets the
16076 intended purpose of the certificate. If a purpose is set then the
16077 new chain verify code is used to check extension consistency.
16081 * Support for the authority information access extension.
16085 * Modify RSA and DSA PEM read routines to transparently handle
16086 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16087 public keys in a format compatible with certificate
16088 SubjectPublicKeyInfo structures. Unfortunately there were already
16089 functions called *_PublicKey_* which used various odd formats so
16090 these are retained for compatibility: however the DSA variants were
16091 never in a public release so they have been deleted. Changed dsa/rsa
16092 utilities to handle the new format: note no releases ever handled public
16093 keys so we should be OK.
16095 The primary motivation for this change is to avoid the same fiasco
16096 that dogs private keys: there are several incompatible private key
16097 formats some of which are standard and some OpenSSL specific and
16098 require various evil hacks to allow partial transparent handling and
16099 even then it doesn't work with DER formats. Given the option anything
16100 other than PKCS#8 should be dumped: but the other formats have to
16101 stay in the name of compatibility.
16103 With public keys and the benefit of hindsight one standard format
16104 is used which works with EVP_PKEY, RSA or DSA structures: though
16105 it clearly returns an error if you try to read the wrong kind of key.
16107 Added a -pubkey option to the 'x509' utility to output the public key.
16108 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16109 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16110 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16111 that do the same as the `EVP_PKEY_assign_*()` except they up the
16112 reference count of the added key (they don't "swallow" the
16117 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16118 CRLs would fail if the file contained no certificates or no CRLs:
16119 added a new function to read in both types and return the number
16120 read: this means that if none are read it will be an error. The
16121 DER versions of the certificate and CRL reader would always fail
16122 because it isn't possible to mix certificates and CRLs in DER format
16123 without choking one or the other routine. Changed this to just read
16124 a certificate: this is the best we can do. Also modified the code
16125 in apps/verify.c to take notice of return codes: it was previously
16126 attempting to read in certificates from NULL pointers and ignoring
16127 any errors: this is one reason why the cert and CRL reader seemed
16128 to work. It doesn't check return codes from the default certificate
16129 routines: these may well fail if the certificates aren't installed.
16133 * Code to support otherName option in GeneralName.
16137 * First update to verify code. Change the verify utility
16138 so it warns if it is passed a self signed certificate:
16139 for consistency with the normal behaviour. X509_verify
16140 has been modified to it will now verify a self signed
16141 certificate if *exactly* the same certificate appears
16142 in the store: it was previously impossible to trust a
16143 single self signed certificate. This means that:
16144 openssl verify ss.pem
16145 now gives a warning about a self signed certificate but
16146 openssl verify -CAfile ss.pem ss.pem
16151 * For servers, store verify_result in SSL_SESSION data structure
16152 (and add it to external session representation).
16153 This is needed when client certificate verifications fails,
16154 but an application-provided verification callback (set by
16155 SSL_CTX_set_cert_verify_callback) allows accepting the session
16156 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16157 but returns 1): When the session is reused, we have to set
16158 ssl->verify_result to the appropriate error code to avoid
16161 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16163 * Fix a bug in the new PKCS#7 code: it didn't consider the
16164 case in PKCS7_dataInit() where the signed PKCS7 structure
16165 didn't contain any existing data because it was being created.
16167 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16169 * Add a salt to the key derivation routines in enc.c. This
16170 forms the first 8 bytes of the encrypted file. Also add a
16171 -S option to allow a salt to be input on the command line.
16175 * New function X509_cmp(). Oddly enough there wasn't a function
16176 to compare two certificates. We do this by working out the SHA1
16177 hash and comparing that. X509_cmp() will be needed by the trust
16182 * SSL_get1_session() is like SSL_get_session(), but increments
16183 the reference count in the SSL_SESSION returned.
16185 *Geoff Thorpe <geoff@eu.c2.net>*
16187 * Fix for 'req': it was adding a null to request attributes.
16188 Also change the X509_LOOKUP and X509_INFO code to handle
16189 certificate auxiliary information.
16193 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16198 * Add the possibility to add extra information to the memory leak
16199 detecting output, to form tracebacks, showing from where each
16200 allocation was originated: CRYPTO_push_info("constant string") adds
16201 the string plus current file name and line number to a per-thread
16202 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16203 is like calling CYRPTO_pop_info() until the stack is empty.
16204 Also updated memory leak detection code to be multi-thread-safe.
16208 * Add options -text and -noout to pkcs7 utility and delete the
16209 encryption options which never did anything. Update docs.
16213 * Add options to some of the utilities to allow the pass phrase
16214 to be included on either the command line (not recommended on
16215 OSes like Unix) or read from the environment. Update the
16216 manpages and fix a few bugs.
16220 * Add a few manpages for some of the openssl commands.
16224 * Fix the -revoke option in ca. It was freeing up memory twice,
16225 leaking and not finding already revoked certificates.
16229 * Extensive changes to support certificate auxiliary information.
16230 This involves the use of X509_CERT_AUX structure and X509_AUX
16231 functions. An X509_AUX function such as PEM_read_X509_AUX()
16232 can still read in a certificate file in the usual way but it
16233 will also read in any additional "auxiliary information". By
16234 doing things this way a fair degree of compatibility can be
16235 retained: existing certificates can have this information added
16236 using the new 'x509' options.
16238 Current auxiliary information includes an "alias" and some trust
16239 settings. The trust settings will ultimately be used in enhanced
16240 certificate chain verification routines: currently a certificate
16241 can only be trusted if it is self signed and then it is trusted
16246 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16247 The problem was that one of the replacement routines had not been working
16248 since SSLeay releases. For now the offending routine has been replaced
16249 with non-optimised assembler. Even so, this now gives around 95%
16250 performance improvement for 1024 bit RSA signs.
16254 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16255 handling. Most clients have the effective key size in bits equal to
16256 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16257 A few however don't do this and instead use the size of the decrypted key
16258 to determine the RC2 key length and the AlgorithmIdentifier to determine
16259 the effective key length. In this case the effective key length can still
16260 be 40 bits but the key length can be 168 bits for example. This is fixed
16261 by manually forcing an RC2 key into the EVP_PKEY structure because the
16262 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16263 the key length and effective key length are equal.
16267 * Add a bunch of functions that should simplify the creation of
16268 X509_NAME structures. Now you should be able to do:
16269 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16270 and have it automatically work out the correct field type and fill in
16271 the structures. The more adventurous can try:
16272 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16273 and it will (hopefully) work out the correct multibyte encoding.
16277 * Change the 'req' utility to use the new field handling and multibyte
16278 copy routines. Before the DN field creation was handled in an ad hoc
16279 way in req, ca, and x509 which was rather broken and didn't support
16280 BMPStrings or UTF8Strings. Since some software doesn't implement
16281 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16282 using the dirstring_type option. See the new comment in the default
16283 openssl.cnf for more info.
16287 * Make crypto/rand/md_rand.c more robust:
16288 - Assure unique random numbers after fork().
16289 - Make sure that concurrent threads access the global counter and
16290 md serializably so that we never lose entropy in them
16291 or use exactly the same state in multiple threads.
16292 Access to the large state is not always serializable because
16293 the additional locking could be a performance killer, and
16294 md should be large enough anyway.
16298 * New file apps/app_rand.c with commonly needed functionality
16299 for handling the random seed file.
