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 that applications as well as other
116 libraries can use to form a separate context within which libcrypto
117 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 `OPENSSL_CTX_set0_default`.
127 All public OpenSSL functions that take an `OPENSSL_CTX` pointer,
128 apart from the functions directly related to `OPENSSL_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 `OPENSSL_CTX_set0_default` should take care to restore it with a
133 second call before returning to the caller.
137 * Handshake now fails if Extended Master Secret extension is dropped
142 * Dropped interactive mode from the 'openssl' program. From now on,
143 the `openssl` command without arguments is equivalent to `openssl
148 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
149 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
150 While the old function names have been retained for backward compatibility
151 they should not be used in new developments
152 because their return values are confusing: Unlike other `_cmp()` functions
153 they do not return 0 in case their arguments are equal.
157 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
158 EC_GROUP_get_field_type().
162 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
163 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
164 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
165 Applications should rely on the library automatically assigning a suitable
166 EC_METHOD internally upon EC_GROUP construction.
170 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
171 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
172 assigned internally without application intervention.
173 Users of EC_GROUP_new() should switch to a different suitable constructor.
177 * Add CAdES-BES signature verification support, mostly derived
178 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
180 *Filipe Raimundo da Silva*
182 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
186 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
187 functions are not widely used and now OpenSSL automatically perform this
188 conversion when needed.
192 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
193 EC_KEY_precompute_mult(). These functions are not widely used and
194 applications should instead switch to named curves which OpenSSL has
195 hardcoded lookup tables for.
199 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
200 should instead use the L<EC_POINT_mul(3)> function.
204 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
205 that are not applicable to the new provider model. Applications should
206 instead use EVP_default_properties_is_fips_enabled() and
207 EVP_default_properties_enable_fips().
211 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
212 is set, an unexpected EOF is ignored, it pretends a close notify was received
213 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
217 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
218 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
219 used and applications should instead use the
220 L<EC_POINT_set_affine_coordinates(3)> and
221 L<EC_POINT_get_affine_coordinates(3)> functions.
225 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
226 arrays to be more easily constructed via a series of utility functions.
227 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
228 the various push functions and finally convert to a passable OSSL_PARAM
229 array using OSSL_PARAM_BLD_to_param().
233 * The security strength of SHA1 and MD5 based signatures in TLS has been
234 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
235 working at the default security level of 1 and instead requires security
236 level 0. The security level can be changed either using the cipher string
237 with @SECLEVEL, or calling SSL_CTX_set_security_level().
241 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
242 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
243 internal keys, if they correspond to one of those built in types.
247 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
248 contain a provider side internal key.
252 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
253 They are old functions that we don't use, and that you could disable with
254 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
258 * Project text documents not yet having a proper file name extension
259 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
260 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
264 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
265 have been converted to Markdown with the goal to produce documents
266 which not only look pretty when viewed online in the browser, but
267 remain well readable inside a plain text editor.
269 To achieve this goal, a 'minimalistic' Markdown style has been applied
270 which avoids formatting elements that interfere too much with the
271 reading flow in the text file. For example, it
273 * avoids [ATX headings][] and uses [setext headings][] instead
274 (which works for `<h1>` and `<h2>` headings only).
275 * avoids [inline links][] and uses [reference links][] instead.
276 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
278 [ATX headings]: https://github.github.com/gfm/#atx-headings
279 [setext headings]: https://github.github.com/gfm/#setext-headings
280 [inline links]: https://github.github.com/gfm/#inline-link
281 [reference links]: https://github.github.com/gfm/#reference-link
282 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
283 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
285 *Matthias St. Pierre*
287 * The test suite is changed to preserve results of each test recipe.
288 A new directory test-runs/ with subdirectories named like the
289 test recipes are created in the build tree for this purpose.
293 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
294 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
295 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
297 *David von Oheimb, Martin Peylo*
299 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
300 The legacy OCSP-focused and only partly documented API is retained.
301 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
305 * Added `util/check-format.pl`, a tool for checking adherence to the
306 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
307 The checks performed are incomplete and yield some false positives.
308 Still the tool should be useful for detecting most typical glitches.
312 * BIO_do_connect and BIO_do_handshake have been extended:
313 If domain name resolution yields multiple IP addresses all of them are tried
314 after connect() failures.
318 * All of the low level RSA functions have been deprecated including:
320 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
321 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
322 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
323 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
324 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
325 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
326 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
327 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
328 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
329 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
330 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
331 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
332 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
333 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
334 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
335 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
336 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
337 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
338 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
339 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
340 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
341 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
342 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
343 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
344 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
345 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
346 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
347 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
349 Use of these low level functions has been informally discouraged for a long
350 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
351 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
352 L<EVP_PKEY_decrypt(3)>.
356 * X509 certificates signed using SHA1 are no longer allowed at security
358 In TLS/SSL the default security level is 1. It can be set either
359 using the cipher string with @SECLEVEL, or calling
360 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
361 a call to SSL_CTX_use_certificate() will fail if the security level is not
363 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
364 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
369 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
370 modified to use PKEY APIs. These commands are now in maintenance mode
371 and no new features will be added to them.
375 * The command line utility rsautl has been deprecated.
376 Instead use the pkeyutl program.
380 * The command line utilities genrsa and rsa have been modified to use PKEY
381 APIs. They now write PKCS#8 keys by default. These commands are now in
382 maintenance mode and no new features will be added to them.
386 * All of the low level DH functions have been deprecated including:
388 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
389 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
390 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
391 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
392 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
393 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
394 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
395 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
396 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
397 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
398 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
399 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
400 DH_meth_get_generate_params and DH_meth_set_generate_params.
402 Use of these low level functions has been informally discouraged for a long
403 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
404 and L<EVP_PKEY_derive(3)>.
408 * All of the low level DSA functions have been deprecated including:
410 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
411 DSA_get_default_method, DSA_set_method, DSA_get_method,
412 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
413 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
414 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
415 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
416 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
417 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
418 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
419 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
420 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
421 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
422 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
425 Use of these low level functions has been informally discouraged for a long
426 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
427 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
431 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
432 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
433 This means that applications don't have to look at the curve NID and
434 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
435 However, they still can, that EVP_PKEY_set_alias_type() call acts as
436 a no-op when the EVP_PKEY is already of the given type.
438 Parameter and key generation is also reworked to make it possible
439 to generate EVP_PKEY_SM2 parameters and keys without having to go
440 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
441 However, code that does the latter will still work as before.
445 * Deprecated low level ECDH and ECDSA functions. These include:
447 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
448 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
451 Use of these low level functions has been informally discouraged for a long
452 time. Instead applications should use the EVP_PKEY_derive(3),
453 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
457 * Deprecated the EC_KEY_METHOD functions. These include:
459 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
460 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
461 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
462 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
463 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
464 EC_KEY_METHOD_get_verify.
466 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
470 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
471 and EVP_PKEY_decrypt() instead.
472 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
473 and EVP_PKEY_encrypt() instead.
477 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
478 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
479 a new formulation to include all the things it can be used for,
480 as well as words of caution.
484 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
485 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
489 * All of the low level HMAC functions have been deprecated including:
491 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
492 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
495 Use of these low level functions has been informally discouraged for a long
496 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
497 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
498 and L<EVP_MAC_final(3)>.
502 * Over two thousand fixes were made to the documentation, including:
503 - Common options (such as -rand/-writerand, TLS version control, etc)
504 were refactored and point to newly-enhanced descriptions in openssl.pod.
505 - Added style conformance for all options (with help from Richard Levitte),
506 documented all reported missing options, added a CI build to check
507 that all options are documented and that no unimplemented options
509 - Documented some internals, such as all use of environment variables.
510 - Addressed all internal broken L<> references.
514 * All of the low level CMAC functions have been deprecated including:
516 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
517 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
519 Use of these low level functions has been informally discouraged for a long
520 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
521 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
522 and L<EVP_MAC_final(3)>.
526 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
527 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
530 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
531 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
532 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
533 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
534 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
535 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
536 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
537 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
538 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
539 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
541 Use of these low level functions has been informally discouraged
542 for a long time. Applications should use the EVP_DigestInit_ex(3),
543 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
547 * Corrected the documentation of the return values from the `EVP_DigestSign*`
548 set of functions. The documentation mentioned negative values for some
549 errors, but this was never the case, so the mention of negative values
552 Code that followed the documentation and thereby check with something
553 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
557 * All of the low level cipher functions have been deprecated including:
559 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
560 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
561 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
562 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
563 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
564 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
565 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
566 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
567 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
568 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
569 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
570 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
571 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
572 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
573 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
574 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
575 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
576 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
577 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
578 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
579 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
580 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
581 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
582 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
583 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
584 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
585 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
586 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
587 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
589 Use of these low level functions has been informally discouraged for
590 a long time. Applications should use the high level EVP APIs, e.g.
591 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
592 equivalently named decrypt functions instead.
594 *Matt Caswell and Paul Dale*
596 * Removed include/openssl/opensslconf.h.in and replaced it with
597 include/openssl/configuration.h.in, which differs in not including
598 <openssl/macros.h>. A short header include/openssl/opensslconf.h
599 was added to include both.
601 This allows internal hacks where one might need to modify the set
602 of configured macros, for example this if deprecated symbols are
603 still supposed to be available internally:
605 #include <openssl/configuration.h>
607 #undef OPENSSL_NO_DEPRECATED
608 #define OPENSSL_SUPPRESS_DEPRECATED
610 #include <openssl/macros.h>
612 This should not be used by applications that use the exported
613 symbols, as that will lead to linking errors.
617 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
618 used in exponentiation with 512-bit moduli. No EC algorithms are
619 affected. Analysis suggests that attacks against 2-prime RSA1024,
620 3-prime RSA1536, and DSA1024 as a result of this defect would be very
621 difficult to perform and are not believed likely. Attacks against DH512
622 are considered just feasible. However, for an attack the target would
623 have to re-use the DH512 private key, which is not recommended anyway.
624 Also applications directly using the low level API BN_mod_exp may be
625 affected if they use BN_FLG_CONSTTIME.
630 * Most memory-debug features have been deprecated, and the functionality
631 replaced with no-ops.
635 * Added documentation for the STACK API.
639 * Introduced a new method type and API, OSSL_ENCODER, to
640 represent generic encoders. An implementation is expected to
641 be able to encode an object associated with a given name (such
642 as an algorithm name for an asymmetric key) into forms given by
643 implementation properties.
645 Encoders are primarily used from inside libcrypto, through
646 calls to functions like EVP_PKEY_print_private(),
647 PEM_write_bio_PrivateKey() and similar.
649 Encoders are specified in such a way that they can be made to
650 directly handle the provider side portion of an object, if this
651 provider side part comes from the same provider as the encoder
652 itself, but can also be made to handle objects in parametrized
653 form (as an OSSL_PARAM array of data). This allows a provider to
654 offer generic encoders as a service for any other provider.
658 * Added a .pragma directive to the syntax of configuration files, to
659 allow varying behavior in a supported and predictable manner.
660 Currently added pragma:
664 This allows dollar signs to be a keyword character unless it's
665 followed by a opening brace or parenthesis. This is useful for
666 platforms where dollar signs are commonly used in names, such as
667 volume names and system directory names on VMS.
671 * Added functionality to create an EVP_PKEY from user data. This
672 is effectively the same as creating a RSA, DH or DSA object and
673 then assigning them to an EVP_PKEY, but directly using algorithm
674 agnostic EVP functions. A benefit is that this should be future
675 proof for public key algorithms to come.
679 * Change the interpretation of the '--api' configuration option to
680 mean that this is a desired API compatibility level with no
681 further meaning. The previous interpretation, that this would
682 also mean to remove all deprecated symbols up to and including
683 the given version, no requires that 'no-deprecated' is also used
684 in the configuration.
686 When building applications, the desired API compatibility level
687 can be set with the OPENSSL_API_COMPAT macro like before. For
688 API compatibility version below 3.0, the old style numerical
689 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
690 For version 3.0 and on, the value is expected to be the decimal
691 value calculated from the major and minor version like this:
693 MAJOR * 10000 + MINOR * 100
697 -DOPENSSL_API_COMPAT=30000 For 3.0
698 -DOPENSSL_API_COMPAT=30200 For 3.2
700 To hide declarations that are deprecated up to and including the
701 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
702 given when building the application as well.
706 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
707 access to certificate and CRL stores via URIs and OSSL_STORE
710 This adds the following functions:
712 - X509_LOOKUP_store()
713 - X509_STORE_load_file()
714 - X509_STORE_load_path()
715 - X509_STORE_load_store()
716 - SSL_add_store_cert_subjects_to_stack()
717 - SSL_CTX_set_default_verify_store()
718 - SSL_CTX_load_verify_file()
719 - SSL_CTX_load_verify_dir()
720 - SSL_CTX_load_verify_store()
724 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
725 The presence of this system service is determined at run-time.
729 * Added functionality to create an EVP_PKEY context based on data
730 for methods from providers. This takes an algorithm name and a
731 property query string and simply stores them, with the intent
732 that any operation that uses this context will use those strings
733 to fetch the needed methods implicitly, thereby making the port
734 of application written for pre-3.0 OpenSSL easier.
738 * The undocumented function NCONF_WIN32() has been deprecated; for
739 conversion details see the HISTORY section of doc/man5/config.pod
743 * Introduced the new functions EVP_DigestSignInit_ex() and
744 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
745 EVP_DigestVerifyUpdate() have been converted to functions. See the man
746 pages for further details.
750 * Over two thousand fixes were made to the documentation, including:
751 adding missing command flags, better style conformance, documentation
754 *Rich Salz, Richard Levitte*
756 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
757 X25519, X448, Ed25519 and Ed448.
761 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
766 * Deprecated the public definition of ERR_STATE as well as the function
767 ERR_get_state(). This is done in preparation of making ERR_STATE an
772 * Added ERR functionality to give callers access to the stored function
773 names that have replaced the older function code based functions.
775 New functions are ERR_get_error_func(), ERR_peek_error_func(),
776 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
777 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
778 and ERR_peek_last_error_all().
780 These functions have become deprecated: ERR_get_error_line_data(),
781 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
782 ERR_func_error_string().
786 * Extended testing to be verbose for failing tests only. The make variables
787 VERBOSE_FAILURE or VF can be used to enable this:
789 $ make VF=1 test # Unix
790 $ mms /macro=(VF=1) test ! OpenVMS
791 $ nmake VF=1 test # Windows
795 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
796 used even when parsing explicit parameters, when loading a encoded key
797 or calling `EC_GROUP_new_from_ecpkparameters()`/
798 `EC_GROUP_new_from_ecparameters()`.
799 This prevents bypass of security hardening and performance gains,
800 especially for curves with specialized EC_METHODs.
801 By default, if a key encoded with explicit parameters is loaded and later
802 encoded, the output is still encoded with explicit parameters, even if
803 internally a "named" EC_GROUP is used for computation.
807 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
808 this change, EC_GROUP_set_generator would accept order and/or cofactor as
809 NULL. After this change, only the cofactor parameter can be NULL. It also
810 does some minimal sanity checks on the passed order.
815 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
816 An attack is simple, if the first CMS_recipientInfo is valid but the
817 second CMS_recipientInfo is chosen ciphertext. If the second
818 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
819 encryption key will be replaced by garbage, and the message cannot be
820 decoded, but if the RSA decryption fails, the correct encryption key is
821 used and the recipient will not notice the attack.
822 As a work around for this potential attack the length of the decrypted
823 key must be equal to the cipher default key length, in case the
824 certifiate is not given and all recipientInfo are tried out.
825 The old behaviour can be re-enabled in the CMS code by setting the
826 CMS_DEBUG_DECRYPT flag.
830 * Early start up entropy quality from the DEVRANDOM seed source has been
831 improved for older Linux systems. The RAND subsystem will wait for
832 /dev/random to be producing output before seeding from /dev/urandom.
833 The seeded state is stored for future library initialisations using
834 a system global shared memory segment. The shared memory identifier
835 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
836 the desired value. The default identifier is 114.
840 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
841 when primes for RSA keys are computed.
842 Since we previously always generated primes == 2 (mod 3) for RSA keys,
843 the 2-prime and 3-prime RSA modules were easy to distinguish, since
844 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
845 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
846 This avoids possible fingerprinting of newly generated RSA modules.
850 * Correct the extended master secret constant on EBCDIC systems. Without this
851 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
852 negotiate EMS will fail. Unfortunately this also means that TLS connections
853 between EBCDIC systems with this fix, and EBCDIC systems without this
854 fix will fail if they negotiate EMS.
858 * Changed the library initialisation so that the config file is now loaded
859 by default. This was already the case for libssl. It now occurs for both
860 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
861 OPENSSL_init_crypto() to suppress automatic loading of a config file.
865 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
866 where the former acts as a replacement for ERR_put_error(), and the
867 latter replaces the combination ERR_put_error()+ERR_add_error_data().
868 ERR_raise_data() adds more flexibility by taking a format string and
869 an arbitrary number of arguments following it, to be processed with
874 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
875 to check if a named provider is loaded and available. When called, it
876 will also activate all fallback providers if such are still present.
880 * Enforce a minimum DH modulus size of 512 bits.
884 * Changed DH parameters to generate the order q subgroup instead of 2q.
885 Previously generated DH parameters are still accepted by DH_check
886 but DH_generate_key works around that by clearing bit 0 of the
887 private key for those. This avoids leaking bit 0 of the private key.
891 * Significantly reduce secure memory usage by the randomness pools.
895 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
900 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
901 algorithms. An implementation of a key exchange algorithm can be obtained
902 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
903 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
904 the older EVP_PKEY_derive_init() function. See the man pages for the new
905 functions for further details.
909 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
913 * Removed the function names from error messages and deprecated the
916 * Removed NextStep support and the macro OPENSSL_UNISTD
920 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
921 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
922 Also removed "export var as function" capability; we do not export
923 variables, only functions.
927 * RC5_32_set_key has been changed to return an int type, with 0 indicating
928 an error and 1 indicating success. In previous versions of OpenSSL this
929 was a void type. If a key was set longer than the maximum possible this
934 * Support SM2 signing and verification schemes with X509 certificate.
938 * Use SHA256 as the default digest for TS query in the ts app.
942 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
943 This checks that the salt length is at least 128 bits, the derived key
944 length is at least 112 bits, and that the iteration count is at least 1000.
945 For backwards compatibility these checks are disabled by default in the
946 default provider, but are enabled by default in the fips provider.
947 To enable or disable these checks use the control
948 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
952 * Default cipher lists/suites are now available via a function, the
953 #defines are deprecated.
957 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
958 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
959 for Windows Store apps easier. Also, the "no-uplink" option has been added.
963 * Join the directories crypto/x509 and crypto/x509v3
967 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
968 This changes the size when using the genpkey app when no size is given. It
969 fixes an omission in earlier changes that changed all RSA, DSA and DH
970 generation apps to use 2048 bits by default.
974 * Added command 'openssl kdf' that uses the EVP_KDF API.
978 * Added command 'openssl mac' that uses the EVP_MAC API.
982 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
983 as default directories. Also added the command 'openssl info'
984 for scripting purposes.
988 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
989 deprecated. These undocumented functions were never integrated into the EVP
990 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
991 Bi-directional IGE mode. These modes were never formally standardised and
992 usage of these functions is believed to be very small. In particular
993 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
994 is ever used. The security implications are believed to be minimal, but
995 this issue was never fixed for backwards compatibility reasons. New code
996 should not use these modes.
1000 * Add prediction resistance to the DRBG reseeding process.
1004 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1005 mandated by IEEE Std 1619-2018.
1009 * Added newline escaping functionality to a filename when using openssl dgst.
1010 This output format is to replicate the output format found in the `*sum`
1011 checksum programs. This aims to preserve backward compatibility.
1013 *Matt Eaton, Richard Levitte, and Paul Dale*
1015 * Removed the heartbeat message in DTLS feature, as it has very
1016 little usage and doesn't seem to fulfill a valuable purpose.
1017 The configuration option is now deprecated.
1021 * Changed the output of 'openssl {digestname} < file' to display the
1022 digest name in its output.
1026 * Added a new generic trace API which provides support for enabling
1027 instrumentation through trace output. This feature is mainly intended
1028 as an aid for developers and is disabled by default. To utilize it,
1029 OpenSSL needs to be configured with the `enable-trace` option.
1031 If the tracing API is enabled, the application can activate trace output
1032 by registering BIOs as trace channels for a number of tracing and debugging
1035 The 'openssl' application has been expanded to enable any of the types
1036 available via environment variables defined by the user, and serves as
1037 one possible example on how to use this functionality.
1039 *Richard Levitte & Matthias St. Pierre*
1041 * Added build tests for C++. These are generated files that only do one
1042 thing, to include one public OpenSSL head file each. This tests that
1043 the public header files can be usefully included in a C++ application.
1045 This test isn't enabled by default. It can be enabled with the option
1046 'enable-buildtest-c++'.
1050 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1054 * Add KMAC to EVP_MAC.
1058 * Added property based algorithm implementation selection framework to
1063 * Added SCA hardening for modular field inversion in EC_GROUP through
1064 a new dedicated field_inv() pointer in EC_METHOD.
1065 This also addresses a leakage affecting conversions from projective
1066 to affine coordinates.
1068 *Billy Bob Brumley, Nicola Tuveri*
1070 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1071 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1072 those algorithms that were already supported through the EVP_PKEY API
1073 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1074 and scrypt are now wrappers that call EVP_KDF.
1078 * Build devcrypto engine as a dynamic engine.
1080 *Eneas U de Queiroz*
1082 * Add keyed BLAKE2 to EVP_MAC.
1086 * Fix a bug in the computation of the endpoint-pair shared secret used
1087 by DTLS over SCTP. This breaks interoperability with older versions
1088 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1089 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1090 interoperability with such broken implementations. However, enabling
1091 this switch breaks interoperability with correct implementations.
1093 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1094 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1098 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1102 * Change the license to the Apache License v2.0.
1106 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1108 - Major releases (indicated by incrementing the MAJOR release number)
1109 may introduce incompatible API/ABI changes.
1110 - Minor releases (indicated by incrementing the MINOR release number)
1111 may introduce new features but retain API/ABI compatibility.
1112 - Patch releases (indicated by incrementing the PATCH number)
1113 are intended for bug fixes and other improvements of existing
1114 features only (like improving performance or adding documentation)
1115 and retain API/ABI compatibility.
1119 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1123 * Remove the 'dist' target and add a tarball building script. The
1124 'dist' target has fallen out of use, and it shouldn't be
1125 necessary to configure just to create a source distribution.
1129 * Recreate the OS390-Unix config target. It no longer relies on a
1130 special script like it did for OpenSSL pre-1.1.0.
1134 * Instead of having the source directories listed in Configure, add
1135 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1140 * Add GMAC to EVP_MAC.
1144 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1148 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1149 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1150 to facilitate the continued use of MACs through raw private keys in
1151 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1155 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1156 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1160 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1161 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1162 are retained for backwards compatibility.
1166 * AES-XTS mode now enforces that its two keys are different to mitigate
1167 the attacked described in "Efficient Instantiations of Tweakable
1168 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1169 Details of this attack can be obtained from:
1170 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1174 * Rename the object files, i.e. give them other names than in previous
1175 versions. Their names now include the name of the final product, as
1176 well as its type mnemonic (bin, lib, shlib).
1180 * Added new option for 'openssl list', '-objects', which will display the
1181 list of built in objects, i.e. OIDs with names.
1185 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1186 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1191 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1192 improves application performance by removing data copies and providing
1193 applications with zero-copy system calls such as sendfile and splice.
1197 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1198 option is set, openssl cleanses (zeroize) plaintext bytes from
1199 internal buffers after delivering them to the application. Note,
1200 the application is still responsible for cleansing other copies
1201 (e.g.: data received by SSL_read(3)).
1205 * `PKCS12_parse` now maintains the order of the parsed certificates
1206 when outputting them via `*ca` (rather than reversing it).
1213 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1217 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1219 * Certificates with explicit curve parameters are now disallowed in
1220 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1224 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1225 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1226 conversely, silently ignore DTLS protocol version bounds when configuring
1227 TLS-based contexts. The commands can be repeated to set bounds of both
1228 types. The same applies with the corresponding "min_protocol" and
1229 "max_protocol" command-line switches, in case some application uses both TLS
1232 SSL_CTX instances that are created for a fixed protocol version (e.g.
1233 TLSv1_server_method()) also silently ignore version bounds. Previously
1234 attempts to apply bounds to these protocol versions would result in an
1235 error. Now only the "version-flexible" SSL_CTX instances are subject to
1236 limits in configuration files in command-line options.
1240 * Handshake now fails if Extended Master Secret extension is dropped
1245 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1247 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1249 * Fixed segmentation fault in SSL_check_chain()
1250 Server or client applications that call the SSL_check_chain() function
1251 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1252 dereference as a result of incorrect handling of the
1253 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1254 or unrecognised signature algorithm is received from the peer. This could
1255 be exploited by a malicious peer in a Denial of Service attack.
1260 * Added AES consttime code for no-asm configurations
1261 an optional constant time support for AES was added
1262 when building openssl for no-asm.
1263 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1264 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1265 At this time this feature is by default disabled.
1266 It will be enabled by default in 3.0.
1270 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1272 * Revert the change of EOF detection while reading in libssl to avoid
1273 regressions in applications depending on the current way of reporting
1274 the EOF. As the existing method is not fully accurate the change to
1275 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1276 branch and will be present in the 3.0 release.
1280 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1281 when primes for RSA keys are computed.
1282 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1283 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1284 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1285 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1286 This avoids possible fingerprinting of newly generated RSA modules.
1290 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1292 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1293 while reading in libssl then we would report an error back to the
1294 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1295 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1296 therefore give a hint as to what went wrong.
1300 * Check that ed25519 and ed448 are allowed by the security level. Previously
1301 signature algorithms not using an MD were not being checked that they were
1302 allowed by the security level.
1306 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1307 was not quite right. The behaviour was not consistent between resumption
1308 and normal handshakes, and also not quite consistent with historical
1309 behaviour. The behaviour in various scenarios has been clarified and
1310 it has been updated to make it match historical behaviour as closely as
1315 * *[VMS only]* The header files that the VMS compilers include automatically,
1316 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1317 that the C++ compiler doesn't understand. This is a shortcoming in the
1318 compiler, but can be worked around with `__cplusplus` guards.
1320 C++ applications that use OpenSSL libraries must be compiled using the
1321 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1322 functions. Otherwise, only functions with symbols of less than 31
1323 characters can be used, as the linker will not be able to successfully
1324 resolve symbols with longer names.
1328 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1329 The presence of this system service is determined at run-time.
1333 * Added newline escaping functionality to a filename when using openssl dgst.
1334 This output format is to replicate the output format found in the `*sum`
1335 checksum programs. This aims to preserve backward compatibility.
1337 *Matt Eaton, Richard Levitte, and Paul Dale*
1339 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1344 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1346 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1347 number generator (RNG). This was intended to include protection in the
1348 event of a fork() system call in order to ensure that the parent and child
1349 processes did not share the same RNG state. However this protection was not
1350 being used in the default case.
1352 A partial mitigation for this issue is that the output from a high
1353 precision timer is mixed into the RNG state so the likelihood of a parent
1354 and child process sharing state is significantly reduced.
1356 If an application already calls OPENSSL_init_crypto() explicitly using
1357 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1360 *Matthias St. Pierre*
1362 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1363 used even when parsing explicit parameters, when loading a encoded key
1364 or calling `EC_GROUP_new_from_ecpkparameters()`/
1365 `EC_GROUP_new_from_ecparameters()`.
1366 This prevents bypass of security hardening and performance gains,
1367 especially for curves with specialized EC_METHODs.
1368 By default, if a key encoded with explicit parameters is loaded and later
1369 encoded, the output is still encoded with explicit parameters, even if
1370 internally a "named" EC_GROUP is used for computation.
1374 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1375 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1376 NULL. After this change, only the cofactor parameter can be NULL. It also
1377 does some minimal sanity checks on the passed order.
1382 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1383 An attack is simple, if the first CMS_recipientInfo is valid but the
1384 second CMS_recipientInfo is chosen ciphertext. If the second
1385 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1386 encryption key will be replaced by garbage, and the message cannot be
1387 decoded, but if the RSA decryption fails, the correct encryption key is
1388 used and the recipient will not notice the attack.
1389 As a work around for this potential attack the length of the decrypted
1390 key must be equal to the cipher default key length, in case the
1391 certifiate is not given and all recipientInfo are tried out.
1392 The old behaviour can be re-enabled in the CMS code by setting the
1393 CMS_DEBUG_DECRYPT flag.
1398 * Early start up entropy quality from the DEVRANDOM seed source has been
1399 improved for older Linux systems. The RAND subsystem will wait for
1400 /dev/random to be producing output before seeding from /dev/urandom.
1401 The seeded state is stored for future library initialisations using
1402 a system global shared memory segment. The shared memory identifier
1403 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1404 the desired value. The default identifier is 114.
1408 * Correct the extended master secret constant on EBCDIC systems. Without this
1409 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1410 negotiate EMS will fail. Unfortunately this also means that TLS connections
1411 between EBCDIC systems with this fix, and EBCDIC systems without this
1412 fix will fail if they negotiate EMS.
1416 * Use Windows installation paths in the mingw builds
1418 Mingw isn't a POSIX environment per se, which means that Windows
1419 paths should be used for installation.
1424 * Changed DH_check to accept parameters with order q and 2q subgroups.
1425 With order 2q subgroups the bit 0 of the private key is not secret
1426 but DH_generate_key works around that by clearing bit 0 of the
1427 private key for those. This avoids leaking bit 0 of the private key.
1431 * Significantly reduce secure memory usage by the randomness pools.
1435 * Revert the DEVRANDOM_WAIT feature for Linux systems
1437 The DEVRANDOM_WAIT feature added a select() call to wait for the
1438 /dev/random device to become readable before reading from the
1439 /dev/urandom device.
1441 It turned out that this change had negative side effects on
1442 performance which were not acceptable. After some discussion it
1443 was decided to revert this feature and leave it up to the OS
1444 resp. the platform maintainer to ensure a proper initialization
1445 during early boot time.
1447 *Matthias St. Pierre*
1449 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1451 * Add build tests for C++. These are generated files that only do one
1452 thing, to include one public OpenSSL head file each. This tests that
1453 the public header files can be usefully included in a C++ application.
1455 This test isn't enabled by default. It can be enabled with the option
1456 'enable-buildtest-c++'.
1460 * Enable SHA3 pre-hashing for ECDSA and DSA.
1464 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1465 This changes the size when using the genpkey app when no size is given. It
1466 fixes an omission in earlier changes that changed all RSA, DSA and DH
1467 generation apps to use 2048 bits by default.
1471 * Reorganize the manual pages to consistently have RETURN VALUES,
1472 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1473 util/fix-doc-nits accordingly.
1475 *Paul Yang, Joshua Lock*
1477 * Add the missing accessor EVP_PKEY_get0_engine()
1481 * Have apps like 's_client' and 's_server' output the signature scheme
1482 along with other cipher suite parameters when debugging.
1486 * Make OPENSSL_config() error agnostic again.
1490 * Do the error handling in RSA decryption constant time.
1494 * Prevent over long nonces in ChaCha20-Poly1305.
1496 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1497 for every encryption operation. RFC 7539 specifies that the nonce value
1498 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1499 and front pads the nonce with 0 bytes if it is less than 12
1500 bytes. However it also incorrectly allows a nonce to be set of up to 16
1501 bytes. In this case only the last 12 bytes are significant and any
1502 additional leading bytes are ignored.
1504 It is a requirement of using this cipher that nonce values are
1505 unique. Messages encrypted using a reused nonce value are susceptible to
1506 serious confidentiality and integrity attacks. If an application changes
1507 the default nonce length to be longer than 12 bytes and then makes a
1508 change to the leading bytes of the nonce expecting the new value to be a
1509 new unique nonce then such an application could inadvertently encrypt
1510 messages with a reused nonce.
1512 Additionally the ignored bytes in a long nonce are not covered by the
1513 integrity guarantee of this cipher. Any application that relies on the
1514 integrity of these ignored leading bytes of a long nonce may be further
1515 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1516 is safe because no such use sets such a long nonce value. However user
1517 applications that use this cipher directly and set a non-default nonce
1518 length to be longer than 12 bytes may be vulnerable.
1520 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1526 * Add DEVRANDOM_WAIT feature for Linux systems
1528 On older Linux systems where the getrandom() system call is not available,
1529 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1530 Contrary to getrandom(), the /dev/urandom device will not block during
1531 early boot when the kernel CSPRNG has not been seeded yet.
1533 To mitigate this known weakness, use select() to wait for /dev/random to
1534 become readable before reading from /dev/urandom.
1536 * Ensure that SM2 only uses SM3 as digest algorithm
1540 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1542 * Change the info callback signals for the start and end of a post-handshake
1543 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1544 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1545 confused by this and assume that a TLSv1.2 renegotiation has started. This
1546 can break KeyUpdate handling. Instead we no longer signal the start and end
1547 of a post handshake message exchange (although the messages themselves are
1548 still signalled). This could break some applications that were expecting
1549 the old signals. However without this KeyUpdate is not usable for many
1554 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1556 * Timing vulnerability in DSA signature generation
1558 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1559 timing side channel attack. An attacker could use variations in the signing
1560 algorithm to recover the private key.
1562 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1567 * Timing vulnerability in ECDSA signature generation
1569 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1570 timing side channel attack. An attacker could use variations in the signing
1571 algorithm to recover the private key.
1573 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1578 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1579 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1580 of two gigabytes and the error handling improved.
1582 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1583 categorized as a normal bug, not a security issue, because the DRBG reseeds
1584 automatically and is fully functional even without additional randomness
1585 provided by the application.
1587 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1589 * Add a new ClientHello callback. Provides a callback interface that gives
1590 the application the ability to adjust the nascent SSL object at the
1591 earliest stage of ClientHello processing, immediately after extensions have
1592 been collected but before they have been processed. In particular, this
1593 callback can adjust the supported TLS versions in response to the contents
1598 * Add SM2 base algorithm support.
1602 * s390x assembly pack: add (improved) hardware-support for the following
1603 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1604 aes-cfb/cfb8, aes-ecb.
1608 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1609 parameter is no longer accepted, as it leads to a corrupt table. NULL
1610 pem_str is reserved for alias entries only.
1614 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1615 step for prime curves. The new implementation is based on formulae from
1616 differential addition-and-doubling in homogeneous projective coordinates
1617 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1618 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1619 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1620 to work in projective coordinates.
1622 *Billy Bob Brumley, Nicola Tuveri*
1624 * Change generating and checking of primes so that the error rate of not
1625 being prime depends on the intended use based on the size of the input.
1626 For larger primes this will result in more rounds of Miller-Rabin.
1627 The maximal error rate for primes with more than 1080 bits is lowered
1630 *Kurt Roeckx, Annie Yousar*
1632 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1636 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1637 moving between systems, and to avoid confusion when a Windows build is
1638 done with mingw vs with MSVC. For POSIX installs, there's still a
1639 symlink or copy named 'tsget' to avoid that confusion as well.
1643 * Revert blinding in ECDSA sign and instead make problematic addition
1644 length-invariant. Switch even to fixed-length Montgomery multiplication.
1648 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1649 step for binary curves. The new implementation is based on formulae from
1650 differential addition-and-doubling in mixed Lopez-Dahab projective
1651 coordinates, modified to independently blind the operands.
1653 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1655 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1656 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1657 EC_METHODs to implement their own specialized "ladder step", to take
1658 advantage of more favorable coordinate systems or more efficient
1659 differential addition-and-doubling algorithms.
1661 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1663 * Modified the random device based seed sources to keep the relevant
1664 file descriptors open rather than reopening them on each access.
1665 This allows such sources to operate in a chroot() jail without
1666 the associated device nodes being available. This behaviour can be
1667 controlled using RAND_keep_random_devices_open().
1671 * Numerous side-channel attack mitigations have been applied. This may have
1672 performance impacts for some algorithms for the benefit of improved
1673 security. Specific changes are noted in this change log by their respective
1678 * AIX shared library support overhaul. Switch to AIX "natural" way of
1679 handling shared libraries, which means collecting shared objects of
1680 different versions and bitnesses in one common archive. This allows to
1681 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1682 doesn't affect the way 3rd party applications are linked, only how
1683 multi-version installation is managed.
1687 * Make ec_group_do_inverse_ord() more robust and available to other
1688 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1689 mitigations are applied to the fallback BN_mod_inverse().
1690 When using this function rather than BN_mod_inverse() directly, new
1691 EC cryptosystem implementations are then safer-by-default.
1695 * Add coordinate blinding for EC_POINT and implement projective
1696 coordinate blinding for generic prime curves as a countermeasure to
1697 chosen point SCA attacks.
1699 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1701 * Add blinding to ECDSA and DSA signatures to protect against side channel
1702 attacks discovered by Keegan Ryan (NCC Group).
1706 * Enforce checking in the pkeyutl command line app to ensure that the input
1707 length does not exceed the maximum supported digest length when performing
1708 a sign, verify or verifyrecover operation.
1712 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1713 I/O in combination with something like select() or poll() will hang. This
1714 can be turned off again using SSL_CTX_clear_mode().
1715 Many applications do not properly handle non-application data records, and
1716 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1717 around the problems in those applications, but can also break some.
1718 It's recommended to read the manpages about SSL_read(), SSL_write(),
1719 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1720 SSL_CTX_set_read_ahead() again.
1724 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1725 now allow empty (zero character) pass phrases.
1729 * Apply blinding to binary field modular inversion and remove patent
1730 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1734 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1735 binary and prime elliptic curves.
1739 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1740 constant time fixed point multiplication.
1744 * Revise elliptic curve scalar multiplication with timing attack
1745 defenses: ec_wNAF_mul redirects to a constant time implementation
1746 when computing fixed point and variable point multiplication (which
1747 in OpenSSL are mostly used with secret scalars in keygen, sign,
1748 ECDH derive operations).
1749 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1752 * Updated CONTRIBUTING
1756 * Updated DRBG / RAND to request nonce and additional low entropy
1757 randomness from the system.
1759 *Matthias St. Pierre*
1761 * Updated 'openssl rehash' to use OpenSSL consistent default.
1765 * Moved the load of the ssl_conf module to libcrypto, which helps
1766 loading engines that libssl uses before libssl is initialised.
1770 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1774 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1776 *Ingo Schwarze, Rich Salz*
1778 * Added output of accepting IP address and port for 'openssl s_server'
1782 * Added a new API for TLSv1.3 ciphersuites:
1783 SSL_CTX_set_ciphersuites()
1784 SSL_set_ciphersuites()
1788 * Memory allocation failures consistently add an error to the error
1793 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1794 in libcrypto when run as setuid/setgid.
1798 * Load any config file by default when libssl is used.
1802 * Added new public header file <openssl/rand_drbg.h> and documentation
1803 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1805 *Matthias St. Pierre*
1807 * QNX support removed (cannot find contributors to get their approval
1808 for the license change).
1812 * TLSv1.3 replay protection for early data has been implemented. See the
1813 SSL_read_early_data() man page for further details.
1817 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1818 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1819 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1820 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1821 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1822 configuration has been separated out. See the ciphers man page or the
1823 SSL_CTX_set_ciphersuites() man page for more information.
1827 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1828 in responder mode now supports the new "-multi" option, which
1829 spawns the specified number of child processes to handle OCSP
1830 requests. The "-timeout" option now also limits the OCSP
1831 responder's patience to wait to receive the full client request
1832 on a newly accepted connection. Child processes are respawned
1833 as needed, and the CA index file is automatically reloaded
1834 when changed. This makes it possible to run the "ocsp" responder
1835 as a long-running service, making the OpenSSL CA somewhat more
1836 feature-complete. In this mode, most diagnostic messages logged
1837 after entering the event loop are logged via syslog(3) rather than
1842 * Added support for X448 and Ed448. Heavily based on original work by
1847 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1848 objects loaded. This adds the functions OSSL_STORE_expect() and
1849 OSSL_STORE_find() as well as needed tools to construct searches and
1850 get the search data out of them.
1854 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1855 version of OpenSSL should review their configuration settings to ensure
1856 that they are still appropriate for TLSv1.3. For further information see:
1857 <https://wiki.openssl.org/index.php/TLS1.3>
1861 * Grand redesign of the OpenSSL random generator
1863 The default RAND method now utilizes an AES-CTR DRBG according to
1864 NIST standard SP 800-90Ar1. The new random generator is essentially
1865 a port of the default random generator from the OpenSSL FIPS 2.0
1866 object module. It is a hybrid deterministic random bit generator
1867 using an AES-CTR bit stream and which seeds and reseeds itself
1868 automatically using trusted system entropy sources.
1870 Some of its new features are:
1871 - Support for multiple DRBG instances with seed chaining.
1872 - The default RAND method makes use of a DRBG.
1873 - There is a public and private DRBG instance.
1874 - The DRBG instances are fork-safe.
1875 - Keep all global DRBG instances on the secure heap if it is enabled.
1876 - The public and private DRBG instance are per thread for lock free
1879 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1881 * Changed Configure so it only says what it does and doesn't dump
1882 so much data. Instead, ./configdata.pm should be used as a script
1883 to display all sorts of configuration data.
1887 * Added processing of "make variables" to Configure.
1891 * Added SHA512/224 and SHA512/256 algorithm support.
1895 * The last traces of Netware support, first removed in 1.1.0, have
1900 * Get rid of Makefile.shared, and in the process, make the processing
1901 of certain files (rc.obj, or the .def/.map/.opt files produced from
1902 the ordinal files) more visible and hopefully easier to trace and
1903 debug (or make silent).
1907 * Make it possible to have environment variable assignments as
1908 arguments to config / Configure.
1912 * Add multi-prime RSA (RFC 8017) support.
1916 * Add SM3 implemented according to GB/T 32905-2016
1917 *Jack Lloyd <jack.lloyd@ribose.com>,*
1918 *Ronald Tse <ronald.tse@ribose.com>,*
1919 *Erick Borsboom <erick.borsboom@ribose.com>*
1921 * Add 'Maximum Fragment Length' TLS extension negotiation and support
1922 as documented in RFC6066.
1923 Based on a patch from Tomasz Moń
1925 *Filipe Raimundo da Silva*
1927 * Add SM4 implemented according to GB/T 32907-2016.
1928 *Jack Lloyd <jack.lloyd@ribose.com>,*
1929 *Ronald Tse <ronald.tse@ribose.com>,*
1930 *Erick Borsboom <erick.borsboom@ribose.com>*
1932 * Reimplement -newreq-nodes and ERR_error_string_n; the
1933 original author does not agree with the license change.
1937 * Add ARIA AEAD TLS support.
1941 * Some macro definitions to support VS6 have been removed. Visual
1942 Studio 6 has not worked since 1.1.0
1946 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
1947 without clearing the errors.
1951 * Add "atfork" functions. If building on a system that without
1952 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
1953 requirements. The RAND facility now uses/requires this.
1961 * The UI API becomes a permanent and integral part of libcrypto, i.e.
1962 not possible to disable entirely. However, it's still possible to
1963 disable the console reading UI method, UI_OpenSSL() (use UI_null()
1966 To disable, configure with 'no-ui-console'. 'no-ui' is still
1967 possible to use as an alias. Check at compile time with the
1968 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
1969 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
1973 * Add a STORE module, which implements a uniform and URI based reader of
1974 stores that can contain keys, certificates, CRLs and numerous other
1975 objects. The main API is loosely based on a few stdio functions,
1976 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
1977 OSSL_STORE_error and OSSL_STORE_close.
1978 The implementation uses backends called "loaders" to implement arbitrary
1979 URI schemes. There is one built in "loader" for the 'file' scheme.
1983 * Add devcrypto engine. This has been implemented against cryptodev-linux,
1984 then adjusted to work on FreeBSD 8.4 as well.
1985 Enable by configuring with 'enable-devcryptoeng'. This is done by default
1986 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
1990 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
1991 util/mkerr.pl, which is adapted to allow those prefixes, leading to
1992 error code calls like this:
1994 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
1996 With this change, we claim the namespaces OSSL and OPENSSL in a manner
1997 that can be encoded in C. For the foreseeable future, this will only
2000 *Richard Levitte and Tim Hudson*
2002 * Removed BSD cryptodev engine.
2006 * Add a build target 'build_all_generated', to build all generated files
2007 and only that. This can be used to prepare everything that requires
2008 things like perl for a system that lacks perl and then move everything
2009 to that system and do the rest of the build there.
2013 * In the UI interface, make it possible to duplicate the user data. This
2014 can be used by engines that need to retain the data for a longer time
2015 than just the call where this user data is passed.
2019 * Ignore the '-named_curve auto' value for compatibility of applications
2022 *Tomas Mraz <tmraz@fedoraproject.org>*
2024 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2025 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2026 alerts across multiple records (some of which could be empty). In practice
2027 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2028 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2029 support this at all. Supporting it adds significant complexity to the
2030 record layer, and its removal is unlikely to cause interoperability
2035 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2036 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2037 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2042 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2043 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2045 *Richard Levitte, Andy Polyakov*
2047 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2052 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2053 platform rather than 'mingw'.
2057 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2058 success if they are asked to add an object which already exists
2059 in the store. This change cascades to other functions which load
2060 certificates and CRLs.
2064 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2065 facilitate stack unwinding even from assembly subroutines.
2069 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2070 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2074 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2075 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2076 which is the minimum version we support.
2080 * Certificate time validation (X509_cmp_time) enforces stricter
2081 compliance with RFC 5280. Fractional seconds and timezone offsets
2082 are no longer allowed.
2086 * Add support for ARIA
2090 * s_client will now send the Server Name Indication (SNI) extension by
2091 default unless the new "-noservername" option is used. The server name is
2092 based on the host provided to the "-connect" option unless overridden by
2093 using "-servername".
2097 * Add support for SipHash
2101 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2102 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2103 prevent issues where no progress is being made and the peer continually
2104 sends unrecognised record types, using up resources processing them.
2108 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2109 using the algorithm defined in
2110 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2114 * Heartbeat support has been removed; the ABI is changed for now.
2116 *Richard Levitte, Rich Salz*
2118 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2122 * The RSA "null" method, which was partially supported to avoid patent
2123 issues, has been replaced to always returns NULL.
2130 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2132 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2133 used even when parsing explicit parameters, when loading a encoded key
2134 or calling `EC_GROUP_new_from_ecpkparameters()`/
2135 `EC_GROUP_new_from_ecparameters()`.
2136 This prevents bypass of security hardening and performance gains,
2137 especially for curves with specialized EC_METHODs.