16301 Use the random seed file in some applications that previously did not:
16303 dsaparam -genkey (which also ignored its '-rand' option),
16306 x509 (when signing).
16307 Except on systems with /dev/urandom, it is crucial to have a random
16308 seed file at least for key creation, DSA signing, and for DH exchanges;
16309 for RSA signatures we could do without one.
16311 gendh and gendsa (unlike genrsa) used to read only the first byte
16312 of each file listed in the '-rand' option. The function as previously
16313 found in genrsa is now in app_rand.c and is used by all programs
16314 that support '-rand'.
16318 * In RAND_write_file, use mode 0600 for creating files;
16319 don't just chmod when it may be too late.
16323 * Report an error from X509_STORE_load_locations
16324 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16328 * New function ASN1_mbstring_copy() this copies a string in either
16329 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16330 into an ASN1_STRING type. A mask of permissible types is passed
16331 and it chooses the "minimal" type to use or an error if not type
16336 * Add function equivalents to the various macros in asn1.h. The old
16337 macros are retained with an `M_` prefix. Code inside the library can
16338 use the `M_` macros. External code (including the openssl utility)
16339 should *NOT* in order to be "shared library friendly".
16343 * Add various functions that can check a certificate's extensions
16344 to see if it usable for various purposes such as SSL client,
16345 server or S/MIME and CAs of these types. This is currently
16346 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16347 verification. Also added a -purpose flag to x509 utility to
16348 print out all the purposes.
16352 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16357 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16358 for, obtain and decode and extension and obtain its critical flag.
16359 This allows all the necessary extension code to be handled in a
16360 single function call.
16364 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16365 platforms. See crypto/rc4/rc4_enc.c for further details.
16369 * New -noout option to asn1parse. This causes no output to be produced
16370 its main use is when combined with -strparse and -out to extract data
16371 from a file (which may not be in ASN.1 format).
16375 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16376 when producing the local key id.
16378 *Richard Levitte <levitte@stacken.kth.se>*
16380 * New option -dhparam in s_server. This allows a DH parameter file to be
16381 stated explicitly. If it is not stated then it tries the first server
16382 certificate file. The previous behaviour hard coded the filename
16387 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16388 a public key to be input or output. For example:
16389 openssl rsa -in key.pem -pubout -out pubkey.pem
16390 Also added necessary DSA public key functions to handle this.
16394 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16395 in the message. This was handled by allowing
16396 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16398 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16400 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16401 to the end of the strings whereas this didn't. This would cause problems
16402 if strings read with d2i_ASN1_bytes() were later modified.
16404 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16406 * Fix for base64 decode bug. When a base64 bio reads only one line of
16407 data and it contains EOF it will end up returning an error. This is
16408 caused by input 46 bytes long. The cause is due to the way base64
16409 BIOs find the start of base64 encoded data. They do this by trying a
16410 trial decode on each line until they find one that works. When they
16411 do a flag is set and it starts again knowing it can pass all the
16412 data directly through the decoder. Unfortunately it doesn't reset
16413 the context it uses. This means that if EOF is reached an attempt
16414 is made to pass two EOFs through the context and this causes the
16415 resulting error. This can also cause other problems as well. As is
16416 usual with these problems it takes *ages* to find and the fix is
16417 trivial: move one line.
16419 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16421 * Ugly workaround to get s_client and s_server working under Windows. The
16422 old code wouldn't work because it needed to select() on sockets and the
16423 tty (for keypresses and to see if data could be written). Win32 only
16424 supports select() on sockets so we select() with a 1s timeout on the
16425 sockets and then see if any characters are waiting to be read, if none
16426 are present then we retry, we also assume we can always write data to
16427 the tty. This isn't nice because the code then blocks until we've
16428 received a complete line of data and it is effectively polling the
16429 keyboard at 1s intervals: however it's quite a bit better than not
16430 working at all :-) A dedicated Windows application might handle this
16431 with an event loop for example.
16435 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16436 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16437 will be called when RSA_sign() and RSA_verify() are used. This is useful
16438 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16439 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16440 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16441 This necessitated the support of an extra signature type NID_md5_sha1
16442 for SSL signatures and modifications to the SSL library to use it instead
16443 of calling RSA_public_decrypt() and RSA_private_encrypt().
16447 * Add new -verify -CAfile and -CApath options to the crl program, these
16448 will lookup a CRL issuers certificate and verify the signature in a
16449 similar way to the verify program. Tidy up the crl program so it
16450 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16451 less strict. It will now permit CRL extensions even if it is not
16452 a V2 CRL: this will allow it to tolerate some broken CRLs.
16456 * Initialize all non-automatic variables each time one of the openssl
16457 sub-programs is started (this is necessary as they may be started
16458 multiple times from the "OpenSSL>" prompt).
16460 *Lennart Bang, Bodo Moeller*
16462 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16463 removing all other RSA functionality (this is what NO_RSA does). This
16464 is so (for example) those in the US can disable those operations covered
16465 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16470 * Non-copying interface to BIO pairs.
16471 (still largely untested)
16475 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16476 ASCII string. This was handled independently in various places before.
16480 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16481 UTF8 strings a character at a time.
16485 * Use client_version from client hello to select the protocol
16486 (s23_srvr.c) and for RSA client key exchange verification
16487 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16491 * Add various utility functions to handle SPKACs, these were previously
16492 handled by poking round in the structure internals. Added new function
16493 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16494 print, verify and generate SPKACs. Based on an original idea from
16495 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16499 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16503 * Allow the config file extension section to be overwritten on the
16504 command line. Based on an original idea from Massimiliano Pala
16505 <madwolf@comune.modena.it>. The new option is called -extensions
16506 and can be applied to ca, req and x509. Also -reqexts to override
16507 the request extensions in req and -crlexts to override the crl extensions
16512 * Add new feature to the SPKAC handling in ca. Now you can include
16513 the same field multiple times by preceding it by "XXXX." for example:
16516 this is the same syntax as used in the req config file.
16520 * Allow certificate extensions to be added to certificate requests. These
16521 are specified in a 'req_extensions' option of the req section of the
16522 config file. They can be printed out with the -text option to req but
16523 are otherwise ignored at present.
16527 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16528 data read consists of only the final block it would not decrypted because
16529 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16530 A misplaced 'break' also meant the decrypted final block might not be
16531 copied until the next read.
16535 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16536 a few extra parameters to the DH structure: these will be useful if
16537 for example we want the value of 'q' or implement X9.42 DH.
16541 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16542 provides hooks that allow the default DSA functions or functions on a
16543 "per key" basis to be replaced. This allows hardware acceleration and
16544 hardware key storage to be handled without major modification to the
16545 library. Also added low level modexp hooks and CRYPTO_EX structure and
16546 associated functions.
16550 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16551 as "read only": it can't be written to and the buffer it points to will
16552 not be freed. Reading from a read only BIO is much more efficient than
16553 a normal memory BIO. This was added because there are several times when
16554 an area of memory needs to be read from a BIO. The previous method was
16555 to create a memory BIO and write the data to it, this results in two
16556 copies of the data and an O(n^2) reading algorithm. There is a new
16557 function BIO_new_mem_buf() which creates a read only memory BIO from
16558 an area of memory. Also modified the PKCS#7 routines to use read only
16563 * Bugfix: ssl23_get_client_hello did not work properly when called in
16564 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16565 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16566 but a retry condition occurred while trying to read the rest.
16570 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16571 NID_pkcs7_encrypted by default: this was wrong since this should almost
16572 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16573 the encrypted data type: this is a more sensible place to put it and it
16574 allows the PKCS#12 code to be tidied up that duplicated this
16579 * Changed obj_dat.pl script so it takes its input and output files on
16580 the command line. This should avoid shell escape redirection problems
16585 * Initial support for certificate extension requests, these are included
16586 in things like Xenroll certificate requests. Included functions to allow
16587 extensions to be obtained and added.
16591 * -crlf option to s_client and s_server for sending newlines as
16592 CRLF (as required by many protocols).
16596 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16598 * Install libRSAglue.a when OpenSSL is built with RSAref.
16600 *Ralf S. Engelschall*
16602 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16604 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16606 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16611 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16612 DH parameters/keys (q is lost during that conversion, but the resulting
16613 DH parameters contain its length).
16615 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16616 much faster than DH_generate_parameters (which creates parameters
16617 where `p = 2*q + 1`), and also the smaller q makes DH computations
16618 much more efficient (160-bit exponentiation instead of 1024-bit
16619 exponentiation); so this provides a convenient way to support DHE
16620 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16621 utter importance to use
16622 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16624 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16625 when such DH parameters are used, because otherwise small subgroup
16626 attacks may become possible!
16630 * Avoid memory leak in i2d_DHparams.
16634 * Allow the -k option to be used more than once in the enc program:
16635 this allows the same encrypted message to be read by multiple recipients.
16639 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16640 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16641 it will always use the numerical form of the OID, even if it has a short
16646 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16647 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16648 otherwise bn_mod_exp was called. In the case of hardware keys for example
16649 no private key components need be present and it might store extra data
16650 in the RSA structure, which cannot be accessed from bn_mod_exp.
16651 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16652 private key operations.
16656 * Added support for SPARC Linux.
16660 * pem_password_cb function type incompatibly changed from
16661 typedef int pem_password_cb(char *buf, int size, int rwflag);
16663 ....(char *buf, int size, int rwflag, void *userdata);
16664 so that applications can pass data to their callbacks:
16665 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16666 additional void * argument, which is just handed through whenever
16667 the password callback is called.
16669 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16671 New function SSL_CTX_set_default_passwd_cb_userdata.
16673 Compatibility note: As many C implementations push function arguments
16674 onto the stack in reverse order, the new library version is likely to
16675 interoperate with programs that have been compiled with the old
16676 pem_password_cb definition (PEM_whatever takes some data that
16677 happens to be on the stack as its last argument, and the callback
16678 just ignores this garbage); but there is no guarantee whatsoever that
16681 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16682 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16683 problems not only on Windows, but also on some Unix platforms.
16684 To avoid problematic command lines, these definitions are now in an
16685 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16686 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16690 * MIPS III/IV assembler module is reimplemented.
16694 * More DES library cleanups: remove references to srand/rand and
16695 delete an unused file.
16699 * Add support for the free Netwide assembler (NASM) under Win32,
16700 since not many people have MASM (ml) and it can be hard to obtain.
16701 This is currently experimental but it seems to work OK and pass all
16702 the tests. Check out INSTALL.W32 for info.
16706 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16707 without temporary keys kept an extra copy of the server key,
16708 and connections with temporary keys did not free everything in case
16713 * New function RSA_check_key and new openssl rsa option -check
16714 for verifying the consistency of RSA keys.
16716 *Ulf Moeller, Bodo Moeller*
16718 * Various changes to make Win32 compile work:
16719 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16720 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16721 comparison" warnings.
16722 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16726 * Add a debugging option to PKCS#5 v2 key generation function: when
16727 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16728 derived keys are printed to stderr.
16732 * Copy the flags in ASN1_STRING_dup().
16734 *Roman E. Pavlov <pre@mo.msk.ru>*
16736 * The x509 application mishandled signing requests containing DSA
16737 keys when the signing key was also DSA and the parameters didn't match.
16739 It was supposed to omit the parameters when they matched the signing key:
16740 the verifying software was then supposed to automatically use the CA's
16741 parameters if they were absent from the end user certificate.
16743 Omitting parameters is no longer recommended. The test was also
16744 the wrong way round! This was probably due to unusual behaviour in
16745 EVP_cmp_parameters() which returns 1 if the parameters match.
16746 This meant that parameters were omitted when they *didn't* match and
16747 the certificate was useless. Certificates signed with 'ca' didn't have
16750 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16752 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16753 The interface is as follows:
16754 Applications can use
16755 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16756 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16757 "off" is now the default.
16758 The library internally uses
16759 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16760 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16761 to disable memory-checking temporarily.
16763 Some inconsistent states that previously were possible (and were
16764 even the default) are now avoided.
16766 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16767 with each memory chunk allocated; this is occasionally more helpful
16768 than just having a counter.
16770 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16772 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16777 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16778 which largely parallels "options", but is for changing API behaviour,
16779 whereas "options" are about protocol behaviour.
16780 Initial "mode" flags are:
16782 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16783 a single record has been written.
16784 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16785 retries use the same buffer location.
16786 (But all of the contents must be
16791 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16794 * Fix problems with no-hmac etc.
16796 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16798 * New functions RSA_get_default_method(), RSA_set_method() and
16799 RSA_get_method(). These allows replacement of RSA_METHODs without having
16800 to mess around with the internals of an RSA structure.
16804 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16805 Also really enable memory leak checks in openssl.c and in some
16808 *Chad C. Mulligan, Bodo Moeller*
16810 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16811 up the length of negative integers. This has now been simplified to just
16812 store the length when it is first determined and use it later, rather
16813 than trying to keep track of where data is copied and updating it to
16815 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16817 * Add a new function PKCS7_signatureVerify. This allows the verification
16818 of a PKCS#7 signature but with the signing certificate passed to the
16819 function itself. This contrasts with PKCS7_dataVerify which assumes the
16820 certificate is present in the PKCS#7 structure. This isn't always the
16821 case: certificates can be omitted from a PKCS#7 structure and be
16822 distributed by "out of band" means (such as a certificate database).
16826 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16827 function prototypes in pem.h, also change util/mkdef.pl to add the
16828 necessary function names.
16832 * mk1mf.pl (used by Windows builds) did not properly read the
16833 options set by Configure in the top level Makefile, and Configure
16834 was not even able to write more than one option correctly.
16835 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16839 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16840 file to be loaded from a BIO or FILE pointer. The BIO version will
16841 for example allow memory BIOs to contain config info.
16845 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16846 Whoever hopes to achieve shared-library compatibility across versions
16847 must use this, not the compile-time macro.
16848 (Exercise 0.9.4: Which is the minimum library version required by
16850 Note: All this applies only to multi-threaded programs, others don't
16855 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16856 through a BIO pair triggered the default case, i.e.
16857 SSLerr(...,SSL_R_UNKNOWN_STATE).
16861 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16862 can use the SSL library even if none of the specific BIOs is
16867 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16868 for the encoded length.
16870 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16872 * Add initial documentation of the X509V3 functions.
16876 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16877 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16878 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16879 secure PKCS#8 private key format with a high iteration count.
16883 * Fix determination of Perl interpreter: A perl or perl5
16884 *directory* in $PATH was also accepted as the interpreter.
16886 *Ralf S. Engelschall*
16888 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16889 wrong with it but it was very old and did things like calling
16890 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16891 unusual formatting.
16895 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16896 to use the new extension code.