2138 By default, if a key encoded with explicit parameters is loaded and later
2139 encoded, the output is still encoded with explicit parameters, even if
2140 internally a "named" EC_GROUP is used for computation.
2144 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2145 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2146 NULL. After this change, only the cofactor parameter can be NULL. It also
2147 does some minimal sanity checks on the passed order.
2152 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2153 An attack is simple, if the first CMS_recipientInfo is valid but the
2154 second CMS_recipientInfo is chosen ciphertext. If the second
2155 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2156 encryption key will be replaced by garbage, and the message cannot be
2157 decoded, but if the RSA decryption fails, the correct encryption key is
2158 used and the recipient will not notice the attack.
2159 As a work around for this potential attack the length of the decrypted
2160 key must be equal to the cipher default key length, in case the
2161 certifiate is not given and all recipientInfo are tried out.
2162 The old behaviour can be re-enabled in the CMS code by setting the
2163 CMS_DEBUG_DECRYPT flag.
2168 * Use Windows installation paths in the mingw builds
2170 Mingw isn't a POSIX environment per se, which means that Windows
2171 paths should be used for installation.
2176 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2178 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2179 This changes the size when using the genpkey app when no size is given. It
2180 fixes an omission in earlier changes that changed all RSA, DSA and DH
2181 generation apps to use 2048 bits by default.
2185 * Prevent over long nonces in ChaCha20-Poly1305.
2187 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2188 for every encryption operation. RFC 7539 specifies that the nonce value
2189 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2190 and front pads the nonce with 0 bytes if it is less than 12
2191 bytes. However it also incorrectly allows a nonce to be set of up to 16
2192 bytes. In this case only the last 12 bytes are significant and any
2193 additional leading bytes are ignored.
2195 It is a requirement of using this cipher that nonce values are
2196 unique. Messages encrypted using a reused nonce value are susceptible to
2197 serious confidentiality and integrity attacks. If an application changes
2198 the default nonce length to be longer than 12 bytes and then makes a
2199 change to the leading bytes of the nonce expecting the new value to be a
2200 new unique nonce then such an application could inadvertently encrypt
2201 messages with a reused nonce.
2203 Additionally the ignored bytes in a long nonce are not covered by the
2204 integrity guarantee of this cipher. Any application that relies on the
2205 integrity of these ignored leading bytes of a long nonce may be further
2206 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2207 is safe because no such use sets such a long nonce value. However user
2208 applications that use this cipher directly and set a non-default nonce
2209 length to be longer than 12 bytes may be vulnerable.
2211 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2217 * Added SCA hardening for modular field inversion in EC_GROUP through
2218 a new dedicated field_inv() pointer in EC_METHOD.
2219 This also addresses a leakage affecting conversions from projective
2220 to affine coordinates.
2222 *Billy Bob Brumley, Nicola Tuveri*
2224 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2225 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2229 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2233 * Remove the 'dist' target and add a tarball building script. The
2234 'dist' target has fallen out of use, and it shouldn't be
2235 necessary to configure just to create a source distribution.
2239 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2241 * Timing vulnerability in DSA signature generation
2243 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2244 timing side channel attack. An attacker could use variations in the signing
2245 algorithm to recover the private key.
2247 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2252 * Timing vulnerability in ECDSA signature generation
2254 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2255 timing side channel attack. An attacker could use variations in the signing
2256 algorithm to recover the private key.
2258 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2263 * Add coordinate blinding for EC_POINT and implement projective
2264 coordinate blinding for generic prime curves as a countermeasure to
2265 chosen point SCA attacks.
2267 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2269 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2271 * Client DoS due to large DH parameter
2273 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2274 malicious server can send a very large prime value to the client. This will
2275 cause the client to spend an unreasonably long period of time generating a
2276 key for this prime resulting in a hang until the client has finished. This
2277 could be exploited in a Denial Of Service attack.
2279 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2284 * Cache timing vulnerability in RSA Key Generation
2286 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2287 a cache timing side channel attack. An attacker with sufficient access to
2288 mount cache timing attacks during the RSA key generation process could
2289 recover the private key.
2291 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2292 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2297 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2298 parameter is no longer accepted, as it leads to a corrupt table. NULL
2299 pem_str is reserved for alias entries only.
2303 * Revert blinding in ECDSA sign and instead make problematic addition
2304 length-invariant. Switch even to fixed-length Montgomery multiplication.
2308 * Change generating and checking of primes so that the error rate of not
2309 being prime depends on the intended use based on the size of the input.
2310 For larger primes this will result in more rounds of Miller-Rabin.
2311 The maximal error rate for primes with more than 1080 bits is lowered
2314 *Kurt Roeckx, Annie Yousar*
2316 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2320 * Add blinding to ECDSA and DSA signatures to protect against side channel
2321 attacks discovered by Keegan Ryan (NCC Group).
2325 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2326 now allow empty (zero character) pass phrases.
2330 * Certificate time validation (X509_cmp_time) enforces stricter
2331 compliance with RFC 5280. Fractional seconds and timezone offsets
2332 are no longer allowed.
2336 * Fixed a text canonicalisation bug in CMS
2338 Where a CMS detached signature is used with text content the text goes
2339 through a canonicalisation process first prior to signing or verifying a
2340 signature. This process strips trailing space at the end of lines, converts
2341 line terminators to CRLF and removes additional trailing line terminators
2342 at the end of a file. A bug in the canonicalisation process meant that
2343 some characters, such as form-feed, were incorrectly treated as whitespace
2344 and removed. This is contrary to the specification (RFC5485). This fix
2345 could mean that detached text data signed with an earlier version of
2346 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2347 signed with a fixed OpenSSL may fail to verify with an earlier version of
2348 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2349 and use the "-binary" flag (for the "cms" command line application) or set
2350 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2354 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2356 * Constructed ASN.1 types with a recursive definition could exceed the stack
2358 Constructed ASN.1 types with a recursive definition (such as can be found
2359 in PKCS7) could eventually exceed the stack given malicious input with
2360 excessive recursion. This could result in a Denial Of Service attack. There
2361 are no such structures used within SSL/TLS that come from untrusted sources
2362 so this is considered safe.
2364 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2370 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2372 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2373 effectively reduced to only comparing the least significant bit of each
2374 byte. This allows an attacker to forge messages that would be considered as
2375 authenticated in an amount of tries lower than that guaranteed by the
2376 security claims of the scheme. The module can only be compiled by the
2377 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2379 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2385 * Add a build target 'build_all_generated', to build all generated files
2386 and only that. This can be used to prepare everything that requires
2387 things like perl for a system that lacks perl and then move everything
2388 to that system and do the rest of the build there.
2392 * Backport SSL_OP_NO_RENGOTIATION
2394 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2395 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2396 changes this is no longer possible in 1.1.0. Therefore the new
2397 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2398 1.1.0 to provide equivalent functionality.
2400 Note that if an application built against 1.1.0h headers (or above) is run
2401 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2402 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2406 * Removed the OS390-Unix config target. It relied on a script that doesn't
2411 * rsaz_1024_mul_avx2 overflow bug on x86_64
2413 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2414 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2415 Analysis suggests that attacks against RSA and DSA as a result of this
2416 defect would be very difficult to perform and are not believed likely.
2417 Attacks against DH1024 are considered just feasible, because most of the
2418 work necessary to deduce information about a private key may be performed
2419 offline. The amount of resources required for such an attack would be
2420 significant. However, for an attack on TLS to be meaningful, the server
2421 would have to share the DH1024 private key among multiple clients, which is
2422 no longer an option since CVE-2016-0701.
2424 This only affects processors that support the AVX2 but not ADX extensions
2425 like Intel Haswell (4th generation).
2427 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2428 was originally found via the OSS-Fuzz project.
2433 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2435 * bn_sqrx8x_internal carry bug on x86_64
2437 There is a carry propagating bug in the x86_64 Montgomery squaring
2438 procedure. No EC algorithms are affected. Analysis suggests that attacks
2439 against RSA and DSA as a result of this defect would be very difficult to
2440 perform and are not believed likely. Attacks against DH are considered just
2441 feasible (although very difficult) because most of the work necessary to
2442 deduce information about a private key may be performed offline. The amount
2443 of resources required for such an attack would be very significant and
2444 likely only accessible to a limited number of attackers. An attacker would
2445 additionally need online access to an unpatched system using the target
2446 private key in a scenario with persistent DH parameters and a private
2447 key that is shared between multiple clients.
2449 This only affects processors that support the BMI1, BMI2 and ADX extensions
2450 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2452 This issue was reported to OpenSSL by the OSS-Fuzz project.
2457 * Malformed X.509 IPAddressFamily could cause OOB read
2459 If an X.509 certificate has a malformed IPAddressFamily extension,
2460 OpenSSL could do a one-byte buffer overread. The most likely result
2461 would be an erroneous display of the certificate in text format.
2463 This issue was reported to OpenSSL by the OSS-Fuzz project.
2468 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2470 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2471 platform rather than 'mingw'.
2475 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2476 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2477 which is the minimum version we support.
2481 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2483 * Encrypt-Then-Mac renegotiation crash
2485 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2486 negotiated where it was not in the original handshake (or vice-versa) then
2487 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2488 and servers are affected.
2490 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2495 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2497 * Truncated packet could crash via OOB read
2499 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2500 cipher is being used, then a truncated packet can cause that host to
2501 perform an out-of-bounds read, usually resulting in a crash.
2503 This issue was reported to OpenSSL by Robert Święcki of Google.
2508 * Bad (EC)DHE parameters cause a client crash
2510 If a malicious server supplies bad parameters for a DHE or ECDHE key
2511 exchange then this can result in the client attempting to dereference a
2512 NULL pointer leading to a client crash. This could be exploited in a Denial
2515 This issue was reported to OpenSSL by Guido Vranken.
2520 * BN_mod_exp may produce incorrect results on x86_64
2522 There is a carry propagating bug in the x86_64 Montgomery squaring
2523 procedure. No EC algorithms are affected. Analysis suggests that attacks
2524 against RSA and DSA as a result of this defect would be very difficult to
2525 perform and are not believed likely. Attacks against DH are considered just
2526 feasible (although very difficult) because most of the work necessary to
2527 deduce information about a private key may be performed offline. The amount
2528 of resources required for such an attack would be very significant and
2529 likely only accessible to a limited number of attackers. An attacker would
2530 additionally need online access to an unpatched system using the target
2531 private key in a scenario with persistent DH parameters and a private
2532 key that is shared between multiple clients. For example this can occur by
2533 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2534 similar to CVE-2015-3193 but must be treated as a separate problem.
2536 This issue was reported to OpenSSL by the OSS-Fuzz project.
2541 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2543 * ChaCha20/Poly1305 heap-buffer-overflow
2545 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2546 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2547 crash. This issue is not considered to be exploitable beyond a DoS.
2549 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2554 * CMS Null dereference
2556 Applications parsing invalid CMS structures can crash with a NULL pointer
2557 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2558 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2559 structure callback if an attempt is made to free certain invalid encodings.
2560 Only CHOICE structures using a callback which do not handle NULL value are
2563 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2568 * Montgomery multiplication may produce incorrect results
2570 There is a carry propagating bug in the Broadwell-specific Montgomery
2571 multiplication procedure that handles input lengths divisible by, but
2572 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2573 and DH private keys are impossible. This is because the subroutine in
2574 question is not used in operations with the private key itself and an input
2575 of the attacker's direct choice. Otherwise the bug can manifest itself as
2576 transient authentication and key negotiation failures or reproducible
2577 erroneous outcome of public-key operations with specially crafted input.
2578 Among EC algorithms only Brainpool P-512 curves are affected and one
2579 presumably can attack ECDH key negotiation. Impact was not analyzed in
2580 detail, because pre-requisites for attack are considered unlikely. Namely
2581 multiple clients have to choose the curve in question and the server has to
2582 share the private key among them, neither of which is default behaviour.
2583 Even then only clients that chose the curve will be affected.
2585 This issue was publicly reported as transient failures and was not
2586 initially recognized as a security issue. Thanks to Richard Morgan for
2587 providing reproducible case.
2592 * Removed automatic addition of RPATH in shared libraries and executables,
2593 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2597 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2599 * Fix Use After Free for large message sizes
2601 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2602 message larger than approx 16k is received then the underlying buffer to
2603 store the incoming message is reallocated and moved. Unfortunately a
2604 dangling pointer to the old location is left which results in an attempt to
2605 write to the previously freed location. This is likely to result in a
2606 crash, however it could potentially lead to execution of arbitrary code.
2608 This issue only affects OpenSSL 1.1.0a.
2610 This issue was reported to OpenSSL by Robert Święcki.
2615 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2617 * OCSP Status Request extension unbounded memory growth
2619 A malicious client can send an excessively large OCSP Status Request
2620 extension. If that client continually requests renegotiation, sending a
2621 large OCSP Status Request extension each time, then there will be unbounded
2622 memory growth on the server. This will eventually lead to a Denial Of
2623 Service attack through memory exhaustion. Servers with a default
2624 configuration are vulnerable even if they do not support OCSP. Builds using
2625 the "no-ocsp" build time option are not affected.
2627 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2632 * SSL_peek() hang on empty record
2634 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2635 sends an empty record. This could be exploited by a malicious peer in a
2636 Denial Of Service attack.
2638 This issue was reported to OpenSSL by Alex Gaynor.
2643 * Excessive allocation of memory in tls_get_message_header() and
2644 dtls1_preprocess_fragment()
2646 A (D)TLS message includes 3 bytes for its length in the header for the
2647 message. This would allow for messages up to 16Mb in length. Messages of
2648 this length are excessive and OpenSSL includes a check to ensure that a
2649 peer is sending reasonably sized messages in order to avoid too much memory
2650 being consumed to service a connection. A flaw in the logic of version
2651 1.1.0 means that memory for the message is allocated too early, prior to
2652 the excessive message length check. Due to way memory is allocated in
2653 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2654 to service a connection. This could lead to a Denial of Service through
2655 memory exhaustion. However, the excessive message length check still takes
2656 place, and this would cause the connection to immediately fail. Assuming
2657 that the application calls SSL_free() on the failed connection in a timely
2658 manner then the 21Mb of allocated memory will then be immediately freed
2659 again. Therefore the excessive memory allocation will be transitory in
2660 nature. This then means that there is only a security impact if:
2662 1) The application does not call SSL_free() in a timely manner in the event
2663 that the connection fails
2665 2) The application is working in a constrained environment where there is
2666 very little free memory
2668 3) The attacker initiates multiple connection attempts such that there are
2669 multiple connections in a state where memory has been allocated for the
2670 connection; SSL_free() has not yet been called; and there is insufficient
2671 memory to service the multiple requests.
2673 Except in the instance of (1) above any Denial Of Service is likely to be
2674 transitory because as soon as the connection fails the memory is
2675 subsequently freed again in the SSL_free() call. However there is an
2676 increased risk during this period of application crashes due to the lack of
2677 memory - which would then mean a more serious Denial of Service.
2679 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2680 (CVE-2016-6307 and CVE-2016-6308)
2684 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2685 had to be removed. Primary reason is that vendor assembler can't
2686 assemble our modules with -KPIC flag. As result it, assembly
2687 support, was not even available as option. But its lack means
2688 lack of side-channel resistant code, which is incompatible with
2689 security by todays standards. Fortunately gcc is readily available
2690 prepackaged option, which we firmly point at...
2694 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2696 * Windows command-line tool supports UTF-8 opt-in option for arguments
2697 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2698 (to any value) allows Windows user to access PKCS#12 file generated
2699 with Windows CryptoAPI and protected with non-ASCII password, as well
2700 as files generated under UTF-8 locale on Linux also protected with
2705 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2706 have been disabled by default and removed from DEFAULT, just like RC4.
2707 See the RC4 item below to re-enable both.
2711 * The method for finding the storage location for the Windows RAND seed file
2712 has changed. First we check %RANDFILE%. If that is not set then we check
2713 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2714 all else fails we fall back to C:\.
2718 * The EVP_EncryptUpdate() function has had its return type changed from void
2719 to int. A return of 0 indicates and error while a return of 1 indicates
2724 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2725 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2726 off the constant time implementation for RSA, DSA and DH have been made
2727 no-ops and deprecated.
2731 * Windows RAND implementation was simplified to only get entropy by
2732 calling CryptGenRandom(). Various other RAND-related tickets
2735 *Joseph Wylie Yandle, Rich Salz*
2737 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2738 and `OPENSSL_LH_`, respectively. The old names are available
2739 with API compatibility. They new names are now completely documented.
2743 * Unify TYPE_up_ref(obj) methods signature.
2744 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2745 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2746 int (instead of void) like all others TYPE_up_ref() methods.
2747 So now these methods also check the return value of CRYPTO_atomic_add(),
2748 and the validity of object reference counter.
2750 *fdasilvayy@gmail.com*
2752 * With Windows Visual Studio builds, the .pdb files are installed
2753 alongside the installed libraries and executables. For a static
2754 library installation, ossl_static.pdb is the associate compiler
2755 generated .pdb file to be used when linking programs.
2759 * Remove openssl.spec. Packaging files belong with the packagers.
2763 * Automatic Darwin/OSX configuration has had a refresh, it will now
2764 recognise x86_64 architectures automatically. You can still decide
2765 to build for a different bitness with the environment variable
2766 KERNEL_BITS (can be 32 or 64), for example:
2768 KERNEL_BITS=32 ./config
2772 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2773 256 bit AES and HMAC with SHA256.
2777 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2781 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2785 * To enable users to have their own config files and build file templates,
2786 Configure looks in the directory indicated by the environment variable
2787 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2788 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2789 name and is used as is.
2793 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2794 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2795 X509_CERT_FILE_CTX was removed.
2799 * "shared" builds are now the default. To create only static libraries use
2800 the "no-shared" Configure option.
2804 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2805 All of these option have not worked for some while and are fundamental
2810 * Make various cleanup routines no-ops and mark them as deprecated. Most
2811 global cleanup functions are no longer required because they are handled
2812 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2813 Explicitly de-initing can cause problems (e.g. where a library that uses
2814 OpenSSL de-inits, but an application is still using it). The affected
2815 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2816 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2817 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2818 COMP_zlib_cleanup().
2822 * --strict-warnings no longer enables runtime debugging options
2823 such as REF_DEBUG. Instead, debug options are automatically
2824 enabled with '--debug' builds.
2826 *Andy Polyakov, Emilia Käsper*
2828 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2829 have been moved out of the public header files. New functions for managing
2830 these have been added.
2834 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2835 objects have been moved out of the public header files. New
2836 functions for managing these have been added.
2840 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2841 have been moved out of the public header files. New functions for managing
2842 these have been added.
2846 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2847 moved out of the public header files. New functions for managing these
2852 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2856 * Removed the mk1mf build scripts.
2860 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2861 it is always safe to #include a header now.
2865 * Removed the aged BC-32 config and all its supporting scripts
2869 * Removed support for Ultrix, Netware, and OS/2.
2873 * Add support for HKDF.
2875 *Alessandro Ghedini*
2877 * Add support for blake2b and blake2s
2881 * Added support for "pipelining". Ciphers that have the
2882 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2883 encryptions/decryptions simultaneously. There are currently no built-in
2884 ciphers with this property but the expectation is that engines will be able
2885 to offer it to significantly improve throughput. Support has been extended
2886 into libssl so that multiple records for a single connection can be
2887 processed in one go (for >=TLS 1.1).
2891 * Added the AFALG engine. This is an async capable engine which is able to
2892 offload work to the Linux kernel. In this initial version it only supports
2893 AES128-CBC. The kernel must be version 4.1.0 or greater.
2897 * OpenSSL now uses a new threading API. It is no longer necessary to
2898 set locking callbacks to use OpenSSL in a multi-threaded environment. There
2899 are two supported threading models: pthreads and windows threads. It is
2900 also possible to configure OpenSSL at compile time for "no-threads". The
2901 old threading API should no longer be used. The functions have been
2902 replaced with "no-op" compatibility macros.
2904 *Alessandro Ghedini, Matt Caswell*
2906 * Modify behavior of ALPN to invoke callback after SNI/servername
2907 callback, such that updates to the SSL_CTX affect ALPN.
2911 * Add SSL_CIPHER queries for authentication and key-exchange.
2915 * Changes to the DEFAULT cipherlist:
2916 - Prefer (EC)DHE handshakes over plain RSA.
2917 - Prefer AEAD ciphers over legacy ciphers.
2918 - Prefer ECDSA over RSA when both certificates are available.
2919 - Prefer TLSv1.2 ciphers/PRF.
2920 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
2925 * Change the ECC default curve list to be this, in order: x25519,
2926 secp256r1, secp521r1, secp384r1.
2930 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
2931 disabled by default. They can be re-enabled using the
2932 enable-weak-ssl-ciphers option to Configure.
2936 * If the server has ALPN configured, but supports no protocols that the
2937 client advertises, send a fatal "no_application_protocol" alert.
2938 This behaviour is SHALL in RFC 7301, though it isn't universally
2939 implemented by other servers.
2943 * Add X25519 support.
2944 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
2945 for public and private key encoding using the format documented in
2946 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
2947 key generation and key derivation.
2949 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
2954 * Deprecate SRP_VBASE_get_by_user.
2955 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2956 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
2957 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
2958 seed, even if the seed is configured.
2960 Users should use SRP_VBASE_get1_by_user instead. Note that in
2961 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2962 also that even though configuring the SRP seed attempts to hide
2963 invalid usernames by continuing the handshake with fake
2964 credentials, this behaviour is not constant time and no strong
2965 guarantees are made that the handshake is indistinguishable from
2966 that of a valid user.
2970 * Configuration change; it's now possible to build dynamic engines
2971 without having to build shared libraries and vice versa. This
2972 only applies to the engines in engines/, those in crypto/engine/
2973 will always be built into libcrypto (i.e. "static").
2975 Building dynamic engines is enabled by default; to disable, use
2976 the configuration option "disable-dynamic-engine".
2978 The only requirements for building dynamic engines are the
2979 presence of the DSO module and building with position independent
2980 code, so they will also automatically be disabled if configuring
2981 with "disable-dso" or "disable-pic".
2983 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
2984 are also taken away from openssl/opensslconf.h, as they are
2989 * Configuration change; if there is a known flag to compile
2990 position independent code, it will always be applied on the
2991 libcrypto and libssl object files, and never on the application
2992 object files. This means other libraries that use routines from
2993 libcrypto / libssl can be made into shared libraries regardless
2994 of how OpenSSL was configured.
2996 If this isn't desirable, the configuration options "disable-pic"
2997 or "no-pic" can be used to disable the use of PIC. This will
2998 also disable building shared libraries and dynamic engines.
3002 * Removed JPAKE code. It was experimental and has no wide use.
3006 * The INSTALL_PREFIX Makefile variable has been renamed to
3007 DESTDIR. That makes for less confusion on what this variable
3008 is for. Also, the configuration option --install_prefix is
3013 * Heartbeat for TLS has been removed and is disabled by default
3014 for DTLS; configure with enable-heartbeats. Code that uses the
3015 old #define's might need to be updated.
3017 *Emilia Käsper, Rich Salz*
3019 * Rename REF_CHECK to REF_DEBUG.
3023 * New "unified" build system
3025 The "unified" build system is aimed to be a common system for all
3026 platforms we support. With it comes new support for VMS.
3028 This system builds supports building in a different directory tree
3029 than the source tree. It produces one Makefile (for unix family
3030 or lookalikes), or one descrip.mms (for VMS).
3032 The source of information to make the Makefile / descrip.mms is
3033 small files called 'build.info', holding the necessary
3034 information for each directory with source to compile, and a
3035 template in Configurations, like unix-Makefile.tmpl or
3038 With this change, the library names were also renamed on Windows
3039 and on VMS. They now have names that are closer to the standard
3040 on Unix, and include the major version number, and in certain
3041 cases, the architecture they are built for. See "Notes on shared
3042 libraries" in INSTALL.
3044 We rely heavily on the perl module Text::Template.
3048 * Added support for auto-initialisation and de-initialisation of the library.
3049 OpenSSL no longer requires explicit init or deinit routines to be called,
3050 except in certain circumstances. See the OPENSSL_init_crypto() and
3051 OPENSSL_init_ssl() man pages for further information.
3055 * The arguments to the DTLSv1_listen function have changed. Specifically the
3056 "peer" argument is now expected to be a BIO_ADDR object.
3058 * Rewrite of BIO networking library. The BIO library lacked consistent
3059 support of IPv6, and adding it required some more extensive
3060 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3061 which hold all types of addresses and chains of address information.
3062 It also introduces a new API, with functions like BIO_socket,
3063 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3064 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3065 have been adapted accordingly.
3069 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3074 * CRIME protection: disable compression by default, even if OpenSSL is
3075 compiled with zlib enabled. Applications can still enable compression
3076 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3077 using the SSL_CONF library to configure compression.
3081 * The signature of the session callback configured with
3082 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3083 was explicitly marked as `const unsigned char*` instead of
3088 * Always DPURIFY. Remove the use of uninitialized memory in the
3089 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3093 * Removed many obsolete configuration items, including
3094 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3095 MD2_CHAR, MD2_INT, MD2_LONG
3097 IDEA_SHORT, IDEA_LONG
3098 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3100 *Rich Salz, with advice from Andy Polyakov*
3102 * Many BN internals have been moved to an internal header file.
3104 *Rich Salz with help from Andy Polyakov*
3106 * Configuration and writing out the results from it has changed.
3107 Files such as Makefile include/openssl/opensslconf.h and are now
3108 produced through general templates, such as Makefile.in and
3109 crypto/opensslconf.h.in and some help from the perl module
3112 Also, the center of configuration information is no longer
3113 Makefile. Instead, Configure produces a perl module in
3114 configdata.pm which holds most of the config data (in the hash
3115 table %config), the target data that comes from the target
3116 configuration in one of the `Configurations/*.conf` files (in
3121 * To clarify their intended purposes, the Configure options
3122 --prefix and --openssldir change their semantics, and become more
3123 straightforward and less interdependent.
3125 --prefix shall be used exclusively to give the location INSTALLTOP
3126 where programs, scripts, libraries, include files and manuals are
3127 going to be installed. The default is now /usr/local.
3129 --openssldir shall be used exclusively to give the default
3130 location OPENSSLDIR where certificates, private keys, CRLs are
3131 managed. This is also where the default openssl.cnf gets
3133 If the directory given with this option is a relative path, the
3134 values of both the --prefix value and the --openssldir value will
3135 be combined to become OPENSSLDIR.
3136 The default for --openssldir is INSTALLTOP/ssl.
3138 Anyone who uses --openssldir to specify where OpenSSL is to be
3139 installed MUST change to use --prefix instead.
3143 * The GOST engine was out of date and therefore it has been removed. An up
3144 to date GOST engine is now being maintained in an external repository.
3145 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3146 support for GOST ciphersuites (these are only activated if a GOST engine
3151 * EGD is no longer supported by default; use enable-egd when
3154 *Ben Kaduk and Rich Salz*
3156 * The distribution now has Makefile.in files, which are used to
3157 create Makefile's when Configure is run. *Configure must be run
3158 before trying to build now.*
3162 * The return value for SSL_CIPHER_description() for error conditions
3167 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3169 Obtaining and performing DNSSEC validation of TLSA records is
3170 the application's responsibility. The application provides
3171 the TLSA records of its choice to OpenSSL, and these are then
3172 used to authenticate the peer.
3174 The TLSA records need not even come from DNS. They can, for
3175 example, be used to implement local end-entity certificate or
3176 trust-anchor "pinning", where the "pin" data takes the form
3177 of TLSA records, which can augment or replace verification
3178 based on the usual WebPKI public certification authorities.
3182 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3183 continues to support deprecated interfaces in default builds.
3184 However, applications are strongly advised to compile their
3185 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3186 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3187 or the 1.1.0 releases.
3189 In environments in which all applications have been ported to
3190 not use any deprecated interfaces OpenSSL's Configure script
3191 should be used with the --api=1.1.0 option to entirely remove
3192 support for the deprecated features from the library and
3193 unconditionally disable them in the installed headers.
3194 Essentially the same effect can be achieved with the "no-deprecated"
3195 argument to Configure, except that this will always restrict
3196 the build to just the latest API, rather than a fixed API
3199 As applications are ported to future revisions of the API,
3200 they should update their compile-time OPENSSL_API_COMPAT define
3201 accordingly, but in most cases should be able to continue to
3202 compile with later releases.
3204 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3205 0x10000000L and 0x00908000L, respectively. However those
3206 versions did not support the OPENSSL_API_COMPAT feature, and
3207 so applications are not typically tested for explicit support
3208 of just the undeprecated features of either release.
3212 * Add support for setting the minimum and maximum supported protocol.
3213 It can bet set via the SSL_set_min_proto_version() and
3214 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3215 MaxProtocol. It's recommended to use the new APIs to disable
3216 protocols instead of disabling individual protocols using
3217 SSL_set_options() or SSL_CONF's Protocol. This change also
3218 removes support for disabling TLS 1.2 in the OpenSSL TLS
3219 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3223 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3227 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3228 and integrates ECDSA and ECDH functionality into EC. Implementations can
3229 now redirect key generation and no longer need to convert to or from
3232 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3233 include the ec.h header file instead.
3237 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3238 ciphers who are no longer supported and drops support the ephemeral RSA key
3239 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3243 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3244 opaque. For HMAC_CTX, the following constructors and destructors
3247 HMAC_CTX *HMAC_CTX_new(void);
3248 void HMAC_CTX_free(HMAC_CTX *ctx);
3250 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3251 destroy such methods has been added. See EVP_MD_meth_new(3) and
3252 EVP_CIPHER_meth_new(3) for documentation.
3255 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3256 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3257 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3258 an already created structure.
3259 2) For consistency with the majority of our object creators and
3260 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3261 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3262 for deprecated builds.
3266 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3267 cryptographic operations to be performed asynchronously as long as an
3268 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3269 further details. Libssl has also had this capability integrated with the
3270 introduction of the new mode SSL_MODE_ASYNC and associated error
3271 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3272 pages. This work was developed in partnership with Intel Corp.
3276 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3277 always enabled now. If you want to disable the support you should
3278 exclude it using the list of supported ciphers. This also means that the
3279 "-no_ecdhe" option has been removed from s_server.
3283 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3284 SSL_{CTX_}set1_curves() which can set a list.
3288 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3289 curve you want to support using SSL_{CTX_}set1_curves().
3293 * State machine rewrite. The state machine code has been significantly
3294 refactored in order to remove much duplication of code and solve issues
3295 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3296 further details). This change does have some associated API changes.
3297 Notably the SSL_state() function has been removed and replaced by
3298 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3299 SSL_set_state() has been removed altogether. The previous handshake states
3300 defined in ssl.h and ssl3.h have also been removed.
3304 * All instances of the string "ssleay" in the public API were replaced
3305 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3306 Some error codes related to internal RSA_eay API's were renamed.
3310 * The demo files in crypto/threads were moved to demo/threads.
3314 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3317 *Matt Caswell, Rich Salz*
3319 * New ASN.1 embed macro.
3321 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3322 structure is not allocated: it is part of the parent. That is instead of
3330 This reduces memory fragmentation and make it impossible to accidentally
3331 set a mandatory field to NULL.
3333 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3334 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3335 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3340 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3344 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3345 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3346 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3347 DES and RC4 ciphersuites.
3351 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3352 This changes the decoding behaviour for some invalid messages,
3353 though the change is mostly in the more lenient direction, and
3354 legacy behaviour is preserved as much as possible.
3358 * Fix no-stdio build.
3359 *David Woodhouse <David.Woodhouse@intel.com> and also*
3360 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3362 * New testing framework
3363 The testing framework has been largely rewritten and is now using
3364 perl and the perl modules Test::Harness and an extended variant of
3365 Test::More called OpenSSL::Test to do its work. All test scripts in
3366 test/ have been rewritten into test recipes, and all direct calls to
3367 executables in test/Makefile have become individual recipes using the
3368 simplified testing OpenSSL::Test::Simple.
3370 For documentation on our testing modules, do:
3372 perldoc test/testlib/OpenSSL/Test/Simple.pm
3373 perldoc test/testlib/OpenSSL/Test.pm
3377 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3378 are used; the latter aborts on memory leaks (usually checked on exit).
3379 Some undocumented "set malloc, etc., hooks" functions were removed
3380 and others were changed. All are now documented.
3384 * In DSA_generate_parameters_ex, if the provided seed is too short,
3387 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3389 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3390 from RFC4279, RFC4785, RFC5487, RFC5489.
3392 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3393 original RSA_PSK patch.
3397 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3398 era flag was never set throughout the codebase (only read). Also removed
3399 SSL3_FLAGS_POP_BUFFER which was only used if
3400 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3404 * Changed the default name options in the "ca", "crl", "req" and "x509"
3405 to be "oneline" instead of "compat".
3409 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3410 not aware of clients that still exhibit this bug, and the workaround
3411 hasn't been working properly for a while.
3415 * The return type of BIO_number_read() and BIO_number_written() as well as
3416 the corresponding num_read and num_write members in the BIO structure has
3417 changed from unsigned long to uint64_t. On platforms where an unsigned
3418 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3423 * Given the pervasive nature of TLS extensions it is inadvisable to run
3424 OpenSSL without support for them. It also means that maintaining
3425 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3426 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3430 * Removed support for the two export grade static DH ciphersuites
3431 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3432 were newly added (along with a number of other static DH ciphersuites) to
3433 1.0.2. However the two export ones have *never* worked since they were
3434 introduced. It seems strange in any case to be adding new export
3435 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3439 * Version negotiation has been rewritten. In particular SSLv23_method(),
3440 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3441 and turned into macros which simply call the new preferred function names
3442 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3443 should use the new names instead. Also as part of this change the ssl23.h
3444 header file has been removed.
3448 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3449 code and the associated standard is no longer considered fit-for-purpose.
3453 * RT2547 was closed. When generating a private key, try to make the
3454 output file readable only by the owner. This behavior change might
3455 be noticeable when interacting with other software.
3457 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3462 * Added HTTP GET support to the ocsp command.
3466 * Changed default digest for the dgst and enc commands from MD5 to
3471 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3475 * Added support for TLS extended master secret from
3476 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3477 initial patch which was a great help during development.
3481 * All libssl internal structures have been removed from the public header
3482 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3483 now redundant). Users should not attempt to access internal structures
3484 directly. Instead they should use the provided API functions.
3488 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3489 Access to deprecated functions can be re-enabled by running config with
3490 "enable-deprecated". In addition applications wishing to use deprecated
3491 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3492 will, by default, disable some transitive includes that previously existed
3493 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3497 * Added support for OCB mode. OpenSSL has been granted a patent license
3498 compatible with the OpenSSL license for use of OCB. Details are available
3499 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3500 for OCB can be removed by calling config with no-ocb.
3504 * SSLv2 support has been removed. It still supports receiving a SSLv2
3505 compatible client hello.
3509 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3510 done while fixing the error code for the key-too-small case.
3512 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3514 * CA.sh has been removed; use CA.pl instead.
3518 * Removed old DES API.
3522 * Remove various unsupported platforms:
3528 Sinix/ReliantUNIX RM400
3533 16-bit platforms such as WIN16
3537 * Clean up OPENSSL_NO_xxx #define's
3538 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3539 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3540 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3541 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3542 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3543 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3544 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3545 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3546 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3547 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3551 * Cleaned up dead code
3552 Remove all but one '#ifdef undef' which is to be looked at.
3556 * Clean up calling of xxx_free routines.
3557 Just like free(), fix most of the xxx_free routines to accept
3558 NULL. Remove the non-null checks from callers. Save much code.
3562 * Add secure heap for storage of private keys (when possible).
3563 Add BIO_s_secmem(), CBIGNUM, etc.
3564 Contributed by Akamai Technologies under our Corporate CLA.
3568 * Experimental support for a new, fast, unbiased prime candidate generator,
3569 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3571 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3573 * New output format NSS in the sess_id command line tool. This allows
3574 exporting the session id and the master key in NSS keylog format.
3576 *Martin Kaiser <martin@kaiser.cx>*
3578 * Harmonize version and its documentation. -f flag is used to display
3581 *mancha <mancha1@zoho.com>*
3583 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3584 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3586 *mancha <mancha1@zoho.com>*
3588 * Fix some double frees. These are not thought to be exploitable.
3590 *mancha <mancha1@zoho.com>*
3592 * A missing bounds check in the handling of the TLS heartbeat extension
3593 can be used to reveal up to 64k of memory to a connected client or
3596 Thanks for Neel Mehta of Google Security for discovering this bug and to
3597 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3598 preparing the fix ([CVE-2014-0160])
3600 *Adam Langley, Bodo Moeller*
3602 * Fix for the attack described in the paper "Recovering OpenSSL
3603 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3604 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3605 <http://eprint.iacr.org/2014/140>
3607 Thanks to Yuval Yarom and Naomi Benger for discovering this
3608 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3610 *Yuval Yarom and Naomi Benger*
3612 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3613 this fixes a limitation in previous versions of OpenSSL.
3617 * Experimental encrypt-then-mac support.
3619 Experimental support for encrypt then mac from
3620 draft-gutmann-tls-encrypt-then-mac-02.txt
3622 To enable it set the appropriate extension number (0x42 for the test
3623 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3625 For non-compliant peers (i.e. just about everything) this should have no
3628 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3632 * Add EVP support for key wrapping algorithms, to avoid problems with
3633 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3634 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3635 algorithms and include tests cases.
3639 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3644 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3645 MGF1 digest and OAEP label.
3649 * Make openssl verify return errors.
3651 *Chris Palmer <palmer@google.com> and Ben Laurie*
3653 * New function ASN1_TIME_diff to calculate the difference between two
3654 ASN1_TIME structures or one structure and the current time.
3658 * Update fips_test_suite to support multiple command line options. New
3659 test to induce all self test errors in sequence and check expected
3664 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3665 sign or verify all in one operation.
3669 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3670 test programs and fips_test_suite. Includes functionality to parse
3671 the minimal script output of fipsalgest.pl directly.
3675 * Add authorisation parameter to FIPS_module_mode_set().
3679 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3683 * Use separate DRBG fields for internal and external flags. New function
3684 FIPS_drbg_health_check() to perform on demand health checking. Add
3685 generation tests to fips_test_suite with reduced health check interval to
3686 demonstrate periodic health checking. Add "nodh" option to
3687 fips_test_suite to skip very slow DH test.
3691 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3696 * More extensive health check for DRBG checking many more failure modes.
3697 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3698 combination: call this in fips_test_suite.
3702 * Add support for canonical generation of DSA parameter 'g'. See
3705 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3706 POST to handle HMAC cases.
3710 * Add functions FIPS_module_version() and FIPS_module_version_text()
3711 to return numerical and string versions of the FIPS module number.
3715 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3716 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3717 outside the validated module in the FIPS capable OpenSSL.
3721 * Minor change to DRBG entropy callback semantics. In some cases
3722 there is no multiple of the block length between min_len and
3723 max_len. Allow the callback to return more than max_len bytes
3724 of entropy but discard any extra: it is the callback's responsibility
3725 to ensure that the extra data discarded does not impact the
3726 requested amount of entropy.
3730 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3731 information in FIPS186-3, SP800-57 and SP800-131A.
3735 * CCM support via EVP. Interface is very similar to GCM case except we
3736 must supply all data in one chunk (i.e. no update, final) and the
3737 message length must be supplied if AAD is used. Add algorithm test
3742 * Initial version of POST overhaul. Add POST callback to allow the status
3743 of POST to be monitored and/or failures induced. Modify fips_test_suite
3744 to use callback. Always run all selftests even if one fails.
3748 * XTS support including algorithm test driver in the fips_gcmtest program.
3749 Note: this does increase the maximum key length from 32 to 64 bytes but
3750 there should be no binary compatibility issues as existing applications
3751 will never use XTS mode.
3755 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3756 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3757 performs algorithm blocking for unapproved PRNG types. Also do not
3758 set PRNG type in FIPS_mode_set(): leave this to the application.
3759 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3760 the standard OpenSSL PRNG: set additional data to a date time vector.
3764 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3765 This shouldn't present any incompatibility problems because applications
3766 shouldn't be using these directly and any that are will need to rethink
3767 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3771 * Extensive self tests and health checking required by SP800-90 DRBG.
3772 Remove strength parameter from FIPS_drbg_instantiate and always
3773 instantiate at maximum supported strength.
3777 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3781 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3785 * New function DH_compute_key_padded() to compute a DH key and pad with
3786 leading zeroes if needed: this complies with SP800-56A et al.
3790 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3791 anything, incomplete, subject to change and largely untested at present.
3795 * Modify fipscanisteronly build option to only build the necessary object
3796 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3800 * Add experimental option FIPSSYMS to give all symbols in
3801 fipscanister.o and FIPS or fips prefix. This will avoid
3802 conflicts with future versions of OpenSSL. Add perl script
3803 util/fipsas.pl to preprocess assembly language source files
3804 and rename any affected symbols.
3808 * Add selftest checks and algorithm block of non-fips algorithms in
3809 FIPS mode. Remove DES2 from selftests.
3813 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3814 return internal method without any ENGINE dependencies. Add new
3815 tiny fips sign and verify functions.
3819 * New build option no-ec2m to disable characteristic 2 code.
3823 * New build option "fipscanisteronly". This only builds fipscanister.o
3824 and (currently) associated fips utilities. Uses the file Makefile.fips
3825 instead of Makefile.org as the prototype.
3829 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3830 Update fips_gcmtest to use IV generator.
3834 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3835 setting output buffer to NULL. The `*Final` function must be
3836 called although it will not retrieve any additional data. The tag
3837 can be set or retrieved with a ctrl. The IV length is by default 12
3838 bytes (96 bits) but can be set to an alternative value. If the IV
3839 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3844 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3845 underlying do_cipher function handles all cipher semantics itself
3846 including padding and finalisation. This is useful if (for example)
3847 an ENGINE cipher handles block padding itself. The behaviour of
3848 do_cipher is subtly changed if this flag is set: the return value
3849 is the number of characters written to the output buffer (zero is
3850 no longer an error code) or a negative error code. Also if the
3851 input buffer is NULL and length 0 finalisation should be performed.
3855 * If a candidate issuer certificate is already part of the constructed
3856 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3860 * Improve forward-security support: add functions
3862 void SSL_CTX_set_not_resumable_session_callback(
3863 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3864 void SSL_set_not_resumable_session_callback(
3865 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3867 for use by SSL/TLS servers; the callback function will be called whenever a
3868 new session is created, and gets to decide whether the session may be
3869 cached to make it resumable (return 0) or not (return 1). (As by the
3870 SSL/TLS protocol specifications, the session_id sent by the server will be
3871 empty to indicate that the session is not resumable; also, the server will
3872 not generate RFC 4507 (RFC 5077) session tickets.)
3874 A simple reasonable callback implementation is to return is_forward_secure.
3875 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3876 by the SSL/TLS server library, indicating whether it can provide forward
3879 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3881 * New -verify_name option in command line utilities to set verification
3886 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
3887 Add CMAC pkey methods.
3891 * Experimental renegotiation in s_server -www mode. If the client
3892 browses /reneg connection is renegotiated. If /renegcert it is
3893 renegotiated requesting a certificate.
3897 * Add an "external" session cache for debugging purposes to s_server. This
3898 should help trace issues which normally are only apparent in deployed
3899 multi-process servers.
3903 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
3904 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
3905 BIO_set_cipher() and some obscure PEM functions were changed so they
3906 can now return an error. The RAND changes required a change to the
3907 RAND_METHOD structure.
3911 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
3912 a gcc attribute to warn if the result of a function is ignored. This
3913 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
3914 whose return value is often ignored.
3918 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
3919 These allow SCTs (signed certificate timestamps) to be requested and
3920 validated when establishing a connection.
3922 *Rob Percival <robpercival@google.com>*
3927 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
3929 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
3930 used even when parsing explicit parameters, when loading a encoded key
3931 or calling `EC_GROUP_new_from_ecpkparameters()`/
3932 `EC_GROUP_new_from_ecparameters()`.
3933 This prevents bypass of security hardening and performance gains,
3934 especially for curves with specialized EC_METHODs.
3935 By default, if a key encoded with explicit parameters is loaded and later
3936 encoded, the output is still encoded with explicit parameters, even if
3937 internally a "named" EC_GROUP is used for computation.
3941 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
3942 this change, EC_GROUP_set_generator would accept order and/or cofactor as
3943 NULL. After this change, only the cofactor parameter can be NULL. It also
3944 does some minimal sanity checks on the passed order.
3949 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
3950 An attack is simple, if the first CMS_recipientInfo is valid but the
3951 second CMS_recipientInfo is chosen ciphertext. If the second
3952 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
3953 encryption key will be replaced by garbage, and the message cannot be
3954 decoded, but if the RSA decryption fails, the correct encryption key is
3955 used and the recipient will not notice the attack.
3956 As a work around for this potential attack the length of the decrypted
3957 key must be equal to the cipher default key length, in case the
3958 certifiate is not given and all recipientInfo are tried out.
3959 The old behaviour can be re-enabled in the CMS code by setting the
3960 CMS_DEBUG_DECRYPT flag.
3965 * Document issue with installation paths in diverse Windows builds
3967 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
3968 binaries and run-time config file.
3973 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
3975 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
3976 This changes the size when using the genpkey app when no size is given. It
3977 fixes an omission in earlier changes that changed all RSA, DSA and DH
3978 generation apps to use 2048 bits by default.
3982 * Add FIPS support for Android Arm 64-bit
3984 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
3985 Module in Version 2.0.10. For some reason, the corresponding target
3986 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
3987 built with FIPS support on Android Arm 64-bit. This omission has been
3990 *Matthias St. Pierre*
3992 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
3994 * 0-byte record padding oracle
3996 If an application encounters a fatal protocol error and then calls
3997 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
3998 then OpenSSL can respond differently to the calling application if a 0 byte
3999 record is received with invalid padding compared to if a 0 byte record is
4000 received with an invalid MAC. If the application then behaves differently
4001 based on that in a way that is detectable to the remote peer, then this
4002 amounts to a padding oracle that could be used to decrypt data.
4004 In order for this to be exploitable "non-stitched" ciphersuites must be in
4005 use. Stitched ciphersuites are optimised implementations of certain
4006 commonly used ciphersuites. Also the application must call SSL_shutdown()
4007 twice even if a protocol error has occurred (applications should not do
4008 this but some do anyway).
4010 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4011 Aviram, with additional investigation by Steven Collison and Andrew
4012 Hourselt. It was reported to OpenSSL on 10th December 2018.
4017 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4021 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4023 * Microarchitecture timing vulnerability in ECC scalar multiplication
4025 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4026 shown to be vulnerable to a microarchitecture timing side channel attack.