16900 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16901 with macros. This should make it easier to change their form, add extra
16902 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16907 * Add to configuration table a new entry that can specify an alternative
16908 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16909 according to Mark Crispin <MRC@Panda.COM>.
16913 * DES CBC did not update the IV. Weird.
16917 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16918 Changing the behaviour of the former might break existing programs --
16919 where IV updating is needed, des_ncbc_encrypt can be used.
16922 * When bntest is run from "make test" it drives bc to check its
16923 calculations, as well as internally checking them. If an internal check
16924 fails, it needs to cause bc to give a non-zero result or make test carries
16925 on without noticing the failure. Fixed.
16929 * DES library cleanups.
16933 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16934 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16935 ciphers. NOTE: although the key derivation function has been verified
16936 against some published test vectors it has not been extensively tested
16937 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16942 * Instead of "mkdir -p", which is not fully portable, use new
16943 Perl script "util/mkdir-p.pl".
16947 * Rewrite the way password based encryption (PBE) is handled. It used to
16948 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16949 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16950 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16951 the 'parameter' field of the AlgorithmIdentifier is passed to the
16952 underlying key generation function so it must do its own ASN1 parsing.
16953 This has also changed the EVP_PBE_CipherInit() function which now has a
16954 'parameter' argument instead of literal salt and iteration count values
16955 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16959 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16960 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16961 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16962 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16963 value was just used as a "magic string" and not used directly its
16964 value doesn't matter.
16968 * Introduce some semblance of const correctness to BN. Shame C doesn't
16973 * "linux-sparc64" configuration (ultrapenguin).
16975 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16976 "linux-sparc" configuration.
16978 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16980 * config now generates no-xxx options for missing ciphers.
16984 * Support the EBCDIC character set (work in progress).
16985 File ebcdic.c not yet included because it has a different license.
16987 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16989 * Support BS2000/OSD-POSIX.
16991 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16993 * Make callbacks for key generation use `void *` instead of `char *`.
16997 * Make S/MIME samples compile (not yet tested).
17001 * Additional typesafe stacks.
17005 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17009 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17011 * New configuration variant "sco5-gcc".
17013 * Updated some demos.
17015 *Sean O Riordain, Wade Scholine*
17017 * Add missing BIO_free at exit of pkcs12 application.
17021 * Fix memory leak in conf.c.
17025 * Updates for Win32 to assembler version of MD5.
17029 * Set #! path to perl in apps/der_chop to where we found it
17030 instead of using a fixed path.
17034 * SHA library changes for irix64-mips4-cc.
17038 * Improvements for VMS support.
17042 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17044 * Bignum library bug fix. IRIX 6 passes "make test" now!
17045 This also avoids the problems with SC4.2 and unpatched SC5.
17047 *Andy Polyakov <appro@fy.chalmers.se>*
17049 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17050 These are required because of the typesafe stack would otherwise break
17051 existing code. If old code used a structure member which used to be STACK
17052 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17053 sk_num or sk_value it would produce an error because the num, data members
17054 are not present in STACK_OF. Now it just produces a warning. sk_set
17055 replaces the old method of assigning a value to sk_value
17056 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17057 that does this will no longer work (and should use sk_set instead) but
17058 this could be regarded as a "questionable" behaviour anyway.
17062 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17063 correctly handle encrypted S/MIME data.
17067 * Change type of various DES function arguments from des_cblock
17068 (which means, in function argument declarations, pointer to char)
17069 to des_cblock * (meaning pointer to array with 8 char elements),
17070 which allows the compiler to do more typechecking; it was like
17071 that back in SSLeay, but with lots of ugly casts.
17073 Introduce new type const_des_cblock.
17077 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17078 problems: find RecipientInfo structure that matches recipient certificate
17079 and initialise the ASN1 structures properly based on passed cipher.
17083 * Belatedly make the BN tests actually check the results.
17087 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17088 to and from BNs: it was completely broken. New compilation option
17089 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17090 key elements as negative integers.
17094 * Reorganize and speed up MD5.
17096 *Andy Polyakov <appro@fy.chalmers.se>*
17100 *Richard Levitte <richard@levitte.org>*
17102 * New option -out to asn1parse to allow the parsed structure to be
17103 output to a file. This is most useful when combined with the -strparse
17104 option to examine the output of things like OCTET STRINGS.
17108 * Make SSL library a little more fool-proof by not requiring any longer
17109 that `SSL_set_{accept,connect}_state` be called before
17110 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17111 in many applications because usually everything *appeared* to work as
17112 intended anyway -- now it really works as intended).
17116 * Move openssl.cnf out of lib/.
17120 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17121 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17122 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17124 *Ralf S. Engelschall*
17126 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17127 handle PKCS#7 enveloped data properly.
17129 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17131 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17132 copying pointers. The cert_st handling is changed by this in
17133 various ways (and thus what used to be known as ctx->default_cert
17134 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17135 any longer when s->cert does not give us what we need).
17136 ssl_cert_instantiate becomes obsolete by this change.
17137 As soon as we've got the new code right (possibly it already is?),
17138 we have solved a couple of bugs of the earlier code where s->cert
17139 was used as if it could not have been shared with other SSL structures.
17141 Note that using the SSL API in certain dirty ways now will result
17142 in different behaviour than observed with earlier library versions:
17143 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17144 does not influence s as it used to.
17146 In order to clean up things more thoroughly, inside SSL_SESSION
17147 we don't use CERT any longer, but a new structure SESS_CERT
17148 that holds per-session data (if available); currently, this is
17149 the peer's certificate chain and, for clients, the server's certificate
17150 and temporary key. CERT holds only those values that can have
17151 meaningful defaults in an SSL_CTX.
17155 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17156 from the internal representation. Various PKCS#7 fixes: remove some
17157 evil casts and set the enc_dig_alg field properly based on the signing
17162 * Allow PKCS#12 password to be set from the command line or the
17163 environment. Let 'ca' get its config file name from the environment
17164 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17169 * Allow certificate policies extension to use an IA5STRING for the
17170 organization field. This is contrary to the PKIX definition but
17171 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17176 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17177 without disallowing inline assembler and the like for non-pedantic builds.
17181 * Support Borland C++ builder.
17183 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17189 * SHA-1 cleanups and performance enhancements.
17191 *Andy Polyakov <appro@fy.chalmers.se>*
17193 * Sparc v8plus assembler for the bignum library.
17195 *Andy Polyakov <appro@fy.chalmers.se>*
17197 * Accept any -xxx and +xxx compiler options in Configure.
17201 * Update HPUX configuration.
17205 * Add missing `sk_<type>_unshift()` function to safestack.h
17207 *Ralf S. Engelschall*
17209 * New function SSL_CTX_use_certificate_chain_file that sets the
17210 "extra_cert"s in addition to the certificate. (This makes sense
17211 only for "PEM" format files, as chains as a whole are not
17216 * Support verify_depth from the SSL API.
17217 x509_vfy.c had what can be considered an off-by-one-error:
17218 Its depth (which was not part of the external interface)
17219 was actually counting the number of certificates in a chain;
17220 now it really counts the depth.
17224 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17225 instead of X509err, which often resulted in confusing error
17226 messages since the error codes are not globally unique
17227 (e.g. an alleged error in ssl3_accept when a certificate
17228 didn't match the private key).
17230 * New function SSL_CTX_set_session_id_context that allows to set a default
17231 value (so that you don't need SSL_set_session_id_context for each
17232 connection using the SSL_CTX).
17236 * OAEP decoding bug fix.
17240 * Support INSTALL_PREFIX for package builders, as proposed by
17245 * New Configure options "threads" and "no-threads". For systems
17246 where the proper compiler options are known (currently Solaris
17247 and Linux), "threads" is the default.
17251 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17255 * Install various scripts to $(OPENSSLDIR)/misc, not to
17256 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17257 such as /usr/local/bin.
17261 * "make linux-shared" to build shared libraries.
17263 *Niels Poppe <niels@netbox.org>*
17265 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17269 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17270 extension adding in x509 utility.
17274 * Remove NOPROTO sections and error code comments.
17278 * Partial rewrite of the DEF file generator to now parse the ANSI
17283 * New Configure options --prefix=DIR and --openssldir=DIR.
17287 * Complete rewrite of the error code script(s). It is all now handled
17288 by one script at the top level which handles error code gathering,
17289 header rewriting and C source file generation. It should be much better
17290 than the old method: it now uses a modified version of Ulf's parser to
17291 read the ANSI prototypes in all header files (thus the old K&R definitions
17292 aren't needed for error creation any more) and do a better job of
17293 translating function codes into names. The old 'ASN1 error code embedded
17294 in a comment' is no longer necessary and it doesn't use .err files which
17295 have now been deleted. Also the error code call doesn't have to appear all
17296 on one line (which resulted in some large lines...).
17300 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17304 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17305 0 (which usually indicates a closed connection), but continue reading.
17309 * Fix some race conditions.
17313 * Add support for CRL distribution points extension. Add Certificate
17314 Policies and CRL distribution points documentation.
17318 * Move the autogenerated header file parts to crypto/opensslconf.h.
17322 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17323 8 of keying material. Merlin has also confirmed interop with this fix
17324 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17326 *Merlin Hughes <merlin@baltimore.ie>*
17328 * Fix lots of warnings.
17330 *Richard Levitte <levitte@stacken.kth.se>*
17332 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17333 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17335 *Richard Levitte <levitte@stacken.kth.se>*
17337 * Fix problems with sizeof(long) == 8.
17339 *Andy Polyakov <appro@fy.chalmers.se>*
17341 * Change functions to ANSI C.
17345 * Fix typos in error codes.
17347 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17349 * Remove defunct assembler files from Configure.
17353 * SPARC v8 assembler BIGNUM implementation.
17355 *Andy Polyakov <appro@fy.chalmers.se>*
17357 * Support for Certificate Policies extension: both print and set.
17358 Various additions to support the r2i method this uses.
17362 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17363 return a const string when you are expecting an allocated buffer.
17367 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17368 types DirectoryString and DisplayText.
17372 * Add code to allow r2i extensions to access the configuration database,
17373 add an LHASH database driver and add several ctx helper functions.
17377 * Fix an evil bug in bn_expand2() which caused various BN functions to
17378 fail when they extended the size of a BIGNUM.
17382 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17383 support typesafe stack.
17387 * Fix typo in SSL_[gs]et_options().
17389 *Nils Frostberg <nils@medcom.se>*
17391 * Delete various functions and files that belonged to the (now obsolete)
17392 old X509V3 handling code.
17396 * New Configure option "rsaref".
17400 * Don't auto-generate pem.h.
17404 * Introduce type-safe ASN.1 SETs.
17408 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17410 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17412 * Introduce type-safe STACKs. This will almost certainly break lots of code
17413 that links with OpenSSL (well at least cause lots of warnings), but fear
17414 not: the conversion is trivial, and it eliminates loads of evil casts. A
17415 few STACKed things have been converted already. Feel free to convert more.
17416 In the fullness of time, I'll do away with the STACK type altogether.
17420 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17421 specified in `<certfile>` by updating the entry in the index.txt file.
17422 This way one no longer has to edit the index.txt file manually for
17423 revoking a certificate. The -revoke option does the gory details now.
17425 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17427 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17428 `-text` option at all and this way the `-noout -text` combination was
17429 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17431 *Ralf S. Engelschall*
17433 * Make sure a corresponding plain text error message exists for the
17434 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17435 verify callback function determined that a certificate was revoked.
17437 *Ralf S. Engelschall*
17439 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17440 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17441 all available ciphers including rc5, which was forgotten until now.
17442 In order to let the testing shell script know which algorithms
17443 are available, a new (up to now undocumented) command
17444 `openssl list-cipher-commands` is used.
17448 * Bugfix: s_client occasionally would sleep in select() when
17449 it should have checked SSL_pending() first.
17453 * New functions DSA_do_sign and DSA_do_verify to provide access to
17454 the raw DSA values prior to ASN.1 encoding.
17458 * Tweaks to Configure
17460 *Niels Poppe <niels@netbox.org>*
17462 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17467 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17471 * New config option to avoid instructions that are illegal on the 80386.
17472 The default code is faster, but requires at least a 486.
17476 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17477 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17478 same as SSL2_VERSION anyway.
17482 * New "-showcerts" option for s_client.
17486 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17487 application. Various cleanups and fixes.
17491 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17492 modify error routines to work internally. Add error codes and PBE init
17493 to library startup routines.
17497 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17498 packing functions to asn1 and evp. Changed function names and error
17499 codes along the way.
17503 * PKCS12 integration: and so it begins... First of several patches to
17504 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17505 objects to objects.h
17509 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17510 and display support for Thawte strong extranet extension.
17514 * Add LinuxPPC support.
17516 *Jeff Dubrule <igor@pobox.org>*
17518 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17519 bn_div_words in alpha.s.
17521 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17523 * Make sure the RSA OAEP test is skipped under -DRSAref because
17524 OAEP isn't supported when OpenSSL is built with RSAref.
17526 *Ulf Moeller <ulf@fitug.de>*
17528 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17529 so they no longer are missing under -DNOPROTO.
17531 *Soren S. Jorvang <soren@t.dk>*
17533 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17535 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17536 doesn't work when the session is reused. Coming soon!
17540 * Fix a security hole, that allows sessions to be reused in the wrong
17541 context thus bypassing client cert protection! All software that uses
17542 client certs and session caches in multiple contexts NEEDS PATCHING to
17543 allow session reuse! A fuller solution is in the works.
17545 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17547 * Some more source tree cleanups (removed obsolete files
17548 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17549 permission on "config" script to be executable) and a fix for the INSTALL
17552 *Ulf Moeller <ulf@fitug.de>*
17554 * Remove some legacy and erroneous uses of malloc, free instead of
17557 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17559 * Make rsa_oaep_test return non-zero on error.
17561 *Ulf Moeller <ulf@fitug.de>*
17563 * Add support for native Solaris shared libraries. Configure
17564 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17565 if someone would make that last step automatic.
17567 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17569 * ctx_size was not built with the right compiler during "make links". Fixed.
17573 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17574 except NULL ciphers". This means the default cipher list will no longer
17575 enable NULL ciphers. They need to be specifically enabled e.g. with
17576 the string "DEFAULT:eNULL".
17580 * Fix to RSA private encryption routines: if p < q then it would
17581 occasionally produce an invalid result. This will only happen with
17582 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17586 * Be less restrictive and allow also `perl util/perlpath.pl
17587 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17588 because this way one can also use an interpreter named `perl5` (which is
17589 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17590 installed as `perl`).
17592 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17594 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17596 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17598 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17599 advapi32.lib to Win32 build and change the pem test comparison
17600 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17601 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17602 and crypto/des/ede_cbcm_enc.c.