4027 An attacker with sufficient access to mount local timing attacks during
4028 ECDSA signature generation could recover the private key.
4030 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4031 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4037 * Timing vulnerability in DSA signature generation
4039 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4040 timing side channel attack. An attacker could use variations in the signing
4041 algorithm to recover the private key.
4043 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4048 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4049 Module, accidentally introduced while backporting security fixes from the
4050 development branch and hindering the use of ECC in FIPS mode.
4054 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4056 * Client DoS due to large DH parameter
4058 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4059 malicious server can send a very large prime value to the client. This will
4060 cause the client to spend an unreasonably long period of time generating a
4061 key for this prime resulting in a hang until the client has finished. This
4062 could be exploited in a Denial Of Service attack.
4064 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4069 * Cache timing vulnerability in RSA Key Generation
4071 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4072 a cache timing side channel attack. An attacker with sufficient access to
4073 mount cache timing attacks during the RSA key generation process could
4074 recover the private key.
4076 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4077 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4082 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4083 parameter is no longer accepted, as it leads to a corrupt table. NULL
4084 pem_str is reserved for alias entries only.
4088 * Revert blinding in ECDSA sign and instead make problematic addition
4089 length-invariant. Switch even to fixed-length Montgomery multiplication.
4093 * Change generating and checking of primes so that the error rate of not
4094 being prime depends on the intended use based on the size of the input.
4095 For larger primes this will result in more rounds of Miller-Rabin.
4096 The maximal error rate for primes with more than 1080 bits is lowered
4099 *Kurt Roeckx, Annie Yousar*
4101 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4105 * Add blinding to ECDSA and DSA signatures to protect against side channel
4106 attacks discovered by Keegan Ryan (NCC Group).
4110 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4111 now allow empty (zero character) pass phrases.
4115 * Certificate time validation (X509_cmp_time) enforces stricter
4116 compliance with RFC 5280. Fractional seconds and timezone offsets
4117 are no longer allowed.
4121 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4123 * Constructed ASN.1 types with a recursive definition could exceed the stack
4125 Constructed ASN.1 types with a recursive definition (such as can be found
4126 in PKCS7) could eventually exceed the stack given malicious input with
4127 excessive recursion. This could result in a Denial Of Service attack. There
4128 are no such structures used within SSL/TLS that come from untrusted sources
4129 so this is considered safe.
4131 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4137 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4139 * Read/write after SSL object in error state
4141 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4142 mechanism. The intent was that if a fatal error occurred during a handshake
4143 then OpenSSL would move into the error state and would immediately fail if
4144 you attempted to continue the handshake. This works as designed for the
4145 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4146 SSL_connect()), however due to a bug it does not work correctly if
4147 SSL_read() or SSL_write() is called directly. In that scenario, if the
4148 handshake fails then a fatal error will be returned in the initial function
4149 call. If SSL_read()/SSL_write() is subsequently called by the application
4150 for the same SSL object then it will succeed and the data is passed without
4151 being decrypted/encrypted directly from the SSL/TLS record layer.
4153 In order to exploit this issue an application bug would have to be present
4154 that resulted in a call to SSL_read()/SSL_write() being issued after having
4155 already received a fatal error.
4157 This issue was reported to OpenSSL by David Benjamin (Google).
4162 * rsaz_1024_mul_avx2 overflow bug on x86_64
4164 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4165 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4166 Analysis suggests that attacks against RSA and DSA as a result of this
4167 defect would be very difficult to perform and are not believed likely.
4168 Attacks against DH1024 are considered just feasible, because most of the
4169 work necessary to deduce information about a private key may be performed
4170 offline. The amount of resources required for such an attack would be
4171 significant. However, for an attack on TLS to be meaningful, the server
4172 would have to share the DH1024 private key among multiple clients, which is
4173 no longer an option since CVE-2016-0701.
4175 This only affects processors that support the AVX2 but not ADX extensions
4176 like Intel Haswell (4th generation).
4178 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4179 was originally found via the OSS-Fuzz project.
4184 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4186 * bn_sqrx8x_internal carry bug on x86_64
4188 There is a carry propagating bug in the x86_64 Montgomery squaring
4189 procedure. No EC algorithms are affected. Analysis suggests that attacks
4190 against RSA and DSA as a result of this defect would be very difficult to
4191 perform and are not believed likely. Attacks against DH are considered just
4192 feasible (although very difficult) because most of the work necessary to
4193 deduce information about a private key may be performed offline. The amount
4194 of resources required for such an attack would be very significant and
4195 likely only accessible to a limited number of attackers. An attacker would
4196 additionally need online access to an unpatched system using the target
4197 private key in a scenario with persistent DH parameters and a private
4198 key that is shared between multiple clients.
4200 This only affects processors that support the BMI1, BMI2 and ADX extensions
4201 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4203 This issue was reported to OpenSSL by the OSS-Fuzz project.
4208 * Malformed X.509 IPAddressFamily could cause OOB read
4210 If an X.509 certificate has a malformed IPAddressFamily extension,
4211 OpenSSL could do a one-byte buffer overread. The most likely result
4212 would be an erroneous display of the certificate in text format.
4214 This issue was reported to OpenSSL by the OSS-Fuzz project.
4219 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4221 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4222 platform rather than 'mingw'.
4226 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4228 * Truncated packet could crash via OOB read
4230 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4231 cipher is being used, then a truncated packet can cause that host to
4232 perform an out-of-bounds read, usually resulting in a crash.
4234 This issue was reported to OpenSSL by Robert Święcki of Google.
4239 * BN_mod_exp may produce incorrect results on x86_64
4241 There is a carry propagating bug in the x86_64 Montgomery squaring
4242 procedure. No EC algorithms are affected. Analysis suggests that attacks
4243 against RSA and DSA as a result of this defect would be very difficult to
4244 perform and are not believed likely. Attacks against DH are considered just
4245 feasible (although very difficult) because most of the work necessary to
4246 deduce information about a private key may be performed offline. The amount
4247 of resources required for such an attack would be very significant and
4248 likely only accessible to a limited number of attackers. An attacker would
4249 additionally need online access to an unpatched system using the target
4250 private key in a scenario with persistent DH parameters and a private
4251 key that is shared between multiple clients. For example this can occur by
4252 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4253 similar to CVE-2015-3193 but must be treated as a separate problem.
4255 This issue was reported to OpenSSL by the OSS-Fuzz project.
4260 * Montgomery multiplication may produce incorrect results
4262 There is a carry propagating bug in the Broadwell-specific Montgomery
4263 multiplication procedure that handles input lengths divisible by, but
4264 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4265 and DH private keys are impossible. This is because the subroutine in
4266 question is not used in operations with the private key itself and an input
4267 of the attacker's direct choice. Otherwise the bug can manifest itself as
4268 transient authentication and key negotiation failures or reproducible
4269 erroneous outcome of public-key operations with specially crafted input.
4270 Among EC algorithms only Brainpool P-512 curves are affected and one
4271 presumably can attack ECDH key negotiation. Impact was not analyzed in
4272 detail, because pre-requisites for attack are considered unlikely. Namely
4273 multiple clients have to choose the curve in question and the server has to
4274 share the private key among them, neither of which is default behaviour.
4275 Even then only clients that chose the curve will be affected.
4277 This issue was publicly reported as transient failures and was not
4278 initially recognized as a security issue. Thanks to Richard Morgan for
4279 providing reproducible case.
4284 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4285 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4286 prevent issues where no progress is being made and the peer continually
4287 sends unrecognised record types, using up resources processing them.
4291 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4293 * Missing CRL sanity check
4295 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4296 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4297 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4299 This issue only affects the OpenSSL 1.0.2i
4304 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4306 * OCSP Status Request extension unbounded memory growth
4308 A malicious client can send an excessively large OCSP Status Request
4309 extension. If that client continually requests renegotiation, sending a
4310 large OCSP Status Request extension each time, then there will be unbounded
4311 memory growth on the server. This will eventually lead to a Denial Of
4312 Service attack through memory exhaustion. Servers with a default
4313 configuration are vulnerable even if they do not support OCSP. Builds using
4314 the "no-ocsp" build time option are not affected.
4316 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4321 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4324 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4330 * OOB write in MDC2_Update()
4332 An overflow can occur in MDC2_Update() either if called directly or
4333 through the EVP_DigestUpdate() function using MDC2. If an attacker
4334 is able to supply very large amounts of input data after a previous
4335 call to EVP_EncryptUpdate() with a partial block then a length check
4336 can overflow resulting in a heap corruption.
4338 The amount of data needed is comparable to SIZE_MAX which is impractical
4341 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4346 * Malformed SHA512 ticket DoS
4348 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4349 DoS attack where a malformed ticket will result in an OOB read which will
4352 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4353 a custom server callback and ticket lookup mechanism.
4355 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4360 * OOB write in BN_bn2dec()
4362 The function BN_bn2dec() does not check the return value of BN_div_word().
4363 This can cause an OOB write if an application uses this function with an
4364 overly large BIGNUM. This could be a problem if an overly large certificate
4365 or CRL is printed out from an untrusted source. TLS is not affected because
4366 record limits will reject an oversized certificate before it is parsed.
4368 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4373 * OOB read in TS_OBJ_print_bio()
4375 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4376 the total length the OID text representation would use and not the amount
4377 of data written. This will result in OOB reads when large OIDs are
4380 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4385 * Pointer arithmetic undefined behaviour
4387 Avoid some undefined pointer arithmetic
4389 A common idiom in the codebase is to check limits in the following manner:
4392 Where "p" points to some malloc'd data of SIZE bytes and
4395 "len" here could be from some externally supplied data (e.g. from a TLS
4398 The rules of C pointer arithmetic are such that "p + len" is only well
4399 defined where len <= SIZE. Therefore the above idiom is actually
4400 undefined behaviour.
4402 For example this could cause problems if some malloc implementation
4403 provides an address for "p" such that "p + len" actually overflows for
4404 values of len that are too big and therefore p + len < limit.
4406 This issue was reported to OpenSSL by Guido Vranken
4411 * Constant time flag not preserved in DSA signing
4413 Operations in the DSA signing algorithm should run in constant time in
4414 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4415 implementation means that a non-constant time codepath is followed for
4416 certain operations. This has been demonstrated through a cache-timing
4417 attack to be sufficient for an attacker to recover the private DSA key.
4419 This issue was reported by César Pereida (Aalto University), Billy Brumley
4420 (Tampere University of Technology), and Yuval Yarom (The University of
4421 Adelaide and NICTA).
4426 * DTLS buffered message DoS
4428 In a DTLS connection where handshake messages are delivered out-of-order
4429 those messages that OpenSSL is not yet ready to process will be buffered
4430 for later use. Under certain circumstances, a flaw in the logic means that
4431 those messages do not get removed from the buffer even though the handshake
4432 has been completed. An attacker could force up to approx. 15 messages to
4433 remain in the buffer when they are no longer required. These messages will
4434 be cleared when the DTLS connection is closed. The default maximum size for
4435 a message is 100k. Therefore the attacker could force an additional 1500k
4436 to be consumed per connection. By opening many simulataneous connections an
4437 attacker could cause a DoS attack through memory exhaustion.
4439 This issue was reported to OpenSSL by Quan Luo.
4444 * DTLS replay protection DoS
4446 A flaw in the DTLS replay attack protection mechanism means that records
4447 that arrive for future epochs update the replay protection "window" before
4448 the MAC for the record has been validated. This could be exploited by an
4449 attacker by sending a record for the next epoch (which does not have to
4450 decrypt or have a valid MAC), with a very large sequence number. This means
4451 that all subsequent legitimate packets are dropped causing a denial of
4452 service for a specific DTLS connection.
4454 This issue was reported to OpenSSL by the OCAP audit team.
4459 * Certificate message OOB reads
4461 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4462 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4463 theoretical DoS risk but this has not been observed in practice on common
4466 The messages affected are client certificate, client certificate request
4467 and server certificate. As a result the attack can only be performed
4468 against a client or a server which enables client authentication.
4470 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4475 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4477 * Prevent padding oracle in AES-NI CBC MAC check
4479 A MITM attacker can use a padding oracle attack to decrypt traffic
4480 when the connection uses an AES CBC cipher and the server support
4483 This issue was introduced as part of the fix for Lucky 13 padding
4484 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4485 constant time by making sure that always the same bytes are read and
4486 compared against either the MAC or padding bytes. But it no longer
4487 checked that there was enough data to have both the MAC and padding
4490 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4495 * Fix EVP_EncodeUpdate overflow
4497 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4498 Base64 encoding of binary data. If an attacker is able to supply very large
4499 amounts of input data then a length check can overflow resulting in a heap
4502 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4503 the `PEM_write_bio*` family of functions. These are mainly used within the
4504 OpenSSL command line applications, so any application which processes data
4505 from an untrusted source and outputs it as a PEM file should be considered
4506 vulnerable to this issue. User applications that call these APIs directly
4507 with large amounts of untrusted data may also be vulnerable.
4509 This issue was reported by Guido Vranken.
4514 * Fix EVP_EncryptUpdate overflow
4516 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4517 is able to supply very large amounts of input data after a previous call to
4518 EVP_EncryptUpdate() with a partial block then a length check can overflow
4519 resulting in a heap corruption. Following an analysis of all OpenSSL
4520 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4521 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4522 the first called function after an EVP_EncryptInit(), and therefore that
4523 specific call must be safe. The second form is where the length passed to
4524 EVP_EncryptUpdate() can be seen from the code to be some small value and
4525 therefore there is no possibility of an overflow. Since all instances are
4526 one of these two forms, it is believed that there can be no overflows in
4527 internal code due to this problem. It should be noted that
4528 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4529 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4530 of these calls have also been analysed too and it is believed there are no
4531 instances in internal usage where an overflow could occur.
4533 This issue was reported by Guido Vranken.
4538 * Prevent ASN.1 BIO excessive memory allocation
4540 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4541 a short invalid encoding can cause allocation of large amounts of memory
4542 potentially consuming excessive resources or exhausting memory.
4544 Any application parsing untrusted data through d2i BIO functions is
4545 affected. The memory based functions such as d2i_X509() are *not* affected.
4546 Since the memory based functions are used by the TLS library, TLS
4547 applications are not affected.
4549 This issue was reported by Brian Carpenter.
4556 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4557 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4558 in arbitrary stack data being returned in the buffer.
4560 This issue was reported by Guido Vranken.
4565 * Modify behavior of ALPN to invoke callback after SNI/servername
4566 callback, such that updates to the SSL_CTX affect ALPN.
4570 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4575 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4576 methods are enabled and ssl2 is disabled the methods return NULL.
4580 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4582 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4583 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4584 provide any "EXPORT" or "LOW" strength ciphers.
4588 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4589 is by default disabled at build-time. Builds that are not configured with
4590 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4591 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4592 will need to explicitly call either of:
4594 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4596 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4598 as appropriate. Even if either of those is used, or the application
4599 explicitly uses the version-specific SSLv2_method() or its client and
4600 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4601 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4602 ciphers, and SSLv2 56-bit DES are no longer available.
4607 * Fix a double-free in DSA code
4609 A double free bug was discovered when OpenSSL parses malformed DSA private
4610 keys and could lead to a DoS attack or memory corruption for applications
4611 that receive DSA private keys from untrusted sources. This scenario is
4614 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4620 * Disable SRP fake user seed to address a server memory leak.
4622 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4624 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4625 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4626 was changed to ignore the "fake user" SRP seed, even if the seed
4629 Users should use SRP_VBASE_get1_by_user instead. Note that in
4630 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4631 also that even though configuring the SRP seed attempts to hide
4632 invalid usernames by continuing the handshake with fake
4633 credentials, this behaviour is not constant time and no strong
4634 guarantees are made that the handshake is indistinguishable from
4635 that of a valid user.
4640 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4642 In the BN_hex2bn function the number of hex digits is calculated using an
4643 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4644 large values of `i` this can result in `bn_expand` not allocating any
4645 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4646 field as NULL leading to a subsequent NULL ptr deref. For very large values
4647 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4648 In this case memory is allocated to the internal BIGNUM data field, but it
4649 is insufficiently sized leading to heap corruption. A similar issue exists
4650 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4651 is ever called by user applications with very large untrusted hex/dec data.
4652 This is anticipated to be a rare occurrence.
4654 All OpenSSL internal usage of these functions use data that is not expected
4655 to be untrusted, e.g. config file data or application command line
4656 arguments. If user developed applications generate config file data based
4657 on untrusted data then it is possible that this could also lead to security
4658 consequences. This is also anticipated to be rare.
4660 This issue was reported to OpenSSL by Guido Vranken.
4665 * Fix memory issues in `BIO_*printf` functions
4667 The internal `fmtstr` function used in processing a "%s" format string in
4668 the `BIO_*printf` functions could overflow while calculating the length of a
4669 string and cause an OOB read when printing very long strings.
4671 Additionally the internal `doapr_outch` function can attempt to write to an
4672 OOB memory location (at an offset from the NULL pointer) in the event of a
4673 memory allocation failure. In 1.0.2 and below this could be caused where
4674 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4675 could be in processing a very long "%s" format string. Memory leaks can
4678 The first issue may mask the second issue dependent on compiler behaviour.
4679 These problems could enable attacks where large amounts of untrusted data
4680 is passed to the `BIO_*printf` functions. If applications use these functions
4681 in this way then they could be vulnerable. OpenSSL itself uses these
4682 functions when printing out human-readable dumps of ASN.1 data. Therefore
4683 applications that print this data could be vulnerable if the data is from
4684 untrusted sources. OpenSSL command line applications could also be
4685 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4686 as command line arguments.
4688 Libssl is not considered directly vulnerable. Additionally certificates etc
4689 received via remote connections via libssl are also unlikely to be able to
4690 trigger these issues because of message size limits enforced within libssl.
4692 This issue was reported to OpenSSL Guido Vranken.
4697 * Side channel attack on modular exponentiation
4699 A side-channel attack was found which makes use of cache-bank conflicts on
4700 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4701 of RSA keys. The ability to exploit this issue is limited as it relies on
4702 an attacker who has control of code in a thread running on the same
4703 hyper-threaded core as the victim thread which is performing decryptions.
4705 This issue was reported to OpenSSL by Yuval Yarom, The University of
4706 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4707 Nadia Heninger, University of Pennsylvania with more information at
4708 <http://cachebleed.info>.
4713 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4714 if no keysize is specified with default_bits. This fixes an
4715 omission in an earlier change that changed all RSA/DSA key generation
4716 apps to use 2048 bits by default.
4720 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4722 * DH small subgroups
4724 Historically OpenSSL only ever generated DH parameters based on "safe"
4725 primes. More recently (in version 1.0.2) support was provided for
4726 generating X9.42 style parameter files such as those required for RFC 5114
4727 support. The primes used in such files may not be "safe". Where an
4728 application is using DH configured with parameters based on primes that are
4729 not "safe" then an attacker could use this fact to find a peer's private
4730 DH exponent. This attack requires that the attacker complete multiple
4731 handshakes in which the peer uses the same private DH exponent. For example
4732 this could be used to discover a TLS server's private DH exponent if it's
4733 reusing the private DH exponent or it's using a static DH ciphersuite.
4735 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4736 TLS. It is not on by default. If the option is not set then the server
4737 reuses the same private DH exponent for the life of the server process and
4738 would be vulnerable to this attack. It is believed that many popular
4739 applications do set this option and would therefore not be at risk.
4741 The fix for this issue adds an additional check where a "q" parameter is
4742 available (as is the case in X9.42 based parameters). This detects the
4743 only known attack, and is the only possible defense for static DH
4744 ciphersuites. This could have some performance impact.
4746 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4747 default and cannot be disabled. This could have some performance impact.
4749 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4754 * SSLv2 doesn't block disabled ciphers
4756 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4757 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4758 been disabled, provided that the SSLv2 protocol was not also disabled via
4761 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4762 and Sebastian Schinzel.
4767 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4769 * BN_mod_exp may produce incorrect results on x86_64
4771 There is a carry propagating bug in the x86_64 Montgomery squaring
4772 procedure. No EC algorithms are affected. Analysis suggests that attacks
4773 against RSA and DSA as a result of this defect would be very difficult to
4774 perform and are not believed likely. Attacks against DH are considered just
4775 feasible (although very difficult) because most of the work necessary to
4776 deduce information about a private key may be performed offline. The amount
4777 of resources required for such an attack would be very significant and
4778 likely only accessible to a limited number of attackers. An attacker would
4779 additionally need online access to an unpatched system using the target
4780 private key in a scenario with persistent DH parameters and a private
4781 key that is shared between multiple clients. For example this can occur by
4782 default in OpenSSL DHE based SSL/TLS ciphersuites.
4784 This issue was reported to OpenSSL by Hanno Böck.
4789 * Certificate verify crash with missing PSS parameter
4791 The signature verification routines will crash with a NULL pointer
4792 dereference if presented with an ASN.1 signature using the RSA PSS
4793 algorithm and absent mask generation function parameter. Since these
4794 routines are used to verify certificate signature algorithms this can be
4795 used to crash any certificate verification operation and exploited in a
4796 DoS attack. Any application which performs certificate verification is
4797 vulnerable including OpenSSL clients and servers which enable client
4800 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4805 * X509_ATTRIBUTE memory leak
4807 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4808 memory. This structure is used by the PKCS#7 and CMS routines so any
4809 application which reads PKCS#7 or CMS data from untrusted sources is
4810 affected. SSL/TLS is not affected.
4812 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4818 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4819 This changes the decoding behaviour for some invalid messages,
4820 though the change is mostly in the more lenient direction, and
4821 legacy behaviour is preserved as much as possible.
4825 * In DSA_generate_parameters_ex, if the provided seed is too short,
4828 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4830 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4832 * Alternate chains certificate forgery
4834 During certificate verification, OpenSSL will attempt to find an
4835 alternative certificate chain if the first attempt to build such a chain
4836 fails. An error in the implementation of this logic can mean that an
4837 attacker could cause certain checks on untrusted certificates to be
4838 bypassed, such as the CA flag, enabling them to use a valid leaf
4839 certificate to act as a CA and "issue" an invalid certificate.
4841 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4846 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4848 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4849 incompatibility in the handling of HMAC. The previous ABI has now been
4854 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4856 * Malformed ECParameters causes infinite loop
4858 When processing an ECParameters structure OpenSSL enters an infinite loop
4859 if the curve specified is over a specially malformed binary polynomial
4862 This can be used to perform denial of service against any
4863 system which processes public keys, certificate requests or
4864 certificates. This includes TLS clients and TLS servers with
4865 client authentication enabled.
4867 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4872 * Exploitable out-of-bounds read in X509_cmp_time
4874 X509_cmp_time does not properly check the length of the ASN1_TIME
4875 string and can read a few bytes out of bounds. In addition,
4876 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4879 An attacker can use this to craft malformed certificates and CRLs of
4880 various sizes and potentially cause a segmentation fault, resulting in
4881 a DoS on applications that verify certificates or CRLs. TLS clients
4882 that verify CRLs are affected. TLS clients and servers with client
4883 authentication enabled may be affected if they use custom verification
4886 This issue was reported to OpenSSL by Robert Swiecki (Google), and
4887 independently by Hanno Böck.
4892 * PKCS7 crash with missing EnvelopedContent
4894 The PKCS#7 parsing code does not handle missing inner EncryptedContent
4895 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
4896 with missing content and trigger a NULL pointer dereference on parsing.
4898 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
4899 structures from untrusted sources are affected. OpenSSL clients and
4900 servers are not affected.
4902 This issue was reported to OpenSSL by Michal Zalewski (Google).
4907 * CMS verify infinite loop with unknown hash function
4909 When verifying a signedData message the CMS code can enter an infinite loop
4910 if presented with an unknown hash function OID. This can be used to perform
4911 denial of service against any system which verifies signedData messages using
4913 This issue was reported to OpenSSL by Johannes Bauer.
4918 * Race condition handling NewSessionTicket
4920 If a NewSessionTicket is received by a multi-threaded client when attempting to
4921 reuse a previous ticket then a race condition can occur potentially leading to
4922 a double free of the ticket data.
4927 * Only support 256-bit or stronger elliptic curves with the
4928 'ecdh_auto' setting (server) or by default (client). Of supported
4929 curves, prefer P-256 (both).
4933 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
4935 * ClientHello sigalgs DoS fix
4937 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
4938 invalid signature algorithms extension a NULL pointer dereference will
4939 occur. This can be exploited in a DoS attack against the server.
4941 This issue was was reported to OpenSSL by David Ramos of Stanford
4945 *Stephen Henson and Matt Caswell*
4947 * Multiblock corrupted pointer fix
4949 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
4950 feature only applies on 64 bit x86 architecture platforms that support AES
4951 NI instructions. A defect in the implementation of "multiblock" can cause
4952 OpenSSL's internal write buffer to become incorrectly set to NULL when
4953 using non-blocking IO. Typically, when the user application is using a
4954 socket BIO for writing, this will only result in a failed connection.
4955 However if some other BIO is used then it is likely that a segmentation
4956 fault will be triggered, thus enabling a potential DoS attack.
4958 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
4963 * Segmentation fault in DTLSv1_listen fix
4965 The DTLSv1_listen function is intended to be stateless and processes the
4966 initial ClientHello from many peers. It is common for user code to loop
4967 over the call to DTLSv1_listen until a valid ClientHello is received with
4968 an associated cookie. A defect in the implementation of DTLSv1_listen means
4969 that state is preserved in the SSL object from one invocation to the next
4970 that can lead to a segmentation fault. Errors processing the initial
4971 ClientHello can trigger this scenario. An example of such an error could be
4972 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
4975 This issue was reported to OpenSSL by Per Allansson.
4980 * Segmentation fault in ASN1_TYPE_cmp fix
4982 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
4983 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
4984 certificate signature algorithm consistency this can be used to crash any
4985 certificate verification operation and exploited in a DoS attack. Any
4986 application which performs certificate verification is vulnerable including
4987 OpenSSL clients and servers which enable client authentication.
4992 * Segmentation fault for invalid PSS parameters fix
4994 The signature verification routines will crash with a NULL pointer
4995 dereference if presented with an ASN.1 signature using the RSA PSS
4996 algorithm and invalid parameters. Since these routines are used to verify
4997 certificate signature algorithms this can be used to crash any
4998 certificate verification operation and exploited in a DoS attack. Any
4999 application which performs certificate verification is vulnerable including
5000 OpenSSL clients and servers which enable client authentication.
5002 This issue was was reported to OpenSSL by Brian Carpenter.
5007 * ASN.1 structure reuse memory corruption fix
5009 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5010 memory corruption via an invalid write. Such reuse is and has been
5011 strongly discouraged and is believed to be rare.
5013 Applications that parse structures containing CHOICE or ANY DEFINED BY
5014 components may be affected. Certificate parsing (d2i_X509 and related
5015 functions) are however not affected. OpenSSL clients and servers are
5021 * PKCS7 NULL pointer dereferences fix
5023 The PKCS#7 parsing code does not handle missing outer ContentInfo
5024 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5025 missing content and trigger a NULL pointer dereference on parsing.
5027 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5028 otherwise parse PKCS#7 structures from untrusted sources are
5029 affected. OpenSSL clients and servers are not affected.
5031 This issue was reported to OpenSSL by Michal Zalewski (Google).
5036 * DoS via reachable assert in SSLv2 servers fix
5038 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5039 servers that both support SSLv2 and enable export cipher suites by sending
5040 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5042 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5043 (OpenSSL development team).
5048 * Empty CKE with client auth and DHE fix
5050 If client auth is used then a server can seg fault in the event of a DHE
5051 ciphersuite being selected and a zero length ClientKeyExchange message
5052 being sent by the client. This could be exploited in a DoS attack.
5057 * Handshake with unseeded PRNG fix
5059 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5060 with an unseeded PRNG. The conditions are:
5061 - The client is on a platform where the PRNG has not been seeded
5062 automatically, and the user has not seeded manually
5063 - A protocol specific client method version has been used (i.e. not
5064 SSL_client_methodv23)
5065 - A ciphersuite is used that does not require additional random data from
5066 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5068 If the handshake succeeds then the client random that has been used will
5069 have been generated from a PRNG with insufficient entropy and therefore the
5070 output may be predictable.
5072 For example using the following command with an unseeded openssl will
5073 succeed on an unpatched platform:
5075 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5080 * Use After Free following d2i_ECPrivatekey error fix
5082 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5083 could cause a use after free condition. This, in turn, could cause a double
5084 free in several private key parsing functions (such as d2i_PrivateKey
5085 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5086 for applications that receive EC private keys from untrusted
5087 sources. This scenario is considered rare.
5089 This issue was discovered by the BoringSSL project and fixed in their
5095 * X509_to_X509_REQ NULL pointer deref fix
5097 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5098 the certificate key is invalid. This function is rarely used in practice.
5100 This issue was discovered by Brian Carpenter.
5105 * Removed the export ciphers from the DEFAULT ciphers
5109 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5111 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5112 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5113 So far those who have to target multiple platforms would compromise
5114 and argue that binary targeting say ARMv5 would still execute on
5115 ARMv8. "Universal" build resolves this compromise by providing
5116 near-optimal performance even on newer platforms.
5120 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5121 (other platforms pending).
5123 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5125 * Add support for the SignedCertificateTimestampList certificate and
5126 OCSP response extensions from RFC6962.
5130 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5131 for corner cases. (Certain input points at infinity could lead to
5132 bogus results, with non-infinity inputs mapped to infinity too.)
5136 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5137 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5138 common cases are optimized and there still is room for further
5139 improvements. Vector Permutation AES for Altivec is also added.
5143 * Add support for little-endian ppc64 Linux target.
5145 *Marcelo Cerri (IBM)*
5147 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5148 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5149 are optimized and there still is room for further improvements.
5150 Both 32- and 64-bit modes are supported.
5152 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5154 * Improved ARMv7 NEON support.
5158 * Support for SPARC Architecture 2011 crypto extensions, first
5159 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5160 SHA256/512, MD5, GHASH and modular exponentiation.
5162 *Andy Polyakov, David Miller*
5164 * Accelerated modular exponentiation for Intel processors, a.k.a.
5167 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5169 * Support for new and upcoming Intel processors, including AVX2,
5170 BMI and SHA ISA extensions. This includes additional "stitched"
5171 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5174 This work was sponsored by Intel Corp.
5178 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5179 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5180 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5184 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5185 this fixes a limitation in previous versions of OpenSSL.
5189 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5190 MGF1 digest and OAEP label.
5194 * Add EVP support for key wrapping algorithms, to avoid problems with
5195 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5196 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5197 algorithms and include tests cases.
5201 * Add functions to allocate and set the fields of an ECDSA_METHOD
5204 *Douglas E. Engert, Steve Henson*
5206 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5207 difference in days and seconds between two tm or ASN1_TIME structures.
5211 * Add -rev test option to s_server to just reverse order of characters
5212 received by client and send back to server. Also prints an abbreviated
5213 summary of the connection parameters.
5217 * New option -brief for s_client and s_server to print out a brief summary
5218 of connection parameters.
5222 * Add callbacks for arbitrary TLS extensions.
5224 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5226 * New option -crl_download in several openssl utilities to download CRLs
5227 from CRLDP extension in certificates.
5231 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5235 * New function X509_CRL_diff to generate a delta CRL from the difference
5236 of two full CRLs. Add support to "crl" utility.
5240 * New functions to set lookup_crls function and to retrieve
5241 X509_STORE from X509_STORE_CTX.
5245 * Print out deprecated issuer and subject unique ID fields in
5250 * Extend OCSP I/O functions so they can be used for simple general purpose
5251 HTTP as well as OCSP. New wrapper function which can be used to download
5252 CRLs using the OCSP API.
5256 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5260 * `SSL_CONF*` functions. These provide a common framework for application
5261 configuration using configuration files or command lines.
5265 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5266 message callback and prints the results. Needs compile time option
5267 "enable-ssl-trace". New options to s_client and s_server to enable
5272 * New ctrl and macro to retrieve supported points extensions.
5273 Print out extension in s_server and s_client.
5277 * New functions to retrieve certificate signature and signature
5282 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5287 * New Suite B modes for TLS code. These use and enforce the requirements
5288 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5289 only use Suite B curves. The Suite B modes can be set by using the
5290 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5294 * New chain verification flags for Suite B levels of security. Check
5295 algorithms are acceptable when flags are set in X509_verify_cert.
5299 * Make tls1_check_chain return a set of flags indicating checks passed
5300 by a certificate chain. Add additional tests to handle client
5301 certificates: checks for matching certificate type and issuer name
5306 * If an attempt is made to use a signature algorithm not in the peer
5307 preference list abort the handshake. If client has no suitable
5308 signature algorithms in response to a certificate request do not
5309 use the certificate.
5313 * If server EC tmp key is not in client preference list abort handshake.
5317 * Add support for certificate stores in CERT structure. This makes it
5318 possible to have different stores per SSL structure or one store in
5319 the parent SSL_CTX. Include distinct stores for certificate chain
5320 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5321 to build and store a certificate chain in CERT structure: returning
5322 an error if the chain cannot be built: this will allow applications
5323 to test if a chain is correctly configured.
5325 Note: if the CERT based stores are not set then the parent SSL_CTX
5326 store is used to retain compatibility with existing behaviour.
5330 * New function ssl_set_client_disabled to set a ciphersuite disabled
5331 mask based on the current session, check mask when sending client
5332 hello and checking the requested ciphersuite.
5336 * New ctrls to retrieve and set certificate types in a certificate
5337 request message. Print out received values in s_client. If certificate
5338 types is not set with custom values set sensible values based on
5339 supported signature algorithms.
5343 * Support for distinct client and server supported signature algorithms.
5347 * Add certificate callback. If set this is called whenever a certificate
5348 is required by client or server. An application can decide which
5349 certificate chain to present based on arbitrary criteria: for example
5350 supported signature algorithms. Add very simple example to s_server.
5351 This fixes many of the problems and restrictions of the existing client
5352 certificate callback: for example you can now clear an existing
5353 certificate and specify the whole chain.
5357 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5358 the certificate can be used for (if anything). Set valid_flags field
5359 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5360 to have similar checks in it.
5362 Add new "cert_flags" field to CERT structure and include a "strict mode".
5363 This enforces some TLS certificate requirements (such as only permitting
5364 certificate signature algorithms contained in the supported algorithms
5365 extension) which some implementations ignore: this option should be used
5366 with caution as it could cause interoperability issues.
5370 * Update and tidy signature algorithm extension processing. Work out
5371 shared signature algorithms based on preferences and peer algorithms
5372 and print them out in s_client and s_server. Abort handshake if no
5373 shared signature algorithms.
5377 * Add new functions to allow customised supported signature algorithms
5378 for SSL and SSL_CTX structures. Add options to s_client and s_server
5383 * New function SSL_certs_clear() to delete all references to certificates
5384 from an SSL structure. Before this once a certificate had been added
5385 it couldn't be removed.
5389 * Integrate hostname, email address and IP address checking with certificate
5390 verification. New verify options supporting checking in openssl utility.
5394 * Fixes and wildcard matching support to hostname and email checking
5395 functions. Add manual page.
5397 *Florian Weimer (Red Hat Product Security Team)*
5399 * New functions to check a hostname email or IP address against a
5400 certificate. Add options x509 utility to print results of checks against
5405 * Fix OCSP checking.
5407 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5409 * Initial experimental support for explicitly trusted non-root CAs.
5410 OpenSSL still tries to build a complete chain to a root but if an
5411 intermediate CA has a trust setting included that is used. The first
5412 setting is used: whether to trust (e.g., -addtrust option to the x509
5417 * Add -trusted_first option which attempts to find certificates in the
5418 trusted store even if an untrusted chain is also supplied.
5422 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5423 platform support for Linux and Android.
5427 * Support for linux-x32, ILP32 environment in x86_64 framework.
5431 * Experimental multi-implementation support for FIPS capable OpenSSL.
5432 When in FIPS mode the approved implementations are used as normal,
5433 when not in FIPS mode the internal unapproved versions are used instead.
5434 This means that the FIPS capable OpenSSL isn't forced to use the
5435 (often lower performance) FIPS implementations outside FIPS mode.
5439 * Transparently support X9.42 DH parameters when calling
5440 PEM_read_bio_DHparameters. This means existing applications can handle
5441 the new parameter format automatically.
5445 * Initial experimental support for X9.42 DH parameter format: mainly
5446 to support use of 'q' parameter for RFC5114 parameters.
5450 * Add DH parameters from RFC5114 including test data to dhtest.
5454 * Support for automatic EC temporary key parameter selection. If enabled
5455 the most preferred EC parameters are automatically used instead of
5456 hardcoded fixed parameters. Now a server just has to call:
5457 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5458 support ECDH and use the most appropriate parameters.
5462 * Enhance and tidy EC curve and point format TLS extension code. Use
5463 static structures instead of allocation if default values are used.
5464 New ctrls to set curves we wish to support and to retrieve shared curves.
5465 Print out shared curves in s_server. New options to s_server and s_client
5466 to set list of supported curves.
5470 * New ctrls to retrieve supported signature algorithms and
5471 supported curve values as an array of NIDs. Extend openssl utility
5472 to print out received values.
5476 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5477 between NIDs and the more common NIST names such as "P-256". Enhance
5478 ecparam utility and ECC method to recognise the NIST names for curves.
5482 * Enhance SSL/TLS certificate chain handling to support different
5483 chains for each certificate instead of one chain in the parent SSL_CTX.
5487 * Support for fixed DH ciphersuite client authentication: where both
5488 server and client use DH certificates with common parameters.
5492 * Support for fixed DH ciphersuites: those requiring DH server
5497 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5499 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5500 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5501 X509_CINF_get_signature were reverted post internal team review.
5506 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5508 * OCSP Status Request extension unbounded memory growth
5510 A malicious client can send an excessively large OCSP Status Request
5511 extension. If that client continually requests renegotiation, sending a
5512 large OCSP Status Request extension each time, then there will be unbounded
5513 memory growth on the server. This will eventually lead to a Denial Of
5514 Service attack through memory exhaustion. Servers with a default
5515 configuration are vulnerable even if they do not support OCSP. Builds using
5516 the "no-ocsp" build time option are not affected.
5518 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5523 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5526 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5532 * OOB write in MDC2_Update()
5534 An overflow can occur in MDC2_Update() either if called directly or
5535 through the EVP_DigestUpdate() function using MDC2. If an attacker
5536 is able to supply very large amounts of input data after a previous
5537 call to EVP_EncryptUpdate() with a partial block then a length check
5538 can overflow resulting in a heap corruption.
5540 The amount of data needed is comparable to SIZE_MAX which is impractical
5543 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5548 * Malformed SHA512 ticket DoS
5550 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5551 DoS attack where a malformed ticket will result in an OOB read which will
5554 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5555 a custom server callback and ticket lookup mechanism.
5557 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5562 * OOB write in BN_bn2dec()
5564 The function BN_bn2dec() does not check the return value of BN_div_word().
5565 This can cause an OOB write if an application uses this function with an
5566 overly large BIGNUM. This could be a problem if an overly large certificate
5567 or CRL is printed out from an untrusted source. TLS is not affected because
5568 record limits will reject an oversized certificate before it is parsed.
5570 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5575 * OOB read in TS_OBJ_print_bio()
5577 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5578 the total length the OID text representation would use and not the amount
5579 of data written. This will result in OOB reads when large OIDs are
5582 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5587 * Pointer arithmetic undefined behaviour
5589 Avoid some undefined pointer arithmetic
5591 A common idiom in the codebase is to check limits in the following manner:
5594 Where "p" points to some malloc'd data of SIZE bytes and
5597 "len" here could be from some externally supplied data (e.g. from a TLS
5600 The rules of C pointer arithmetic are such that "p + len" is only well
5601 defined where len <= SIZE. Therefore the above idiom is actually
5602 undefined behaviour.
5604 For example this could cause problems if some malloc implementation
5605 provides an address for "p" such that "p + len" actually overflows for
5606 values of len that are too big and therefore p + len < limit.
5608 This issue was reported to OpenSSL by Guido Vranken
5613 * Constant time flag not preserved in DSA signing
5615 Operations in the DSA signing algorithm should run in constant time in
5616 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5617 implementation means that a non-constant time codepath is followed for
5618 certain operations. This has been demonstrated through a cache-timing
5619 attack to be sufficient for an attacker to recover the private DSA key.
5621 This issue was reported by César Pereida (Aalto University), Billy Brumley
5622 (Tampere University of Technology), and Yuval Yarom (The University of
5623 Adelaide and NICTA).
5628 * DTLS buffered message DoS
5630 In a DTLS connection where handshake messages are delivered out-of-order
5631 those messages that OpenSSL is not yet ready to process will be buffered
5632 for later use. Under certain circumstances, a flaw in the logic means that
5633 those messages do not get removed from the buffer even though the handshake
5634 has been completed. An attacker could force up to approx. 15 messages to
5635 remain in the buffer when they are no longer required. These messages will
5636 be cleared when the DTLS connection is closed. The default maximum size for
5637 a message is 100k. Therefore the attacker could force an additional 1500k
5638 to be consumed per connection. By opening many simulataneous connections an
5639 attacker could cause a DoS attack through memory exhaustion.
5641 This issue was reported to OpenSSL by Quan Luo.
5646 * DTLS replay protection DoS
5648 A flaw in the DTLS replay attack protection mechanism means that records
5649 that arrive for future epochs update the replay protection "window" before
5650 the MAC for the record has been validated. This could be exploited by an
5651 attacker by sending a record for the next epoch (which does not have to
5652 decrypt or have a valid MAC), with a very large sequence number. This means
5653 that all subsequent legitimate packets are dropped causing a denial of
5654 service for a specific DTLS connection.
5656 This issue was reported to OpenSSL by the OCAP audit team.
5661 * Certificate message OOB reads
5663 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5664 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5665 theoretical DoS risk but this has not been observed in practice on common
5668 The messages affected are client certificate, client certificate request
5669 and server certificate. As a result the attack can only be performed
5670 against a client or a server which enables client authentication.
5672 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5677 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5679 * Prevent padding oracle in AES-NI CBC MAC check
5681 A MITM attacker can use a padding oracle attack to decrypt traffic
5682 when the connection uses an AES CBC cipher and the server support
5685 This issue was introduced as part of the fix for Lucky 13 padding
5686 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5687 constant time by making sure that always the same bytes are read and
5688 compared against either the MAC or padding bytes. But it no longer
5689 checked that there was enough data to have both the MAC and padding
5692 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5697 * Fix EVP_EncodeUpdate overflow
5699 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5700 Base64 encoding of binary data. If an attacker is able to supply very large
5701 amounts of input data then a length check can overflow resulting in a heap
5704 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5705 the `PEM_write_bio*` family of functions. These are mainly used within the
5706 OpenSSL command line applications, so any application which processes data
5707 from an untrusted source and outputs it as a PEM file should be considered
5708 vulnerable to this issue. User applications that call these APIs directly
5709 with large amounts of untrusted data may also be vulnerable.
5711 This issue was reported by Guido Vranken.
5716 * Fix EVP_EncryptUpdate overflow
5718 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5719 is able to supply very large amounts of input data after a previous call to
5720 EVP_EncryptUpdate() with a partial block then a length check can overflow
5721 resulting in a heap corruption. Following an analysis of all OpenSSL
5722 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5723 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5724 the first called function after an EVP_EncryptInit(), and therefore that
5725 specific call must be safe. The second form is where the length passed to
5726 EVP_EncryptUpdate() can be seen from the code to be some small value and
5727 therefore there is no possibility of an overflow. Since all instances are
5728 one of these two forms, it is believed that there can be no overflows in
5729 internal code due to this problem. It should be noted that
5730 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5731 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5732 of these calls have also been analysed too and it is believed there are no
5733 instances in internal usage where an overflow could occur.
5735 This issue was reported by Guido Vranken.
5740 * Prevent ASN.1 BIO excessive memory allocation
5742 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5743 a short invalid encoding can casuse allocation of large amounts of memory
5744 potentially consuming excessive resources or exhausting memory.
5746 Any application parsing untrusted data through d2i BIO functions is
5747 affected. The memory based functions such as d2i_X509() are *not* affected.
5748 Since the memory based functions are used by the TLS library, TLS
5749 applications are not affected.
5751 This issue was reported by Brian Carpenter.
5758 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5759 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5760 in arbitrary stack data being returned in the buffer.
5762 This issue was reported by Guido Vranken.
5767 * Modify behavior of ALPN to invoke callback after SNI/servername
5768 callback, such that updates to the SSL_CTX affect ALPN.
5772 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5777 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5778 methods are enabled and ssl2 is disabled the methods return NULL.
5782 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5784 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5785 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5786 provide any "EXPORT" or "LOW" strength ciphers.
5790 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5791 is by default disabled at build-time. Builds that are not configured with
5792 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5793 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5794 will need to explicitly call either of:
5796 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5798 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5800 as appropriate. Even if either of those is used, or the application
5801 explicitly uses the version-specific SSLv2_method() or its client and
5802 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5803 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5804 ciphers, and SSLv2 56-bit DES are no longer available.
5809 * Fix a double-free in DSA code
5811 A double free bug was discovered when OpenSSL parses malformed DSA private
5812 keys and could lead to a DoS attack or memory corruption for applications
5813 that receive DSA private keys from untrusted sources. This scenario is
5816 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5822 * Disable SRP fake user seed to address a server memory leak.
5824 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5826 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5827 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5828 was changed to ignore the "fake user" SRP seed, even if the seed
5831 Users should use SRP_VBASE_get1_by_user instead. Note that in
5832 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5833 also that even though configuring the SRP seed attempts to hide
5834 invalid usernames by continuing the handshake with fake
5835 credentials, this behaviour is not constant time and no strong
5836 guarantees are made that the handshake is indistinguishable from
5837 that of a valid user.
5842 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5844 In the BN_hex2bn function the number of hex digits is calculated using an
5845 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5846 large values of `i` this can result in `bn_expand` not allocating any
5847 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5848 field as NULL leading to a subsequent NULL ptr deref. For very large values
5849 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5850 In this case memory is allocated to the internal BIGNUM data field, but it
5851 is insufficiently sized leading to heap corruption. A similar issue exists
5852 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5853 is ever called by user applications with very large untrusted hex/dec data.
5854 This is anticipated to be a rare occurrence.
5856 All OpenSSL internal usage of these functions use data that is not expected
5857 to be untrusted, e.g. config file data or application command line
5858 arguments. If user developed applications generate config file data based
5859 on untrusted data then it is possible that this could also lead to security
5860 consequences. This is also anticipated to be rare.