17606 * DES quad checksum was broken on big-endian architectures. Fixed.
17610 * Comment out two functions in bio.h that aren't implemented. Fix up the
17611 Win32 test batch file so it (might) work again. The Win32 test batch file
17612 is horrible: I feel ill....
17616 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17617 in e_os.h. Audit of header files to check ANSI and non ANSI
17618 sections: 10 functions were absent from non ANSI section and not exported
17619 from Windows DLLs. Fixed up libeay.num for new functions.
17623 * Make `openssl version` output lines consistent.
17625 *Ralf S. Engelschall*
17627 * Fix Win32 symbol export lists for BIO functions: Added
17628 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17629 to ms/libeay{16,32}.def.
17631 *Ralf S. Engelschall*
17633 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17634 fine under Unix and passes some trivial tests I've now added. But the
17635 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17636 added to make sure no one expects that this stuff really works in the
17637 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17638 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17641 *Ralf S. Engelschall*
17643 * Fix the generation of two part addresses in perl.
17645 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17647 * Add config entry for Linux on MIPS.
17649 *John Tobey <jtobey@channel1.com>*
17651 * Make links whenever Configure is run, unless we are on Windoze.
17655 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17656 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17661 * Add a useful kludge to allow package maintainers to specify compiler and
17662 other platforms details on the command line without having to patch the
17663 Configure script every time: One now can use
17664 `perl Configure <id>:<details>`,
17665 i.e. platform ids are allowed to have details appended
17666 to them (separated by colons). This is treated as there would be a static
17667 pre-configured entry in Configure's %table under key `<id>` with value
17668 `<details>` and `perl Configure <id>` is called. So, when you want to
17669 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17670 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17671 now, which overrides the FreeBSD-elf entry on-the-fly.
17673 *Ralf S. Engelschall*
17675 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17679 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17680 on the `perl Configure ...` command line. This way one can compile
17681 OpenSSL libraries with Position Independent Code (PIC) which is needed
17682 for linking it into DSOs.
17684 *Ralf S. Engelschall*
17686 * Remarkably, export ciphers were totally broken and no-one had noticed!
17691 * Cleaned up the LICENSE document: The official contact for any license
17692 questions now is the OpenSSL core team under openssl-core@openssl.org.
17693 And add a paragraph about the dual-license situation to make sure people
17694 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17695 to the OpenSSL toolkit.
17697 *Ralf S. Engelschall*
17699 * General source tree makefile cleanups: Made `making xxx in yyy...`
17700 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17701 Additionally cleaned up the `make links` target: Remove unnecessary
17702 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17703 to speed processing and no longer clutter the display with confusing
17704 stuff. Instead only the actually done links are displayed.
17706 *Ralf S. Engelschall*
17708 * Permit null encryption ciphersuites, used for authentication only. It used
17709 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17710 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17715 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17716 signed attributes when verifying signatures (this would break them),
17717 the detached data encoding was wrong and public keys obtained using
17718 X509_get_pubkey() weren't freed.
17722 * Add text documentation for the BUFFER functions. Also added a work around
17723 to a Win95 console bug. This was triggered by the password read stuff: the
17724 last character typed gets carried over to the next fread(). If you were
17725 generating a new cert request using 'req' for example then the last
17726 character of the passphrase would be CR which would then enter the first
17731 * Added the new 'Includes OpenSSL Cryptography Software' button as
17732 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17733 button and can be used by applications based on OpenSSL to show the
17734 relationship to the OpenSSL project.
17736 *Ralf S. Engelschall*
17738 * Remove confusing variables in function signatures in files
17739 ssl/ssl_lib.c and ssl/ssl.h.
17741 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17743 * Don't install bss_file.c under PREFIX/include/
17745 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17747 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17748 functions that return function pointers and has support for NT specific
17749 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17750 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17751 unsigned to signed types: this was killing the Win32 compile.
17755 * Add new certificate file to stack functions,
17756 SSL_add_dir_cert_subjects_to_stack() and
17757 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17758 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17759 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17760 This means that Apache-SSL and similar packages don't have to mess around
17761 to add as many CAs as they want to the preferred list.
17765 * Experiment with doxygen documentation. Currently only partially applied to
17767 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17768 openssl.doxy as the configuration file.
17772 * Get rid of remaining C++-style comments which strict C compilers hate.
17774 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17776 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17777 compiled in by default: it has problems with large keys.
17781 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17782 DH private keys and/or callback functions which directly correspond to
17783 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17784 is needed for applications which have to configure certificates on a
17785 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17787 For the RSA certificate situation is makes no difference, but
17788 for the DSA certificate situation this fixes the "no shared cipher"
17789 problem where the OpenSSL cipher selection procedure failed because the
17790 temporary keys were not overtaken from the context and the API provided
17791 no way to reconfigure them.
17792 The new functions now let applications reconfigure the stuff and they
17793 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17794 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17795 non-public-API function ssl_cert_instantiate() is used as a helper
17796 function and also to reduce code redundancy inside ssl_rsa.c.
17798 *Ralf S. Engelschall*
17800 * Move s_server -dcert and -dkey options out of the undocumented feature
17801 area because they are useful for the DSA situation and should be
17802 recognized by the users.
17804 *Ralf S. Engelschall*
17806 * Fix the cipher decision scheme for export ciphers: the export bits are
17807 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17808 SSL_EXP_MASK. So, the original variable has to be used instead of the
17809 already masked variable.
17811 *Richard Levitte <levitte@stacken.kth.se>*
17813 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17815 *Richard Levitte <levitte@stacken.kth.se>*
17817 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17818 from `int` to `unsigned int` because it is a length and initialized by
17819 EVP_DigestFinal() which expects an `unsigned int *`.
17821 *Richard Levitte <levitte@stacken.kth.se>*
17823 * Don't hard-code path to Perl interpreter on shebang line of Configure
17824 script. Instead use the usual Shell->Perl transition trick.
17826 *Ralf S. Engelschall*
17828 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17829 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17830 -noout -modulus` as it's already the case for `openssl rsa -noout
17831 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17832 currently the public key is printed (a decision which was already done by
17833 `openssl dsa -modulus` in the past) which serves a similar purpose.
17834 Additionally the NO_RSA no longer completely removes the whole -modulus
17835 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17838 *Ralf S. Engelschall*
17840 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17841 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17843 *Arne Ansper <arne@ats.cyber.ee>*
17845 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17846 to be added. Now both 'req' and 'ca' can use new objects defined in the
17851 * Add cool BIO that does syslog (or event log on NT).
17853 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17855 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17856 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17857 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17858 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17862 * Add preliminary config info for new extension code.
17866 * Make RSA_NO_PADDING really use no padding.
17868 *Ulf Moeller <ulf@fitug.de>*
17870 * Generate errors when private/public key check is done.
17874 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17875 for some CRL extensions and new objects added.
17879 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17880 key usage extension and fuller support for authority key id.
17884 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17885 padding method for RSA, which is recommended for new applications in PKCS
17886 #1 v2.0 (RFC 2437, October 1998).
17887 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17888 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17889 against Bleichbacher's attack on RSA.