5862 This issue was reported to OpenSSL by Guido Vranken.
5867 * Fix memory issues in `BIO_*printf` functions
5869 The internal `fmtstr` function used in processing a "%s" format string in
5870 the `BIO_*printf` functions could overflow while calculating the length of a
5871 string and cause an OOB read when printing very long strings.
5873 Additionally the internal `doapr_outch` function can attempt to write to an
5874 OOB memory location (at an offset from the NULL pointer) in the event of a
5875 memory allocation failure. In 1.0.2 and below this could be caused where
5876 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5877 could be in processing a very long "%s" format string. Memory leaks can
5880 The first issue may mask the second issue dependent on compiler behaviour.
5881 These problems could enable attacks where large amounts of untrusted data
5882 is passed to the `BIO_*printf` functions. If applications use these functions
5883 in this way then they could be vulnerable. OpenSSL itself uses these
5884 functions when printing out human-readable dumps of ASN.1 data. Therefore
5885 applications that print this data could be vulnerable if the data is from
5886 untrusted sources. OpenSSL command line applications could also be
5887 vulnerable where they print out ASN.1 data, or if untrusted data is passed
5888 as command line arguments.
5890 Libssl is not considered directly vulnerable. Additionally certificates etc
5891 received via remote connections via libssl are also unlikely to be able to
5892 trigger these issues because of message size limits enforced within libssl.
5894 This issue was reported to OpenSSL Guido Vranken.
5899 * Side channel attack on modular exponentiation
5901 A side-channel attack was found which makes use of cache-bank conflicts on
5902 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
5903 of RSA keys. The ability to exploit this issue is limited as it relies on
5904 an attacker who has control of code in a thread running on the same
5905 hyper-threaded core as the victim thread which is performing decryptions.
5907 This issue was reported to OpenSSL by Yuval Yarom, The University of
5908 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
5909 Nadia Heninger, University of Pennsylvania with more information at
5910 <http://cachebleed.info>.
5915 * Change the req app to generate a 2048-bit RSA/DSA key by default,
5916 if no keysize is specified with default_bits. This fixes an
5917 omission in an earlier change that changed all RSA/DSA key generation
5918 apps to use 2048 bits by default.
5922 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
5924 * Protection for DH small subgroup attacks
5926 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
5927 switched on by default and cannot be disabled. This could have some
5932 * SSLv2 doesn't block disabled ciphers
5934 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5935 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5936 been disabled, provided that the SSLv2 protocol was not also disabled via
5939 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5940 and Sebastian Schinzel.
5945 * Reject DH handshakes with parameters shorter than 1024 bits.
5949 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
5951 * Certificate verify crash with missing PSS parameter
5953 The signature verification routines will crash with a NULL pointer
5954 dereference if presented with an ASN.1 signature using the RSA PSS
5955 algorithm and absent mask generation function parameter. Since these
5956 routines are used to verify certificate signature algorithms this can be
5957 used to crash any certificate verification operation and exploited in a
5958 DoS attack. Any application which performs certificate verification is
5959 vulnerable including OpenSSL clients and servers which enable client
5962 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5967 * X509_ATTRIBUTE memory leak
5969 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5970 memory. This structure is used by the PKCS#7 and CMS routines so any
5971 application which reads PKCS#7 or CMS data from untrusted sources is
5972 affected. SSL/TLS is not affected.
5974 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5980 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5981 This changes the decoding behaviour for some invalid messages,
5982 though the change is mostly in the more lenient direction, and
5983 legacy behaviour is preserved as much as possible.
5987 * In DSA_generate_parameters_ex, if the provided seed is too short,
5988 use a random seed, as already documented.
5990 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5992 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
5994 * Alternate chains certificate forgery
5996 During certificate verfification, OpenSSL will attempt to find an
5997 alternative certificate chain if the first attempt to build such a chain
5998 fails. An error in the implementation of this logic can mean that an
5999 attacker could cause certain checks on untrusted certificates to be
6000 bypassed, such as the CA flag, enabling them to use a valid leaf
6001 certificate to act as a CA and "issue" an invalid certificate.
6003 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6009 * Race condition handling PSK identify hint
6011 If PSK identity hints are received by a multi-threaded client then
6012 the values are wrongly updated in the parent SSL_CTX structure. This can
6013 result in a race condition potentially leading to a double free of the
6019 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6021 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6022 incompatibility in the handling of HMAC. The previous ABI has now been
6025 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6027 * Malformed ECParameters causes infinite loop
6029 When processing an ECParameters structure OpenSSL enters an infinite loop
6030 if the curve specified is over a specially malformed binary polynomial
6033 This can be used to perform denial of service against any
6034 system which processes public keys, certificate requests or
6035 certificates. This includes TLS clients and TLS servers with
6036 client authentication enabled.
6038 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6043 * Exploitable out-of-bounds read in X509_cmp_time
6045 X509_cmp_time does not properly check the length of the ASN1_TIME
6046 string and can read a few bytes out of bounds. In addition,
6047 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6050 An attacker can use this to craft malformed certificates and CRLs of
6051 various sizes and potentially cause a segmentation fault, resulting in
6052 a DoS on applications that verify certificates or CRLs. TLS clients
6053 that verify CRLs are affected. TLS clients and servers with client
6054 authentication enabled may be affected if they use custom verification
6057 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6058 independently by Hanno Böck.
6063 * PKCS7 crash with missing EnvelopedContent
6065 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6066 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6067 with missing content and trigger a NULL pointer dereference on parsing.
6069 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6070 structures from untrusted sources are affected. OpenSSL clients and
6071 servers are not affected.
6073 This issue was reported to OpenSSL by Michal Zalewski (Google).
6078 * CMS verify infinite loop with unknown hash function
6080 When verifying a signedData message the CMS code can enter an infinite loop
6081 if presented with an unknown hash function OID. This can be used to perform
6082 denial of service against any system which verifies signedData messages using
6084 This issue was reported to OpenSSL by Johannes Bauer.
6089 * Race condition handling NewSessionTicket
6091 If a NewSessionTicket is received by a multi-threaded client when attempting to
6092 reuse a previous ticket then a race condition can occur potentially leading to
6093 a double free of the ticket data.
6098 * Reject DH handshakes with parameters shorter than 768 bits.
6100 *Kurt Roeckx and Emilia Kasper*
6102 * dhparam: generate 2048-bit parameters by default.
6104 *Kurt Roeckx and Emilia Kasper*
6106 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6108 * Segmentation fault in ASN1_TYPE_cmp fix
6110 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6111 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6112 certificate signature algorithm consistency this can be used to crash any
6113 certificate verification operation and exploited in a DoS attack. Any
6114 application which performs certificate verification is vulnerable including
6115 OpenSSL clients and servers which enable client authentication.
6120 * ASN.1 structure reuse memory corruption fix
6122 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6123 memory corruption via an invalid write. Such reuse is and has been
6124 strongly discouraged and is believed to be rare.
6126 Applications that parse structures containing CHOICE or ANY DEFINED BY
6127 components may be affected. Certificate parsing (d2i_X509 and related
6128 functions) are however not affected. OpenSSL clients and servers are
6134 * PKCS7 NULL pointer dereferences fix
6136 The PKCS#7 parsing code does not handle missing outer ContentInfo
6137 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6138 missing content and trigger a NULL pointer dereference on parsing.
6140 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6141 otherwise parse PKCS#7 structures from untrusted sources are
6142 affected. OpenSSL clients and servers are not affected.
6144 This issue was reported to OpenSSL by Michal Zalewski (Google).
6149 * DoS via reachable assert in SSLv2 servers fix
6151 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6152 servers that both support SSLv2 and enable export cipher suites by sending
6153 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6155 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6156 (OpenSSL development team).
6161 * Use After Free following d2i_ECPrivatekey error fix
6163 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6164 could cause a use after free condition. This, in turn, could cause a double
6165 free in several private key parsing functions (such as d2i_PrivateKey
6166 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6167 for applications that receive EC private keys from untrusted
6168 sources. This scenario is considered rare.
6170 This issue was discovered by the BoringSSL project and fixed in their
6176 * X509_to_X509_REQ NULL pointer deref fix
6178 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6179 the certificate key is invalid. This function is rarely used in practice.
6181 This issue was discovered by Brian Carpenter.
6186 * Removed the export ciphers from the DEFAULT ciphers
6190 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6192 * Build fixes for the Windows and OpenVMS platforms
6194 *Matt Caswell and Richard Levitte*
6196 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6198 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6199 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6200 dereference. This could lead to a Denial Of Service attack. Thanks to
6201 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6206 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6207 dtls1_buffer_record function under certain conditions. In particular this
6208 could occur if an attacker sent repeated DTLS records with the same
6209 sequence number but for the next epoch. The memory leak could be exploited
6210 by an attacker in a Denial of Service attack through memory exhaustion.
6211 Thanks to Chris Mueller for reporting this issue.
6216 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6217 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6218 method would be set to NULL which could later result in a NULL pointer
6219 dereference. Thanks to Frank Schmirler for reporting this issue.
6224 * Abort handshake if server key exchange message is omitted for ephemeral
6227 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6228 reporting this issue.
6233 * Remove non-export ephemeral RSA code on client and server. This code
6234 violated the TLS standard by allowing the use of temporary RSA keys in
6235 non-export ciphersuites and could be used by a server to effectively
6236 downgrade the RSA key length used to a value smaller than the server
6237 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6238 INRIA or reporting this issue.
6243 * Fixed issue where DH client certificates are accepted without verification.
6244 An OpenSSL server will accept a DH certificate for client authentication
6245 without the certificate verify message. This effectively allows a client to
6246 authenticate without the use of a private key. This only affects servers
6247 which trust a client certificate authority which issues certificates
6248 containing DH keys: these are extremely rare and hardly ever encountered.
6249 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6255 * Ensure that the session ID context of an SSL is updated when its
6256 SSL_CTX is updated via SSL_set_SSL_CTX.
6258 The session ID context is typically set from the parent SSL_CTX,
6259 and can vary with the CTX.
6263 * Fix various certificate fingerprint issues.
6265 By using non-DER or invalid encodings outside the signed portion of a
6266 certificate the fingerprint can be changed without breaking the signature.
6267 Although no details of the signed portion of the certificate can be changed
6268 this can cause problems with some applications: e.g. those using the
6269 certificate fingerprint for blacklists.
6271 1. Reject signatures with non zero unused bits.
6273 If the BIT STRING containing the signature has non zero unused bits reject
6274 the signature. All current signature algorithms require zero unused bits.
6276 2. Check certificate algorithm consistency.
6278 Check the AlgorithmIdentifier inside TBS matches the one in the
6279 certificate signature. NB: this will result in signature failure
6280 errors for some broken certificates.
6282 Thanks to Konrad Kraszewski from Google for reporting this issue.
6284 3. Check DSA/ECDSA signatures use DER.
6286 Re-encode DSA/ECDSA signatures and compare with the original received
6287 signature. Return an error if there is a mismatch.
6289 This will reject various cases including garbage after signature
6290 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6291 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6292 (negative or with leading zeroes).
6294 Further analysis was conducted and fixes were developed by Stephen Henson
6295 of the OpenSSL core team.
6301 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6302 results on some platforms, including x86_64. This bug occurs at random
6303 with a very low probability, and is not known to be exploitable in any
6304 way, though its exact impact is difficult to determine. Thanks to Pieter
6305 Wuille (Blockstream) who reported this issue and also suggested an initial
6306 fix. Further analysis was conducted by the OpenSSL development team and
6307 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6308 the OpenSSL core team.
6313 * Do not resume sessions on the server if the negotiated protocol
6314 version does not match the session's version. Resuming with a different
6315 version, while not strictly forbidden by the RFC, is of questionable
6316 sanity and breaks all known clients.
6318 *David Benjamin, Emilia Käsper*
6320 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6321 early CCS messages during renegotiation. (Note that because
6322 renegotiation is encrypted, this early CCS was not exploitable.)
6326 * Tighten client-side session ticket handling during renegotiation:
6327 ensure that the client only accepts a session ticket if the server sends
6328 the extension anew in the ServerHello. Previously, a TLS client would
6329 reuse the old extension state and thus accept a session ticket if one was
6330 announced in the initial ServerHello.
6332 Similarly, ensure that the client requires a session ticket if one
6333 was advertised in the ServerHello. Previously, a TLS client would
6334 ignore a missing NewSessionTicket message.
6338 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6342 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6343 sends a carefully crafted handshake message, to cause OpenSSL to fail
6344 to free up to 64k of memory causing a memory leak. This could be
6345 exploited in a Denial Of Service attack. This issue affects OpenSSL
6346 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6347 whether SRTP is used or configured. Implementations of OpenSSL that
6348 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6350 The fix was developed by the OpenSSL team.
6355 * Session Ticket Memory Leak.
6357 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6358 integrity of that ticket is first verified. In the event of a session
6359 ticket integrity check failing, OpenSSL will fail to free memory
6360 causing a memory leak. By sending a large number of invalid session
6361 tickets an attacker could exploit this issue in a Denial Of Service
6367 * Build option no-ssl3 is incomplete.
6369 When OpenSSL is configured with "no-ssl3" as a build option, servers
6370 could accept and complete a SSL 3.0 handshake, and clients could be
6371 configured to send them.
6374 *Akamai and the OpenSSL team*
6376 * Add support for TLS_FALLBACK_SCSV.
6377 Client applications doing fallback retries should call
6378 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6381 *Adam Langley, Bodo Moeller*
6383 * Add additional DigestInfo checks.
6385 Re-encode DigestInto in DER and check against the original when
6386 verifying RSA signature: this will reject any improperly encoded
6387 DigestInfo structures.
6389 Note: this is a precautionary measure and no attacks are currently known.
6393 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6395 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6396 SRP code can be overrun an internal buffer. Add sanity check that
6397 g, A, B < N to SRP code.
6399 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6400 Group for discovering this issue.
6405 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6406 TLS 1.0 instead of higher protocol versions when the ClientHello message
6407 is badly fragmented. This allows a man-in-the-middle attacker to force a
6408 downgrade to TLS 1.0 even if both the server and the client support a
6409 higher protocol version, by modifying the client's TLS records.
6411 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6412 researching this issue.
6417 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6418 to a denial of service attack. A malicious server can crash the client
6419 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6420 ciphersuite and sending carefully crafted handshake messages.
6422 Thanks to Felix Gröbert (Google) for discovering and researching this
6428 * By sending carefully crafted DTLS packets an attacker could cause openssl
6429 to leak memory. This can be exploited through a Denial of Service attack.
6430 Thanks to Adam Langley for discovering and researching this issue.
6435 * An attacker can force openssl to consume large amounts of memory whilst
6436 processing DTLS handshake messages. This can be exploited through a
6437 Denial of Service attack.
6438 Thanks to Adam Langley for discovering and researching this issue.
6443 * An attacker can force an error condition which causes openssl to crash
6444 whilst processing DTLS packets due to memory being freed twice. This
6445 can be exploited through a Denial of Service attack.
6446 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6452 * If a multithreaded client connects to a malicious server using a resumed
6453 session and the server sends an ec point format extension it could write
6454 up to 255 bytes to freed memory.
6456 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6462 * A malicious server can crash an OpenSSL client with a null pointer
6463 dereference (read) by specifying an SRP ciphersuite even though it was not
6464 properly negotiated with the client. This can be exploited through a
6465 Denial of Service attack.
6467 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6468 discovering and researching this issue.
6473 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6474 X509_name_oneline, X509_name_print_ex et al. to leak some information
6475 from the stack. Applications may be affected if they echo pretty printing
6476 output to the attacker.
6478 Thanks to Ivan Fratric (Google) for discovering this issue.
6481 *Emilia Käsper, and Steve Henson*
6483 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6484 for corner cases. (Certain input points at infinity could lead to
6485 bogus results, with non-infinity inputs mapped to infinity too.)
6489 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6491 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6492 handshake can force the use of weak keying material in OpenSSL
6493 SSL/TLS clients and servers.
6495 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6496 researching this issue. ([CVE-2014-0224])
6498 *KIKUCHI Masashi, Steve Henson*
6500 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6501 OpenSSL DTLS client the code can be made to recurse eventually crashing
6504 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6507 *Imre Rad, Steve Henson*
6509 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6510 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6511 client or server. This is potentially exploitable to run arbitrary
6512 code on a vulnerable client or server.
6514 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6516 *Jüri Aedla, Steve Henson*
6518 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6519 are subject to a denial of service attack.
6521 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6522 this issue. ([CVE-2014-3470])
6524 *Felix Gröbert, Ivan Fratric, Steve Henson*
6526 * Harmonize version and its documentation. -f flag is used to display
6529 *mancha <mancha1@zoho.com>*
6531 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6532 in i2d_ECPrivateKey.
6534 *mancha <mancha1@zoho.com>*
6536 * Fix some double frees. These are not thought to be exploitable.
6538 *mancha <mancha1@zoho.com>*
6540 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6542 * A missing bounds check in the handling of the TLS heartbeat extension
6543 can be used to reveal up to 64k of memory to a connected client or
6546 Thanks for Neel Mehta of Google Security for discovering this bug and to
6547 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6548 preparing the fix ([CVE-2014-0160])
6550 *Adam Langley, Bodo Moeller*
6552 * Fix for the attack described in the paper "Recovering OpenSSL
6553 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6554 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6555 <http://eprint.iacr.org/2014/140>
6557 Thanks to Yuval Yarom and Naomi Benger for discovering this
6558 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6560 *Yuval Yarom and Naomi Benger*
6562 * TLS pad extension: draft-agl-tls-padding-03
6564 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6565 TLS client Hello record length value would otherwise be > 255 and
6566 less that 512 pad with a dummy extension containing zeroes so it
6567 is at least 512 bytes long.
6569 *Adam Langley, Steve Henson*
6571 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6573 * Fix for TLS record tampering bug. A carefully crafted invalid
6574 handshake could crash OpenSSL with a NULL pointer exception.
6575 Thanks to Anton Johansson for reporting this issues.
6578 * Keep original DTLS digest and encryption contexts in retransmission
6579 structures so we can use the previous session parameters if they need
6580 to be resent. ([CVE-2013-6450])
6584 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6585 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6586 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6587 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6588 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6589 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6591 *Rob Stradling, Adam Langley*
6593 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6595 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6596 supporting platforms or when small records were transferred.
6598 *Andy Polyakov, Steve Henson*
6600 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6602 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6604 This addresses the flaw in CBC record processing discovered by
6605 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6606 at: <http://www.isg.rhul.ac.uk/tls/>
6608 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6609 Security Group at Royal Holloway, University of London
6610 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6611 Emilia Käsper for the initial patch.
6614 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6616 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6617 ciphersuites which can be exploited in a denial of service attack.
6618 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6619 and detecting this bug and to Wolfgang Ettlinger
6620 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6625 * Return an error when checking OCSP signatures when key is NULL.
6626 This fixes a DoS attack. ([CVE-2013-0166])
6630 * Make openssl verify return errors.
6632 *Chris Palmer <palmer@google.com> and Ben Laurie*
6634 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6635 the right response is stapled. Also change SSL_get_certificate()
6636 so it returns the certificate actually sent.
6637 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6639 *Rob Stradling <rob.stradling@comodo.com>*
6641 * Fix possible deadlock when decoding public keys.
6645 * Don't use TLS 1.0 record version number in initial client hello
6650 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6652 * Sanity check record length before skipping explicit IV in TLS
6653 1.2, 1.1 and DTLS to fix DoS attack.
6655 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6656 fuzzing as a service testing platform.
6661 * Initialise tkeylen properly when encrypting CMS messages.
6662 Thanks to Solar Designer of Openwall for reporting this issue.
6666 * In FIPS mode don't try to use composite ciphers as they are not
6671 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6673 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6674 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6675 mean any application compiled against OpenSSL 1.0.0 headers setting
6676 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6677 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6678 0x10000000L Any application which was previously compiled against
6679 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6680 will need to be recompiled as a result. Letting be results in
6681 inability to disable specifically TLS 1.1 and in client context,
6682 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6686 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6687 disable just protocol X, but all protocols above X *if* there are
6688 protocols *below* X still enabled. In more practical terms it means
6689 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6690 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6691 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6696 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6698 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6699 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6700 in CRYPTO_realloc_clean.
6702 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6703 issue and to Adam Langley <agl@chromium.org> for fixing it.
6706 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6708 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6712 * Workarounds for some broken servers that "hang" if a client hello
6713 record length exceeds 255 bytes.
6715 1. Do not use record version number > TLS 1.0 in initial client
6716 hello: some (but not all) hanging servers will now work.
6717 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6718 the number of ciphers sent in the client hello. This should be
6719 set to an even number, such as 50, for example by passing:
6720 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6721 Most broken servers should now work.
6722 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6723 TLS 1.2 client support entirely.
6727 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6731 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6733 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6734 STRING form instead of a DigestInfo.
6738 * The format used for MDC2 RSA signatures is inconsistent between EVP
6739 and the RSA_sign/RSA_verify functions. This was made more apparent when
6740 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6741 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6742 the correct format in RSA_verify so both forms transparently work.
6746 * Some servers which support TLS 1.0 can choke if we initially indicate
6747 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6748 encrypted premaster secret. As a workaround use the maximum permitted
6749 client version in client hello, this should keep such servers happy
6750 and still work with previous versions of OpenSSL.
6754 * Add support for TLS/DTLS heartbeats.
6756 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6758 * Add support for SCTP.
6760 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6762 * Improved PRNG seeding for VOS.
6764 *Paul Green <Paul.Green@stratus.com>*
6766 * Extensive assembler packs updates, most notably:
6768 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6769 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6770 - x86_64: bit-sliced AES implementation;
6771 - ARM: NEON support, contemporary platforms optimizations;
6772 - s390x: z196 support;
6773 - `*`: GHASH and GF(2^m) multiplication implementations;
6777 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6778 (removal of unnecessary code)
6780 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6782 * Add TLS key material exporter from RFC 5705.
6786 * Add DTLS-SRTP negotiation from RFC 5764.
6790 * Add Next Protocol Negotiation,
6791 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6792 disabled with a no-npn flag to config or Configure. Code donated
6795 *Adam Langley <agl@google.com> and Ben Laurie*
6797 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6798 NIST-P256, NIST-P521, with constant-time single point multiplication on
6799 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6800 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6801 Code made available under Apache License version 2.0.
6803 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6804 line to include this in your build of OpenSSL, and run "make depend" (or
6805 "make update"). This enables the following EC_METHODs:
6807 EC_GFp_nistp224_method()
6808 EC_GFp_nistp256_method()
6809 EC_GFp_nistp521_method()
6811 EC_GROUP_new_by_curve_name() will automatically use these (while
6812 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6815 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6817 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6818 all platforms. Move ssize_t definition from e_os.h to the public
6819 header file e_os2.h as it now appears in public header file cms.h
6823 * New -sigopt option to the ca, req and x509 utilities. Additional
6824 signature parameters can be passed using this option and in
6829 * Add RSA PSS signing function. This will generate and set the
6830 appropriate AlgorithmIdentifiers for PSS based on those in the
6831 corresponding EVP_MD_CTX structure. No application support yet.
6835 * Support for companion algorithm specific ASN1 signing routines.
6836 New function ASN1_item_sign_ctx() signs a pre-initialised
6837 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6838 the appropriate parameters.
6842 * Add new algorithm specific ASN1 verification initialisation function
6843 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6844 handling will be the same no matter what EVP_PKEY_METHOD is used.
6845 Add a PSS handler to support verification of PSS signatures: checked
6846 against a number of sample certificates.
6850 * Add signature printing for PSS. Add PSS OIDs.
6852 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6854 * Add algorithm specific signature printing. An individual ASN1 method
6855 can now print out signatures instead of the standard hex dump.
6857 More complex signatures (e.g. PSS) can print out more meaningful
6858 information. Include DSA version that prints out the signature
6863 * Password based recipient info support for CMS library: implementing
6868 * Split password based encryption into PBES2 and PBKDF2 functions. This
6869 neatly separates the code into cipher and PBE sections and is required
6870 for some algorithms that split PBES2 into separate pieces (such as
6871 password based CMS).
6875 * Session-handling fixes:
6876 - Fix handling of connections that are resuming with a session ID,
6877 but also support Session Tickets.
6878 - Fix a bug that suppressed issuing of a new ticket if the client
6879 presented a ticket with an expired session.
6880 - Try to set the ticket lifetime hint to something reasonable.
6881 - Make tickets shorter by excluding irrelevant information.
6882 - On the client side, don't ignore renewed tickets.
6884 *Adam Langley, Bodo Moeller (Google)*
6886 * Fix PSK session representation.
6890 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
6892 This work was sponsored by Intel.
6896 * Add GCM support to TLS library. Some custom code is needed to split
6897 the IV between the fixed (from PRF) and explicit (from TLS record)
6898 portions. This adds all GCM ciphersuites supported by RFC5288 and
6899 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
6900 add a special AESGCM string for GCM only.
6904 * Expand range of ctrls for AES GCM. Permit setting invocation
6905 field on decrypt and retrieval of invocation field only on encrypt.
6909 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
6910 As required by RFC5289 these ciphersuites cannot be used if for
6911 versions of TLS earlier than 1.2.
6915 * For FIPS capable OpenSSL interpret a NULL default public key method
6916 as unset and return the appropriate default but do *not* set the default.
6917 This means we can return the appropriate method in applications that
6918 switch between FIPS and non-FIPS modes.
6922 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
6923 ENGINE is used then we cannot handle that in the FIPS module so we
6924 keep original code iff non-FIPS operations are allowed.
6928 * Add -attime option to openssl utilities.
6930 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
6932 * Redirect DSA and DH operations to FIPS module in FIPS mode.
6936 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
6937 FIPS EC methods unconditionally for now.
6941 * New build option no-ec2m to disable characteristic 2 code.
6945 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
6946 all cases can be covered as some introduce binary incompatibilities.
6950 * Redirect RSA operations to FIPS module including keygen,
6951 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
6955 * Add similar low level API blocking to ciphers.
6959 * Low level digest APIs are not approved in FIPS mode: any attempt
6960 to use these will cause a fatal error. Applications that *really* want
6961 to use them can use the `private_*` version instead.
6965 * Redirect cipher operations to FIPS module for FIPS builds.
6969 * Redirect digest operations to FIPS module for FIPS builds.
6973 * Update build system to add "fips" flag which will link in fipscanister.o
6974 for static and shared library builds embedding a signature if needed.
6978 * Output TLS supported curves in preference order instead of numerical
6979 order. This is currently hardcoded for the highest order curves first.
6980 This should be configurable so applications can judge speed vs strength.
6984 * Add TLS v1.2 server support for client authentication.
6988 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
6993 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
6994 FIPS modules versions.
6998 * Add TLS v1.2 client side support for client authentication. Keep cache
6999 of handshake records longer as we don't know the hash algorithm to use
7000 until after the certificate request message is received.
7004 * Initial TLS v1.2 client support. Add a default signature algorithms
7005 extension including all the algorithms we support. Parse new signature
7006 format in client key exchange. Relax some ECC signing restrictions for
7007 TLS v1.2 as indicated in RFC5246.
7011 * Add server support for TLS v1.2 signature algorithms extension. Switch
7012 to new signature format when needed using client digest preference.
7013 All server ciphersuites should now work correctly in TLS v1.2. No client
7014 support yet and no support for client certificates.
7018 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7019 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7020 ciphersuites. At present only RSA key exchange ciphersuites work with
7021 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7022 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7023 and version checking.
7027 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7028 with this defined it will not be affected by any changes to ssl internal
7029 structures. Add several utility functions to allow openssl application
7030 to work with OPENSSL_NO_SSL_INTERN defined.
7034 * A long standing patch to add support for SRP from EdelWeb (Peter
7035 Sylvester and Christophe Renou) was integrated.
7036 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7037 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7040 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7044 * Permit abbreviated handshakes when renegotiating using the function
7045 SSL_renegotiate_abbreviated().
7047 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7049 * Add call to ENGINE_register_all_complete() to
7050 ENGINE_load_builtin_engines(), so some implementations get used
7051 automatically instead of needing explicit application support.
7055 * Add support for TLS key exporter as described in RFC5705.
7057 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7059 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7060 a few changes are required:
7062 Add SSL_OP_NO_TLSv1_1 flag.
7063 Add TLSv1_1 methods.
7064 Update version checking logic to handle version 1.1.
7065 Add explicit IV handling (ported from DTLS code).
7066 Add command line options to s_client/s_server.
7073 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7075 * X509_ATTRIBUTE memory leak
7077 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7078 memory. This structure is used by the PKCS#7 and CMS routines so any
7079 application which reads PKCS#7 or CMS data from untrusted sources is
7080 affected. SSL/TLS is not affected.
7082 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7088 * Race condition handling PSK identify hint
7090 If PSK identity hints are received by a multi-threaded client then
7091 the values are wrongly updated in the parent SSL_CTX structure. This can
7092 result in a race condition potentially leading to a double free of the
7098 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7100 * Malformed ECParameters causes infinite loop
7102 When processing an ECParameters structure OpenSSL enters an infinite loop
7103 if the curve specified is over a specially malformed binary polynomial
7106 This can be used to perform denial of service against any
7107 system which processes public keys, certificate requests or
7108 certificates. This includes TLS clients and TLS servers with
7109 client authentication enabled.
7111 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7116 * Exploitable out-of-bounds read in X509_cmp_time
7118 X509_cmp_time does not properly check the length of the ASN1_TIME
7119 string and can read a few bytes out of bounds. In addition,
7120 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7123 An attacker can use this to craft malformed certificates and CRLs of
7124 various sizes and potentially cause a segmentation fault, resulting in
7125 a DoS on applications that verify certificates or CRLs. TLS clients
7126 that verify CRLs are affected. TLS clients and servers with client
7127 authentication enabled may be affected if they use custom verification
7130 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7131 independently by Hanno Böck.
7136 * PKCS7 crash with missing EnvelopedContent
7138 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7139 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7140 with missing content and trigger a NULL pointer dereference on parsing.
7142 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7143 structures from untrusted sources are affected. OpenSSL clients and
7144 servers are not affected.
7146 This issue was reported to OpenSSL by Michal Zalewski (Google).
7151 * CMS verify infinite loop with unknown hash function
7153 When verifying a signedData message the CMS code can enter an infinite loop
7154 if presented with an unknown hash function OID. This can be used to perform
7155 denial of service against any system which verifies signedData messages using
7157 This issue was reported to OpenSSL by Johannes Bauer.
7162 * Race condition handling NewSessionTicket
7164 If a NewSessionTicket is received by a multi-threaded client when attempting to
7165 reuse a previous ticket then a race condition can occur potentially leading to
7166 a double free of the ticket data.
7171 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7173 * Segmentation fault in ASN1_TYPE_cmp fix
7175 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7176 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7177 certificate signature algorithm consistency this can be used to crash any
7178 certificate verification operation and exploited in a DoS attack. Any
7179 application which performs certificate verification is vulnerable including
7180 OpenSSL clients and servers which enable client authentication.
7185 * ASN.1 structure reuse memory corruption fix
7187 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7188 memory corruption via an invalid write. Such reuse is and has been
7189 strongly discouraged and is believed to be rare.
7191 Applications that parse structures containing CHOICE or ANY DEFINED BY
7192 components may be affected. Certificate parsing (d2i_X509 and related
7193 functions) are however not affected. OpenSSL clients and servers are
7199 * PKCS7 NULL pointer dereferences fix
7201 The PKCS#7 parsing code does not handle missing outer ContentInfo
7202 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7203 missing content and trigger a NULL pointer dereference on parsing.
7205 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7206 otherwise parse PKCS#7 structures from untrusted sources are
7207 affected. OpenSSL clients and servers are not affected.
7209 This issue was reported to OpenSSL by Michal Zalewski (Google).
7214 * DoS via reachable assert in SSLv2 servers fix
7216 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7217 servers that both support SSLv2 and enable export cipher suites by sending
7218 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7220 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7221 (OpenSSL development team).
7226 * Use After Free following d2i_ECPrivatekey error fix
7228 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7229 could cause a use after free condition. This, in turn, could cause a double
7230 free in several private key parsing functions (such as d2i_PrivateKey
7231 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7232 for applications that receive EC private keys from untrusted
7233 sources. This scenario is considered rare.
7235 This issue was discovered by the BoringSSL project and fixed in their
7241 * X509_to_X509_REQ NULL pointer deref fix
7243 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7244 the certificate key is invalid. This function is rarely used in practice.
7246 This issue was discovered by Brian Carpenter.
7251 * Removed the export ciphers from the DEFAULT ciphers
7255 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7257 * Build fixes for the Windows and OpenVMS platforms
7259 *Matt Caswell and Richard Levitte*
7261 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7263 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7264 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7265 dereference. This could lead to a Denial Of Service attack. Thanks to
7266 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7271 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7272 dtls1_buffer_record function under certain conditions. In particular this
7273 could occur if an attacker sent repeated DTLS records with the same
7274 sequence number but for the next epoch. The memory leak could be exploited
7275 by an attacker in a Denial of Service attack through memory exhaustion.
7276 Thanks to Chris Mueller for reporting this issue.
7281 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7282 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7283 method would be set to NULL which could later result in a NULL pointer
7284 dereference. Thanks to Frank Schmirler for reporting this issue.
7289 * Abort handshake if server key exchange message is omitted for ephemeral
7292 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7293 reporting this issue.
7298 * Remove non-export ephemeral RSA code on client and server. This code
7299 violated the TLS standard by allowing the use of temporary RSA keys in
7300 non-export ciphersuites and could be used by a server to effectively
7301 downgrade the RSA key length used to a value smaller than the server
7302 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7303 INRIA or reporting this issue.
7308 * Fixed issue where DH client certificates are accepted without verification.
7309 An OpenSSL server will accept a DH certificate for client authentication
7310 without the certificate verify message. This effectively allows a client to
7311 authenticate without the use of a private key. This only affects servers
7312 which trust a client certificate authority which issues certificates
7313 containing DH keys: these are extremely rare and hardly ever encountered.
7314 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7320 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7321 results on some platforms, including x86_64. This bug occurs at random
7322 with a very low probability, and is not known to be exploitable in any
7323 way, though its exact impact is difficult to determine. Thanks to Pieter
7324 Wuille (Blockstream) who reported this issue and also suggested an initial
7325 fix. Further analysis was conducted by the OpenSSL development team and
7326 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7327 the OpenSSL core team.
7332 * Fix various certificate fingerprint issues.
7334 By using non-DER or invalid encodings outside the signed portion of a
7335 certificate the fingerprint can be changed without breaking the signature.
7336 Although no details of the signed portion of the certificate can be changed
7337 this can cause problems with some applications: e.g. those using the
7338 certificate fingerprint for blacklists.
7340 1. Reject signatures with non zero unused bits.
7342 If the BIT STRING containing the signature has non zero unused bits reject
7343 the signature. All current signature algorithms require zero unused bits.
7345 2. Check certificate algorithm consistency.
7347 Check the AlgorithmIdentifier inside TBS matches the one in the
7348 certificate signature. NB: this will result in signature failure
7349 errors for some broken certificates.
7351 Thanks to Konrad Kraszewski from Google for reporting this issue.
7353 3. Check DSA/ECDSA signatures use DER.
7355 Reencode DSA/ECDSA signatures and compare with the original received
7356 signature. Return an error if there is a mismatch.
7358 This will reject various cases including garbage after signature
7359 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7360 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7361 (negative or with leading zeroes).
7363 Further analysis was conducted and fixes were developed by Stephen Henson
7364 of the OpenSSL core team.
7370 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7372 * Session Ticket Memory Leak.
7374 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7375 integrity of that ticket is first verified. In the event of a session
7376 ticket integrity check failing, OpenSSL will fail to free memory
7377 causing a memory leak. By sending a large number of invalid session
7378 tickets an attacker could exploit this issue in a Denial Of Service
7384 * Build option no-ssl3 is incomplete.
7386 When OpenSSL is configured with "no-ssl3" as a build option, servers
7387 could accept and complete a SSL 3.0 handshake, and clients could be
7388 configured to send them.
7391 *Akamai and the OpenSSL team*
7393 * Add support for TLS_FALLBACK_SCSV.
7394 Client applications doing fallback retries should call
7395 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7398 *Adam Langley, Bodo Moeller*
7400 * Add additional DigestInfo checks.
7402 Reencode DigestInto in DER and check against the original when
7403 verifying RSA signature: this will reject any improperly encoded
7404 DigestInfo structures.
7406 Note: this is a precautionary measure and no attacks are currently known.
7410 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7412 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7413 to a denial of service attack. A malicious server can crash the client
7414 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7415 ciphersuite and sending carefully crafted handshake messages.
7417 Thanks to Felix Gröbert (Google) for discovering and researching this
7423 * By sending carefully crafted DTLS packets an attacker could cause openssl
7424 to leak memory. This can be exploited through a Denial of Service attack.
7425 Thanks to Adam Langley for discovering and researching this issue.
7430 * An attacker can force openssl to consume large amounts of memory whilst
7431 processing DTLS handshake messages. This can be exploited through a
7432 Denial of Service attack.
7433 Thanks to Adam Langley for discovering and researching this issue.
7438 * An attacker can force an error condition which causes openssl to crash
7439 whilst processing DTLS packets due to memory being freed twice. This
7440 can be exploited through a Denial of Service attack.
7441 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7447 * If a multithreaded client connects to a malicious server using a resumed
7448 session and the server sends an ec point format extension it could write
7449 up to 255 bytes to freed memory.
7451 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7457 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7458 X509_name_oneline, X509_name_print_ex et al. to leak some information
7459 from the stack. Applications may be affected if they echo pretty printing
7460 output to the attacker.
7462 Thanks to Ivan Fratric (Google) for discovering this issue.
7465 *Emilia Käsper, and Steve Henson*
7467 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7468 for corner cases. (Certain input points at infinity could lead to
7469 bogus results, with non-infinity inputs mapped to infinity too.)
7473 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7475 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7476 handshake can force the use of weak keying material in OpenSSL
7477 SSL/TLS clients and servers.
7479 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7480 researching this issue. ([CVE-2014-0224])
7482 *KIKUCHI Masashi, Steve Henson*
7484 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7485 OpenSSL DTLS client the code can be made to recurse eventually crashing
7488 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7491 *Imre Rad, Steve Henson*
7493 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7494 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7495 client or server. This is potentially exploitable to run arbitrary
7496 code on a vulnerable client or server.
7498 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7500 *Jüri Aedla, Steve Henson*
7502 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7503 are subject to a denial of service attack.
7505 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7506 this issue. ([CVE-2014-3470])
7508 *Felix Gröbert, Ivan Fratric, Steve Henson*
7510 * Harmonize version and its documentation. -f flag is used to display
7513 *mancha <mancha1@zoho.com>*
7515 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7516 in i2d_ECPrivateKey.
7518 *mancha <mancha1@zoho.com>*
7520 * Fix some double frees. These are not thought to be exploitable.
7522 *mancha <mancha1@zoho.com>*
7524 * Fix for the attack described in the paper "Recovering OpenSSL
7525 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7526 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7527 <http://eprint.iacr.org/2014/140>
7529 Thanks to Yuval Yarom and Naomi Benger for discovering this
7530 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7532 *Yuval Yarom and Naomi Benger*
7534 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7536 * Keep original DTLS digest and encryption contexts in retransmission
7537 structures so we can use the previous session parameters if they need
7538 to be resent. ([CVE-2013-6450])
7542 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7543 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7544 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7545 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7546 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7547 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7549 *Rob Stradling, Adam Langley*
7551 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7553 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7555 This addresses the flaw in CBC record processing discovered by
7556 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7557 at: <http://www.isg.rhul.ac.uk/tls/>
7559 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7560 Security Group at Royal Holloway, University of London
7561 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7562 Emilia Käsper for the initial patch.
7565 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7567 * Return an error when checking OCSP signatures when key is NULL.
7568 This fixes a DoS attack. ([CVE-2013-0166])
7572 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7573 the right response is stapled. Also change SSL_get_certificate()
7574 so it returns the certificate actually sent.
7575 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7576 (This is a backport)
7578 *Rob Stradling <rob.stradling@comodo.com>*
7580 * Fix possible deadlock when decoding public keys.
7584 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7586 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7589 * Sanity check record length before skipping explicit IV in DTLS
7592 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7593 fuzzing as a service testing platform.
7598 * Initialise tkeylen properly when encrypting CMS messages.
7599 Thanks to Solar Designer of Openwall for reporting this issue.
7603 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7605 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7606 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7607 in CRYPTO_realloc_clean.
7609 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7610 issue and to Adam Langley <agl@chromium.org> for fixing it.
7613 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7615 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7617 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7618 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7619 content decryption and always return the same error. Note: this attack
7620 needs on average 2^20 messages so it only affects automated senders. The
7621 old behaviour can be re-enabled in the CMS code by setting the
7622 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7623 an MMA defence is not necessary.
7624 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7625 this issue. ([CVE-2012-0884])
7629 * Fix CVE-2011-4619: make sure we really are receiving a
7630 client hello before rejecting multiple SGC restarts. Thanks to
7631 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7635 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7637 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7638 Thanks to Antonio Martin, Enterprise Secure Access Research and
7639 Development, Cisco Systems, Inc. for discovering this bug and
7640 preparing a fix. ([CVE-2012-0050])
7644 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7646 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7647 of the Vaudenay padding oracle attack on CBC mode encryption
7648 which enables an efficient plaintext recovery attack against
7649 the OpenSSL implementation of DTLS. Their attack exploits timing
7650 differences arising during decryption processing. A research
7651 paper describing this attack can be found at:
7652 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7653 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7654 Security Group at Royal Holloway, University of London
7655 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7656 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7657 for preparing the fix. ([CVE-2011-4108])
7659 *Robin Seggelmann, Michael Tuexen*
7661 * Clear bytes used for block padding of SSL 3.0 records.
7664 *Adam Langley (Google)*
7666 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7667 Kadianakis <desnacked@gmail.com> for discovering this issue and
7668 Adam Langley for preparing the fix. ([CVE-2011-4619])
7670 *Adam Langley (Google)*
7672 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7674 *Andrey Kulikov <amdeich@gmail.com>*
7676 * Prevent malformed RFC3779 data triggering an assertion failure.
7677 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7678 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7680 *Rob Austein <sra@hactrn.net>*
7682 * Improved PRNG seeding for VOS.
7684 *Paul Green <Paul.Green@stratus.com>*
7686 * Fix ssl_ciph.c set-up race.
7688 *Adam Langley (Google)*
7690 * Fix spurious failures in ecdsatest.c.
7692 *Emilia Käsper (Google)*
7694 * Fix the BIO_f_buffer() implementation (which was mixing different
7695 interpretations of the `..._len` fields).
7697 *Adam Langley (Google)*
7699 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7700 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7701 threads won't reuse the same blinding coefficients.
7703 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7704 lock to call BN_BLINDING_invert_ex, and avoids one use of
7705 BN_BLINDING_update for each BN_BLINDING structure (previously,
7706 the last update always remained unused).
7708 *Emilia Käsper (Google)*
7710 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7712 *Bob Buckholz (Google)*
7714 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7716 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7717 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7719 *Kaspar Brand <ossl@velox.ch>*
7721 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7722 for multi-threaded use of ECDH. ([CVE-2011-3210])
7724 *Adam Langley (Google)*
7726 * Fix x509_name_ex_d2i memory leak on bad inputs.
7730 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7731 signature public key algorithm by using OID xref utilities instead.
7732 Before this you could only use some ECC ciphersuites with SHA1 only.
7736 * Add protection against ECDSA timing attacks as mentioned in the paper
7737 by Billy Bob Brumley and Nicola Tuveri, see:
7738 <http://eprint.iacr.org/2011/232.pdf>
7740 *Billy Bob Brumley and Nicola Tuveri*
7742 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7744 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7746 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7748 * Fix bug in string printing code: if *any* escaping is enabled we must
7749 escape the escape character (backslash) or the resulting string is
7754 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7756 * Disable code workaround for ancient and obsolete Netscape browsers
7757 and servers: an attacker can use it in a ciphersuite downgrade attack.
7758 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7762 * Fixed J-PAKE implementation error, originally discovered by
7763 Sebastien Martini, further info and confirmation from Stefan
7764 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7768 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7770 * Fix extension code to avoid race conditions which can result in a buffer
7771 overrun vulnerability: resumed sessions must not be modified as they can
7772 be shared by multiple threads. CVE-2010-3864
7776 * Fix WIN32 build system to correctly link an ENGINE directory into
7781 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7783 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7786 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7788 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7790 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7791 context. The operation can be customised via the ctrl mechanism in
7792 case ENGINEs want to include additional functionality.
7796 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7800 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7801 output hashes compatible with older versions of OpenSSL.
7803 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7805 * Fix compression algorithm handling: if resuming a session use the
7806 compression algorithm of the resumed session instead of determining
7807 it from client hello again. Don't allow server to change algorithm.
7811 * Add load_crls() function to apps tidying load_certs() too. Add option
7812 to verify utility to allow additional CRLs to be included.
7816 * Update OCSP request code to permit adding custom headers to the request:
7817 some responders need this.
7821 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7824 *Julia Lawall <julia@diku.dk>*
7826 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7827 needlessly dereferenced structures, used obsolete functions and
7828 didn't handle all updated verify codes correctly.
7832 * Disable MD2 in the default configuration.
7836 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7837 indicate the initial BIO being pushed or popped. This makes it possible
7838 to determine whether the BIO is the one explicitly called or as a result
7839 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7840 it handles reference counts correctly and doesn't zero out the I/O bio
7841 when it is not being explicitly popped. WARNING: applications which
7842 included workarounds for the old buggy behaviour will need to be modified
7843 or they could free up already freed BIOs.
7847 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7848 renaming to all platforms (within the 0.9.8 branch, this was
7849 done conditionally on Netware platforms to avoid a name clash).
7851 *Guenter <lists@gknw.net>*
7853 * Add ECDHE and PSK support to DTLS.
7855 *Michael Tuexen <tuexen@fh-muenster.de>*
7857 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7862 * Add "missing" function EVP_MD_flags() (without this the only way to
7863 retrieve a digest flags is by accessing the structure directly. Update
7864 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7865 or cipher is registered as in the "from" argument. Print out all
7866 registered digests in the dgst usage message instead of manually
7867 attempting to work them out.
7871 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7872 this allows the use of compression and extensions. Change default cipher
7873 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7874 by default unless an application cipher string requests it.
7878 * Alter match criteria in PKCS12_parse(). It used to try to use local
7879 key ids to find matching certificates and keys but some PKCS#12 files
7880 don't follow the (somewhat unwritten) rules and this strategy fails.