17890 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17893 * Updates to the new SSL compression code
17895 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17897 * Fix so that the version number in the master secret, when passed
17898 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17899 (because the server will not accept higher), that the version number
17900 is 0x03,0x01, not 0x03,0x00
17902 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17904 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17905 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17906 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17910 * Support for RAW extensions where an arbitrary extension can be
17911 created by including its DER encoding. See apps/openssl.cnf for
17916 * Make sure latest Perl versions don't interpret some generated C array
17917 code as Perl array code in the crypto/err/err_genc.pl script.
17919 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17921 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17922 not many people have the assembler. Various Win32 compilation fixes and
17923 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17924 build instructions.
17928 * Modify configure script 'Configure' to automatically create crypto/date.h
17929 file under Win32 and also build pem.h from pem.org. New script
17930 util/mkfiles.pl to create the MINFO file on environments that can't do a
17931 'make files': perl util/mkfiles.pl >MINFO should work.
17935 * Major rework of DES function declarations, in the pursuit of correctness
17936 and purity. As a result, many evil casts evaporated, and some weirdness,
17937 too. You may find this causes warnings in your code. Zapping your evil
17938 casts will probably fix them. Mostly.
17942 * Fix for a typo in asn1.h. Bug fix to object creation script
17943 obj_dat.pl. It considered a zero in an object definition to mean
17944 "end of object": none of the objects in objects.h have any zeros
17945 so it wasn't spotted.
17947 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17949 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17950 Masking (CBCM). In the absence of test vectors, the best I have been able
17951 to do is check that the decrypt undoes the encrypt, so far. Send me test
17952 vectors if you have them.
17956 * Correct calculation of key length for export ciphers (too much space was
17957 allocated for null ciphers). This has not been tested!
17961 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17962 message is now correct (it understands "crypto" and "ssl" on its
17963 command line). There is also now an "update" option. This will update
17964 the util/ssleay.num and util/libeay.num files with any new functions.
17966 perl util/mkdef.pl crypto ssl update
17967 it will update them.
17971 * Overhauled the Perl interface:
17972 - ported BN stuff to OpenSSL's different BN library
17973 - made the perl/ source tree CVS-aware
17974 - renamed the package from SSLeay to OpenSSL (the files still contain
17975 their history because I've copied them in the repository)
17976 - removed obsolete files (the test scripts will be replaced
17977 by better Test::Harness variants in the future)
17979 *Ralf S. Engelschall*
17981 * First cut for a very conservative source tree cleanup:
17982 1. merge various obsolete readme texts into doc/ssleay.txt
17983 where we collect the old documents and readme texts.
17984 2. remove the first part of files where I'm already sure that we no
17985 longer need them because of three reasons: either they are just temporary
17986 files which were left by Eric or they are preserved original files where
17987 I've verified that the diff is also available in the CVS via "cvs diff
17988 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17989 the crypto/md/ stuff).
17991 *Ralf S. Engelschall*
17993 * More extension code. Incomplete support for subject and issuer alt
17994 name, issuer and authority key id. Change the i2v function parameters
17995 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17996 what that's for :-) Fix to ASN1 macro which messed up
17997 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18001 * Preliminary support for ENUMERATED type. This is largely copied from the
18006 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18008 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18010 * Make sure `make rehash` target really finds the `openssl` program.
18012 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18014 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18015 like to hear about it if this slows down other processors.
18019 * Add CygWin32 platform information to Configure script.
18021 *Alan Batie <batie@aahz.jf.intel.com>*
18023 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18025 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18027 * New program nseq to manipulate netscape certificate sequences
18031 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18036 * Fixes to BN code. Previously the default was to define BN_RECURSION
18037 but the BN code had some problems that would cause failures when
18038 doing certificate verification and some other functions.
18040 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18042 * Add ASN1 and PEM code to support netscape certificate sequences.
18046 * Add ASN1 and PEM code to support netscape certificate sequences.
18050 * Add several PKIX and private extended key usage OIDs.
18054 * Modify the 'ca' program to handle the new extension code. Modify
18055 openssl.cnf for new extension format, add comments.
18059 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18060 and add a sample to openssl.cnf so req -x509 now adds appropriate
18065 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18066 error code, add initial support to X509_print() and x509 application.
18070 * Takes a deep breath and start adding X509 V3 extension support code. Add
18071 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18072 stuff is currently isolated and isn't even compiled yet.
18076 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18077 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18078 Removed the versions check from X509 routines when loading extensions:
18079 this allows certain broken certificates that don't set the version
18080 properly to be processed.
18084 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18085 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18086 can still be regenerated with "make depend".
18090 * Spelling mistake in C version of CAST-128.
18092 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18094 * Changes to the error generation code. The perl script err-code.pl
18095 now reads in the old error codes and retains the old numbers, only
18096 adding new ones if necessary. It also only changes the .err files if new
18097 codes are added. The makefiles have been modified to only insert errors
18098 when needed (to avoid needlessly modifying header files). This is done
18099 by only inserting errors if the .err file is newer than the auto generated
18100 C file. To rebuild all the error codes from scratch (the old behaviour)
18101 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18102 or delete all the .err files.
18106 * CAST-128 was incorrectly implemented for short keys. The C version has
18107 been fixed, but is untested. The assembler versions are also fixed, but
18108 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18109 to regenerate it if needed.
18110 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18111 Hagino <itojun@kame.net>*
18113 * File was opened incorrectly in randfile.c.
18115 *Ulf Möller <ulf@fitug.de>*
18117 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18118 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18119 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18120 al: it's just almost always a UTCTime. Note this patch adds new error
18121 codes so do a "make errors" if there are problems.
18125 * Correct Linux 1 recognition in config.
18127 *Ulf Möller <ulf@fitug.de>*
18129 * Remove pointless MD5 hash when using DSA keys in ca.
18131 *Anonymous <nobody@replay.com>*
18133 * Generate an error if given an empty string as a cert directory. Also
18134 generate an error if handed NULL (previously returned 0 to indicate an
18135 error, but didn't set one).
18137 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18139 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18143 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18144 parameters. This was causing a warning which killed off the Win32 compile.
18148 * Remove C++ style comments from crypto/bn/bn_local.h.
18150 *Neil Costigan <neil.costigan@celocom.com>*
18152 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18153 based on a text string, looking up short and long names and finally
18154 "dot" format. The "dot" format stuff didn't work. Added new function
18155 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18156 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18157 OID is not part of the table.
18161 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18162 X509_LOOKUP_by_alias().
18166 * Sort openssl functions by name.
18170 * Get the gendsa program working (hopefully) and add it to app list. Remove
18171 encryption from sample DSA keys (in case anyone is interested the password
18176 * Make *all* `*_free` functions accept a NULL pointer.
18178 *Frans Heymans <fheymans@isaserver.be>*
18180 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18183 *Anonymous <nobody@replay.com>*
18185 * s_server should send the CAfile as acceptable CAs, not its own cert.
18187 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18189 * Don't blow it for numeric -newkey arguments to apps/req.
18191 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18193 * Temp key "for export" tests were wrong in s3_srvr.c.
18195 *Anonymous <nobody@replay.com>*
18197 * Add prototype for temp key callback functions
18198 SSL_CTX_set_tmp_{rsa,dh}_callback().
18202 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18203 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18207 * X509_name_add_entry() freed the wrong thing after an error.
18209 *Arne Ansper <arne@ats.cyber.ee>*
18211 * rsa_eay.c would attempt to free a NULL context.
18213 *Arne Ansper <arne@ats.cyber.ee>*
18215 * BIO_s_socket() had a broken should_retry() on Windoze.