7881 Now just gather all certificates together and the first private key
7882 then look for the first certificate that matches the key.
7886 * Support use of registered digest and cipher names for dgst and cipher
7887 commands instead of having to add each one as a special case. So now
7894 openssl dgst -sha256 foo
7896 and this works for ENGINE based algorithms too.
7900 * Update Gost ENGINE to support parameter files.
7902 *Victor B. Wagner <vitus@cryptocom.ru>*
7904 * Support GeneralizedTime in ca utility.
7906 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
7908 * Enhance the hash format used for certificate directory links. The new
7909 form uses the canonical encoding (meaning equivalent names will work
7910 even if they aren't identical) and uses SHA1 instead of MD5. This form
7911 is incompatible with the older format and as a result c_rehash should
7912 be used to rebuild symbolic links.
7916 * Make PKCS#8 the default write format for private keys, replacing the
7917 traditional format. This form is standardised, more secure and doesn't
7918 include an implicit MD5 dependency.
7922 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
7923 committed to OpenSSL should pass this lot as a minimum.
7927 * Add session ticket override functionality for use by EAP-FAST.
7929 *Jouni Malinen <j@w1.fi>*
7931 * Modify HMAC functions to return a value. Since these can be implemented
7932 in an ENGINE errors can occur.
7936 * Type-checked OBJ_bsearch_ex.
7940 * Type-checked OBJ_bsearch. Also some constification necessitated
7941 by type-checking. Still to come: TXT_DB, bsearch(?),
7942 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
7947 * New function OPENSSL_gmtime_adj() to add a specific number of days and
7948 seconds to a tm structure directly, instead of going through OS
7949 specific date routines. This avoids any issues with OS routines such
7950 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
7951 and X509_time_adj_ex() to cover the extended range. The existing
7952 X509_time_adj() is still usable and will no longer have any date issues.
7956 * Delta CRL support. New use deltas option which will attempt to locate
7957 and search any appropriate delta CRLs available.
7959 This work was sponsored by Google.
7963 * Support for CRLs partitioned by reason code. Reorganise CRL processing
7964 code and add additional score elements. Validate alternate CRL paths
7965 as part of the CRL checking and indicate a new error "CRL path validation
7966 error" in this case. Applications wanting additional details can use
7967 the verify callback and check the new "parent" field. If this is not
7968 NULL CRL path validation is taking place. Existing applications won't
7969 see this because it requires extended CRL support which is off by
7972 This work was sponsored by Google.
7976 * Support for freshest CRL extension.
7978 This work was sponsored by Google.
7982 * Initial indirect CRL support. Currently only supported in the CRLs
7983 passed directly and not via lookup. Process certificate issuer
7984 CRL entry extension and lookup CRL entries by bother issuer name
7985 and serial number. Check and process CRL issuer entry in IDP extension.
7987 This work was sponsored by Google.
7991 * Add support for distinct certificate and CRL paths. The CRL issuer
7992 certificate is validated separately in this case. Only enabled if
7993 an extended CRL support flag is set: this flag will enable additional
7994 CRL functionality in future.
7996 This work was sponsored by Google.
8000 * Add support for policy mappings extension.
8002 This work was sponsored by Google.
8006 * Fixes to pathlength constraint, self issued certificate handling,
8007 policy processing to align with RFC3280 and PKITS tests.
8009 This work was sponsored by Google.
8013 * Support for name constraints certificate extension. DN, email, DNS
8014 and URI types are currently supported.
8016 This work was sponsored by Google.
8020 * To cater for systems that provide a pointer-based thread ID rather
8021 than numeric, deprecate the current numeric thread ID mechanism and
8022 replace it with a structure and associated callback type. This
8023 mechanism allows a numeric "hash" to be extracted from a thread ID in
8024 either case, and on platforms where pointers are larger than 'long',
8025 mixing is done to help ensure the numeric 'hash' is usable even if it
8026 can't be guaranteed unique. The default mechanism is to use "&errno"
8027 as a pointer-based thread ID to distinguish between threads.
8029 Applications that want to provide their own thread IDs should now use
8030 CRYPTO_THREADID_set_callback() to register a callback that will call
8031 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8033 Note that ERR_remove_state() is now deprecated, because it is tied
8034 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8035 to free the current thread's error state should be replaced by
8036 ERR_remove_thread_state(NULL).
8038 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8039 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8040 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8041 application was previously providing a numeric thread callback that
8042 was inappropriate for distinguishing threads, then uniqueness might
8043 have been obtained with &errno that happened immediately in the
8044 intermediate development versions of OpenSSL; this is no longer the
8045 case, the numeric thread callback will now override the automatic use
8048 *Geoff Thorpe, with help from Bodo Moeller*
8050 * Initial support for different CRL issuing certificates. This covers a
8051 simple case where the self issued certificates in the chain exist and
8052 the real CRL issuer is higher in the existing chain.
8054 This work was sponsored by Google.
8058 * Removed effectively defunct crypto/store from the build.
8062 * Revamp of STACK to provide stronger type-checking. Still to come:
8063 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8064 ASN1_STRING, CONF_VALUE.
8068 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8069 RAM on SSL connections. This option can save about 34k per idle SSL.
8073 * Revamp of LHASH to provide stronger type-checking. Still to come:
8074 STACK, TXT_DB, bsearch, qsort.
8078 * Initial support for Cryptographic Message Syntax (aka CMS) based
8079 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8080 support for data, signedData, compressedData, digestedData and
8081 encryptedData, envelopedData types included. Scripts to check against
8082 RFC4134 examples draft and interop and consistency checks of many
8083 content types and variants.
8087 * Add options to enc utility to support use of zlib compression BIO.
8091 * Extend mk1mf to support importing of options and assembly language
8092 files from Configure script, currently only included in VC-WIN32.
8093 The assembly language rules can now optionally generate the source
8094 files from the associated perl scripts.
8098 * Implement remaining functionality needed to support GOST ciphersuites.
8099 Interop testing has been performed using CryptoPro implementations.
8101 *Victor B. Wagner <vitus@cryptocom.ru>*
8103 * s390x assembler pack.
8107 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8112 * Implement Opaque PRF Input TLS extension as specified in
8113 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8114 official specification yet and no extension type assignment by
8115 IANA exists, this extension (for now) will have to be explicitly
8116 enabled when building OpenSSL by providing the extension number
8117 to use. For example, specify an option
8119 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8121 to the "config" or "Configure" script to enable the extension,
8122 assuming extension number 0x9527 (which is a completely arbitrary
8123 and unofficial assignment based on the MD5 hash of the Internet
8124 Draft). Note that by doing so, you potentially lose
8125 interoperability with other TLS implementations since these might
8126 be using the same extension number for other purposes.
8128 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8129 opaque PRF input value to use in the handshake. This will create
8130 an internal copy of the length-'len' string at 'src', and will
8131 return non-zero for success.
8133 To get more control and flexibility, provide a callback function
8136 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8137 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8141 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8144 Callback function 'cb' will be called in handshakes, and is
8145 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8146 Argument 'arg' is for application purposes (the value as given to
8147 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8148 be provided to the callback function). The callback function
8149 has to return non-zero to report success: usually 1 to use opaque
8150 PRF input just if possible, or 2 to enforce use of the opaque PRF
8151 input. In the latter case, the library will abort the handshake
8152 if opaque PRF input is not successfully negotiated.
8154 Arguments 'peerinput' and 'len' given to the callback function
8155 will always be NULL and 0 in the case of a client. A server will
8156 see the client's opaque PRF input through these variables if
8157 available (NULL and 0 otherwise). Note that if the server
8158 provides an opaque PRF input, the length must be the same as the
8159 length of the client's opaque PRF input.
8161 Note that the callback function will only be called when creating
8162 a new session (session resumption can resume whatever was
8163 previously negotiated), and will not be called in SSL 2.0
8164 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8165 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8166 for applications that need to enforce opaque PRF input.
8170 * Update ssl code to support digests other than SHA1+MD5 for handshake
8173 *Victor B. Wagner <vitus@cryptocom.ru>*
8175 * Add RFC4507 support to OpenSSL. This includes the corrections in
8176 RFC4507bis. The encrypted ticket format is an encrypted encoded
8177 SSL_SESSION structure, that way new session features are automatically
8180 If a client application caches session in an SSL_SESSION structure
8181 support is transparent because tickets are now stored in the encoded
8184 The SSL_CTX structure automatically generates keys for ticket
8185 protection in servers so again support should be possible
8186 with no application modification.
8188 If a client or server wishes to disable RFC4507 support then the option
8189 SSL_OP_NO_TICKET can be set.
8191 Add a TLS extension debugging callback to allow the contents of any client
8192 or server extensions to be examined.
8194 This work was sponsored by Google.
8198 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8199 OpenSSL should now compile cleanly on gcc 4.2
8201 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8203 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8204 support including streaming MAC support: this is required for GOST
8205 ciphersuite support.
8207 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8209 * Add option -stream to use PKCS#7 streaming in smime utility. New
8210 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8211 to output in BER and PEM format.
8215 * Experimental support for use of HMAC via EVP_PKEY interface. This
8216 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8217 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8218 ENGINE support for HMAC keys which are unextractable. New -mac and
8219 -macopt options to dgst utility.
8223 * New option -sigopt to dgst utility. Update dgst to use
8224 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8225 alternative signing parameters such as X9.31 or PSS in the dgst
8230 * Change ssl_cipher_apply_rule(), the internal function that does
8231 the work each time a ciphersuite string requests enabling
8232 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8233 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8234 the order of disabled ciphersuites such that those ciphersuites
8235 that most recently went from enabled to disabled not only stay
8236 in order with respect to each other, but also have higher priority
8237 than other disabled ciphersuites the next time ciphersuites are
8240 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8241 the same ciphersuites as with "HIGH" alone, but in a specific
8242 order where the PSK ciphersuites come first (since they are the
8243 most recently disabled ciphersuites when "HIGH" is parsed).
8245 Also, change ssl_create_cipher_list() (using this new
8246 functionality) such that between otherwise identical
8247 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8252 * Change ssl_create_cipher_list() so that it automatically
8253 arranges the ciphersuites in reasonable order before starting
8254 to process the rule string. Thus, the definition for "DEFAULT"
8255 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8256 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8257 This makes it much easier to arrive at a reasonable default order
8258 in applications for which anonymous ciphers are OK (meaning
8259 that you can't actually use DEFAULT).
8261 *Bodo Moeller; suggested by Victor Duchovni*
8263 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8264 processing) into multiple integers instead of setting
8265 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8266 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8267 (These masks as well as the individual bit definitions are hidden
8268 away into the non-exported interface ssl/ssl_locl.h, so this
8269 change to the definition of the SSL_CIPHER structure shouldn't
8270 affect applications.) This give us more bits for each of these
8271 categories, so there is no longer a need to coagulate AES128 and
8272 AES256 into a single algorithm bit, and to coagulate Camellia128
8273 and Camellia256 into a single algorithm bit, which has led to all
8276 Thus, among other things, the kludge introduced in 0.9.7m and
8277 0.9.8e for masking out AES256 independently of AES128 or masking
8278 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8280 With the change, we also introduce new ciphersuite aliases that
8281 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8286 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8287 Use the leftmost N bytes of the signature input if the input is
8288 larger than the prime q (with N being the size in bytes of q).
8292 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8293 it yet and it is largely untested.
8297 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8301 * Initial incomplete changes to avoid need for function casts in OpenSSL
8302 some compilers (gcc 4.2 and later) reject their use. Safestack is
8303 reimplemented. Update ASN1 to avoid use of legacy functions.
8307 * Win32/64 targets are linked with Winsock2.
8311 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8312 to external functions. This can be used to increase CRL handling
8313 efficiency especially when CRLs are very large by (for example) storing
8314 the CRL revoked certificates in a database.
8318 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8319 new CRLs added to a directory can be used. New command line option
8320 -verify_return_error to s_client and s_server. This causes real errors
8321 to be returned by the verify callback instead of carrying on no matter
8322 what. This reflects the way a "real world" verify callback would behave.
8326 * GOST engine, supporting several GOST algorithms and public key formats.
8327 Kindly donated by Cryptocom.
8331 * Partial support for Issuing Distribution Point CRL extension. CRLs
8332 partitioned by DP are handled but no indirect CRL or reason partitioning
8333 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8334 selected via a scoring technique which handles IDP and AKID in CRLs.
8338 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8339 will ultimately be used for all verify operations: this will remove the
8340 X509_STORE dependency on certificate verification and allow alternative
8341 lookup methods. X509_STORE based implementations of these two callbacks.
8345 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8346 Modify get_crl() to find a valid (unexpired) CRL if possible.
8350 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8351 this would be called X509_CRL_cmp() but that name is already used by
8352 a function that just compares CRL issuer names. Cache several CRL
8353 extensions in X509_CRL structure and cache CRLDP in X509.
8357 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8358 this maps equivalent X509_NAME structures into a consistent structure.
8359 Name comparison can then be performed rapidly using memcmp().
8363 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8368 * Allow digests to supply their own micalg string for S/MIME type using
8369 the ctrl EVP_MD_CTRL_MICALG.
8373 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8374 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8375 ctrl. It can then customise the structure before and/or after signing
8380 * New function OBJ_add_sigid() to allow application defined signature OIDs
8381 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8382 to free up any added signature OIDs.
8386 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8387 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8388 digest and cipher tables. New options added to openssl utility:
8389 list-message-digest-algorithms and list-cipher-algorithms.
8393 * Change the array representation of binary polynomials: the list
8394 of degrees of non-zero coefficients is now terminated with -1.
8395 Previously it was terminated with 0, which was also part of the
8396 value; thus, the array representation was not applicable to
8397 polynomials where t^0 has coefficient zero. This change makes
8398 the array representation useful in a more general context.
8402 * Various modifications and fixes to SSL/TLS cipher string
8403 handling. For ECC, the code now distinguishes between fixed ECDH
8404 with RSA certificates on the one hand and with ECDSA certificates
8405 on the other hand, since these are separate ciphersuites. The
8406 unused code for Fortezza ciphersuites has been removed.
8408 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8409 (not "ECDHE"). For consistency with the code for DH
8410 certificates, use of ECDH certificates is now considered ECDH
8411 authentication, not RSA or ECDSA authentication (the latter is
8412 merely the CA's signing algorithm and not actively used in the
8415 The temporary ciphersuite alias "ECCdraft" is no longer
8416 available, and ECC ciphersuites are no longer excluded from "ALL"
8417 and "DEFAULT". The following aliases now exist for RFC 4492
8418 ciphersuites, most of these by analogy with the DH case:
8420 kECDHr - ECDH cert, signed with RSA
8421 kECDHe - ECDH cert, signed with ECDSA
8422 kECDH - ECDH cert (signed with either RSA or ECDSA)
8423 kEECDH - ephemeral ECDH
8424 ECDH - ECDH cert or ephemeral ECDH
8430 AECDH - anonymous ECDH
8431 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8435 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8436 Use correct micalg parameters depending on digest(s) in signed message.
8440 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8441 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8445 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8446 an engine to register a method. Add ENGINE lookups for methods and
8447 functional reference processing.
8451 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8452 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8457 * New -resign option to smime utility. This adds one or more signers
8458 to an existing PKCS#7 signedData structure. Also -md option to use an
8459 alternative message digest algorithm for signing.
8463 * Tidy up PKCS#7 routines and add new functions to make it easier to
8464 create PKCS7 structures containing multiple signers. Update smime
8465 application to support multiple signers.
8469 * New -macalg option to pkcs12 utility to allow setting of an alternative
8474 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8475 Reorganize PBE internals to lookup from a static table using NIDs,
8476 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8477 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8478 PRF which will be automatically used with PBES2.
8482 * Replace the algorithm specific calls to generate keys in "req" with the
8487 * Update PKCS#7 enveloped data routines to use new API. This is now
8488 supported by any public key method supporting the encrypt operation. A
8489 ctrl is added to allow the public key algorithm to examine or modify
8490 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8495 * Add a ctrl to asn1 method to allow a public key algorithm to express
8496 a default digest type to use. In most cases this will be SHA1 but some
8497 algorithms (such as GOST) need to specify an alternative digest. The
8498 return value indicates how strong the preference is 1 means optional and
8499 2 is mandatory (that is it is the only supported type). Modify
8500 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8501 use the default md. Update openssl utilities to use the default digest
8502 type for signing if it is not explicitly indicated.
8506 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8507 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8508 signing method from the key type. This effectively removes the link
8509 between digests and public key types.
8513 * Add an OID cross reference table and utility functions. Its purpose is to
8514 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8515 rsaEncryption. This will allow some of the algorithm specific hackery
8516 needed to use the correct OID to be removed.
8520 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8521 structures for PKCS7_sign(). They are now set up by the relevant public
8526 * Add provisional EC pkey method with support for ECDSA and ECDH.
8530 * Add support for key derivation (agreement) in the API, DH method and
8535 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8536 public and private key formats. As a side effect these add additional
8537 command line functionality not previously available: DSA signatures can be
8538 generated and verified using pkeyutl and DH key support and generation in
8545 *Oliver Tappe <zooey@hirschkaefer.de>*
8547 * New make target "install_html_docs" installs HTML renditions of the
8550 *Oliver Tappe <zooey@hirschkaefer.de>*
8552 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8553 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8554 support key and parameter generation and add initial key generation
8555 functionality for RSA.
8559 * Add functions for main EVP_PKEY_method operations. The undocumented
8560 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8561 `EVP_PKEY_{encrypt,decrypt}_old`.
8565 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8566 key API, doesn't do much yet.
8570 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8571 public key algorithms. New option to openssl utility:
8572 "list-public-key-algorithms" to print out info.
8576 * Implement the Supported Elliptic Curves Extension for
8577 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8581 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8582 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8586 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8587 utilities such as rsa, dsa, dsaparam etc except they process any key
8592 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8593 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8594 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8599 * Initial support for pluggable public key ASN1.
8600 De-spaghettify the public key ASN1 handling. Move public and private
8601 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8602 algorithm specific handling to a single module within the relevant
8603 algorithm directory. Add functions to allow (near) opaque processing
8604 of public and private key structures.
8608 * Implement the Supported Point Formats Extension for
8609 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8613 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8614 for the psk identity [hint] and the psk callback functions to the
8615 SSL_SESSION, SSL and SSL_CTX structure.
8618 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8622 SSL_CTX_use_psk_identity_hint
8623 SSL_get_psk_identity_hint
8624 SSL_get_psk_identity
8625 SSL_use_psk_identity_hint
8627 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8629 * Add RFC 3161 compliant time stamp request creation, response generation
8630 and response verification functionality.
8632 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8634 * Add initial support for TLS extensions, specifically for the server_name
8635 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8636 have new members for a host name. The SSL data structure has an
8637 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8638 stored in that context to allow for session resumption, even after the
8639 SSL has been switched to a new SSL_CTX in reaction to a client's
8640 server_name extension.
8642 New functions (subject to change):
8644 SSL_get_servername()
8645 SSL_get_servername_type()
8648 New CTRL codes and macros (subject to change):
8650 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8651 - SSL_CTX_set_tlsext_servername_callback()
8652 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8653 - SSL_CTX_set_tlsext_servername_arg()
8654 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8656 openssl s_client has a new '-servername ...' option.
8658 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8659 '-key2 ...', '-servername_fatal' (subject to change). This allows
8660 testing the HostName extension for a specific single host name ('-cert'
8661 and '-key' remain fallbacks for handshakes without HostName
8662 negotiation). If the unrecognized_name alert has to be sent, this by
8663 default is a warning; it becomes fatal with the '-servername_fatal'
8666 *Peter Sylvester, Remy Allais, Christophe Renou*
8668 * Whirlpool hash implementation is added.
8672 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8673 bn(64,32). Because of instruction set limitations it doesn't have
8674 any negative impact on performance. This was done mostly in order
8675 to make it possible to share assembler modules, such as bn_mul_mont
8676 implementations, between 32- and 64-bit builds without hassle.
8680 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8681 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8686 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8687 dedicated Montgomery multiplication procedure, is introduced.
8688 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8689 "64-bit" performance on certain 32-bit targets.
8693 * New option SSL_OP_NO_COMP to disable use of compression selectively
8694 in SSL structures. New SSL ctrl to set maximum send fragment size.
8695 Save memory by setting the I/O buffer sizes dynamically instead of
8696 using the maximum available value.
8700 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8701 in addition to the text details.
8705 * Very, very preliminary EXPERIMENTAL support for printing of general
8706 ASN1 structures. This currently produces rather ugly output and doesn't
8707 handle several customised structures at all.
8711 * Integrated support for PVK file format and some related formats such
8712 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8713 these in the 'rsa' and 'dsa' utilities.
8717 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8721 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8722 place for the (very old) "NETSCAPE" format certificates which are now
8723 handled using new ASN1 code equivalents.
8727 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8728 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8729 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8733 * Modify CRL distribution points extension code to print out previously
8734 unsupported fields. Enhance extension setting code to allow setting of
8739 * Add print and set support for Issuing Distribution Point CRL extension.
8743 * Change 'Configure' script to enable Camellia by default.
8750 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8752 * When rejecting SSL/TLS records due to an incorrect version number, never
8753 update s->server with a new major version number. As of
8754 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8755 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8756 the previous behavior could result in a read attempt at NULL when
8757 receiving specific incorrect SSL/TLS records once record payload
8758 protection is active. ([CVE-2010-0740])
8760 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8762 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8763 could be crashed if the relevant tables were not present (e.g. chrooted).
8765 *Tomas Hoger <thoger@redhat.com>*
8767 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8769 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8771 *Martin Olsson, Neel Mehta*
8773 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8774 accommodate for stack sorting, always a write lock!).
8778 * On some versions of WIN32 Heap32Next is very slow. This can cause
8779 excessive delays in the RAND_poll(): over a minute. As a workaround
8780 include a time check in the inner Heap32Next loop too.
8784 * The code that handled flushing of data in SSL/TLS originally used the
8785 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8786 the problem outlined in PR#1949. The fix suggested there however can
8787 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8788 of Apache). So instead simplify the code to flush unconditionally.
8789 This should be fine since flushing with no data to flush is a no op.
8793 * Handle TLS versions 2.0 and later properly and correctly use the
8794 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8795 off ancient servers have a habit of sticking around for a while...
8799 * Modify compression code so it frees up structures without using the
8800 ex_data callbacks. This works around a problem where some applications
8801 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8802 restarting) then use compression (e.g. SSL with compression) later.
8803 This results in significant per-connection memory leaks and
8804 has caused some security issues including CVE-2008-1678 and
8809 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8810 change when encrypting or decrypting.
8814 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8815 connect and renegotiate with servers which do not support RI.
8816 Until RI is more widely deployed this option is enabled by default.
8820 * Add "missing" ssl ctrls to clear options and mode.
8824 * If client attempts to renegotiate and doesn't support RI respond with
8825 a no_renegotiation alert as required by RFC5746. Some renegotiating
8826 TLS clients will continue a connection gracefully when they receive
8827 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8828 waiting for a server hello which it will never receive. Now we treat a
8829 received no_renegotiation alert as a fatal error. This is because
8830 applications requesting a renegotiation might well expect it to succeed
8831 and would have no code in place to handle the server denying it so the
8832 only safe thing to do is to terminate the connection.
8836 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8837 peer supports secure renegotiation and 0 otherwise. Print out peer
8838 renegotiation support in s_client/s_server.
8842 * Replace the highly broken and deprecated SPKAC certification method with
8843 the updated NID creation version. This should correctly handle UTF8.
8847 * Implement RFC5746. Re-enable renegotiation but require the extension
8848 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8849 turns out to be a bad idea. It has been replaced by
8850 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8851 SSL_CTX_set_options(). This is really not recommended unless you
8852 know what you are doing.
8854 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8856 * Fixes to stateless session resumption handling. Use initial_ctx when
8857 issuing and attempting to decrypt tickets in case it has changed during
8858 servername handling. Use a non-zero length session ID when attempting
8859 stateless session resumption: this makes it possible to determine if
8860 a resumption has occurred immediately after receiving server hello
8861 (several places in OpenSSL subtly assume this) instead of later in
8866 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8867 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8868 fixes for a few places where the return code is not checked
8871 *Julia Lawall <julia@diku.dk>*
8873 * Add --strict-warnings option to Configure script to include devteam
8874 warnings in other configurations.
8878 * Add support for --libdir option and LIBDIR variable in makefiles. This
8879 makes it possible to install openssl libraries in locations which
8880 have names other than "lib", for example "/usr/lib64" which some
8883 *Steve Henson, based on patch from Jeremy Utley*
8885 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8886 X690 8.9.12 and can produce some misleading textual output of OIDs.
8888 *Steve Henson, reported by Dan Kaminsky*
8890 * Delete MD2 from algorithm tables. This follows the recommendation in
8891 several standards that it is not used in new applications due to
8892 several cryptographic weaknesses. For binary compatibility reasons
8893 the MD2 API is still compiled in by default.
8897 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
8902 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
8903 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
8906 *Guenter <lists@gknw.net>*
8908 * Fix the server certificate chain building code to use X509_verify_cert(),
8909 it used to have an ad-hoc builder which was unable to cope with anything
8910 other than a simple chain.
8912 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
8914 * Don't check self signed certificate signatures in X509_verify_cert()
8915 by default (a flag can override this): it just wastes time without
8916 adding any security. As a useful side effect self signed root CAs
8917 with non-FIPS digests are now usable in FIPS mode.
8921 * In dtls1_process_out_of_seq_message() the check if the current message
8922 is already buffered was missing. For every new message was memory
8923 allocated, allowing an attacker to perform an denial of service attack
8924 with sending out of seq handshake messages until there is no memory
8925 left. Additionally every future message was buffered, even if the
8926 sequence number made no sense and would be part of another handshake.
8927 So only messages with sequence numbers less than 10 in advance will be
8928 buffered. ([CVE-2009-1378])
8930 *Robin Seggelmann, discovered by Daniel Mentz*
8932 * Records are buffered if they arrive with a future epoch to be
8933 processed after finishing the corresponding handshake. There is
8934 currently no limitation to this buffer allowing an attacker to perform
8935 a DOS attack with sending records with future epochs until there is no
8936 memory left. This patch adds the pqueue_size() function to determine
8937 the size of a buffer and limits the record buffer to 100 entries.
8940 *Robin Seggelmann, discovered by Daniel Mentz*
8942 * Keep a copy of frag->msg_header.frag_len so it can be used after the
8943 parent structure is freed. ([CVE-2009-1379])
8947 * Handle non-blocking I/O properly in SSL_shutdown() call.
8949 *Darryl Miles <darryl-mailinglists@netbauds.net>*
8951 * Add `2.5.4.*` OIDs
8953 *Ilya O. <vrghost@gmail.com>*
8955 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
8957 * Disable renegotiation completely - this fixes a severe security
8958 problem ([CVE-2009-3555]) at the cost of breaking all
8959 renegotiation. Renegotiation can be re-enabled by setting
8960 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
8961 run-time. This is really not recommended unless you know what
8966 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
8968 * Don't set val to NULL when freeing up structures, it is freed up by
8969 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
8970 zeroing past the valid field. ([CVE-2009-0789])
8972 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
8974 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
8975 checked correctly. This would allow some invalid signed attributes to
8976 appear to verify correctly. ([CVE-2009-0591])
8978 *Ivan Nestlerode <inestlerode@us.ibm.com>*
8980 * Reject UniversalString and BMPString types with invalid lengths. This
8981 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
8982 a legal length. ([CVE-2009-0590])
8986 * Set S/MIME signing as the default purpose rather than setting it
8987 unconditionally. This allows applications to override it at the store
8992 * Permit restricted recursion of ASN1 strings. This is needed in practice
8993 to handle some structures.
8997 * Improve efficiency of mem_gets: don't search whole buffer each time
9000 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9002 * New -hex option for openssl rand.
9006 * Print out UTF8String and NumericString when parsing ASN1.
9010 * Support NumericString type for name components.
9014 * Allow CC in the environment to override the automatically chosen
9015 compiler. Note that nothing is done to ensure flags work with the
9020 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9022 * Properly check EVP_VerifyFinal() and similar return values
9025 *Ben Laurie, Bodo Moeller, Google Security Team*
9027 * Enable TLS extensions by default.
9031 * Allow the CHIL engine to be loaded, whether the application is
9032 multithreaded or not. (This does not release the developer from the
9033 obligation to set up the dynamic locking callbacks.)
9035 *Sander Temme <sander@temme.net>*
9037 * Use correct exit code if there is an error in dgst command.
9039 *Steve Henson; problem pointed out by Roland Dirlewanger*
9041 * Tweak Configure so that you need to say "experimental-jpake" to enable
9042 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9046 * Add experimental JPAKE support, including demo authentication in
9047 s_client and s_server.
9051 * Set the comparison function in v3_addr_canonize().
9053 *Rob Austein <sra@hactrn.net>*
9055 * Add support for XMPP STARTTLS in s_client.
9057 *Philip Paeps <philip@freebsd.org>*
9059 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9060 to ensure that even with this option, only ciphersuites in the
9061 server's preference list will be accepted. (Note that the option
9062 applies only when resuming a session, so the earlier behavior was
9063 just about the algorithm choice for symmetric cryptography.)
9067 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9069 * Fix NULL pointer dereference if a DTLS server received
9070 ChangeCipherSpec as first record ([CVE-2009-1386]).
9074 * Fix a state transition in s3_srvr.c and d1_srvr.c
9075 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9079 * The fix in 0.9.8c that supposedly got rid of unsafe
9080 double-checked locking was incomplete for RSA blinding,
9081 addressing just one layer of what turns out to have been
9082 doubly unsafe triple-checked locking.
9084 So now fix this for real by retiring the MONT_HELPER macro
9085 in crypto/rsa/rsa_eay.c.
9087 *Bodo Moeller; problem pointed out by Marius Schilder*
9089 * Various precautionary measures:
9091 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9093 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9094 (NB: This would require knowledge of the secret session ticket key
9095 to exploit, in which case you'd be SOL either way.)
9097 - Change bn_nist.c so that it will properly handle input BIGNUMs
9098 outside the expected range.
9100 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9103 *Neel Mehta, Bodo Moeller*
9105 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9106 the load fails. Useful for distros.
9108 *Ben Laurie and the FreeBSD team*
9110 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9114 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9118 * Expand ENGINE to support engine supplied SSL client certificate functions.
9120 This work was sponsored by Logica.
9124 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9125 keystores. Support for SSL/TLS client authentication too.
9126 Not compiled unless enable-capieng specified to Configure.
9128 This work was sponsored by Logica.
9132 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9133 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9134 attribute creation routines such as certificate requests and PKCS#12
9139 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9141 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9142 handshake which could lead to a client crash as found using the
9143 Codenomicon TLS test suite ([CVE-2008-1672])
9145 *Steve Henson, Mark Cox*
9147 * Fix double free in TLS server name extensions which could lead to
9148 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9152 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9154 Clear the error queue to ensure that error entries left from
9155 older function calls do not interfere with the correct operation.
9157 *Lutz Jaenicke, Erik de Castro Lopo*
9159 * Remove root CA certificates of commercial CAs:
9161 The OpenSSL project does not recommend any specific CA and does not
9162 have any policy with respect to including or excluding any CA.
9163 Therefore it does not make any sense to ship an arbitrary selection
9164 of root CA certificates with the OpenSSL software.
9168 * RSA OAEP patches to fix two separate invalid memory reads.
9169 The first one involves inputs when 'lzero' is greater than
9170 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9171 before the beginning of from). The second one involves inputs where
9172 the 'db' section contains nothing but zeroes (there is a one-byte
9173 invalid read after the end of 'db').
9175 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9177 * Partial backport from 0.9.9-dev:
9179 Introduce bn_mul_mont (dedicated Montgomery multiplication
9180 procedure) as a candidate for BIGNUM assembler implementation.
9181 While 0.9.9-dev uses assembler for various architectures, only
9182 x86_64 is available by default here in the 0.9.8 branch, and
9183 32-bit x86 is available through a compile-time setting.
9185 To try the 32-bit x86 assembler implementation, use Configure
9186 option "enable-montasm" (which exists only for this backport).
9188 As "enable-montasm" for 32-bit x86 disclaims code stability
9189 anyway, in this constellation we activate additional code
9190 backported from 0.9.9-dev for further performance improvements,
9191 namely BN_from_montgomery_word. (To enable this otherwise,
9192 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9194 *Andy Polyakov (backport partially by Bodo Moeller)*
9196 * Add TLS session ticket callback. This allows an application to set
9197 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9198 values. This is useful for key rollover for example where several key
9199 sets may exist with different names.
9203 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9204 This was broken until now in 0.9.8 releases, such that the only way
9205 a registered ENGINE could be used (assuming it initialises
9206 successfully on the host) was to explicitly set it as the default
9207 for the relevant algorithms. This is in contradiction with 0.9.7
9208 behaviour and the documentation. With this fix, when an ENGINE is
9209 registered into a given algorithm's table of implementations, the
9210 'uptodate' flag is reset so that auto-discovery will be used next
9211 time a new context for that algorithm attempts to select an
9214 *Ian Lister (tweaked by Geoff Thorpe)*
9216 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9217 implementation in the following ways:
9219 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9222 Lack of BER streaming support means one pass streaming processing is
9223 only supported if data is detached: setting the streaming flag is
9224 ignored for embedded content.
9226 CMS support is disabled by default and must be explicitly enabled
9227 with the enable-cms configuration option.
9231 * Update the GMP engine glue to do direct copies between BIGNUM and
9232 mpz_t when openssl and GMP use the same limb size. Otherwise the
9233 existing "conversion via a text string export" trick is still used.
9235 *Paul Sheer <paulsheer@gmail.com>*
9237 * Zlib compression BIO. This is a filter BIO which compressed and
9238 uncompresses any data passed through it.
9242 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9243 RFC3394 compatible AES key wrapping.
9247 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9248 sets string data without copying. X509_ALGOR_set0() and
9249 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9250 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9251 from an X509_ATTRIBUTE structure optionally checking it occurs only
9252 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9257 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9258 to get the expected BN_FLG_CONSTTIME behavior.
9260 *Bodo Moeller (Google)*
9264 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9265 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9266 - added some more tests to do_tests.pl
9267 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9268 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9269 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9270 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9271 - various changes to netware.pl to enable gcc-cross builds on Win32
9273 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9274 - various changes to fix missing prototype warnings
9275 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9276 - added AES, WHIRLPOOL and CPUID assembler code to build files
9277 - added missing AES assembler make rules to mk1mf.pl
9278 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9280 *Guenter Knauf <eflash@gmx.net>*
9282 * Implement certificate status request TLS extension defined in RFC3546.
9283 A client can set the appropriate parameters and receive the encoded
9284 OCSP response via a callback. A server can query the supplied parameters
9285 and set the encoded OCSP response in the callback. Add simplified examples
9286 to s_client and s_server.
9290 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9293 + Binary incompatibility of ssl_ctx_st structure
9294 + DTLS interoperation with non-compliant servers
9295 + Don't call get_session_cb() without proposed session
9296 + Fix ia64 assembler code
9298 *Andy Polyakov, Steve Henson*
9300 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9302 * DTLS Handshake overhaul. There were longstanding issues with
9303 OpenSSL DTLS implementation, which were making it impossible for
9304 RFC 4347 compliant client to communicate with OpenSSL server.
9305 Unfortunately just fixing these incompatibilities would "cut off"
9306 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9307 server keeps tolerating non RFC compliant syntax. The opposite is
9308 not true, 0.9.8f client can not communicate with earlier server.
9309 This update even addresses CVE-2007-4995.
9313 * Changes to avoid need for function casts in OpenSSL: some compilers
9314 (gcc 4.2 and later) reject their use.
9315 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9318 * Add RFC4507 support to OpenSSL. This includes the corrections in
9319 RFC4507bis. The encrypted ticket format is an encrypted encoded
9320 SSL_SESSION structure, that way new session features are automatically
9323 If a client application caches session in an SSL_SESSION structure
9324 support is transparent because tickets are now stored in the encoded
9327 The SSL_CTX structure automatically generates keys for ticket
9328 protection in servers so again support should be possible
9329 with no application modification.
9331 If a client or server wishes to disable RFC4507 support then the option
9332 SSL_OP_NO_TICKET can be set.
9334 Add a TLS extension debugging callback to allow the contents of any client
9335 or server extensions to be examined.
9337 This work was sponsored by Google.
9341 * Add initial support for TLS extensions, specifically for the server_name
9342 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9343 have new members for a host name. The SSL data structure has an
9344 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9345 stored in that context to allow for session resumption, even after the
9346 SSL has been switched to a new SSL_CTX in reaction to a client's
9347 server_name extension.
9349 New functions (subject to change):
9351 SSL_get_servername()
9352 SSL_get_servername_type()
9355 New CTRL codes and macros (subject to change):
9357 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9358 - SSL_CTX_set_tlsext_servername_callback()
9359 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9360 - SSL_CTX_set_tlsext_servername_arg()
9361 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9363 openssl s_client has a new '-servername ...' option.
9365 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9366 '-key2 ...', '-servername_fatal' (subject to change). This allows
9367 testing the HostName extension for a specific single host name ('-cert'
9368 and '-key' remain fallbacks for handshakes without HostName
9369 negotiation). If the unrecognized_name alert has to be sent, this by
9370 default is a warning; it becomes fatal with the '-servername_fatal'
9373 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9375 * Add AES and SSE2 assembly language support to VC++ build.
9379 * Mitigate attack on final subtraction in Montgomery reduction.
9383 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9384 (which previously caused an internal error).
9388 * Squeeze another 10% out of IGE mode when in != out.
9392 * AES IGE mode speedup.
9394 *Dean Gaudet (Google)*
9396 * Add the Korean symmetric 128-bit cipher SEED (see
9397 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9398 add SEED ciphersuites from RFC 4162:
9400 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9401 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9402 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9403 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9405 To minimize changes between patchlevels in the OpenSSL 0.9.8
9406 series, SEED remains excluded from compilation unless OpenSSL
9407 is configured with 'enable-seed'.
9409 *KISA, Bodo Moeller*
9411 * Mitigate branch prediction attacks, which can be practical if a
9412 single processor is shared, allowing a spy process to extract
9413 information. For detailed background information, see
9414 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9415 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9416 and Necessary Software Countermeasures"). The core of the change
9417 are new versions BN_div_no_branch() and
9418 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9419 respectively, which are slower, but avoid the security-relevant
9420 conditional branches. These are automatically called by BN_div()
9421 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9422 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9423 remove a conditional branch.
9425 BN_FLG_CONSTTIME is the new name for the previous
9426 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9427 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9428 in the exponent causes BN_mod_exp_mont() to use the alternative
9429 implementation in BN_mod_exp_mont_consttime().) The old name
9430 remains as a deprecated alias.
9432 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9433 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9434 constant-time implementations for more than just exponentiation.
9435 Here too the old name is kept as a deprecated alias.
9437 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9438 the BN_BLINDING structure gets an independent copy of the
9439 modulus. This means that the previous `BIGNUM *m` argument to
9440 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9441 essentially becomes `const BIGNUM *m`, although we can't actually
9442 change this in the header file before 0.9.9. It allows
9443 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9444 enable BN_FLG_CONSTTIME.
9446 *Matthew D Wood (Intel Corp)*
9448 * In the SSL/TLS server implementation, be strict about session ID
9449 context matching (which matters if an application uses a single
9450 external cache for different purposes). Previously,
9451 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9452 set. This did ensure strict client verification, but meant that,
9453 with applications using a single external cache for quite
9454 different requirements, clients could circumvent ciphersuite
9455 restrictions for a given session ID context by starting a session
9456 in a different context.
9460 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9461 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9462 authentication-only ciphersuites.
9466 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9467 not complete and could lead to a possible single byte overflow
9468 ([CVE-2007-5135]) [Ben Laurie]
9470 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9472 * Since AES128 and AES256 (and similarly Camellia128 and
9473 Camellia256) share a single mask bit in the logic of
9474 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9475 kludge to work properly if AES128 is available and AES256 isn't
9476 (or if Camellia128 is available and Camellia256 isn't).
9480 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9481 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9482 When a point or a seed is encoded in a BIT STRING, we need to
9483 prevent the removal of trailing zero bits to get the proper DER
9484 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9485 of a NamedBitList, for which trailing 0 bits need to be removed.)
9489 * Have SSL/TLS server implementation tolerate "mismatched" record
9490 protocol version while receiving ClientHello even if the
9491 ClientHello is fragmented. (The server can't insist on the
9492 particular protocol version it has chosen before the ServerHello
9493 message has informed the client about his choice.)
9497 * Add RFC 3779 support.
9499 *Rob Austein for ARIN, Ben Laurie*
9501 * Load error codes if they are not already present instead of using a
9502 static variable. This allows them to be cleanly unloaded and reloaded.
9503 Improve header file function name parsing.
9507 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9508 or CAPABILITY handshake as required by RFCs.
9512 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9514 * Introduce limits to prevent malicious keys being able to
9515 cause a denial of service. ([CVE-2006-2940])
9517 *Steve Henson, Bodo Moeller*
9519 * Fix ASN.1 parsing of certain invalid structures that can result
9520 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9522 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9523 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9525 * Fix SSL client code which could crash if connecting to a
9526 malicious SSLv2 server. ([CVE-2006-4343])
9528 *Tavis Ormandy and Will Drewry, Google Security Team*
9530 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9531 match only those. Before that, "AES256-SHA" would be interpreted
9532 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9533 the same strength classification in 0.9.7h) as we currently only
9534 have a single AES bit in the ciphersuite description bitmap.
9535 That change, however, also applied to ciphersuite strings such as
9536 "RC4-MD5" that intentionally matched multiple ciphersuites --
9537 namely, SSL 2.0 ciphersuites in addition to the more common ones
9538 from SSL 3.0/TLS 1.0.
9540 So we change the selection algorithm again: Naming an explicit
9541 ciphersuite selects this one ciphersuite, and any other similar
9542 ciphersuite (same bitmap) from *other* protocol versions.
9543 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9544 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9546 Since SSL 2.0 does not have any ciphersuites for which the
9547 128/256 bit distinction would be relevant, this works for now.
9548 The proper fix will be to use different bits for AES128 and
9549 AES256, which would have avoided the problems from the beginning;
9550 however, bits are scarce, so we can only do this in a new release
9551 (not just a patchlevel) when we can change the SSL_CIPHER
9552 definition to split the single 'unsigned long mask' bitmap into
9553 multiple values to extend the available space.
9557 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9559 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9560 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9562 * Add AES IGE and biIGE modes.
9566 * Change the Unix randomness entropy gathering to use poll() when
9567 possible instead of select(), since the latter has some
9568 undesirable limitations.
9570 *Darryl Miles via Richard Levitte and Bodo Moeller*
9572 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9573 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9574 cannot be implicitly activated as part of, e.g., the "AES" alias.
9575 However, please upgrade to OpenSSL 0.9.9[-dev] for
9576 non-experimental use of the ECC ciphersuites to get TLS extension
9577 support, which is required for curve and point format negotiation
9578 to avoid potential handshake problems.
9582 * Disable rogue ciphersuites:
9584 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9585 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9586 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9588 The latter two were purportedly from
9589 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9592 Also deactivate the remaining ciphersuites from
9593 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9594 unofficial, and the ID has long expired.
9598 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9599 dual-core machines) and other potential thread-safety issues.
9603 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9604 versions), which is now available for royalty-free use
9605 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9606 Also, add Camellia TLS ciphersuites from RFC 4132.
9608 To minimize changes between patchlevels in the OpenSSL 0.9.8
9609 series, Camellia remains excluded from compilation unless OpenSSL
9610 is configured with 'enable-camellia'.
9614 * Disable the padding bug check when compression is in use. The padding
9615 bug check assumes the first packet is of even length, this is not
9616 necessarily true if compression is enabled and can result in false
9617 positives causing handshake failure. The actual bug test is ancient
9618 code so it is hoped that implementations will either have fixed it by
9619 now or any which still have the bug do not support compression.
9623 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9625 * When applying a cipher rule check to see if string match is an explicit
9626 cipher suite and only match that one cipher suite if it is.
9630 * Link in manifests for VC++ if needed.
9632 *Austin Ziegler <halostatue@gmail.com>*
9634 * Update support for ECC-based TLS ciphersuites according to
9635 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9636 TLS extensions, which are supported starting with the 0.9.9
9637 branch, not in the OpenSSL 0.9.8 branch).
9641 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9642 opaque EVP_CIPHER_CTX handling.
9646 * Fixes and enhancements to zlib compression code. We now only use
9647 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9648 to conform with the standards mentioned here:
9649 <http://www.zlib.net/DLL_FAQ.txt>
9650 Static zlib linking now works on Windows and the new --with-zlib-include
9651 --with-zlib-lib options to Configure can be used to supply the location
9652 of the headers and library. Gracefully handle case where zlib library
9657 * Several fixes and enhancements to the OID generation code. The old code
9658 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9659 handle numbers larger than ULONG_MAX, truncated printing and had a
9660 non standard OBJ_obj2txt() behaviour.
9664 * Add support for building of engines under engine/ as shared libraries
9665 under VC++ build system.
9669 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9670 Hopefully, we will not see any false combination of paths any more.
9674 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9676 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9677 (part of SSL_OP_ALL). This option used to disable the
9678 countermeasure against man-in-the-middle protocol-version
9679 rollback in the SSL 2.0 server implementation, which is a bad
9680 idea. ([CVE-2005-2969])
9682 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9683 for Information Security, National Institute of Advanced Industrial
9684 Science and Technology [AIST], Japan)*
9686 * Add two function to clear and return the verify parameter flags.
9690 * Keep cipherlists sorted in the source instead of sorting them at
9691 runtime, thus removing the need for a lock.
9695 * Avoid some small subgroup attacks in Diffie-Hellman.
9697 *Nick Mathewson and Ben Laurie*
9699 * Add functions for well-known primes.
9703 * Extended Windows CE support.
9705 *Satoshi Nakamura and Andy Polyakov*
9707 * Initialize SSL_METHOD structures at compile time instead of during
9708 runtime, thus removing the need for a lock.
9712 * Make PKCS7_decrypt() work even if no certificate is supplied by
9713 attempting to decrypt each encrypted key in turn. Add support to
9718 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9720 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9723 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9727 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9728 key into the same file any more.
9732 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9736 * Add -utf8 command line and config file option to 'ca'.
9738 *Stefan <stf@udoma.org*
9740 * Removed the macro des_crypt(), as it seems to conflict with some
9741 libraries. Use DES_crypt().
9745 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9746 involves renaming the source and generated shared-libs for
9747 both. The engines will accept the corrected or legacy ids
9748 ('ncipher' and '4758_cca' respectively) when binding. NB,
9749 this only applies when building 'shared'.
9751 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9753 * Add attribute functions to EVP_PKEY structure. Modify
9754 PKCS12_create() to recognize a CSP name attribute and
9755 use it. Make -CSP option work again in pkcs12 utility.
9759 * Add new functionality to the bn blinding code:
9760 - automatic re-creation of the BN_BLINDING parameters after
9761 a fixed number of uses (currently 32)
9762 - add new function for parameter creation
9763 - introduce flags to control the update behaviour of the
9764 BN_BLINDING parameters
9765 - hide BN_BLINDING structure
9766 Add a second BN_BLINDING slot to the RSA structure to improve
9767 performance when a single RSA object is shared among several
9772 * Add support for DTLS.
9774 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9776 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9777 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9781 * Remove buggy and incomplete DH cert support from
9782 ssl/ssl_rsa.c and ssl/s3_both.c
9786 * Use SHA-1 instead of MD5 as the default digest algorithm for
9787 the apps/openssl applications.
9791 * Compile clean with "-Wall -Wmissing-prototypes
9792 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9793 DEBUG_SAFESTACK must also be set.
9797 * Change ./Configure so that certain algorithms can be disabled by default.
9798 The new counterpiece to "no-xxx" is "enable-xxx".
9800 The patented RC5 and MDC2 algorithms will now be disabled unless
9801 "enable-rc5" and "enable-mdc2", respectively, are specified.
9803 (IDEA remains enabled despite being patented. This is because IDEA
9804 is frequently required for interoperability, and there is no license
9805 fee for non-commercial use. As before, "no-idea" can be used to
9806 avoid this algorithm.)
9810 * Add processing of proxy certificates (see RFC 3820). This work was
9811 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9812 EGEE (Enabling Grids for E-science in Europe).
9816 * RC4 performance overhaul on modern architectures/implementations, such
9817 as Intel P4, IA-64 and AMD64.
9821 * New utility extract-section.pl. This can be used specify an alternative
9822 section number in a pod file instead of having to treat each file as
9823 a separate case in Makefile. This can be done by adding two lines to the
9826 =for comment openssl_section:XXX
9828 The blank line is mandatory.
9832 * New arguments -certform, -keyform and -pass for s_client and s_server
9833 to allow alternative format key and certificate files and passphrase
9838 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9839 update associated structures and add various utility functions.
9841 Add new policy related verify parameters, include policy checking in
9842 standard verify code. Enhance 'smime' application with extra parameters
9843 to support policy checking and print out.
9847 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9848 Nehemiah processors. These extensions support AES encryption in hardware
9849 as well as RNG (though RNG support is currently disabled).
9851 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9853 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9857 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9859 *Andy Polyakov and a number of other people*
9861 * Improved PowerPC platform support. Most notably BIGNUM assembler
9862 implementation contributed by IBM.
9864 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9866 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9867 exponent rather than 'unsigned long'. There is a corresponding change to
9868 the new 'rsa_keygen' element of the RSA_METHOD structure.
9870 *Jelte Jansen, Geoff Thorpe*
9872 * Functionality for creating the initial serial number file is now
9873 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9875 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9876 number file to 1, which is bound to cause problems. To avoid
9877 the problems while respecting compatibility between different 0.9.7
9878 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9879 CA.pl for serial number initialization. With the new release 0.9.8,
9880 we can fix the problem directly in the 'ca' utility.)
9884 * Reduced header interdependencies by declaring more opaque objects in
9885 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9886 give fewer recursive includes, which could break lazy source code - so
9887 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
9888 developers should define this symbol when building and using openssl to
9889 ensure they track the recommended behaviour, interfaces, [etc], but
9890 backwards-compatible behaviour prevails when this isn't defined.
9894 * New function X509_POLICY_NODE_print() which prints out policy nodes.
9898 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
9899 This will generate a random key of the appropriate length based on the
9900 cipher context. The EVP_CIPHER can provide its own random key generation
9901 routine to support keys of a specific form. This is used in the des and
9902 3des routines to generate a key of the correct parity. Update S/MIME
9903 code to use new functions and hence generate correct parity DES keys.
9904 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
9905 valid (weak or incorrect parity).
9909 * Add a local set of CRLs that can be used by X509_verify_cert() as well
9910 as looking them up. This is useful when the verified structure may contain
9911 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
9912 present unless the new PKCS7_NO_CRL flag is asserted.
9916 * Extend ASN1 oid configuration module. It now additionally accepts the
9919 shortName = some long name, 1.2.3.4
9923 * Reimplemented the BN_CTX implementation. There is now no more static
9924 limitation on the number of variables it can handle nor the depth of the
9925 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
9926 information can now expand as required, and rather than having a single
9927 static array of bignums, BN_CTX now uses a linked-list of such arrays
9928 allowing it to expand on demand whilst maintaining the usefulness of
9929 BN_CTX's "bundling".
9933 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
9934 to allow all RSA operations to function using a single BN_CTX.
9938 * Preliminary support for certificate policy evaluation and checking. This
9939 is initially intended to pass the tests outlined in "Conformance Testing
9940 of Relying Party Client Certificate Path Processing Logic" v1.07.
9944 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
9945 remained unused and not that useful. A variety of other little bignum
9946 tweaks and fixes have also been made continuing on from the audit (see
9951 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
9952 associated ASN1, EVP and SSL functions and old ASN1 macros.
9956 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
9957 and this should never fail. So the return value from the use of
9958 BN_set_word() (which can fail due to needless expansion) is now deprecated;
9959 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
9963 * BN_CTX_get() should return zero-valued bignums, providing the same
9964 initialised value as BN_new().
9966 *Geoff Thorpe, suggested by Ulf Möller*
9968 * Support for inhibitAnyPolicy certificate extension.
9972 * An audit of the BIGNUM code is underway, for which debugging code is
9973 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
9974 is considered valid when processing BIGNUMs, and causes execution to
9975 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
9976 further steps are taken to deliberately pollute unused data in BIGNUM
9977 structures to try and expose faulty code further on. For now, openssl will
9978 (in its default mode of operation) continue to tolerate the inconsistent
9979 forms that it has tolerated in the past, but authors and packagers should
9980 consider trying openssl and their own applications when compiled with
9981 these debugging symbols defined. It will help highlight potential bugs in
9982 their own code, and will improve the test coverage for OpenSSL itself. At
9983 some point, these tighter rules will become openssl's default to improve
9984 maintainability, though the assert()s and other overheads will remain only
9985 in debugging configurations. See bn.h for more details.
9987 *Geoff Thorpe, Nils Larsch, Ulf Möller*
9989 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
9990 that can only be obtained through BN_CTX_new() (which implicitly
9991 initialises it). The presence of this function only made it possible
9992 to overwrite an existing structure (and cause memory leaks).
9996 * Because of the callback-based approach for implementing LHASH as a
9997 template type, lh_insert() adds opaque objects to hash-tables and
9998 lh_doall() or lh_doall_arg() are typically used with a destructor callback
9999 to clean up those corresponding objects before destroying the hash table
10000 (and losing the object pointers). So some over-zealous constifications in
10001 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10002 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10003 prototyped to have "const" restrictions on the object pointers they are
10004 given (and so aren't required to cast them away any more).
10008 * The tmdiff.h API was so ugly and minimal that our own timing utility
10009 (speed) prefers to use its own implementation. The two implementations
10010 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10011 its object type properly exposed (MS_TM) instead of casting to/from
10012 `char *`. This may still change yet if someone realises MS_TM and
10014 aren't necessarily the greatest nomenclatures - but this is what was used
10015 internally to the implementation so I've used that for now.
10019 * Ensure that deprecated functions do not get compiled when
10020 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10021 the self-tests were still using deprecated key-generation functions so
10022 these have been updated also.
10026 * Reorganise PKCS#7 code to separate the digest location functionality
10027 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10028 New function PKCS7_set_digest() to set the digest type for PKCS#7
10029 digestedData type. Add additional code to correctly generate the
10030 digestedData type and add support for this type in PKCS7 initialization
10035 * New function PKCS7_set0_type_other() this initializes a PKCS7
10036 structure of type "other".
10040 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10041 sure the loop does correctly stop and breaking ("division by zero")
10042 modulus operations are not performed. The (pre-generated) prime
10043 table crypto/bn/bn_prime.h was already correct, but it could not be
10044 re-generated on some platforms because of the "division by zero"
10045 situation in the script.
10047 *Ralf S. Engelschall*
10049 * Update support for ECC-based TLS ciphersuites according to
10050 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10051 SHA-1 now is only used for "small" curves (where the
10052 representation of a field element takes up to 24 bytes); for
10053 larger curves, the field element resulting from ECDH is directly
10054 used as premaster secret.
10056 *Douglas Stebila (Sun Microsystems Laboratories)*
10058 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10059 curve secp160r1 to the tests.
10061 *Douglas Stebila (Sun Microsystems Laboratories)*
10063 * Add the possibility to load symbols globally with DSO.
10065 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10067 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10068 control of the error stack.
10072 * Add support for STORE in ENGINE.
10076 * Add the STORE type. The intention is to provide a common interface
10077 to certificate and key stores, be they simple file-based stores, or
10078 HSM-type store, or LDAP stores, or...
10079 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10083 * Add a generic structure called OPENSSL_ITEM. This can be used to
10084 pass a list of arguments to any function as well as provide a way
10085 for a function to pass data back to the caller.
10089 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10090 works like BUF_strdup() but can be used to duplicate a portion of
10091 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10096 * Add the function sk_find_ex() which works like sk_find(), but will
10097 return an index to an element even if an exact match couldn't be
10098 found. The index is guaranteed to point at the element where the
10099 searched-for key would be inserted to preserve sorting order.
10103 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10104 takes an extra flags argument for optional functionality. Currently,
10105 the following flags are defined:
10107 OBJ_BSEARCH_VALUE_ON_NOMATCH
10108 This one gets OBJ_bsearch_ex() to return a pointer to the first
10109 element where the comparing function returns a negative or zero
10112 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10113 This one gets OBJ_bsearch_ex() to return a pointer to the first
10114 element where the comparing function returns zero. This is useful
10115 if there are more than one element where the comparing function
10120 * Make it possible to create self-signed certificates with 'openssl ca'
10121 in such a way that the self-signed certificate becomes part of the
10122 CA database and uses the same mechanisms for serial number generation
10123 as all other certificate signing. The new flag '-selfsign' enables
10124 this functionality. Adapt CA.sh and CA.pl.in.
10128 * Add functionality to check the public key of a certificate request
10129 against a given private. This is useful to check that a certificate
10130 request can be signed by that key (self-signing).
10134 * Make it possible to have multiple active certificates with the same
10135 subject in the CA index file. This is done only if the keyword
10136 'unique_subject' is set to 'no' in the main CA section (default
10137 if 'CA_default') of the configuration file. The value is saved
10138 with the database itself in a separate index attribute file,
10139 named like the index file with '.attr' appended to the name.
10143 * Generate multi-valued AVAs using '+' notation in config files for
10148 * Support for nameConstraints certificate extension.
10152 * Support for policyConstraints certificate extension.
10156 * Support for policyMappings certificate extension.
10160 * Make sure the default DSA_METHOD implementation only uses its
10161 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10162 and change its own handlers to be NULL so as to remove unnecessary
10163 indirection. This lets alternative implementations fallback to the
10164 default implementation more easily.
10168 * Support for directoryName in GeneralName related extensions
10173 * Make it possible to link applications using Makefile.shared.
10174 Make that possible even when linking against static libraries!
10178 * Support for single pass processing for S/MIME signing. This now
10179 means that S/MIME signing can be done from a pipe, in addition
10180 cleartext signing (multipart/signed type) is effectively streaming
10181 and the signed data does not need to be all held in memory.
10183 This is done with a new flag PKCS7_STREAM. When this flag is set
10184 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10185 is done after the data is output (and digests calculated) in
10186 SMIME_write_PKCS7().
10190 * Add full support for -rpath/-R, both in shared libraries and
10191 applications, at least on the platforms where it's known how
10196 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10197 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10198 will now compute a table of multiples of the generator that
10199 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10200 faster (notably in the case of a single point multiplication,
10201 scalar * generator).
10203 *Nils Larsch, Bodo Moeller*
10205 * IPv6 support for certificate extensions. The various extensions
10206 which use the IP:a.b.c.d can now take IPv6 addresses using the
10207 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10212 * Added an ENGINE that implements RSA by performing private key
10213 exponentiations with the GMP library. The conversions to and from
10214 GMP's mpz_t format aren't optimised nor are any montgomery forms
10215 cached, and on x86 it appears OpenSSL's own performance has caught up.
10216 However there are likely to be other architectures where GMP could
10217 provide a boost. This ENGINE is not built in by default, but it can be
10218 specified at Configure time and should be accompanied by the necessary
10219 linker additions, eg;
10220 ./config -DOPENSSL_USE_GMP -lgmp
10224 * "openssl engine" will not display ENGINE/DSO load failure errors when
10225 testing availability of engines with "-t" - the old behaviour is
10226 produced by increasing the feature's verbosity with "-tt".
10230 * ECDSA routines: under certain error conditions uninitialized BN objects
10231 could be freed. Solution: make sure initialization is performed early
10232 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10237 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10238 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10239 software implementations. For DSA and DH, parameter generation can
10240 also be overridden by providing the appropriate method callbacks.
10244 * Change the "progress" mechanism used in key-generation and
10245 primality testing to functions that take a new BN_GENCB pointer in
10246 place of callback/argument pairs. The new API functions have `_ex`
10247 postfixes and the older functions are reimplemented as wrappers for
10248 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10249 declarations of the old functions to help (graceful) attempts to
10250 migrate to the new functions. Also, the new key-generation API
10251 functions operate on a caller-supplied key-structure and return
10252 success/failure rather than returning a key or NULL - this is to
10253 help make "keygen" another member function of RSA_METHOD etc.
10255 Example for using the new callback interface:
10257 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10258 void *my_arg = ...;
10261 BN_GENCB_set(&my_cb, my_callback, my_arg);
10263 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10264 /* For the meaning of a, b in calls to my_callback(), see the
10265 * documentation of the function that calls the callback.
10266 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10267 * my_callback should return 1 if it wants BN_is_prime_ex()
10268 * to continue, or 0 to stop.
10273 * Change the ZLIB compression method to be stateful, and make it
10274 available to TLS with the number defined in
10275 draft-ietf-tls-compression-04.txt.
10279 * Add the ASN.1 structures and functions for CertificatePair, which
10280 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10282 CertificatePair ::= SEQUENCE {
10283 forward [0] Certificate OPTIONAL,
10284 reverse [1] Certificate OPTIONAL,
10285 -- at least one of the pair shall be present -- }
10287 Also implement the PEM functions to read and write certificate
10288 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10290 This needed to be defined, mostly for the sake of the LDAP
10291 attribute crossCertificatePair, but may prove useful elsewhere as
10296 * Make it possible to inhibit symlinking of shared libraries in
10297 Makefile.shared, for Cygwin's sake.
10301 * Extend the BIGNUM API by creating a function
10302 void BN_set_negative(BIGNUM *a, int neg);
10303 and a macro that behave like
10304 int BN_is_negative(const BIGNUM *a);
10306 to avoid the need to access 'a->neg' directly in applications.
10310 * Implement fast modular reduction for pseudo-Mersenne primes
10311 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10312 EC_GROUP_new_curve_GFp() will now automatically use this
10315 *Nils Larsch <nla@trustcenter.de>*
10317 * Add new lock type (CRYPTO_LOCK_BN).
10321 * Change the ENGINE framework to automatically load engines
10322 dynamically from specific directories unless they could be
10323 found to already be built in or loaded. Move all the
10324 current engines except for the cryptodev one to a new
10325 directory engines/.
10326 The engines in engines/ are built as shared libraries if
10327 the "shared" options was given to ./Configure or ./config.
10328 Otherwise, they are inserted in libcrypto.a.
10329 /usr/local/ssl/engines is the default directory for dynamic
10330 engines, but that can be overridden at configure time through
10331 the usual use of --prefix and/or --openssldir, and at run
10332 time with the environment variable OPENSSL_ENGINES.
10334 *Geoff Thorpe and Richard Levitte*
10336 * Add Makefile.shared, a helper makefile to build shared
10337 libraries. Adapt Makefile.org.
10341 * Add version info to Win32 DLLs.
10343 *Peter 'Luna' Runestig" <peter@runestig.com>*
10345 * Add new 'medium level' PKCS#12 API. Certificates and keys
10346 can be added using this API to created arbitrary PKCS#12
10347 files while avoiding the low level API.
10349 New options to PKCS12_create(), key or cert can be NULL and
10350 will then be omitted from the output file. The encryption
10351 algorithm NIDs can be set to -1 for no encryption, the mac
10352 iteration count can be set to 0 to omit the mac.
10354 Enhance pkcs12 utility by making the -nokeys and -nocerts
10355 options work when creating a PKCS#12 file. New option -nomac
10356 to omit the mac, NONE can be set for an encryption algorithm.
10357 New code is modified to use the enhanced PKCS12_create()
10358 instead of the low level API.
10362 * Extend ASN1 encoder to support indefinite length constructed
10363 encoding. This can output sequences tags and octet strings in
10364 this form. Modify pk7_asn1.c to support indefinite length
10365 encoding. This is experimental and needs additional code to
10366 be useful, such as an ASN1 bio and some enhanced streaming
10369 Extend template encode functionality so that tagging is passed
10370 down to the template encoder.
10374 * Let 'openssl req' fail if an argument to '-newkey' is not
10375 recognized instead of using RSA as a default.
10379 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10380 As these are not official, they are not included in "ALL";
10381 the "ECCdraft" ciphersuite group alias can be used to select them.
10383 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10385 * Add ECDH engine support.
10387 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10389 * Add ECDH in new directory crypto/ecdh/.
10391 *Douglas Stebila (Sun Microsystems Laboratories)*
10393 * Let BN_rand_range() abort with an error after 100 iterations
10394 without success (which indicates a broken PRNG).
10398 * Change BN_mod_sqrt() so that it verifies that the input value
10399 is really the square of the return value. (Previously,
10400 BN_mod_sqrt would show GIGO behaviour.)
10404 * Add named elliptic curves over binary fields from X9.62, SECG,
10405 and WAP/WTLS; add OIDs that were still missing.
10407 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10409 * Extend the EC library for elliptic curves over binary fields
10410 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10413 EC_GF2m_simple_method
10417 EC_GROUP_new_curve_GF2m
10418 EC_GROUP_set_curve_GF2m
10419 EC_GROUP_get_curve_GF2m
10420 EC_POINT_set_affine_coordinates_GF2m
10421 EC_POINT_get_affine_coordinates_GF2m
10422 EC_POINT_set_compressed_coordinates_GF2m
10424 Point compression for binary fields is disabled by default for
10425 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10428 As binary polynomials are represented as BIGNUMs, various members
10429 of the EC_GROUP and EC_POINT data structures can be shared
10430 between the implementations for prime fields and binary fields;
10431 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10432 are essentially identical to their `..._GFp` counterparts.
10433 (For simplicity, the `..._GFp` prefix has been dropped from
10434 various internal method names.)
10436 An internal 'field_div' method (similar to 'field_mul' and
10437 'field_sqr') has been added; this is used only for binary fields.
10439 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10441 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10442 through methods ('mul', 'precompute_mult').
10444 The generic implementations (now internally called 'ec_wNAF_mul'
10445 and 'ec_wNAF_precomputed_mult') remain the default if these
10446 methods are undefined.
10448 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10450 * New function EC_GROUP_get_degree, which is defined through
10451 EC_METHOD. For curves over prime fields, this returns the bit
10452 length of the modulus.
10454 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10456 * New functions EC_GROUP_dup, EC_POINT_dup.
10457 (These simply call ..._new and ..._copy).
10459 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10461 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10462 Polynomials are represented as BIGNUMs (where the sign bit is not
10463 used) in the following functions [macros]:
10466 BN_GF2m_sub [= BN_GF2m_add]
10467 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10468 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10469 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10471 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10472 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10473 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10474 BN_GF2m_cmp [= BN_ucmp]
10476 (Note that only the 'mod' functions are actually for fields GF(2^m).
10477 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10479 For some functions, an the irreducible polynomial defining a
10480 field can be given as an 'unsigned int[]' with strictly
10481 decreasing elements giving the indices of those bits that are set;
10482 i.e., p[] represents the polynomial
10483 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10485 p[0] > p[1] > ... > p[k] = 0.
10486 This applies to the following functions:
10489 BN_GF2m_mod_mul_arr
10490 BN_GF2m_mod_sqr_arr
10491 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10492 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10493 BN_GF2m_mod_exp_arr
10494 BN_GF2m_mod_sqrt_arr
10495 BN_GF2m_mod_solve_quad_arr
10499 Conversion can be performed by the following functions:
10504 bntest.c has additional tests for binary polynomial arithmetic.
10506 Two implementations for BN_GF2m_mod_div() are available.
10507 The default algorithm simply uses BN_GF2m_mod_inv() and
10508 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10509 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10510 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10512 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10514 * Add new error code 'ERR_R_DISABLED' that can be used when some
10515 functionality is disabled at compile-time.
10517 *Douglas Stebila <douglas.stebila@sun.com>*
10519 * Change default behaviour of 'openssl asn1parse' so that more
10520 information is visible when viewing, e.g., a certificate:
10522 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10523 mode the content of non-printable OCTET STRINGs is output in a
10524 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10525 avoid the appearance of a printable string.
10527 *Nils Larsch <nla@trustcenter.de>*
10529 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10531 EC_GROUP_set_asn1_flag()
10532 EC_GROUP_get_asn1_flag()
10533 EC_GROUP_set_point_conversion_form()
10534 EC_GROUP_get_point_conversion_form()
10535 These control ASN1 encoding details:
10536 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10537 has been set to OPENSSL_EC_NAMED_CURVE.
10538 - Points are encoded in uncompressed form by default; options for
10539 asn1_for are as for point2oct, namely
10540 POINT_CONVERSION_COMPRESSED
10541 POINT_CONVERSION_UNCOMPRESSED
10542 POINT_CONVERSION_HYBRID
10544 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10546 EC_GROUP_set_seed()
10547 EC_GROUP_get0_seed()
10548 EC_GROUP_get_seed_len()
10549 This is used only for ASN1 purposes (so far).
10551 *Nils Larsch <nla@trustcenter.de>*
10553 * Add 'field_type' member to EC_METHOD, which holds the NID
10554 of the appropriate field type OID. The new function
10555 EC_METHOD_get_field_type() returns this value.
10557 *Nils Larsch <nla@trustcenter.de>*
10560 EC_POINT_point2bn()
10561 EC_POINT_bn2point()
10562 EC_POINT_point2hex()
10563 EC_POINT_hex2point()
10564 providing useful interfaces to EC_POINT_point2oct() and
10565 EC_POINT_oct2point().
10567 *Nils Larsch <nla@trustcenter.de>*
10569 * Change internals of the EC library so that the functions
10570 EC_GROUP_set_generator()
10571 EC_GROUP_get_generator()
10572 EC_GROUP_get_order()
10573 EC_GROUP_get_cofactor()
10574 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10575 to methods, which would lead to unnecessary code duplication when
10576 adding different types of curves.
10578 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10580 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10581 arithmetic, and such that modified wNAFs are generated
10582 (which avoid length expansion in many cases).
10586 * Add a function EC_GROUP_check_discriminant() (defined via
10587 EC_METHOD) that verifies that the curve discriminant is non-zero.
10589 Add a function EC_GROUP_check() that makes some sanity tests
10590 on a EC_GROUP, its generator and order. This includes
10591 EC_GROUP_check_discriminant().
10593 *Nils Larsch <nla@trustcenter.de>*
10595 * Add ECDSA in new directory crypto/ecdsa/.
10597 Add applications 'openssl ecparam' and 'openssl ecdsa'
10598 (these are based on 'openssl dsaparam' and 'openssl dsa').
10600 ECDSA support is also included in various other files across the
10601 library. Most notably,
10602 - 'openssl req' now has a '-newkey ecdsa:file' option;
10603 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10604 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10605 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10606 them suitable for ECDSA where domain parameters must be
10607 extracted before the specific public key;
10608 - ECDSA engine support has been added.
10610 *Nils Larsch <nla@trustcenter.de>*
10612 * Include some named elliptic curves, and add OIDs from X9.62,
10613 SECG, and WAP/WTLS. Each curve can be obtained from the new
10615 EC_GROUP_new_by_curve_name(),
10616 and the list of available named curves can be obtained with
10617 EC_get_builtin_curves().
10618 Also add a 'curve_name' member to EC_GROUP objects, which can be
10620 EC_GROUP_set_curve_name()
10621 EC_GROUP_get_curve_name()
10623 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10625 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10626 was actually never needed) and in BN_mul(). The removal in BN_mul()
10627 required a small change in bn_mul_part_recursive() and the addition
10628 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10629 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10630 bn_sub_words() and bn_add_words() except they take arrays with
10635 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10637 * Cleanse PEM buffers before freeing them since they may contain
10640 *Benjamin Bennett <ben@psc.edu>*
10642 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10643 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10644 authentication-only ciphersuites.
10648 * Since AES128 and AES256 share a single mask bit in the logic of
10649 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10650 kludge to work properly if AES128 is available and AES256 isn't.
10654 * Expand security boundary to match 1.1.1 module.
10658 * Remove redundant features: hash file source, editing of test vectors
10659 modify fipsld to use external fips_premain.c signature.
10663 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10664 run algorithm test programs.
10668 * Make algorithm test programs more tolerant of whitespace.
10672 * Have SSL/TLS server implementation tolerate "mismatched" record
10673 protocol version while receiving ClientHello even if the
10674 ClientHello is fragmented. (The server can't insist on the
10675 particular protocol version it has chosen before the ServerHello
10676 message has informed the client about his choice.)
10680 * Load error codes if they are not already present instead of using a
10681 static variable. This allows them to be cleanly unloaded and reloaded.
10685 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10687 * Introduce limits to prevent malicious keys being able to
10688 cause a denial of service. ([CVE-2006-2940])
10690 *Steve Henson, Bodo Moeller*
10692 * Fix ASN.1 parsing of certain invalid structures that can result
10693 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10695 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10696 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10698 * Fix SSL client code which could crash if connecting to a
10699 malicious SSLv2 server. ([CVE-2006-4343])
10701 *Tavis Ormandy and Will Drewry, Google Security Team*
10703 * Change ciphersuite string processing so that an explicit
10704 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10705 will no longer include "AES128-SHA"), and any other similar
10706 ciphersuite (same bitmap) from *other* protocol versions (so that
10707 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10708 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10709 changes from 0.9.8b and 0.9.8d.
10713 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10715 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10716 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10718 * Change the Unix randomness entropy gathering to use poll() when
10719 possible instead of select(), since the latter has some
10720 undesirable limitations.
10722 *Darryl Miles via Richard Levitte and Bodo Moeller*
10724 * Disable rogue ciphersuites:
10726 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10727 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10728 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10730 The latter two were purportedly from
10731 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10734 Also deactivate the remaining ciphersuites from
10735 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10736 unofficial, and the ID has long expired.
10740 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10741 dual-core machines) and other potential thread-safety issues.
10745 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10747 * Adapt fipsld and the build system to link against the validated FIPS
10748 module in FIPS mode.
10752 * Fixes for VC++ 2005 build under Windows.
10756 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10757 from a Windows bash shell such as MSYS. It is autodetected from the
10758 "config" script when run from a VC++ environment. Modify standard VC++
10759 build to use fipscanister.o from the GNU make build.
10763 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10765 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10766 The value now differs depending on if you build for FIPS or not.
10767 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10768 safely run with a non-FIPSed libcrypto, as it may crash because of
10769 the difference induced by this change.
10773 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10775 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10776 (part of SSL_OP_ALL). This option used to disable the
10777 countermeasure against man-in-the-middle protocol-version
10778 rollback in the SSL 2.0 server implementation, which is a bad
10779 idea. ([CVE-2005-2969])
10781 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10782 for Information Security, National Institute of Advanced Industrial
10783 Science and Technology [AIST, Japan)]*
10785 * Minimal support for X9.31 signatures and PSS padding modes. This is
10786 mainly for FIPS compliance and not fully integrated at this stage.
10790 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10791 the exponentiation using a fixed-length exponent. (Otherwise,
10792 the information leaked through timing could expose the secret key
10793 after many signatures; cf. Bleichenbacher's attack on DSA with
10798 * Make a new fixed-window mod_exp implementation the default for
10799 RSA, DSA, and DH private-key operations so that the sequence of
10800 squares and multiplies and the memory access pattern are
10801 independent of the particular secret key. This will mitigate
10802 cache-timing and potential related attacks.
10804 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10805 and this is automatically used by BN_mod_exp_mont() if the new flag
10806 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10807 will use this BN flag for private exponents unless the flag
10808 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10809 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10811 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10813 * Change the client implementation for SSLv23_method() and
10814 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10815 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10816 (Previously, the SSL 2.0 backwards compatible Client Hello
10817 message format would be used even with SSL_OP_NO_SSLv2.)
10821 * Add support for smime-type MIME parameter in S/MIME messages which some
10826 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10827 a threadsafe manner. Modify rsa code to use new function and add calls
10828 to dsa and dh code (which had race conditions before).
10832 * Include the fixed error library code in the C error file definitions
10833 instead of fixing them up at runtime. This keeps the error code
10834 structures constant.
10838 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10840 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10843 * Fixes for newer kerberos headers. NB: the casts are needed because
10844 the 'length' field is signed on one version and unsigned on another
10845 with no (?) obvious way to tell the difference, without these VC++
10846 complains. Also the "definition" of FAR (blank) is no longer included
10847 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10848 some needed definitions.
10852 * Undo Cygwin change.
10856 * Added support for proxy certificates according to RFC 3820.
10857 Because they may be a security thread to unaware applications,
10858 they must be explicitly allowed in run-time. See
10859 docs/HOWTO/proxy_certificates.txt for further information.
10863 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10865 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10866 server and client random values. Previously
10867 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10868 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10870 This change has negligible security impact because:
10872 1. Server and client random values still have 24 bytes of pseudo random
10875 2. Server and client random values are sent in the clear in the initial
10878 3. The master secret is derived using the premaster secret (48 bytes in
10879 size for static RSA ciphersuites) as well as client server and random
10882 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10885 *Stephen Henson, reported by UK NISCC*
10887 * Use Windows randomness collection on Cygwin.
10891 * Fix hang in EGD/PRNGD query when communication socket is closed
10892 prematurely by EGD/PRNGD.
10894 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
10896 * Prompt for pass phrases when appropriate for PKCS12 input format.
10900 * Back-port of selected performance improvements from development
10901 branch, as well as improved support for PowerPC platforms.
10905 * Add lots of checks for memory allocation failure, error codes to indicate
10906 failure and freeing up memory if a failure occurs.
10908 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
10910 * Add new -passin argument to dgst.
10914 * Perform some character comparisons of different types in X509_NAME_cmp:
10915 this is needed for some certificates that re-encode DNs into UTF8Strings
10916 (in violation of RFC3280) and can't or won't issue name rollover
10921 * Make an explicit check during certificate validation to see that
10922 the CA setting in each certificate on the chain is correct. As a
10923 side effect always do the following basic checks on extensions,
10924 not just when there's an associated purpose to the check:
10926 - if there is an unhandled critical extension (unless the user
10927 has chosen to ignore this fault)
10928 - if the path length has been exceeded (if one is set at all)
10929 - that certain extensions fit the associated purpose (if one has
10934 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
10936 * Avoid a race condition when CRLs are checked in a multi threaded
10937 environment. This would happen due to the reordering of the revoked
10938 entries during signature checking and serial number lookup. Now the
10939 encoding is cached and the serial number sort performed under a lock.
10940 Add new STACK function sk_is_sorted().
10944 * Add Delta CRL to the extension code.
10948 * Various fixes to s3_pkt.c so alerts are sent properly.
10950 *David Holmes <d.holmes@f5.com>*
10952 * Reduce the chances of duplicate issuer name and serial numbers (in
10953 violation of RFC3280) using the OpenSSL certificate creation utilities.
10954 This is done by creating a random 64 bit value for the initial serial
10955 number when a serial number file is created or when a self signed
10956 certificate is created using 'openssl req -x509'. The initial serial
10957 number file is created using 'openssl x509 -next_serial' in CA.pl
10958 rather than being initialized to 1.
10962 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
10964 * Fix null-pointer assignment in do_change_cipher_spec() revealed
10965 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
10967 *Joe Orton, Steve Henson*
10969 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
10972 *Joe Orton, Steve Henson*
10974 * Make it possible to have multiple active certificates with the same
10975 subject in the CA index file. This is done only if the keyword
10976 'unique_subject' is set to 'no' in the main CA section (default
10977 if 'CA_default') of the configuration file. The value is saved
10978 with the database itself in a separate index attribute file,
10979 named like the index file with '.attr' appended to the name.
10983 * X509 verify fixes. Disable broken certificate workarounds when
10984 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
10985 keyUsage extension present. Don't accept CRLs with unhandled critical
10986 extensions: since verify currently doesn't process CRL extensions this
10987 rejects a CRL with *any* critical extensions. Add new verify error codes
10992 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
10993 A clarification of RFC2560 will require the use of OCTET STRINGs and
10994 some implementations cannot handle the current raw format. Since OpenSSL
10995 copies and compares OCSP nonces as opaque blobs without any attempt at
10996 parsing them this should not create any compatibility issues.
11000 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11001 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11002 this HMAC (and other) operations are several times slower than OpenSSL
11007 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11009 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11011 * Use the correct content when signing type "other".
11015 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11017 * Fix various bugs revealed by running the NISCC test suite:
11019 Stop out of bounds reads in the ASN1 code when presented with
11020 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11022 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11024 If verify callback ignores invalid public key errors don't try to check
11025 certificate signature with the NULL public key.
11029 * New -ignore_err option in ocsp application to stop the server
11030 exiting on the first error in a request.
11034 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11035 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11040 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11041 extra data after the compression methods not only for TLS 1.0
11042 but also for SSL 3.0 (as required by the specification).
11044 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11046 * Change X509_certificate_type() to mark the key as exported/exportable
11047 when it's 512 *bits* long, not 512 bytes.
11051 * Change AES_cbc_encrypt() so it outputs exact multiple of
11052 blocks during encryption.
11056 * Various fixes to base64 BIO and non blocking I/O. On write
11057 flushes were not handled properly if the BIO retried. On read
11058 data was not being buffered properly and had various logic bugs.
11059 This also affects blocking I/O when the data being decoded is a
11064 * Various S/MIME bugfixes and compatibility changes:
11065 output correct application/pkcs7 MIME type if
11066 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11067 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11068 of files as .eml work). Correctly handle very long lines in MIME
11073 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11075 * Countermeasure against the Klima-Pokorny-Rosa extension of
11076 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11077 a protocol version number mismatch like a decryption error
11078 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11082 * Turn on RSA blinding by default in the default implementation
11083 to avoid a timing attack. Applications that don't want it can call
11084 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11085 They would be ill-advised to do so in most cases.
11087 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11089 * Change RSA blinding code so that it works when the PRNG is not
11090 seeded (in this case, the secret RSA exponent is abused as
11091 an unpredictable seed -- if it is not unpredictable, there
11092 is no point in blinding anyway). Make RSA blinding thread-safe
11093 by remembering the creator's thread ID in rsa->blinding and
11094 having all other threads use local one-time blinding factors
11095 (this requires more computation than sharing rsa->blinding, but
11096 avoids excessive locking; and if an RSA object is not shared
11097 between threads, blinding will still be very fast).
11101 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11102 ENGINE as defaults for all supported algorithms irrespective of
11103 the 'flags' parameter. 'flags' is now honoured, so applications
11104 should make sure they are passing it correctly.
11108 * Target "mingw" now allows native Windows code to be generated in
11109 the Cygwin environment as well as with the MinGW compiler.
11113 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11115 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11116 via timing by performing a MAC computation even if incorrect
11117 block cipher padding has been found. This is a countermeasure
11118 against active attacks where the attacker has to distinguish
11119 between bad padding and a MAC verification error. ([CVE-2003-0078])
11121 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11122 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11123 Martin Vuagnoux (EPFL, Ilion)*
11125 * Make the no-err option work as intended. The intention with no-err
11126 is not to have the whole error stack handling routines removed from
11127 libcrypto, it's only intended to remove all the function name and
11128 reason texts, thereby removing some of the footprint that may not
11129 be interesting if those errors aren't displayed anyway.
11131 NOTE: it's still possible for any application or module to have its
11132 own set of error texts inserted. The routines are there, just not
11133 used by default when no-err is given.
11137 * Add support for FreeBSD on IA64.
11139 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11141 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11142 Kerberos function mit_des_cbc_cksum(). Before this change,
11143 the value returned by DES_cbc_cksum() was like the one from
11144 mit_des_cbc_cksum(), except the bytes were swapped.
11146 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11148 * Allow an application to disable the automatic SSL chain building.
11149 Before this a rather primitive chain build was always performed in
11150 ssl3_output_cert_chain(): an application had no way to send the
11151 correct chain if the automatic operation produced an incorrect result.
11153 Now the chain builder is disabled if either:
11155 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11157 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11159 The reasoning behind this is that an application would not want the
11160 auto chain building to take place if extra chain certificates are
11161 present and it might also want a means of sending no additional
11162 certificates (for example the chain has two certificates and the
11167 * Add the possibility to build without the ENGINE framework.
11169 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11171 * Under Win32 gmtime() can return NULL: check return value in
11172 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11176 * DSA routines: under certain error conditions uninitialized BN objects
11177 could be freed. Solution: make sure initialization is performed early
11178 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11179 Nils Larsch <nla@trustcenter.de> via PR#459)
11183 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11184 checked on reconnect on the client side, therefore session resumption
11185 could still fail with a "ssl session id is different" error. This
11186 behaviour is masked when SSL_OP_ALL is used due to
11187 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11188 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11189 followup to PR #377.
11193 * IA-32 assembler support enhancements: unified ELF targets, support
11194 for SCO/Caldera platforms, fix for Cygwin shared build.
11198 * Add support for FreeBSD on sparc64. As a consequence, support for
11199 FreeBSD on non-x86 processors is separate from x86 processors on
11200 the config script, much like the NetBSD support.
11202 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11204 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11206 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11209 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11210 code (06) was taken as the first octet of the session ID and the last
11211 octet was ignored consequently. As a result SSLv2 client side session
11212 caching could not have worked due to the session ID mismatch between
11214 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11219 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11220 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11225 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11226 seems that in spite of existing for more than a year, many application
11227 author have done nothing to provide the necessary callbacks, which
11228 means that this particular engine will not work properly anywhere.
11229 This is a very unfortunate situation which forces us, in the name
11230 of usability, to give the hw_ncipher.c a static lock, which is part
11232 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11233 appear in 0.9.8 or later. We EXPECT application authors to have
11234 dealt properly with this when 0.9.8 is released (unless we actually
11235 make such changes in the libcrypto locking code that changes will
11236 have to be made anyway).
11240 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11241 octets have been read, EOF or an error occurs. Without this change
11242 some truncated ASN1 structures will not produce an error.
11246 * Disable Heimdal support, since it hasn't been fully implemented.
11247 Still give the possibility to force the use of Heimdal, but with
11248 warnings and a request that patches get sent to openssl-dev.
11252 * Add the VC-CE target, introduce the WINCE sysname, and add
11253 INSTALL.WCE and appropriate conditionals to make it build.
11255 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11257 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11258 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11259 edit numbers of the version.
11261 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11263 * Introduce safe string copy and catenation functions
11264 (BUF_strlcpy() and BUF_strlcat()).
11266 *Ben Laurie (CHATS) and Richard Levitte*
11268 * Avoid using fixed-size buffers for one-line DNs.
11270 *Ben Laurie (CHATS)*
11272 * Add BUF_MEM_grow_clean() to avoid information leakage when
11273 resizing buffers containing secrets, and use where appropriate.
11275 *Ben Laurie (CHATS)*
11277 * Avoid using fixed size buffers for configuration file location.
11279 *Ben Laurie (CHATS)*
11281 * Avoid filename truncation for various CA files.
11283 *Ben Laurie (CHATS)*
11285 * Use sizeof in preference to magic numbers.
11287 *Ben Laurie (CHATS)*
11289 * Avoid filename truncation in cert requests.
11291 *Ben Laurie (CHATS)*
11293 * Add assertions to check for (supposedly impossible) buffer
11296 *Ben Laurie (CHATS)*
11298 * Don't cache truncated DNS entries in the local cache (this could
11299 potentially lead to a spoofing attack).
11301 *Ben Laurie (CHATS)*
11303 * Fix various buffers to be large enough for hex/decimal
11304 representations in a platform independent manner.
11306 *Ben Laurie (CHATS)*
11308 * Add CRYPTO_realloc_clean() to avoid information leakage when
11309 resizing buffers containing secrets, and use where appropriate.
11311 *Ben Laurie (CHATS)*
11313 * Add BIO_indent() to avoid much slightly worrying code to do
11316 *Ben Laurie (CHATS)*
11318 * Convert sprintf()/BIO_puts() to BIO_printf().
11320 *Ben Laurie (CHATS)*
11322 * buffer_gets() could terminate with the buffer only half
11325 *Ben Laurie (CHATS)*
11327 * Add assertions to prevent user-supplied crypto functions from
11328 overflowing internal buffers by having large block sizes, etc.
11330 *Ben Laurie (CHATS)*
11332 * New OPENSSL_assert() macro (similar to assert(), but enabled
11335 *Ben Laurie (CHATS)*
11337 * Eliminate unused copy of key in RC4.
11339 *Ben Laurie (CHATS)*
11341 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11343 *Ben Laurie (CHATS)*
11345 * Fix off-by-one error in EGD path.
11347 *Ben Laurie (CHATS)*
11349 * If RANDFILE path is too long, ignore instead of truncating.
11351 *Ben Laurie (CHATS)*
11353 * Eliminate unused and incorrectly sized X.509 structure
11356 *Ben Laurie (CHATS)*
11358 * Eliminate unused and dangerous function knumber().
11360 *Ben Laurie (CHATS)*
11362 * Eliminate unused and dangerous structure, KSSL_ERR.
11364 *Ben Laurie (CHATS)*
11366 * Protect against overlong session ID context length in an encoded
11367 session object. Since these are local, this does not appear to be
11370 *Ben Laurie (CHATS)*
11372 * Change from security patch (see 0.9.6e below) that did not affect
11373 the 0.9.6 release series:
11375 Remote buffer overflow in SSL3 protocol - an attacker could
11376 supply an oversized master key in Kerberos-enabled versions.
11379 *Ben Laurie (CHATS)*
11381 * Change the SSL kerb5 codes to match RFC 2712.
11385 * Make -nameopt work fully for req and add -reqopt switch.
11387 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11389 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11391 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11393 * Make sure tests can be performed even if the corresponding algorithms
11394 have been removed entirely. This was also the last step to make
11395 OpenSSL compilable with DJGPP under all reasonable conditions.
11397 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11399 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11400 to allow version independent disabling of normally unselected ciphers,
11401 which may be activated as a side-effect of selecting a single cipher.
11403 (E.g., cipher list string "RSA" enables ciphersuites that are left
11404 out of "ALL" because they do not provide symmetric encryption.
11405 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11407 *Lutz Jaenicke, Bodo Moeller*
11409 * Add appropriate support for separate platform-dependent build
11410 directories. The recommended way to make a platform-dependent
11411 build directory is the following (tested on Linux), maybe with
11414 # Place yourself outside of the OpenSSL source tree. In
11415 # this example, the environment variable OPENSSL_SOURCE
11416 # is assumed to contain the absolute OpenSSL source directory.
11417 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11418 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11419 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11420 mkdir -p `dirname $F`
11421 ln -s $OPENSSL_SOURCE/$F $F
11424 To be absolutely sure not to disturb the source tree, a "make clean"
11425 is a good thing. If it isn't successful, don't worry about it,
11426 it probably means the source directory is very clean.
11430 * Make sure any ENGINE control commands make local copies of string
11431 pointers passed to them whenever necessary. Otherwise it is possible
11432 the caller may have overwritten (or deallocated) the original string
11433 data when a later ENGINE operation tries to use the stored values.
11435 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11437 * Improve diagnostics in file reading and command-line digests.
11439 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11441 * Add AES modes CFB and OFB to the object database. Correct an
11442 error in AES-CFB decryption.
11446 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11447 allows existing EVP_CIPHER_CTX structures to be reused after
11448 calling `EVP_*Final()`. This behaviour is used by encryption
11449 BIOs and some applications. This has the side effect that
11450 applications must explicitly clean up cipher contexts with
11451 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11455 * Check the values of dna and dnb in bn_mul_recursive before calling
11456 bn_mul_comba (a non zero value means the a or b arrays do not contain
11457 n2 elements) and fallback to bn_mul_normal if either is not zero.
11461 * Fix escaping of non-ASCII characters when using the -subj option
11462 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11466 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11467 form for "surname", serialNumber has no short form.
11468 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11469 therefore remove "mail" short name for "internet 7".
11470 The OID for unique identifiers in X509 certificates is
11471 x500UniqueIdentifier, not uniqueIdentifier.
11472 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11476 * Add an "init" command to the ENGINE config module and auto initialize
11477 ENGINEs. Without any "init" command the ENGINE will be initialized
11478 after all ctrl commands have been executed on it. If init=1 the
11479 ENGINE is initialized at that point (ctrls before that point are run
11480 on the uninitialized ENGINE and after on the initialized one). If
11481 init=0 then the ENGINE will not be initialized at all.
11485 * Fix the 'app_verify_callback' interface so that the user-defined
11486 argument is actually passed to the callback: In the
11487 SSL_CTX_set_cert_verify_callback() prototype, the callback
11488 declaration has been changed from
11491 int (*cb)(X509_STORE_CTX *,void *);
11492 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11493 i=s->ctx->app_verify_callback(&ctx)
11494 has been changed into
11495 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11497 To update applications using SSL_CTX_set_cert_verify_callback(),
11498 a dummy argument can be added to their callback functions.
11500 *D. K. Smetters <smetters@parc.xerox.com>*
11502 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11504 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11506 * Add and OPENSSL_LOAD_CONF define which will cause
11507 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11508 This allows older applications to transparently support certain
11509 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11510 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11511 load the config file and OPENSSL_add_all_algorithms_conf() which will
11512 always load it have also been added.