18217 *Arne Ansper <arne@ats.cyber.ee>*
18219 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18221 *Arne Ansper <arne@ats.cyber.ee>*
18223 * Make sure the already existing X509_STORE->depth variable is initialized
18224 in X509_STORE_new(), but document the fact that this variable is still
18225 unused in the certificate verification process.
18227 *Ralf S. Engelschall*
18229 * Fix the various library and apps files to free up pkeys obtained from
18230 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18234 * Fix reference counting in X509_PUBKEY_get(). This makes
18235 demos/maurice/example2.c work, amongst others, probably.
18237 *Steve Henson and Ben Laurie*
18239 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18240 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18241 are no longer created. This way we have a single and consistent command
18242 line interface `openssl <command>`, similar to `cvs <command>`.
18244 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18246 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18247 BIT STRING wrapper always have zero unused bits.
18251 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18255 * Make the top-level INSTALL documentation easier to understand.
18259 * Makefiles updated to exit if an error occurs in a sub-directory
18260 make (including if user presses ^C) [Paul Sutton]
18262 * Make Montgomery context stuff explicit in RSA data structure.
18266 * Fix build order of pem and err to allow for generated pem.h.
18270 * Fix renumbering bug in X509_NAME_delete_entry().
18274 * Enhanced the err-ins.pl script so it makes the error library number
18275 global and can add a library name. This is needed for external ASN1 and
18276 other error libraries.
18280 * Fixed sk_insert which never worked properly.
18284 * Fix ASN1 macros so they can handle indefinite length constructed
18285 EXPLICIT tags. Some non standard certificates use these: they can now
18290 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18291 into a single doc/ssleay.txt bundle. This way the information is still
18292 preserved but no longer messes up this directory. Now it's new room for
18293 the new set of documentation files.
18295 *Ralf S. Engelschall*
18297 * SETs were incorrectly DER encoded. This was a major pain, because they
18298 shared code with SEQUENCEs, which aren't coded the same. This means that
18299 almost everything to do with SETs or SEQUENCEs has either changed name or
18300 number of arguments.
18302 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18304 * Fix test data to work with the above.
18308 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18309 was already fixed by Eric for 0.9.1 it seems.
18311 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18313 * Autodetect FreeBSD3.
18317 * Fix various bugs in Configure. This affects the following platforms:
18321 unixware-2.0-pentium
18326 * Eliminate generated files from CVS. Reorder tests to regenerate files
18327 before they are needed.
18331 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18335 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18337 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18338 changed SSLeay to OpenSSL in version strings.
18340 *Ralf S. Engelschall*
18342 * Some fixups to the top-level documents.
18346 * Fixed the nasty bug where rsaref.h was not found under compile-time
18347 because the symlink to include/ was missing.
18349 *Ralf S. Engelschall*
18351 * Incorporated the popular no-RSA/DSA-only patches
18352 which allow to compile a RSA-free SSLeay.
18354 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18356 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18357 when "ssleay" is still not found.
18359 *Ralf S. Engelschall*
18361 * Added more platforms to Configure: Cray T3E, HPUX 11,
18363 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18365 * Updated the README file.
18367 *Ralf S. Engelschall*
18369 * Added various .cvsignore files in the CVS repository subdirs
18370 to make a "cvs update" really silent.
18372 *Ralf S. Engelschall*
18374 * Recompiled the error-definition header files and added
18375 missing symbols to the Win32 linker tables.
18377 *Ralf S. Engelschall*
18379 * Cleaned up the top-level documents;
18380 o new files: CHANGES and LICENSE
18381 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18382 o merged COPYRIGHT into LICENSE
18383 o removed obsolete TODO file
18384 o renamed MICROSOFT to INSTALL.W32
18386 *Ralf S. Engelschall*
18388 * Removed dummy files from the 0.9.1b source tree:
18389 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18390 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18391 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18392 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18393 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18395 *Ralf S. Engelschall*
18397 * Added various platform portability fixes.
18401 * The Genesis of the OpenSSL rpject:
18402 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18403 Young and Tim J. Hudson created while they were working for C2Net until
18406 *The OpenSSL Project*
18408 ### Changes between 0.9.0b and 0.9.1b [not released]
18410 * Updated a few CA certificates under certs/
18414 * Changed some BIGNUM api stuff.
18418 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18419 DGUX x86, Linux Alpha, etc.
18423 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18424 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18429 * Add -strparse option to asn1pars program which parses nested
18432 *Dr Stephen Henson <shenson@bigfoot.com>*
18434 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18438 * DSA fix for "ca" program.
18442 * Added "-genkey" option to "dsaparam" program.
18446 * Added RIPE MD160 (rmd160) message digest.
18450 * Added -a (all) option to "ssleay version" command.
18454 * Added PLATFORM define which is the id given to Configure.
18458 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18462 * Extended the ASN.1 parser routines.
18466 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18470 * Added a BN_CTX to the BN library.
18474 * Fixed the weak key values in DES library
18478 * Changed API in EVP library for cipher aliases.
18482 * Added support for RC2/64bit cipher.
18486 * Converted the lhash library to the crypto/mem.c functions.
18490 * Added more recognized ASN.1 object ids.
18494 * Added more RSA padding checks for SSL/TLS.
18498 * Added BIO proxy/filter functionality.
18502 * Added extra_certs to SSL_CTX which can be used
18503 send extra CA certificates to the client in the CA cert chain sending
18504 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18508 * Now Fortezza is denied in the authentication phase because
18509 this is key exchange mechanism is not supported by SSLeay at all.
18513 * Additional PKCS1 checks.
18517 * Support the string "TLSv1" for all TLS v1 ciphers.
18521 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18522 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18526 * Fixed a few memory leaks.
18530 * Fixed various code and comment typos.
18534 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18535 bytes sent in the client random.
18537 *Edward Bishop <ebishop@spyglass.com>*
18541 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18542 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18543 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18544 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18545 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18546 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18547 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18548 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18549 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18550 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18551 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18552 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18553 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18554 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18555 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18556 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18557 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18558 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18559 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18560 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18561 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18562 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18563 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18564 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18565 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18566 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18567 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18568 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18569 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18570 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18571 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18572 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18573 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18574 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18575 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18576 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18577 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18578 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18579 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18580 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18581 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18582 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18583 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18584 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18585 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18586 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18587 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18588 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18589 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18590 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18591 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18592 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18593 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18594 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18595 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18596 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18597 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18598 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18599 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18600 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18601 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18602 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18603 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18604 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18605 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18606 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18607 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18608 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18609 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18610 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18611 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18612 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18613 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18614 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18615 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18616 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18617 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18618 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18619 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18620 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18621 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18622 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18623 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18624 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18625 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18626 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18627 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18628 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18629 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18630 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18631 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18632 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18633 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18634 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18635 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18636 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18637 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18638 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18639 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18640 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18641 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18642 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18643 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18644 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18645 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18646 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18647 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18648 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18649 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18650 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18651 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18652 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18653 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18654 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18655 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18656 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18657 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18658 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18659 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18660 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18661 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18662 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18663 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18664 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18665 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18666 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18667 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18668 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18669 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18670 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18671 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18672 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18673 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18674 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18675 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18676 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18677 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18678 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18679 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18680 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18681 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18682 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18683 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18684 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18685 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18686 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18687 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18688 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18689 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18690 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18691 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18692 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18693 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18694 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18695 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18696 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18697 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18698 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18699 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18700 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18701 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655