11516 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11517 Adjust NIDs and EVP layer.
11519 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11521 * Config modules support in openssl utility.
11523 Most commands now load modules from the config file,
11524 though in a few (such as version) this isn't done
11525 because it couldn't be used for anything.
11527 In the case of ca and req the config file used is
11528 the same as the utility itself: that is the -config
11529 command line option can be used to specify an
11534 * Move default behaviour from OPENSSL_config(). If appname is NULL
11535 use "openssl_conf" if filename is NULL use default openssl config file.
11539 * Add an argument to OPENSSL_config() to allow the use of an alternative
11540 config section name. Add a new flag to tolerate a missing config file
11541 and move code to CONF_modules_load_file().
11545 * Support for crypto accelerator cards from Accelerated Encryption
11546 Processing, www.aep.ie. (Use engine 'aep')
11547 The support was copied from 0.9.6c [engine] and adapted/corrected
11548 to work with the new engine framework.
11550 *AEP Inc. and Richard Levitte*
11552 * Support for SureWare crypto accelerator cards from Baltimore
11553 Technologies. (Use engine 'sureware')
11554 The support was copied from 0.9.6c [engine] and adapted
11555 to work with the new engine framework.
11559 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11560 make the newer ENGINE framework commands for the CHIL engine work.
11562 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11564 * Make it possible to produce shared libraries on ReliantUNIX.
11566 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11568 * Add the configuration target debug-linux-ppro.
11569 Make 'openssl rsa' use the general key loading routines
11570 implemented in apps.c, and make those routines able to
11571 handle the key format FORMAT_NETSCAPE and the variant
11574 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11576 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11578 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11580 * Add -keyform to rsautl, and document -engine.
11582 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11584 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11585 BIO_R_NO_SUCH_FILE error code rather than the generic
11586 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11590 * Add new functions
11591 ERR_peek_last_error
11592 ERR_peek_last_error_line
11593 ERR_peek_last_error_line_data.
11594 These are similar to
11596 ERR_peek_error_line
11597 ERR_peek_error_line_data,
11598 but report on the latest error recorded rather than the first one
11599 still in the error queue.
11601 *Ben Laurie, Bodo Moeller*
11603 * default_algorithms option in ENGINE config module. This allows things
11605 default_algorithms = ALL
11606 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11610 * Preliminary ENGINE config module.
11614 * New experimental application configuration code.
11618 * Change the AES code to follow the same name structure as all other
11619 symmetric ciphers, and behave the same way. Move everything to
11620 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11622 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11624 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11626 *Ben Laurie and Theo de Raadt*
11628 * Add option to output public keys in req command.
11630 *Massimiliano Pala madwolf@openca.org*
11632 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11633 (up to about 10% better than before for P-192 and P-224).
11637 * New functions/macros
11639 SSL_CTX_set_msg_callback(ctx, cb)
11640 SSL_CTX_set_msg_callback_arg(ctx, arg)
11641 SSL_set_msg_callback(ssl, cb)
11642 SSL_set_msg_callback_arg(ssl, arg)
11644 to request calling a callback function
11646 void cb(int write_p, int version, int content_type,
11647 const void *buf, size_t len, SSL *ssl, void *arg)
11649 whenever a protocol message has been completely received
11650 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11651 protocol version according to which the SSL library interprets
11652 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11653 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11654 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11655 specification (change_cipher_spec(20), alert(21), handshake(22)).
11656 'buf' and 'len' point to the actual message, 'ssl' to the
11657 SSL object, and 'arg' is the application-defined value set by
11658 SSL[_CTX]_set_msg_callback_arg().
11660 'openssl s_client' and 'openssl s_server' have new '-msg' options
11661 to enable a callback that displays all protocol messages.
11665 * Change the shared library support so shared libraries are built as
11666 soon as the corresponding static library is finished, and thereby get
11667 openssl and the test programs linked against the shared library.
11668 This still only happens when the keyword "shard" has been given to
11669 the configuration scripts.
11671 NOTE: shared library support is still an experimental thing, and
11672 backward binary compatibility is still not guaranteed.
11674 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11676 * Add support for Subject Information Access extension.
11678 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11680 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11681 additional bytes when new memory had to be allocated, not just
11682 when reusing an existing buffer.
11686 * New command line and configuration option 'utf8' for the req command.
11687 This allows field values to be specified as UTF8 strings.
11691 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11692 runs for the former and machine-readable output for the latter.
11696 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11697 of the e-mail address in the DN (i.e., it will go into a certificate
11698 extension only). The new configuration file option 'email_in_dn = no'
11699 has the same effect.
11701 *Massimiliano Pala madwolf@openca.org*
11703 * Change all functions with names starting with `des_` to be starting
11704 with `DES_` instead. Add wrappers that are compatible with libdes,
11705 but are named `_ossl_old_des_*`. Finally, add macros that map the
11706 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11707 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11708 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11711 Since we provide two compatibility mappings, the user needs to
11712 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11713 compatibility is desired. The default (i.e., when that macro
11714 isn't defined) is OpenSSL 0.9.6c compatibility.
11716 There are also macros that enable and disable the support of old
11717 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11718 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11719 are defined, the default will apply: to support the old des routines.
11721 In either case, one must include openssl/des.h to get the correct
11722 definitions. Do not try to just include openssl/des_old.h, that
11725 NOTE: This is a major break of an old API into a new one. Software
11726 authors are encouraged to switch to the `DES_` style functions. Some
11727 time in the future, des_old.h and the libdes compatibility functions
11728 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11729 default), and then completely removed.
11733 * Test for certificates which contain unsupported critical extensions.
11734 If such a certificate is found during a verify operation it is
11735 rejected by default: this behaviour can be overridden by either
11736 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11737 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11738 X509_supported_extension() has also been added which returns 1 if a
11739 particular extension is supported.
11743 * Modify the behaviour of EVP cipher functions in similar way to digests
11744 to retain compatibility with existing code.
11748 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11749 compatibility with existing code. In particular the 'ctx' parameter does
11750 not have to be to be initialized before the call to EVP_DigestInit() and
11751 it is tidied up after a call to EVP_DigestFinal(). New function
11752 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11753 EVP_MD_CTX_copy() changed to not require the destination to be
11754 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11755 requires the destination to be valid.
11757 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11758 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11762 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11763 so that complete 'Handshake' protocol structures are kept in memory
11764 instead of overwriting 'msg_type' and 'length' with 'body' data.
11768 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11770 *Massimo Santin via Richard Levitte*
11772 * Major restructuring to the underlying ENGINE code. This includes
11773 reduction of linker bloat, separation of pure "ENGINE" manipulation
11774 (initialisation, etc) from functionality dealing with implementations
11775 of specific crypto interfaces. This change also introduces integrated
11776 support for symmetric ciphers and digest implementations - so ENGINEs
11777 can now accelerate these by providing EVP_CIPHER and EVP_MD
11778 implementations of their own. This is detailed in
11779 [crypto/engine/README.md](crypto/engine/README.md)
11780 as it couldn't be adequately described here. However, there are a few
11781 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11782 were changed in the original introduction of ENGINE code have now
11783 reverted back - the hooking from this code to ENGINE is now a good
11784 deal more passive and at run-time, operations deal directly with
11785 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11786 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11787 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11788 they were not being used by the framework as there is no concept of a
11789 BIGNUM_METHOD and they could not be generalised to the new
11790 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11791 ENGINE_cpy() has been removed as it cannot be consistently defined in
11796 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11800 * Change mkdef.pl to sort symbols that get the same entry number,
11801 and make sure the automatically generated functions `ERR_load_*`
11802 become part of libeay.num as well.
11806 * New function SSL_renegotiate_pending(). This returns true once
11807 renegotiation has been requested (either SSL_renegotiate() call
11808 or HelloRequest/ClientHello received from the peer) and becomes
11809 false once a handshake has been completed.
11810 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11811 sends a HelloRequest, but does not ensure that a handshake takes
11812 place. SSL_renegotiate_pending() is useful for checking if the
11813 client has followed the request.)
11817 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11818 By default, clients may request session resumption even during
11819 renegotiation (if session ID contexts permit); with this option,
11820 session resumption is possible only in the first handshake.
11822 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11823 more bits available for options that should not be part of
11824 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11828 * Add some demos for certificate and certificate request creation.
11832 * Make maximum certificate chain size accepted from the peer application
11833 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11834 "Douglas E. Engert" <deengert@anl.gov>.
11838 * Add support for shared libraries for Unixware-7
11839 (Boyd Lynn Gerber <gerberb@zenez.com>).
11843 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11844 be done prior to destruction. Use this to unload error strings from
11845 ENGINEs that load their own error strings. NB: This adds two new API
11846 functions to "get" and "set" this destroy handler in an ENGINE.
11850 * Alter all existing ENGINE implementations (except "openssl" and
11851 "openbsd") to dynamically instantiate their own error strings. This
11852 makes them more flexible to be built both as statically-linked ENGINEs
11853 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11854 Also, add stub code to each that makes building them as self-contained
11855 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11859 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11860 implementations into applications that are completely implemented in
11861 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11862 commands that can be used to configure what shared-library to load and
11863 to control aspects of the way it is handled. Also, made an update to
11864 the [README-Engine.md](README-Engine.md) file
11865 that brings its information up-to-date and
11866 provides some information and instructions on the "dynamic" ENGINE
11867 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11871 * Make it possible to unload ranges of ERR strings with a new
11872 "ERR_unload_strings" function.
11876 * Add a copy() function to EVP_MD.
11880 * Make EVP_MD routines take a context pointer instead of just the
11881 md_data void pointer.
11885 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11886 that the digest can only process a single chunk of data
11887 (typically because it is provided by a piece of
11888 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
11889 is only going to provide a single chunk of data, and hence the
11890 framework needn't accumulate the data for oneshot drivers.
11894 * As with "ERR", make it possible to replace the underlying "ex_data"
11895 functions. This change also alters the storage and management of global
11896 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
11897 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
11898 index counters. The API functions that use this state have been changed
11899 to take a "class_index" rather than pointers to the class's local STACK
11900 and counter, and there is now an API function to dynamically create new
11901 classes. This centralisation allows us to (a) plug a lot of the
11902 thread-safety problems that existed, and (b) makes it possible to clean
11903 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
11904 such data would previously have always leaked in application code and
11905 workarounds were in place to make the memory debugging turn a blind eye
11906 to it. Application code that doesn't use this new function will still
11907 leak as before, but their memory debugging output will announce it now
11908 rather than letting it slide.
11910 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
11911 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
11912 has a return value to indicate success or failure.
11916 * Make it possible to replace the underlying "ERR" functions such that the
11917 global state (2 LHASH tables and 2 locks) is only used by the "default"
11918 implementation. This change also adds two functions to "get" and "set"
11919 the implementation prior to it being automatically set the first time
11920 any other ERR function takes place. Ie. an application can call "get",
11921 pass the return value to a module it has just loaded, and that module
11922 can call its own "set" function using that value. This means the
11923 module's "ERR" operations will use (and modify) the error state in the
11924 application and not in its own statically linked copy of OpenSSL code.
11928 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
11929 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
11930 the operation, and provides a more encapsulated way for external code
11931 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
11932 to use these functions rather than manually incrementing the counts.
11934 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
11938 * Add EVP test program.
11942 * Add symmetric cipher support to ENGINE. Expect the API to change!
11946 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
11947 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
11948 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
11949 These allow a CRL to be built without having to access X509_CRL fields
11950 directly. Modify 'ca' application to use new functions.
11954 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
11955 bug workarounds. Rollback attack detection is a security feature.
11956 The problem will only arise on OpenSSL servers when TLSv1 is not
11957 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
11958 Software authors not wanting to support TLSv1 will have special reasons
11959 for their choice and can explicitly enable this option.
11961 *Bodo Moeller, Lutz Jaenicke*
11963 * Rationalise EVP so it can be extended: don't include a union of
11964 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
11965 (similar to those existing for EVP_CIPHER_CTX).
11970 EVP_MD_CTX_init(&md); /* new function call */
11971 EVP_DigestInit(&md, EVP_sha1());
11972 EVP_DigestUpdate(&md, in, len);
11973 EVP_DigestFinal(&md, out, NULL);
11974 EVP_MD_CTX_cleanup(&md); /* new function call */
11978 * Make DES key schedule conform to the usual scheme, as well as
11979 correcting its structure. This means that calls to DES functions
11980 now have to pass a pointer to a des_key_schedule instead of a
11981 plain des_key_schedule (which was actually always a pointer
11984 des_key_schedule ks;
11986 des_set_key_checked(..., &ks);
11987 des_ncbc_encrypt(..., &ks, ...);
11989 (Note that a later change renames 'des_...' into 'DES_...'.)
11993 * Initial reduction of linker bloat: the use of some functions, such as
11994 PEM causes large amounts of unused functions to be linked in due to
11995 poor organisation. For example pem_all.c contains every PEM function
11996 which has a knock on effect of linking in large amounts of (unused)
11997 ASN1 code. Grouping together similar functions and splitting unrelated
11998 functions prevents this.
12002 * Cleanup of EVP macros.
12006 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12007 correct `_ecb suffix`.
12011 * Add initial OCSP responder support to ocsp application. The
12012 revocation information is handled using the text based index
12013 use by the ca application. The responder can either handle
12014 requests generated internally, supplied in files (for example
12015 via a CGI script) or using an internal minimal server.
12019 * Add configuration choices to get zlib compression for TLS.
12023 * Changes to Kerberos SSL for RFC 2712 compliance:
12024 1. Implemented real KerberosWrapper, instead of just using
12025 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12026 2. Implemented optional authenticator field of KerberosWrapper.
12028 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12029 and authenticator structs; see crypto/krb5/.
12031 Generalized Kerberos calls to support multiple Kerberos libraries.
12032 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12033 via Richard Levitte*
12035 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12036 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12037 values for each of the key sizes rather than having just
12038 parameters (and 'speed' generating keys each time).
12042 * Speed up EVP routines.
12045 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12046 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12047 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12048 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12050 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12051 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12052 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12055 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12057 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12061 * Added the OS2-EMX target.
12063 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12065 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12066 to support NCONF routines in extension code. New function CONF_set_nconf()
12067 to allow functions which take an NCONF to also handle the old LHASH
12068 structure: this means that the old CONF compatible routines can be
12069 retained (in particular wrt extensions) without having to duplicate the
12070 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12074 * Enhance the general user interface with mechanisms for inner control
12075 and with possibilities to have yes/no kind of prompts.
12079 * Change all calls to low level digest routines in the library and
12080 applications to use EVP. Add missing calls to HMAC_cleanup() and
12081 don't assume HMAC_CTX can be copied using memcpy().
12083 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12085 * Add the possibility to control engines through control names but with
12086 arbitrary arguments instead of just a string.
12087 Change the key loaders to take a UI_METHOD instead of a callback
12088 function pointer. NOTE: this breaks binary compatibility with earlier
12089 versions of OpenSSL [engine].
12090 Adapt the nCipher code for these new conditions and add a card insertion
12095 * Enhance the general user interface with mechanisms to better support
12096 dialog box interfaces, application-defined prompts, the possibility
12097 to use defaults (for example default passwords from somewhere else)
12098 and interrupts/cancellations.
12102 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12103 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12107 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12108 tidy up some unnecessarily weird code in 'sk_new()').
12110 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12112 * Change the key loading routines for ENGINEs to use the same kind
12113 callback (pem_password_cb) as all other routines that need this
12118 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12119 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12120 than this minimum value is recommended.
12124 * New random seeder for OpenVMS, using the system process statistics
12125 that are easily reachable.
12129 * Windows apparently can't transparently handle global
12130 variables defined in DLLs. Initialisations such as:
12132 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12134 won't compile. This is used by the any applications that need to
12135 declare their own ASN1 modules. This was fixed by adding the option
12136 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12137 needed for static libraries under Win32.
12141 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12142 setting of purpose and trust fields. New X509_STORE trust and
12143 purpose functions and tidy up setting in other SSL functions.
12147 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12148 structure. These are inherited by X509_STORE_CTX when it is
12149 initialised. This allows various defaults to be set in the
12150 X509_STORE structure (such as flags for CRL checking and custom
12151 purpose or trust settings) for functions which only use X509_STORE_CTX
12152 internally such as S/MIME.
12154 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12155 trust settings if they are not set in X509_STORE. This allows X509_STORE
12156 purposes and trust (in S/MIME for example) to override any set by default.
12158 Add command line options for CRL checking to smime, s_client and s_server
12163 * Initial CRL based revocation checking. If the CRL checking flag(s)
12164 are set then the CRL is looked up in the X509_STORE structure and
12165 its validity and signature checked, then if the certificate is found
12166 in the CRL the verify fails with a revoked error.
12168 Various new CRL related callbacks added to X509_STORE_CTX structure.
12170 Command line options added to 'verify' application to support this.
12172 This needs some additional work, such as being able to handle multiple
12173 CRLs with different times, extension based lookup (rather than just
12174 by subject name) and ultimately more complete V2 CRL extension
12179 * Add a general user interface API (crypto/ui/). This is designed
12180 to replace things like des_read_password and friends (backward
12181 compatibility functions using this new API are provided).
12182 The purpose is to remove prompting functions from the DES code
12183 section as well as provide for prompting through dialog boxes in
12184 a window system and the like.
12188 * Add "ex_data" support to ENGINE so implementations can add state at a
12189 per-structure level rather than having to store it globally.
12193 * Make it possible for ENGINE structures to be copied when retrieved by
12194 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12195 This causes the "original" ENGINE structure to act like a template,
12196 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12197 operational state can be localised to each ENGINE structure, despite the
12198 fact they all share the same "methods". New ENGINE structures returned in
12199 this case have no functional references and the return value is the single
12200 structural reference. This matches the single structural reference returned
12201 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12206 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12207 needs to match any other type at all we need to manually clear the
12212 * Changes to the "openssl engine" utility to include;
12213 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12214 about an ENGINE's available control commands.
12215 - executing control commands from command line arguments using the
12216 '-pre' and '-post' switches. '-post' is only used if '-t' is
12217 specified and the ENGINE is successfully initialised. The syntax for
12218 the individual commands are colon-separated, for example;
12219 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12223 * New dynamic control command support for ENGINEs. ENGINEs can now
12224 declare their own commands (numbers), names (strings), descriptions,
12225 and input types for run-time discovery by calling applications. A
12226 subset of these commands are implicitly classed as "executable"
12227 depending on their input type, and only these can be invoked through
12228 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12229 can be based on user input, config files, etc). The distinction is
12230 that "executable" commands cannot return anything other than a boolean
12231 result and can only support numeric or string input, whereas some
12232 discoverable commands may only be for direct use through
12233 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12234 pointers, or other custom uses. The "executable" commands are to
12235 support parameterisations of ENGINE behaviour that can be
12236 unambiguously defined by ENGINEs and used consistently across any
12237 OpenSSL-based application. Commands have been added to all the
12238 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12239 control over shared-library paths without source code alterations.
12243 * Changed all ENGINE implementations to dynamically allocate their
12244 ENGINEs rather than declaring them statically. Apart from this being
12245 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12246 this also allows the implementations to compile without using the
12247 internal engine_int.h header.
12251 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12252 'const' value. Any code that should be able to modify a RAND_METHOD
12253 should already have non-const pointers to it (ie. they should only
12254 modify their own ones).
12258 * Made a variety of little tweaks to the ENGINE code.
12259 - "atalla" and "ubsec" string definitions were moved from header files
12260 to C code. "nuron" string definitions were placed in variables
12261 rather than hard-coded - allowing parameterisation of these values
12262 later on via ctrl() commands.
12263 - Removed unused "#if 0"'d code.
12264 - Fixed engine list iteration code so it uses ENGINE_free() to release
12265 structural references.
12266 - Constified the RAND_METHOD element of ENGINE structures.
12267 - Constified various get/set functions as appropriate and added
12268 missing functions (including a catch-all ENGINE_cpy that duplicates
12269 all ENGINE values onto a new ENGINE except reference counts/state).
12270 - Removed NULL parameter checks in get/set functions. Setting a method
12271 or function to NULL is a way of cancelling out a previously set
12272 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12273 and doesn't justify the extra error symbols and code.
12274 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12275 flags from engine_int.h to engine.h.
12276 - Changed prototypes for ENGINE handler functions (init(), finish(),
12277 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12281 * Implement binary inversion algorithm for BN_mod_inverse in addition
12282 to the algorithm using long division. The binary algorithm can be
12283 used only if the modulus is odd. On 32-bit systems, it is faster
12284 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12285 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12286 up to 450 bits. In 64-bit environments, the binary algorithm
12287 appears to be advantageous for much longer moduli; here we use it
12288 for moduli up to 2048 bits.
12292 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12293 could not support the combine flag in choice fields.
12297 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12298 extensions from a certificate request to the certificate.
12302 * Allow multiple 'certopt' and 'nameopt' options to be separated
12303 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12304 file: this allows the display of the certificate about to be
12305 signed to be customised, to allow certain fields to be included
12306 or excluded and extension details. The old system didn't display
12307 multicharacter strings properly, omitted fields not in the policy
12308 and couldn't display additional details such as extensions.
12312 * Function EC_POINTs_mul for multiple scalar multiplication
12313 of an arbitrary number of elliptic curve points
12314 \sum scalars[i]*points[i],
12315 optionally including the generator defined for the EC_GROUP:
12316 scalar*generator + \sum scalars[i]*points[i].
12318 EC_POINT_mul is a simple wrapper function for the typical case
12319 that the point list has just one item (besides the optional
12324 * First EC_METHODs for curves over GF(p):
12326 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12327 operations and provides various method functions that can also
12328 operate with faster implementations of modular arithmetic.
12330 EC_GFp_mont_method() reuses most functions that are part of
12331 EC_GFp_simple_method, but uses Montgomery arithmetic.
12333 *Bodo Moeller; point addition and point doubling
12334 implementation directly derived from source code provided by
12335 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12337 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12338 crypto/ec/ec_lib.c):
12340 Curves are EC_GROUP objects (with an optional group generator)
12341 based on EC_METHODs that are built into the library.
12343 Points are EC_POINT objects based on EC_GROUP objects.
12345 Most of the framework would be able to handle curves over arbitrary
12346 finite fields, but as there are no obvious types for fields other
12347 than GF(p), some functions are limited to that for now.
12351 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12352 that the file contains a complete HTTP response.
12356 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12357 change the def and num file printf format specifier from "%-40sXXX"
12358 to "%-39s XXX". The latter will always guarantee a space after the
12359 field while the former will cause them to run together if the field
12360 is 40 of more characters long.
12364 * Constify the cipher and digest 'method' functions and structures
12365 and modify related functions to take constant EVP_MD and EVP_CIPHER
12370 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12371 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12375 * Modify `EVP_Digest*()` routines so they now return values. Although the
12376 internal software routines can never fail additional hardware versions
12381 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12383 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12384 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12387 ERR_R_NESTED_ASN1_ERROR
12389 ERR_R_MISSING_ASN1_EOS
12390 were 4 .. 9, conflicting with
12391 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12393 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12394 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12396 Add new error code 'ERR_R_INTERNAL_ERROR'.
12400 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12405 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12406 sets the subject name for a new request or supersedes the
12407 subject name in a given request. Formats that can be parsed are
12408 'CN=Some Name, OU=myOU, C=IT'
12410 'CN=Some Name/OU=myOU/C=IT'.
12412 Add options '-batch' and '-verbose' to 'openssl req'.
12414 *Massimiliano Pala <madwolf@hackmasters.net>*
12416 * Introduce the possibility to access global variables through
12417 functions on platform were that's the best way to handle exporting
12418 global variables in shared libraries. To enable this functionality,
12419 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12420 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12421 is normally done by Configure or something similar).
12423 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12424 in the source file (foo.c) like this:
12426 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12427 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12429 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12430 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12432 OPENSSL_DECLARE_GLOBAL(int,foo);
12433 #define foo OPENSSL_GLOBAL_REF(foo)
12434 OPENSSL_DECLARE_GLOBAL(double,bar);
12435 #define bar OPENSSL_GLOBAL_REF(bar)
12437 The #defines are very important, and therefore so is including the
12438 header file everywhere where the defined globals are used.
12440 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12441 of ASN.1 items, but that structure is a bit different.
12443 The largest change is in util/mkdef.pl which has been enhanced with
12444 better and easier to understand logic to choose which symbols should
12445 go into the Windows .def files as well as a number of fixes and code
12446 cleanup (among others, algorithm keywords are now sorted
12447 lexicographically to avoid constant rewrites).
12451 * In BN_div() keep a copy of the sign of 'num' before writing the
12452 result to 'rm' because if rm==num the value will be overwritten
12453 and produce the wrong result if 'num' is negative: this caused
12454 problems with BN_mod() and BN_nnmod().
12458 * Function OCSP_request_verify(). This checks the signature on an
12459 OCSP request and verifies the signer certificate. The signer
12460 certificate is just checked for a generic purpose and OCSP request
12465 * Add OCSP_check_validity() function to check the validity of OCSP
12466 responses. OCSP responses are prepared in real time and may only
12467 be a few seconds old. Simply checking that the current time lies
12468 between thisUpdate and nextUpdate max reject otherwise valid responses
12469 caused by either OCSP responder or client clock inaccuracy. Instead
12470 we allow thisUpdate and nextUpdate to fall within a certain period of
12471 the current time. The age of the response can also optionally be
12472 checked. Two new options -validity_period and -status_age added to
12477 * If signature or public key algorithm is unrecognized print out its
12478 OID rather that just UNKNOWN.
12482 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12483 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12484 ID to be generated from the issuer certificate alone which can then be
12485 passed to OCSP_id_issuer_cmp().
12489 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12490 ASN1 modules to export functions returning ASN1_ITEM pointers
12491 instead of the ASN1_ITEM structures themselves. This adds several
12492 new macros which allow the underlying ASN1 function/structure to
12493 be accessed transparently. As a result code should not use ASN1_ITEM
12494 references directly (such as &X509_it) but instead use the relevant
12495 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12496 use of the new ASN1 code on platforms where exporting structures
12497 is problematical (for example in shared libraries) but exporting
12498 functions returning pointers to structures is not.
12502 * Add support for overriding the generation of SSL/TLS session IDs.
12503 These callbacks can be registered either in an SSL_CTX or per SSL.
12504 The purpose of this is to allow applications to control, if they wish,
12505 the arbitrary values chosen for use as session IDs, particularly as it
12506 can be useful for session caching in multiple-server environments. A
12507 command-line switch for testing this (and any client code that wishes
12508 to use such a feature) has been added to "s_server".
12510 *Geoff Thorpe, Lutz Jaenicke*
12512 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12513 of the form `#if defined(...) || defined(...) || ...` and
12514 `#if !defined(...) && !defined(...) && ...`. This also avoids
12515 the growing number of special cases it was previously handling.
12519 * Make all configuration macros available for application by making
12520 sure they are available in opensslconf.h, by giving them names starting
12521 with `OPENSSL_` to avoid conflicts with other packages and by making
12522 sure e_os2.h will cover all platform-specific cases together with
12524 Additionally, it is now possible to define configuration/platform-
12525 specific names (called "system identities"). In the C code, these
12526 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12527 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12528 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12533 * New option -set_serial to 'req' and 'x509' this allows the serial
12534 number to use to be specified on the command line. Previously self
12535 signed certificates were hard coded with serial number 0 and the
12536 CA options of 'x509' had to use a serial number in a file which was
12541 * New options to 'ca' utility to support V2 CRL entry extensions.
12542 Currently CRL reason, invalidity date and hold instruction are
12543 supported. Add new CRL extensions to V3 code and some new objects.
12547 * New function EVP_CIPHER_CTX_set_padding() this is used to
12548 disable standard block padding (aka PKCS#5 padding) in the EVP
12549 API, which was previously mandatory. This means that the data is
12550 not padded in any way and so the total length much be a multiple
12551 of the block size, otherwise an error occurs.
12555 * Initial (incomplete) OCSP SSL support.
12559 * New function OCSP_parse_url(). This splits up a URL into its host,
12560 port and path components: primarily to parse OCSP URLs. New -url
12561 option to ocsp utility.
12565 * New nonce behavior. The return value of OCSP_check_nonce() now
12566 reflects the various checks performed. Applications can decide
12567 whether to tolerate certain situations such as an absent nonce
12568 in a response when one was present in a request: the ocsp application
12569 just prints out a warning. New function OCSP_add1_basic_nonce()
12570 this is to allow responders to include a nonce in a response even if
12571 the request is nonce-less.
12575 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12576 skipped when using openssl x509 multiple times on a single input file,
12577 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12581 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12582 set string type: to handle setting ASN1_TIME structures. Fix ca
12583 utility to correctly initialize revocation date of CRLs.
12587 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12588 the clients preferred ciphersuites and rather use its own preferences.
12589 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12590 Internet Explorer by ensuring unchanged hash method during stepup.
12591 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12595 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12596 to aes and add a new 'exist' option to print out symbols that don't
12601 * Additional options to ocsp utility to allow flags to be set and
12602 additional certificates supplied.
12606 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12607 OCSP client a number of certificate to only verify the response
12612 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12613 handle the new API. Currently only ECB, CBC modes supported. Add new
12616 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12617 Encryption Standard (AES) Ciphersuites for Transport Layer
12618 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12619 not enabled by default and were not part of the "ALL" ciphersuite
12620 alias because they were not yet official; they could be
12621 explicitly requested by specifying the "AESdraft" ciphersuite
12622 group alias. In the final release of OpenSSL 0.9.7, the group
12623 alias is called "AES" and is part of "ALL".)
12625 *Ben Laurie, Steve Henson, Bodo Moeller*
12627 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12628 request to response.
12632 * Functions for OCSP responders. OCSP_request_onereq_count(),
12633 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12634 extract information from a certificate request. OCSP_response_create()
12635 creates a response and optionally adds a basic response structure.
12636 OCSP_basic_add1_status() adds a complete single response to a basic
12637 response and returns the OCSP_SINGLERESP structure just added (to allow
12638 extensions to be included for example). OCSP_basic_add1_cert() adds a
12639 certificate to a basic response and OCSP_basic_sign() signs a basic
12640 response with various flags. New helper functions ASN1_TIME_check()
12641 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12642 (converts ASN1_TIME to GeneralizedTime).
12646 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12647 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12648 structure from a certificate. X509_pubkey_digest() digests the public_key
12649 contents: this is used in various key identifiers.
12653 * Make sk_sort() tolerate a NULL argument.
12655 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12657 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12658 passed by the function are trusted implicitly. If any of them signed the
12659 response then it is assumed to be valid and is not verified.
12663 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12664 to data. This was previously part of the PKCS7 ASN1 code. This
12665 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12666 *Steve Henson, reported by Kenneth R. Robinette
12667 <support@securenetterm.com>*
12669 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12670 routines: without these tracing memory leaks is very painful.
12671 Fix leaks in PKCS12 and PKCS7 routines.
12675 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12676 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12677 effectively meant GeneralizedTime would never be used. Now it
12678 is initialised to -1 but X509_time_adj() now has to check the value
12679 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12680 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12681 *Steve Henson, reported by Kenneth R. Robinette
12682 <support@securenetterm.com>*
12684 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12685 result in a zero length in the ASN1_INTEGER structure which was
12686 not consistent with the structure when d2i_ASN1_INTEGER() was used
12687 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12688 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12689 where it did not print out a minus for negative ASN1_INTEGER.
12693 * Add summary printout to ocsp utility. The various functions which
12694 convert status values to strings have been renamed to:
12695 OCSP_response_status_str(), OCSP_cert_status_str() and
12696 OCSP_crl_reason_str() and are no longer static. New options
12697 to verify nonce values and to disable verification. OCSP response
12698 printout format cleaned up.
12702 * Add additional OCSP certificate checks. These are those specified
12703 in RFC2560. This consists of two separate checks: the CA of the
12704 certificate being checked must either be the OCSP signer certificate
12705 or the issuer of the OCSP signer certificate. In the latter case the
12706 OCSP signer certificate must contain the OCSP signing extended key
12707 usage. This check is performed by attempting to match the OCSP
12708 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12709 in the OCSP_CERTID structures of the response.
12713 * Initial OCSP certificate verification added to OCSP_basic_verify()
12714 and related routines. This uses the standard OpenSSL certificate
12715 verify routines to perform initial checks (just CA validity) and
12716 to obtain the certificate chain. Then additional checks will be
12717 performed on the chain. Currently the root CA is checked to see
12718 if it is explicitly trusted for OCSP signing. This is used to set
12719 a root CA as a global signing root: that is any certificate that
12720 chains to that CA is an acceptable OCSP signing certificate.
12724 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12725 extensions from a separate configuration file.
12726 As when reading extensions from the main configuration file,
12727 the '-extensions ...' option may be used for specifying the
12730 *Massimiliano Pala <madwolf@comune.modena.it>*
12732 * New OCSP utility. Allows OCSP requests to be generated or
12733 read. The request can be sent to a responder and the output
12734 parsed, outputted or printed in text form. Not complete yet:
12735 still needs to check the OCSP response validity.
12739 * New subcommands for 'openssl ca':
12740 `openssl ca -status <serial>` prints the status of the cert with
12741 the given serial number (according to the index file).
12742 `openssl ca -updatedb` updates the expiry status of certificates
12745 *Massimiliano Pala <madwolf@comune.modena.it>*
12747 * New '-newreq-nodes' command option to CA.pl. This is like
12748 '-newreq', but calls 'openssl req' with the '-nodes' option
12749 so that the resulting key is not encrypted.
12751 *Damien Miller <djm@mindrot.org>*
12753 * New configuration for the GNU Hurd.
12755 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12757 * Initial code to implement OCSP basic response verify. This
12758 is currently incomplete. Currently just finds the signer's
12759 certificate and verifies the signature on the response.
12763 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12764 value of OPENSSLDIR. This is available via the new '-d' option
12765 to 'openssl version', and is also included in 'openssl version -a'.
12769 * Allowing defining memory allocation callbacks that will be given
12770 file name and line number information in additional arguments
12771 (a `const char*` and an int). The basic functionality remains, as
12772 well as the original possibility to just replace malloc(),
12773 realloc() and free() by functions that do not know about these
12774 additional arguments. To register and find out the current
12775 settings for extended allocation functions, the following
12776 functions are provided:
12778 CRYPTO_set_mem_ex_functions
12779 CRYPTO_set_locked_mem_ex_functions
12780 CRYPTO_get_mem_ex_functions
12781 CRYPTO_get_locked_mem_ex_functions
12783 These work the same way as CRYPTO_set_mem_functions and friends.
12784 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12785 extended allocation function is enabled.
12786 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12787 a conventional allocation function is enabled.
12789 *Richard Levitte, Bodo Moeller*
12791 * Finish off removing the remaining LHASH function pointer casts.
12792 There should no longer be any prototype-casting required when using
12793 the LHASH abstraction, and any casts that remain are "bugs". See
12794 the callback types and macros at the head of lhash.h for details
12795 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12799 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12800 If /dev/[u]random devices are not available or do not return enough
12801 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12803 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12804 /etc/entropy will be queried once each in this sequence, querying stops
12805 when enough entropy was collected without querying more sockets.
12809 * Change the Unix RAND_poll() variant to be able to poll several
12810 random devices, as specified by DEVRANDOM, until a sufficient amount
12811 of data has been collected. We spend at most 10 ms on each file
12812 (select timeout) and read in non-blocking mode. DEVRANDOM now
12813 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12814 (previously it was just the string "/dev/urandom"), so on typical
12815 platforms the 10 ms delay will never occur.
12816 Also separate out the Unix variant to its own file, rand_unix.c.
12817 For VMS, there's a currently-empty rand_vms.c.
12821 * Move OCSP client related routines to ocsp_cl.c. These
12822 provide utility functions which an application needing
12823 to issue a request to an OCSP responder and analyse the
12824 response will typically need: as opposed to those which an
12825 OCSP responder itself would need which will be added later.
12827 OCSP_request_sign() signs an OCSP request with an API similar
12828 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12829 response. OCSP_response_get1_basic() extracts basic response
12830 from response. OCSP_resp_find_status(): finds and extracts status
12831 information from an OCSP_CERTID structure (which will be created
12832 when the request structure is built). These are built from lower
12833 level functions which work on OCSP_SINGLERESP structures but
12834 won't normally be used unless the application wishes to examine
12835 extensions in the OCSP response for example.
12837 Replace nonce routines with a pair of functions.
12838 OCSP_request_add1_nonce() adds a nonce value and optionally
12839 generates a random value. OCSP_check_nonce() checks the
12840 validity of the nonce in an OCSP response.
12844 * Change function OCSP_request_add() to OCSP_request_add0_id().
12845 This doesn't copy the supplied OCSP_CERTID and avoids the
12846 need to free up the newly created id. Change return type
12847 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12848 This can then be used to add extensions to the request.
12849 Deleted OCSP_request_new(), since most of its functionality
12850 is now in OCSP_REQUEST_new() (and the case insensitive name
12851 clash) apart from the ability to set the request name which
12852 will be added elsewhere.
12856 * Update OCSP API. Remove obsolete extensions argument from
12857 various functions. Extensions are now handled using the new
12858 OCSP extension code. New simple OCSP HTTP function which
12859 can be used to send requests and parse the response.
12863 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12864 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12865 uses the special reorder version of SET OF to sort the attributes
12866 and reorder them to match the encoded order. This resolves a long
12867 standing problem: a verify on a PKCS7 structure just after signing
12868 it used to fail because the attribute order did not match the
12869 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12870 it uses the received order. This is necessary to tolerate some broken
12871 software that does not order SET OF. This is handled by encoding
12872 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12873 to produce the required SET OF.
12877 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12878 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12879 files to get correct declarations of the ASN.1 item variables.
12883 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12884 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12885 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12886 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
12887 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
12888 ASN1_ITEM and no wrapper functions.
12892 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
12893 replace the old function pointer based I/O routines. Change most of
12894 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
12898 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
12899 lines, recognize more "algorithms" that can be deselected, and make
12900 it complain about algorithm deselection that isn't recognised.
12904 * New ASN1 functions to handle dup, sign, verify, digest, pack and
12905 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
12906 to use new functions. Add NO_ASN1_OLD which can be set to remove
12907 some old style ASN1 functions: this can be used to determine if old
12908 code will still work when these eventually go away.
12912 * New extension functions for OCSP structures, these follow the
12913 same conventions as certificates and CRLs.
12917 * New function X509V3_add1_i2d(). This automatically encodes and
12918 adds an extension. Its behaviour can be customised with various
12919 flags to append, replace or delete. Various wrappers added for
12920 certificates and CRLs.
12924 * Fix to avoid calling the underlying ASN1 print routine when
12925 an extension cannot be parsed. Correct a typo in the
12926 OCSP_SERVICELOC extension. Tidy up print OCSP format.
12930 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
12931 entries for variables.
12935 * Add functionality to apps/openssl.c for detecting locking
12936 problems: As the program is single-threaded, all we have
12937 to do is register a locking callback using an array for
12938 storing which locks are currently held by the program.
12942 * Use a lock around the call to CRYPTO_get_ex_new_index() in
12943 SSL_get_ex_data_X509_STORE_idx(), which is used in
12944 ssl_verify_cert_chain() and thus can be called at any time
12945 during TLS/SSL handshakes so that thread-safety is essential.
12946 Unfortunately, the ex_data design is not at all suited
12947 for multi-threaded use, so it probably should be abolished.
12951 * Added Broadcom "ubsec" ENGINE to OpenSSL.
12953 *Broadcom, tweaked and integrated by Geoff Thorpe*
12955 * Move common extension printing code to new function
12956 X509V3_print_extensions(). Reorganise OCSP print routines and
12957 implement some needed OCSP ASN1 functions. Add OCSP extensions.
12961 * New function X509_signature_print() to remove duplication in some
12966 * Add a special meaning when SET OF and SEQUENCE OF flags are both
12967 set (this was treated exactly the same as SET OF previously). This
12968 is used to reorder the STACK representing the structure to match the
12969 encoding. This will be used to get round a problem where a PKCS7
12970 structure which was signed could not be verified because the STACK
12971 order did not reflect the encoded order.
12975 * Reimplement the OCSP ASN1 module using the new code.
12979 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
12980 for its ASN1 operations. The old style function pointers still exist
12981 for now but they will eventually go away.
12985 * Merge in replacement ASN1 code from the ASN1 branch. This almost
12986 completely replaces the old ASN1 functionality with a table driven
12987 encoder and decoder which interprets an ASN1_ITEM structure describing
12988 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
12989 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
12990 has also been converted to the new form.
12994 * Change BN_mod_exp_recp so that negative moduli are tolerated
12995 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
12996 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
12997 for negative moduli.
13001 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13002 of not touching the result's sign bit.
13006 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13011 * Changed the LHASH code to use prototypes for callbacks, and created
13012 macros to declare and implement thin (optionally static) functions
13013 that provide type-safety and avoid function pointer casting for the
13014 type-specific callbacks.
13018 * Added Kerberos Cipher Suites to be used with TLS, as written in
13020 *Veers Staats <staatsvr@asc.hpc.mil>,
13021 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13023 * Reformat the FAQ so the different questions and answers can be divided
13024 in sections depending on the subject.
13028 * Have the zlib compression code load ZLIB.DLL dynamically under
13033 * New function BN_mod_sqrt for computing square roots modulo a prime
13034 (using the probabilistic Tonelli-Shanks algorithm unless
13035 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13036 be handled deterministically).
13038 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13040 * Make BN_mod_inverse faster by explicitly handling small quotients
13041 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13042 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13046 * New function BN_kronecker.
13050 * Fix BN_gcd so that it works on negative inputs; the result is
13051 positive unless both parameters are zero.
13052 Previously something reasonably close to an infinite loop was
13053 possible because numbers could be growing instead of shrinking
13054 in the implementation of Euclid's algorithm.
13058 * Fix BN_is_word() and BN_is_one() macros to take into account the
13059 sign of the number in question.
13061 Fix BN_is_word(a,w) to work correctly for w == 0.
13063 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13064 because its test if the absolute value of 'a' equals 'w'.
13065 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13066 it exists mostly for use in the implementations of BN_is_zero(),
13067 BN_is_one(), and BN_is_word().
13071 * New function BN_swap.
13075 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13076 the exponentiation functions are more likely to produce reasonable
13077 results on negative inputs.
13081 * Change BN_mod_mul so that the result is always non-negative.
13082 Previously, it could be negative if one of the factors was negative;
13083 I don't think anyone really wanted that behaviour.
13087 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13088 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13089 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13090 and add new functions:
13099 BN_mod_lshift1_quick
13101 BN_mod_lshift_quick
13103 These functions always generate non-negative results.
13105 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13106 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13108 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13109 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13110 be reduced modulo `m`.
13112 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13115 The following entry accidentally appeared in the CHANGES file
13116 distributed with OpenSSL 0.9.7. The modifications described in
13117 it do *not* apply to OpenSSL 0.9.7.
13119 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13120 was actually never needed) and in BN_mul(). The removal in BN_mul()
13121 required a small change in bn_mul_part_recursive() and the addition
13122 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13123 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13124 bn_sub_words() and bn_add_words() except they take arrays with
13130 * In 'openssl passwd', verify passwords read from the terminal
13131 unless the '-salt' option is used (which usually means that
13132 verification would just waste user's time since the resulting
13133 hash is going to be compared with some given password hash)
13134 or the new '-noverify' option is used.
13136 This is an incompatible change, but it does not affect
13137 non-interactive use of 'openssl passwd' (passwords on the command
13138 line, '-stdin' option, '-in ...' option) and thus should not
13139 cause any problems.
13143 * Remove all references to RSAref, since there's no more need for it.
13147 * Make DSO load along a path given through an environment variable
13148 (SHLIB_PATH) with shl_load().
13152 * Constify the ENGINE code as a result of BIGNUM constification.
13153 Also constify the RSA code and most things related to it. In a
13154 few places, most notable in the depth of the ASN.1 code, ugly
13155 casts back to non-const were required (to be solved at a later
13160 * Make it so the openssl application has all engines loaded by default.
13164 * Constify the BIGNUM routines a little more.
13168 * Add the following functions:
13170 ENGINE_load_cswift()
13172 ENGINE_load_atalla()
13173 ENGINE_load_nuron()
13174 ENGINE_load_builtin_engines()
13176 That way, an application can itself choose if external engines that
13177 are built-in in OpenSSL shall ever be used or not. The benefit is
13178 that applications won't have to be linked with libdl or other dso
13179 libraries unless it's really needed.
13181 Changed 'openssl engine' to load all engines on demand.
13182 Changed the engine header files to avoid the duplication of some
13183 declarations (they differed!).
13187 * 'openssl engine' can now list capabilities.
13191 * Better error reporting in 'openssl engine'.
13195 * Never call load_dh_param(NULL) in s_server.
13199 * Add engine application. It can currently list engines by name and
13200 identity, and test if they are actually available.
13204 * Improve RPM specification file by forcing symbolic linking and making
13205 sure the installed documentation is also owned by root.root.
13207 *Damien Miller <djm@mindrot.org>*
13209 * Give the OpenSSL applications more possibilities to make use of
13210 keys (public as well as private) handled by engines.
13214 * Add OCSP code that comes from CertCo.
13218 * Add VMS support for the Rijndael code.
13222 * Added untested support for Nuron crypto accelerator.
13226 * Add support for external cryptographic devices. This code was
13227 previously distributed separately as the "engine" branch.
13229 *Geoff Thorpe, Richard Levitte*
13231 * Rework the filename-translation in the DSO code. It is now possible to
13232 have far greater control over how a "name" is turned into a filename
13233 depending on the operating environment and any oddities about the
13234 different shared library filenames on each system.
13238 * Support threads on FreeBSD-elf in Configure.
13242 * Fix for SHA1 assembly problem with MASM: it produces
13243 warnings about corrupt line number information when assembling
13244 with debugging information. This is caused by the overlapping
13247 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13250 NCONF_get_number() has no error checking at all. As a replacement,
13251 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13252 promoted strongly. The old NCONF_get_number is kept around for
13253 binary backward compatibility.
13254 Make it possible for methods to load from something other than a BIO,
13255 by providing a function pointer that is given a name instead of a BIO.
13256 For example, this could be used to load configuration data from an
13261 * Fix for non blocking accept BIOs. Added new I/O special reason
13262 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13263 with non blocking I/O was not possible because no retry code was
13264 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13269 * Added the beginnings of Rijndael support.
13273 * Fix for bug in DirectoryString mask setting. Add support for
13274 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13275 to allow certificate printing to more controllable, additional
13276 'certopt' option to 'x509' to allow new printing options to be
13281 * Clean old EAY MD5 hack from e_os.h.
13285 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13287 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13288 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13290 *Joe Orton, Steve Henson*
13292 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13294 * Fix additional bug revealed by the NISCC test suite:
13296 Stop bug triggering large recursion when presented with
13297 certain ASN.1 tags ([CVE-2003-0851])
13301 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13303 * Fix various bugs revealed by running the NISCC test suite:
13305 Stop out of bounds reads in the ASN1 code when presented with
13306 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13308 If verify callback ignores invalid public key errors don't try to check
13309 certificate signature with the NULL public key.
13313 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13314 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13319 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13320 extra data after the compression methods not only for TLS 1.0
13321 but also for SSL 3.0 (as required by the specification).
13323 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13325 * Change X509_certificate_type() to mark the key as exported/exportable
13326 when it's 512 *bits* long, not 512 bytes.
13330 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13332 * Countermeasure against the Klima-Pokorny-Rosa extension of
13333 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13334 a protocol version number mismatch like a decryption error
13335 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13339 * Turn on RSA blinding by default in the default implementation
13340 to avoid a timing attack. Applications that don't want it can call
13341 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13342 They would be ill-advised to do so in most cases.
13344 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13346 * Change RSA blinding code so that it works when the PRNG is not
13347 seeded (in this case, the secret RSA exponent is abused as
13348 an unpredictable seed -- if it is not unpredictable, there
13349 is no point in blinding anyway). Make RSA blinding thread-safe
13350 by remembering the creator's thread ID in rsa->blinding and
13351 having all other threads use local one-time blinding factors
13352 (this requires more computation than sharing rsa->blinding, but
13353 avoids excessive locking; and if an RSA object is not shared
13354 between threads, blinding will still be very fast).
13358 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13360 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13361 via timing by performing a MAC computation even if incorrect
13362 block cipher padding has been found. This is a countermeasure
13363 against active attacks where the attacker has to distinguish
13364 between bad padding and a MAC verification error. ([CVE-2003-0078])
13366 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13367 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13368 Martin Vuagnoux (EPFL, Ilion)*
13370 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13372 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13373 memory from its contents. This is done with a counter that will
13374 place alternating values in each byte. This can be used to solve
13375 two issues: 1) the removal of calls to memset() by highly optimizing
13376 compilers, and 2) cleansing with other values than 0, since those can
13377 be read through on certain media, for example a swap space on disk.
13381 * Bugfix: client side session caching did not work with external caching,
13382 because the session->cipher setting was not restored when reloading
13383 from the external cache. This problem was masked, when
13384 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13385 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13389 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13390 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13392 *Zeev Lieber <zeev-l@yahoo.com>*
13394 * Undo an undocumented change introduced in 0.9.6e which caused
13395 repeated calls to OpenSSL_add_all_ciphers() and
13396 OpenSSL_add_all_digests() to be ignored, even after calling
13401 * Change the default configuration reader to deal with last line not
13402 being properly terminated.
13406 * Change X509_NAME_cmp() so it applies the special rules on handling
13407 DN values that are of type PrintableString, as well as RDNs of type
13408 emailAddress where the value has the type ia5String.
13410 *stefank@valicert.com via Richard Levitte*
13412 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13413 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13414 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13415 the bitwise-OR of the two for use by the majority of applications
13416 wanting this behaviour, and update the docs. The documented
13417 behaviour and actual behaviour were inconsistent and had been
13418 changing anyway, so this is more a bug-fix than a behavioural
13421 *Geoff Thorpe, diagnosed by Nadav Har'El*
13423 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13424 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13428 * Fix initialization code race conditions in
13429 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13430 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13431 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13432 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13433 ssl2_get_cipher_by_char(),
13434 ssl3_get_cipher_by_char().
13436 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13438 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13439 the cached sessions are flushed, as the remove_cb() might use ex_data
13440 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13441 (see [openssl.org #212]).
13443 *Geoff Thorpe, Lutz Jaenicke*
13445 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13446 length, instead of the encoding length to d2i_ASN1_OBJECT.
13450 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13452 * [In 0.9.6g-engine release:]
13453 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13455 *Lynn Gazis <lgazis@rainbow.com>*
13457 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13459 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13460 and get fix the header length calculation.
13461 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13462 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13464 * Use proper error handling instead of 'assertions' in buffer
13465 overflow checks added in 0.9.6e. This prevents DoS (the
13466 assertions could call abort()).
13468 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13470 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13472 * Add various sanity checks to asn1_get_length() to reject
13473 the ASN1 length bytes if they exceed sizeof(long), will appear
13474 negative or the content length exceeds the length of the
13477 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13479 * Fix cipher selection routines: ciphers without encryption had no flags
13480 for the cipher strength set and where therefore not handled correctly
13481 by the selection routines (PR #130).
13485 * Fix EVP_dsa_sha macro.
13490 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13491 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13492 that was added in OpenSSL 0.9.6d.
13494 As the countermeasure turned out to be incompatible with some
13495 broken SSL implementations, the new option is part of SSL_OP_ALL.
13496 SSL_OP_ALL is usually employed when compatibility with weird SSL
13497 implementations is desired (e.g. '-bugs' option to 's_client' and
13498 's_server'), so the new option is automatically set in many
13503 * Changes in security patch:
13505 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13506 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13507 Air Force Materiel Command, USAF, under agreement number
13510 * Add various sanity checks to asn1_get_length() to reject
13511 the ASN1 length bytes if they exceed sizeof(long), will appear
13512 negative or the content length exceeds the length of the
13513 supplied buffer. ([CVE-2002-0659])
13515 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13517 * Assertions for various potential buffer overflows, not known to
13518 happen in practice.
13520 *Ben Laurie (CHATS)*
13522 * Various temporary buffers to hold ASCII versions of integers were
13523 too small for 64 bit platforms. ([CVE-2002-0655])
13524 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13526 * Remote buffer overflow in SSL3 protocol - an attacker could
13527 supply an oversized session ID to a client. ([CVE-2002-0656])
13529 *Ben Laurie (CHATS)*
13531 * Remote buffer overflow in SSL2 protocol - an attacker could
13532 supply an oversized client master key. ([CVE-2002-0656])
13534 *Ben Laurie (CHATS)*
13536 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13538 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13539 encoded as NULL) with id-dsa-with-sha1.
13541 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13543 * Check various `X509_...()` return values in apps/req.c.
13545 *Nils Larsch <nla@trustcenter.de>*
13547 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13548 an end-of-file condition would erroneously be flagged, when the CRLF
13549 was just at the end of a processed block. The bug was discovered when
13550 processing data through a buffering memory BIO handing the data to a
13551 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13552 <ptsekov@syntrex.com> and Nedelcho Stanev.
13556 * Implement a countermeasure against a vulnerability recently found
13557 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13558 before application data chunks to avoid the use of known IVs
13559 with data potentially chosen by the attacker.
13563 * Fix length checks in ssl3_get_client_hello().
13567 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13568 to prevent ssl3_read_internal() from incorrectly assuming that
13569 ssl3_read_bytes() found application data while handshake
13570 processing was enabled when in fact s->s3->in_read_app_data was
13571 merely automatically cleared during the initial handshake.
13573 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13575 * Fix object definitions for Private and Enterprise: they were not
13576 recognized in their shortname (=lowercase) representation. Extend
13577 obj_dat.pl to issue an error when using undefined keywords instead
13578 of silently ignoring the problem (Svenning Sorensen
13579 <sss@sss.dnsalias.net>).
13583 * Fix DH_generate_parameters() so that it works for 'non-standard'
13584 generators, i.e. generators other than 2 and 5. (Previously, the
13585 code did not properly initialise the 'add' and 'rem' values to
13586 BN_generate_prime().)
13588 In the new general case, we do not insist that 'generator' is
13589 actually a primitive root: This requirement is rather pointless;
13590 a generator of the order-q subgroup is just as good, if not
13595 * Map new X509 verification errors to alerts. Discovered and submitted by
13596 Tom Wu <tom@arcot.com>.
13600 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13601 returning non-zero before the data has been completely received
13602 when using non-blocking I/O.
13604 *Bodo Moeller; problem pointed out by John Hughes*
13606 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13608 *Ben Laurie, Lutz Jaenicke*
13610 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13611 Yoram Zahavi <YoramZ@gilian.com>).
13615 * Add information about CygWin 1.3 and on, and preserve proper
13616 configuration for the versions before that.
13618 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13620 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13621 check whether we deal with a copy of a session and do not delete from
13622 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13623 <izhar@checkpoint.com>.
13627 * Do not store session data into the internal session cache, if it
13628 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13629 flag is set). Proposed by Aslam <aslam@funk.com>.
13633 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13638 * [In 0.9.6d-engine release:]
13639 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13641 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13643 * Add the configuration target linux-s390x.
13645 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13647 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13648 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13649 variable as an indication that a ClientHello message has been
13650 received. As the flag value will be lost between multiple
13651 invocations of ssl3_accept when using non-blocking I/O, the
13652 function may not be aware that a handshake has actually taken
13653 place, thus preventing a new session from being added to the
13656 To avoid this problem, we now set s->new_session to 2 instead of
13657 using a local variable.
13659 *Lutz Jaenicke, Bodo Moeller*
13661 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13662 if the SSL_R_LENGTH_MISMATCH error is detected.
13664 *Geoff Thorpe, Bodo Moeller*
13666 * New 'shared_ldflag' column in Configure platform table.
13670 * Fix EVP_CIPHER_mode macro.
13672 *"Dan S. Camper" <dan@bti.net>*
13674 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13675 type, we must throw them away by setting rr->length to 0.
13677 *D P Chang <dpc@qualys.com>*
13679 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13681 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13682 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13683 worked incorrectly for those cases where range = `10..._2` and
13684 `3*range` is two bits longer than range.)
13688 * Only add signing time to PKCS7 structures if it is not already
13693 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13694 OBJ_ld_ce should be OBJ_id_ce.
13695 Also some ip-pda OIDs in crypto/objects/objects.txt were
13696 incorrect (cf. RFC 3039).
13698 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13700 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13701 returns early because it has nothing to do.
13703 *Andy Schneider <andy.schneider@bjss.co.uk>*
13705 * [In 0.9.6c-engine release:]
13706 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13708 *Andy Schneider <andy.schneider@bjss.co.uk>*
13710 * [In 0.9.6c-engine release:]
13711 Add support for Cryptographic Appliance's keyserver technology.
13712 (Use engine 'keyclient')
13714 *Cryptographic Appliances and Geoff Thorpe*
13716 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13717 is called via tools/c89.sh because arguments have to be
13718 rearranged (all '-L' options must appear before the first object
13721 *Richard Shapiro <rshapiro@abinitio.com>*
13723 * [In 0.9.6c-engine release:]
13724 Add support for Broadcom crypto accelerator cards, backported
13727 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13729 * [In 0.9.6c-engine release:]
13730 Add support for SureWare crypto accelerator cards from
13731 Baltimore Technologies. (Use engine 'sureware')
13733 *Baltimore Technologies and Mark Cox*
13735 * [In 0.9.6c-engine release:]
13736 Add support for crypto accelerator cards from Accelerated
13737 Encryption Processing, www.aep.ie. (Use engine 'aep')
13739 *AEP Inc. and Mark Cox*
13741 * Add a configuration entry for gcc on UnixWare.
13743 *Gary Benson <gbenson@redhat.com>*
13745 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13746 messages are stored in a single piece (fixed-length part and
13747 variable-length part combined) and fix various bugs found on the way.
13751 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13752 instead. BIO_gethostbyname() does not know what timeouts are
13753 appropriate, so entries would stay in cache even when they have
13755 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13757 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13758 faced with a pathologically small ClientHello fragment that does
13759 not contain client_version: Instead of aborting with an error,
13760 simply choose the highest available protocol version (i.e.,
13761 TLS 1.0 unless it is disabled). In practice, ClientHello
13762 messages are never sent like this, but this change gives us
13763 strictly correct behaviour at least for TLS.
13767 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13768 never resets s->method to s->ctx->method when called from within
13769 one of the SSL handshake functions.
13771 *Bodo Moeller; problem pointed out by Niko Baric*
13773 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13774 (sent using the client's version number) if client_version is
13775 smaller than the protocol version in use. Also change
13776 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13777 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13778 the client will at least see that alert.
13782 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13787 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13788 client receives HelloRequest while in a handshake.
13790 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13792 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13793 should end in 'break', not 'goto end' which circumvents various
13794 cleanups done in state SSL_ST_OK. But session related stuff
13795 must be disabled for SSL_ST_OK in the case that we just sent a
13798 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13799 before just sending a HelloRequest.
13801 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13803 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13804 reveal whether illegal block cipher padding was found or a MAC
13805 verification error occurred. (Neither SSLerr() codes nor alerts
13806 are directly visible to potential attackers, but the information
13807 may leak via logfiles.)
13809 Similar changes are not required for the SSL 2.0 implementation
13810 because the number of padding bytes is sent in clear for SSL 2.0,
13811 and the extra bytes are just ignored. However ssl/s2_pkt.c
13812 failed to verify that the purported number of padding bytes is in
13817 * Add OpenUNIX-8 support including shared libraries
13818 (Boyd Lynn Gerber <gerberb@zenez.com>).
13822 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13823 'wristwatch attack' using huge encoding parameters (cf.
13824 James H. Manger's CRYPTO 2001 paper). Note that the
13825 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13826 encoding parameters and hence was not vulnerable.
13830 * BN_sqr() bug fix.
13832 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13834 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13835 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13836 followed by modular reduction.
13838 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13840 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13841 equivalent based on BN_pseudo_rand() instead of BN_rand().
13845 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13846 This function was broken, as the check for a new client hello message
13847 to handle SGC did not allow these large messages.
13848 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13852 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13856 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13857 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13861 * Rework the configuration and shared library support for Tru64 Unix.
13862 The configuration part makes use of modern compiler features and
13863 still retains old compiler behavior for those that run older versions
13864 of the OS. The shared library support part includes a variant that
13865 uses the RPATH feature, and is available through the special
13866 configuration target "alpha-cc-rpath", which will never be selected
13869 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13871 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13872 with the same message size as in ssl3_get_certificate_request().
13873 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13874 messages might inadvertently be reject as too long.
13876 *Petr Lampa <lampa@fee.vutbr.cz>*
13878 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13882 * Modified SSL library such that the verify_callback that has been set
13883 specifically for an SSL object with SSL_set_verify() is actually being
13884 used. Before the change, a verify_callback set with this function was
13885 ignored and the verify_callback() set in the SSL_CTX at the time of
13886 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
13887 to allow the necessary settings.
13891 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
13892 explicitly to NULL, as at least on Solaris 8 this seems not always to be
13893 done automatically (in contradiction to the requirements of the C
13894 standard). This made problems when used from OpenSSH.
13898 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
13899 dh->length and always used
13901 BN_rand_range(priv_key, dh->p).
13903 BN_rand_range() is not necessary for Diffie-Hellman, and this
13904 specific range makes Diffie-Hellman unnecessarily inefficient if
13905 dh->length (recommended exponent length) is much smaller than the
13906 length of dh->p. We could use BN_rand_range() if the order of
13907 the subgroup was stored in the DH structure, but we only have
13912 BN_rand(priv_key, l, ...)
13914 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
13921 RSA_eay_public_encrypt
13922 RSA_eay_private_decrypt
13923 RSA_eay_private_encrypt (signing)
13924 RSA_eay_public_decrypt (signature verification)
13926 (default implementations for RSA_public_encrypt,
13927 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
13928 always reject numbers >= n.
13932 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
13933 to synchronize access to 'locking_thread'. This is necessary on
13934 systems where access to 'locking_thread' (an 'unsigned long'
13935 variable) is not atomic.
13939 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
13940 *before* setting the 'crypto_lock_rand' flag. The previous code had
13941 a race condition if 0 is a valid thread ID.
13943 *Travis Vitek <vitek@roguewave.com>*
13945 * Add support for shared libraries under Irix.
13947 *Albert Chin-A-Young <china@thewrittenword.com>*
13949 * Add configuration option to build on Linux on both big-endian and
13950 little-endian MIPS.
13952 *Ralf Baechle <ralf@uni-koblenz.de>*
13954 * Add the possibility to create shared libraries on HP-UX.
13958 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
13960 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
13961 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
13962 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
13963 PRNG state recovery was possible based on the output of
13964 one PRNG request appropriately sized to gain knowledge on
13965 'md' followed by enough consecutive 1-byte PRNG requests
13966 to traverse all of 'state'.
13968 1. When updating 'md_local' (the current thread's copy of 'md')
13969 during PRNG output generation, hash all of the previous
13970 'md_local' value, not just the half used for PRNG output.
13972 2. Make the number of bytes from 'state' included into the hash
13973 independent from the number of PRNG bytes requested.
13975 The first measure alone would be sufficient to avoid
13976 Markku-Juhani's attack. (Actually it had never occurred
13977 to me that the half of 'md_local' used for chaining was the
13978 half from which PRNG output bytes were taken -- I had always
13979 assumed that the secret half would be used.) The second
13980 measure makes sure that additional data from 'state' is never
13981 mixed into 'md_local' in small portions; this heuristically
13982 further strengthens the PRNG.
13986 * Fix crypto/bn/asm/mips3.s.
13990 * When only the key is given to "enc", the IV is undefined. Print out
13991 an error message in this case.
13995 * Handle special case when X509_NAME is empty in X509 printing routines.
13999 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14000 positive and less than q.
14004 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14005 used: it isn't thread safe and the add_lock_callback should handle
14008 *Paul Rose <Paul.Rose@bridge.com>*
14010 * Verify that incoming data obeys the block size in
14011 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14017 *Ulf Möller, Bodo Möller*
14019 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14020 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14021 when fixing the server behaviour for backwards-compatible 'client
14022 hello' messages. (Note that the attack is impractical against
14023 SSL 3.0 and TLS 1.0 anyway because length and version checking
14024 means that the probability of guessing a valid ciphertext is
14025 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14028 Before 0.9.5, the countermeasure (hide the error by generating a
14029 random 'decryption result') did not work properly because
14030 ERR_clear_error() was missing, meaning that SSL_get_error() would
14031 detect the supposedly ignored error.
14033 Both problems are now fixed.
14037 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14038 (previously it was 1024).
14042 * Fix for compatibility mode trust settings: ignore trust settings
14043 unless some valid trust or reject settings are present.
14047 * Fix for blowfish EVP: its a variable length cipher.
14051 * Fix various bugs related to DSA S/MIME verification. Handle missing
14052 parameters in DSA public key structures and return an error in the
14053 DSA routines if parameters are absent.
14057 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14058 in the current directory if neither $RANDFILE nor $HOME was set.
14059 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14060 caused some confusion to Windows users who haven't defined $HOME.
14061 Thus RAND_file_name() is changed again: e_os.h can define a
14062 DEFAULT_HOME, which will be used if $HOME is not set.
14063 For Windows, we use "C:"; on other platforms, we still require
14064 environment variables.
14066 * Move 'if (!initialized) RAND_poll()' into regions protected by
14067 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14068 having multiple threads call RAND_poll() concurrently.
14072 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14073 combination of a flag and a thread ID variable.
14074 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14075 flag), *other* threads can enter ssleay_add_bytes without obeying
14076 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14077 that they do not hold after the first thread unsets add_do_not_lock).
14081 * Change bctest again: '-x' expressions are not available in all
14082 versions of 'test'.
14086 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14088 * Fix a couple of memory leaks in PKCS7_dataDecode()
14090 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14092 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14093 the default extension for executables, if any. Also, make the perl
14094 scripts that use symlink() to test if it really exists and use "cp"
14095 if it doesn't. All this made OpenSSL compilable and installable in
14100 * Fix for asn1_GetSequence() for indefinite length constructed data.
14101 If SEQUENCE is length is indefinite just set c->slen to the total
14102 amount of data available.
14104 *Steve Henson, reported by shige@FreeBSD.org*
14106 *This change does not apply to 0.9.7.*
14108 * Change bctest to avoid here-documents inside command substitution
14109 (workaround for FreeBSD /bin/sh bug).
14110 For compatibility with Ultrix, avoid shell functions (introduced
14111 in the bctest version that searches along $PATH).
14115 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14116 with des_encrypt() defined on some operating systems, like Solaris
14121 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14122 On the Importance of Eliminating Errors in Cryptographic
14123 Computations, J. Cryptology 14 (2001) 2, 101-119,
14124 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14128 * MIPS assembler BIGNUM division bug fix.
14132 * Disabled incorrect Alpha assembler code.
14136 * Fix PKCS#7 decode routines so they correctly update the length
14137 after reading an EOC for the EXPLICIT tag.
14141 *This change does not apply to 0.9.7.*
14143 * Fix bug in PKCS#12 key generation routines. This was triggered
14144 if a 3DES key was generated with a 0 initial byte. Include
14145 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14146 (but broken) behaviour.
14150 * Enhance bctest to search for a working bc along $PATH and print
14153 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14155 * Fix memory leaks in err.c: free err_data string if necessary;
14156 don't write to the wrong index in ERR_set_error_data.
14160 * Implement ssl23_peek (analogous to ssl23_read), which previously
14165 * Replace rdtsc with `_emit` statements for VC++ version 5.
14167 *Jeremy Cooper <jeremy@baymoo.org>*
14169 * Make it possible to reuse SSLv2 sessions.
14173 * In copy_email() check for >= 0 as a return value for
14174 X509_NAME_get_index_by_NID() since 0 is a valid index.
14176 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14178 * Avoid coredump with unsupported or invalid public keys by checking if
14179 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14180 PKCS7_verify() fails with non detached data.
14184 * Don't use getenv in library functions when run as setuid/setgid.
14185 New function OPENSSL_issetugid().
14189 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14190 due to incorrect handling of multi-threading:
14192 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14194 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14196 3. Count how many times MemCheck_off() has been called so that
14197 nested use can be treated correctly. This also avoids
14198 inband-signalling in the previous code (which relied on the
14199 assumption that thread ID 0 is impossible).
14203 * Add "-rand" option also to s_client and s_server.
14207 * Fix CPU detection on Irix 6.x.
14208 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14209 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14211 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14216 *This change does not apply to 0.9.7.*
14218 * Use the cached encoding of an X509_NAME structure rather than
14219 copying it. This is apparently the reason for the libsafe "errors"
14220 but the code is actually correct.
14224 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14225 Bleichenbacher's DSA attack.
14226 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14227 to be set and top=0 forces the highest bit to be set; top=-1 is new
14228 and leaves the highest bit random.
14230 *Ulf Moeller, Bodo Moeller*
14232 * In the `NCONF_...`-based implementations for `CONF_...` queries
14233 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14234 a temporary CONF structure with the data component set to NULL
14235 (which gives segmentation faults in lh_retrieve).
14236 Instead, use NULL for the CONF pointer in CONF_get_string and
14237 CONF_get_number (which may use environment variables) and directly
14238 return NULL from CONF_get_section.
14242 * Fix potential buffer overrun for EBCDIC.
14246 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14247 keyUsage if basicConstraints absent for a CA.
14251 * Make SMIME_write_PKCS7() write mail header values with a format that
14252 is more generally accepted (no spaces before the semicolon), since
14253 some programs can't parse those values properly otherwise. Also make
14254 sure BIO's that break lines after each write do not create invalid
14259 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14260 macros previously used would not encode an empty SEQUENCE OF
14261 and break the signature.
14265 *This change does not apply to 0.9.7.*
14267 * Zero the premaster secret after deriving the master secret in
14272 * Add some EVP_add_digest_alias registrations (as found in
14273 OpenSSL_add_all_digests()) to SSL_library_init()
14274 aka OpenSSL_add_ssl_algorithms(). This provides improved
14275 compatibility with peers using X.509 certificates
14276 with unconventional AlgorithmIdentifier OIDs.
14280 * Fix for Irix with NO_ASM.
14282 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14284 * ./config script fixes.
14286 *Ulf Moeller, Richard Levitte*
14288 * Fix 'openssl passwd -1'.
14292 * Change PKCS12_key_gen_asc() so it can cope with non null
14293 terminated strings whose length is passed in the passlen
14294 parameter, for example from PEM callbacks. This was done
14295 by adding an extra length parameter to asc2uni().
14297 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14299 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14300 call failed, free the DSA structure.
14304 * Fix to uni2asc() to cope with zero length Unicode strings.
14305 These are present in some PKCS#12 files.
14309 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14310 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14311 when writing a 32767 byte record.
14313 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14315 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14316 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14318 (RSA objects have a reference count access to which is protected
14319 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14320 so they are meant to be shared between threads.)
14321 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14322 "Reddie, Steven" <Steven.Reddie@ca.com>*
14324 * Fix a deadlock in CRYPTO_mem_leaks().
14328 * Use better test patterns in bntest.
14332 * rand_win.c fix for Borland C.
14336 * BN_rshift bugfix for n == 0.
14340 * Add a 'bctest' script that checks for some known 'bc' bugs
14341 so that 'make test' does not abort just because 'bc' is broken.
14345 * Store verify_result within SSL_SESSION also for client side to
14346 avoid potential security hole. (Re-used sessions on the client side
14347 always resulted in verify_result==X509_V_OK, not using the original
14348 result of the server certificate verification.)
14352 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14353 SSL3_RT_APPLICATION_DATA, return 0.
14354 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14359 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14360 releases, have been re-implemented by renaming the previous
14361 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14362 and ssl3_read_internal, respectively, and adding 'peek' parameters
14363 to them. The new ssl[23]_{read,peek} functions are calls to
14364 ssl[23]_read_internal with the 'peek' flag set appropriately.
14365 A 'peek' parameter has also been added to ssl3_read_bytes, which
14366 does the actual work for ssl3_read_internal.
14370 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14371 the method-specific "init()" handler. Also clean up ex_data after
14372 calling the method-specific "finish()" handler. Previously, this was
14373 happening the other way round.
14377 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14378 The previous value, 12, was not always sufficient for BN_mod_exp().
14382 * Make sure that shared libraries get the internal name engine with
14383 the full version number and not just 0. This should mark the
14384 shared libraries as not backward compatible. Of course, this should
14385 be changed again when we can guarantee backward binary compatibility.
14389 * Fix typo in get_cert_by_subject() in by_dir.c
14391 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14393 * Rework the system to generate shared libraries:
14395 - Make note of the expected extension for the shared libraries and
14396 if there is a need for symbolic links from for example libcrypto.so.0
14397 to libcrypto.so.0.9.7. There is extended info in Configure for
14400 - Make as few rebuilds of the shared libraries as possible.
14402 - Still avoid linking the OpenSSL programs with the shared libraries.
14404 - When installing, install the shared libraries separately from the
14409 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14411 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14412 and not in SSL_clear because the latter is also used by the
14413 accept/connect functions; previously, the settings made by
14414 SSL_set_read_ahead would be lost during the handshake.
14416 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14418 * Correct util/mkdef.pl to be selective about disabled algorithms.
14419 Previously, it would create entries for disabled algorithms no
14424 * Added several new manual pages for SSL_* function.
14428 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14430 * In ssl23_get_client_hello, generate an error message when faced
14431 with an initial SSL 3.0/TLS record that is too small to contain the
14432 first two bytes of the ClientHello message, i.e. client_version.
14433 (Note that this is a pathologic case that probably has never happened
14434 in real life.) The previous approach was to use the version number
14435 from the record header as a substitute; but our protocol choice
14436 should not depend on that one because it is not authenticated
14437 by the Finished messages.
14441 * More robust randomness gathering functions for Windows.
14443 *Jeffrey Altman <jaltman@columbia.edu>*
14445 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14446 not set then we don't setup the error code for issuer check errors
14447 to avoid possibly overwriting other errors which the callback does
14448 handle. If an application does set the flag then we assume it knows
14449 what it is doing and can handle the new informational codes
14454 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14455 a general "ANY" type, as such it should be able to decode anything
14456 including tagged types. However it didn't check the class so it would
14457 wrongly interpret tagged types in the same way as their universal
14458 counterpart and unknown types were just rejected. Changed so that the
14459 tagged and unknown types are handled in the same way as a SEQUENCE:
14460 that is the encoding is stored intact. There is also a new type
14461 "V_ASN1_OTHER" which is used when the class is not universal, in this
14462 case we have no idea what the actual type is so we just lump them all
14467 * On VMS, stdout may very well lead to a file that is written to
14468 in a record-oriented fashion. That means that every write() will
14469 write a separate record, which will be read separately by the
14470 programs trying to read from it. This can be very confusing.
14472 The solution is to put a BIO filter in the way that will buffer
14473 text until a linefeed is reached, and then write everything a
14474 line at a time, so every record written will be an actual line,
14475 not chunks of lines and not (usually doesn't happen, but I've
14476 seen it once) several lines in one record. BIO_f_linebuffer() is
14479 Currently, it's a VMS-only method, because that's where it has
14480 been tested well enough.
14484 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14485 it can return incorrect results.
14486 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14487 but it was in 0.9.6-beta[12].)
14491 * Disable the check for content being present when verifying detached
14492 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14493 include zero length content when signing messages.
14497 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14498 BIO_ctrl (for BIO pairs).
14502 * Add DSO method for VMS.
14506 * Bug fix: Montgomery multiplication could produce results with the
14511 * Add RPM specification openssl.spec and modify it to build three
14512 packages. The default package contains applications, application
14513 documentation and run-time libraries. The devel package contains
14514 include files, static libraries and function documentation. The
14515 doc package contains the contents of the doc directory. The original
14516 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14520 * Add a large number of documentation files for many SSL routines.
14522 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14524 * Add a configuration entry for Sony News 4.
14526 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14528 * Don't set the two most significant bits to one when generating a
14529 random number < q in the DSA library.
14533 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14534 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14535 the underlying transport is blocking) if a handshake took place.
14536 (The default behaviour is needed by applications such as s_client
14537 and s_server that use select() to determine when to use SSL_read;
14538 but for applications that know in advance when to expect data, it
14539 just makes things more complicated.)
14543 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14548 * Add a few more EBCDIC conditionals that make `req` and `x509`
14549 work better on such systems.
14551 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14553 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14554 Update PKCS12_parse() so it copies the friendlyName and the
14555 keyid to the certificates aux info.
14559 * Fix bug in PKCS7_verify() which caused an infinite loop
14560 if there was more than one signature.
14562 *Sven Uszpelkat <su@celocom.de>*
14564 * Major change in util/mkdef.pl to include extra information
14565 about each symbol, as well as presenting variables as well
14566 as functions. This change means that there's n more need
14567 to rebuild the .num files when some algorithms are excluded.
14571 * Allow the verify time to be set by an application,
14572 rather than always using the current time.
14576 * Phase 2 verify code reorganisation. The certificate
14577 verify code now looks up an issuer certificate by a
14578 number of criteria: subject name, authority key id
14579 and key usage. It also verifies self signed certificates
14580 by the same criteria. The main comparison function is
14581 X509_check_issued() which performs these checks.
14583 Lot of changes were necessary in order to support this
14584 without completely rewriting the lookup code.
14586 Authority and subject key identifier are now cached.
14588 The LHASH 'certs' is X509_STORE has now been replaced
14589 by a STACK_OF(X509_OBJECT). This is mainly because an
14590 LHASH can't store or retrieve multiple objects with
14591 the same hash value.
14593 As a result various functions (which were all internal
14594 use only) have changed to handle the new X509_STORE
14595 structure. This will break anything that messed round
14596 with X509_STORE internally.
14598 The functions X509_STORE_add_cert() now checks for an
14599 exact match, rather than just subject name.
14601 The X509_STORE API doesn't directly support the retrieval
14602 of multiple certificates matching a given criteria, however
14603 this can be worked round by performing a lookup first
14604 (which will fill the cache with candidate certificates)
14605 and then examining the cache for matches. This is probably
14606 the best we can do without throwing out X509_LOOKUP
14607 entirely (maybe later...).
14609 The X509_VERIFY_CTX structure has been enhanced considerably.
14611 All certificate lookup operations now go via a get_issuer()
14612 callback. Although this currently uses an X509_STORE it
14613 can be replaced by custom lookups. This is a simple way
14614 to bypass the X509_STORE hackery necessary to make this
14615 work and makes it possible to use more efficient techniques
14616 in future. A very simple version which uses a simple
14617 STACK for its trusted certificate store is also provided
14618 using X509_STORE_CTX_trusted_stack().
14620 The verify_cb() and verify() callbacks now have equivalents
14621 in the X509_STORE_CTX structure.
14623 X509_STORE_CTX also has a 'flags' field which can be used
14624 to customise the verify behaviour.
14628 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14629 excludes S/MIME capabilities.
14633 * When a certificate request is read in keep a copy of the
14634 original encoding of the signed data and use it when outputting
14635 again. Signatures then use the original encoding rather than
14636 a decoded, encoded version which may cause problems if the
14637 request is improperly encoded.
14641 * For consistency with other BIO_puts implementations, call
14642 buffer_write(b, ...) directly in buffer_puts instead of calling
14645 In BIO_puts, increment b->num_write as in BIO_write.
14647 *Peter.Sylvester@EdelWeb.fr*
14649 * Fix BN_mul_word for the case where the word is 0. (We have to use
14650 BN_zero, we may not return a BIGNUM with an array consisting of
14651 words set to zero.)
14655 * Avoid calling abort() from within the library when problems are
14656 detected, except if preprocessor symbols have been defined
14657 (such as REF_CHECK, BN_DEBUG etc.).
14661 * New openssl application 'rsautl'. This utility can be
14662 used for low level RSA operations. DER public key
14663 BIO/fp routines also added.
14667 * New Configure entry and patches for compiling on QNX 4.
14669 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14671 * A demo state-machine implementation was sponsored by
14672 Nuron (<http://www.nuron.com/>) and is now available in
14673 demos/state_machine.
14677 * New options added to the 'dgst' utility for signature
14678 generation and verification.
14682 * Unrecognized PKCS#7 content types are now handled via a
14683 catch all ASN1_TYPE structure. This allows unsupported
14684 types to be stored as a "blob" and an application can
14685 encode and decode it manually.
14689 * Fix various signed/unsigned issues to make a_strex.c
14690 compile under VC++.
14692 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14694 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14695 length if passed a buffer. ASN1_INTEGER_to_BN failed
14696 if passed a NULL BN and its argument was negative.
14698 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14700 * Modification to PKCS#7 encoding routines to output definite
14701 length encoding. Since currently the whole structures are in
14702 memory there's not real point in using indefinite length
14703 constructed encoding. However if OpenSSL is compiled with
14704 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14708 * Added BIO_vprintf() and BIO_vsnprintf().
14712 * Added more prefixes to parse for in the strings written
14713 through a logging bio, to cover all the levels that are available
14714 through syslog. The prefixes are now:
14716 PANIC, EMERG, EMR => LOG_EMERG
14717 ALERT, ALR => LOG_ALERT
14718 CRIT, CRI => LOG_CRIT
14719 ERROR, ERR => LOG_ERR
14720 WARNING, WARN, WAR => LOG_WARNING
14721 NOTICE, NOTE, NOT => LOG_NOTICE
14722 INFO, INF => LOG_INFO
14723 DEBUG, DBG => LOG_DEBUG
14725 and as before, if none of those prefixes are present at the
14726 beginning of the string, LOG_ERR is chosen.
14728 On Win32, the `LOG_*` levels are mapped according to this:
14730 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14731 LOG_WARNING => EVENTLOG_WARNING_TYPE
14732 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14736 * Made it possible to reconfigure with just the configuration
14737 argument "reconf" or "reconfigure". The command line arguments
14738 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14739 and are retrieved from there when reconfiguring.
14745 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14747 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14751 * The obj_dat.pl script was messing up the sorting of object
14752 names. The reason was that it compared the quoted version
14753 of strings as a result "OCSP" > "OCSP Signing" because
14754 " > SPACE. Changed script to store unquoted versions of
14755 names and add quotes on output. It was also omitting some
14756 names from the lookup table if they were given a default
14757 value (that is if SN is missing it is given the same
14758 value as LN and vice versa), these are now added on the
14759 grounds that if an object has a name we should be able to
14760 look it up. Finally added warning output when duplicate
14761 short or long names are found.
14765 * Changes needed for Tandem NSK.
14767 *Scott Uroff <scott@xypro.com>*
14769 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14770 RSA_padding_check_SSLv23(), special padding was never detected
14771 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14772 version rollback attacks was not effective.
14774 In s23_clnt.c, don't use special rollback-attack detection padding
14775 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14776 client; similarly, in s23_srvr.c, don't do the rollback check if
14777 SSL 2.0 is the only protocol enabled in the server.
14781 * Make it possible to get hexdumps of unprintable data with 'openssl
14782 asn1parse'. By implication, the functions ASN1_parse_dump() and
14783 BIO_dump_indent() are added.
14787 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14788 these print out strings and name structures based on various
14789 flags including RFC2253 support and proper handling of
14790 multibyte characters. Added options to the 'x509' utility
14791 to allow the various flags to be set.
14795 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14796 Also change the functions X509_cmp_current_time() and
14797 X509_gmtime_adj() work with an ASN1_TIME structure,
14798 this will enable certificates using GeneralizedTime in validity
14799 dates to be checked.
14803 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14804 negative public key encodings) on by default,
14805 NO_NEG_PUBKEY_BUG can be set to disable it.
14809 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14810 content octets. An i2c_ASN1_OBJECT is unnecessary because
14811 the encoding can be trivially obtained from the structure.
14815 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14816 not read locks (`CRYPTO_r_[un]lock`).
14820 * A first attempt at creating official support for shared
14821 libraries through configuration. I've kept it so the
14822 default is static libraries only, and the OpenSSL programs
14823 are always statically linked for now, but there are
14824 preparations for dynamic linking in place.
14825 This has been tested on Linux and Tru64.
14829 * Randomness polling function for Win9x, as described in:
14830 Peter Gutmann, Software Generation of Practically Strong
14835 * Fix so PRNG is seeded in req if using an already existing
14840 * New options to smime application. -inform and -outform
14841 allow alternative formats for the S/MIME message including
14842 PEM and DER. The -content option allows the content to be
14843 specified separately. This should allow things like Netscape
14844 form signing output easier to verify.
14848 * Fix the ASN1 encoding of tags using the 'long form'.
14852 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14853 STRING types. These convert content octets to and from the
14854 underlying type. The actual tag and length octets are
14855 already assumed to have been read in and checked. These
14856 are needed because all other string types have virtually
14857 identical handling apart from the tag. By having versions
14858 of the ASN1 functions that just operate on content octets
14859 IMPLICIT tagging can be handled properly. It also allows
14860 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14861 and ASN1_INTEGER are identical apart from the tag.
14865 * Change the handling of OID objects as follows:
14867 - New object identifiers are inserted in objects.txt, following
14868 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14869 - objects.pl is used to process obj_mac.num and create a new
14871 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14874 This is currently kind of a hack, and the perl code in objects.pl
14875 isn't very elegant, but it works as I intended. The simplest way
14876 to check that it worked correctly is to look in obj_dat.h and
14877 check the array nid_objs and make sure the objects haven't moved
14878 around (this is important!). Additions are OK, as well as
14879 consistent name changes.
14883 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
14887 * Addition of the command line parameter '-rand file' to 'openssl req'.
14888 The given file adds to whatever has already been seeded into the
14889 random pool through the RANDFILE configuration file option or
14890 environment variable, or the default random state file.
14894 * mkstack.pl now sorts each macro group into lexical order.
14895 Previously the output order depended on the order the files
14896 appeared in the directory, resulting in needless rewriting
14901 * Patches to make OpenSSL compile under Win32 again. Mostly
14902 work arounds for the VC++ problem that it treats func() as
14903 func(void). Also stripped out the parts of mkdef.pl that
14904 added extra typesafe functions: these no longer exist.
14908 * Reorganisation of the stack code. The macros are now all
14909 collected in safestack.h . Each macro is defined in terms of
14910 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
14911 DEBUG_SAFESTACK is now handled in terms of function casts,
14912 this has the advantage of retaining type safety without the
14913 use of additional functions. If DEBUG_SAFESTACK is not defined
14914 then the non typesafe macros are used instead. Also modified the
14915 mkstack.pl script to handle the new form. Needs testing to see
14916 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
14917 the default if no major problems. Similar behaviour for ASN1_SET_OF
14918 and PKCS12_STACK_OF.
14922 * When some versions of IIS use the 'NET' form of private key the
14923 key derivation algorithm is different. Normally MD5(password) is
14924 used as a 128 bit RC4 key. In the modified case
14925 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
14926 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
14927 as the old Netscape_RSA functions except they have an additional
14928 'sgckey' parameter which uses the modified algorithm. Also added
14929 an -sgckey command line option to the rsa utility. Thanks to
14930 Adrian Peck <bertie@ncipher.com> for posting details of the modified
14931 algorithm to openssl-dev.
14935 * The evp_local.h macros were using 'c.##kname' which resulted in
14936 invalid expansion on some systems (SCO 5.0.5 for example).
14937 Corrected to 'c.kname'.
14939 *Phillip Porch <root@theporch.com>*
14941 * New X509_get1_email() and X509_REQ_get1_email() functions that return
14942 a STACK of email addresses from a certificate or request, these look
14943 in the subject name and the subject alternative name extensions and
14944 omit any duplicate addresses.
14948 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
14949 This makes DSA verification about 2 % faster.
14953 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
14954 (meaning that now 2^5 values will be precomputed, which is only 4 KB
14955 plus overhead for 1024 bit moduli).
14956 This makes exponentiations about 0.5 % faster for 1024 bit
14957 exponents (as measured by "openssl speed rsa2048").
14961 * Rename memory handling macros to avoid conflicts with other
14963 Malloc => OPENSSL_malloc
14964 Malloc_locked => OPENSSL_malloc_locked
14965 Realloc => OPENSSL_realloc
14966 Free => OPENSSL_free
14970 * New function BN_mod_exp_mont_word for small bases (roughly 15%
14971 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
14975 * CygWin32 support.
14977 *John Jarvie <jjarvie@newsguy.com>*
14979 * The type-safe stack code has been rejigged. It is now only compiled
14980 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
14981 by default all type-specific stack functions are "#define"d back to
14982 standard stack functions. This results in more streamlined output
14983 but retains the type-safety checking possibilities of the original
14988 * The STACK code has been cleaned up, and certain type declarations
14989 that didn't make a lot of sense have been brought in line. This has
14990 also involved a cleanup of sorts in safestack.h to more correctly
14991 map type-safe stack functions onto their plain stack counterparts.
14992 This work has also resulted in a variety of "const"ifications of
14993 lots of the code, especially `_cmp` operations which should normally
14994 be prototyped with "const" parameters anyway.
14998 * When generating bytes for the first time in md_rand.c, 'stir the pool'
14999 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15000 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15001 where all of 'md' is used each time the PRNG is used, but 'state'
15002 is used only indexed by a cyclic counter. As entropy may not be
15003 well distributed from the beginning, 'md' is important as a
15004 chaining variable. However, the output function chains only half
15005 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15006 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15007 in all of 'state' being rewritten, with the new values depending
15008 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15012 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15013 the handshake is continued after ssl_verify_cert_chain();
15014 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15015 can lead to 'unexplainable' connection aborts later.
15017 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15019 * Major EVP API cipher revision.
15020 Add hooks for extra EVP features. This allows various cipher
15021 parameters to be set in the EVP interface. Support added for variable
15022 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15023 setting of RC2 and RC5 parameters.
15025 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15028 Remove lots of duplicated code from the EVP library. For example *every*
15029 cipher init() function handles the 'iv' in the same way according to the
15030 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15031 for CFB and OFB modes they zero ctx->num.
15033 New functionality allows removal of S/MIME code RC2 hack.
15035 Most of the routines have the same form and so can be declared in terms
15038 By shifting this to the top level EVP_CipherInit() it can be removed from
15039 all individual ciphers. If the cipher wants to handle IVs or keys
15040 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15043 Change lots of functions like EVP_EncryptUpdate() to now return a
15044 value: although software versions of the algorithms cannot fail
15045 any installed hardware versions can.
15049 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15050 this option is set, tolerate broken clients that send the negotiated
15051 protocol version number instead of the requested protocol version
15056 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15057 i.e. non-zero for export ciphersuites, zero otherwise.
15058 Previous versions had this flag inverted, inconsistent with
15059 rsa_tmp_cb (..._TMP_RSA_CB).
15061 *Bodo Moeller; problem reported by Amit Chopra*
15063 * Add missing DSA library text string. Work around for some IIS
15064 key files with invalid SEQUENCE encoding.
15068 * Add a document (doc/standards.txt) that list all kinds of standards
15069 and so on that are implemented in OpenSSL.
15073 * Enhance c_rehash script. Old version would mishandle certificates
15074 with the same subject name hash and wouldn't handle CRLs at all.
15075 Added -fingerprint option to crl utility, to support new c_rehash
15080 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15084 * Fix for SSL server purpose checking. Server checking was
15085 rejecting certificates which had extended key usage present
15086 but no ssl client purpose.
15088 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15090 * Make PKCS#12 code work with no password. The PKCS#12 spec
15091 is a little unclear about how a blank password is handled.
15092 Since the password in encoded as a BMPString with terminating
15093 double NULL a zero length password would end up as just the
15094 double NULL. However no password at all is different and is
15095 handled differently in the PKCS#12 key generation code. NS
15096 treats a blank password as zero length. MSIE treats it as no
15097 password on export: but it will try both on import. We now do
15098 the same: PKCS12_parse() tries zero length and no password if
15099 the password is set to "" or NULL (NULL is now a valid password:
15100 it wasn't before) as does the pkcs12 application.
15104 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15105 perror when PEM_read_bio_X509_REQ fails, the error message must
15106 be obtained from the error queue.
15110 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15111 it in ERR_remove_state if appropriate, and change ERR_get_state
15112 accordingly to avoid race conditions (this is necessary because
15113 thread_hash is no longer constant once set).
15117 * Bugfix for linux-elf makefile.one.
15121 * RSA_get_default_method() will now cause a default
15122 RSA_METHOD to be chosen if one doesn't exist already.
15123 Previously this was only set during a call to RSA_new()
15124 or RSA_new_method(NULL) meaning it was possible for
15125 RSA_get_default_method() to return NULL.
15129 * Added native name translation to the existing DSO code
15130 that will convert (if the flag to do so is set) filenames
15131 that are sufficiently small and have no path information
15132 into a canonical native form. Eg. "blah" converted to
15133 "libblah.so" or "blah.dll" etc.
15137 * New function ERR_error_string_n(e, buf, len) which is like
15138 ERR_error_string(e, buf), but writes at most 'len' bytes
15139 including the 0 terminator. For ERR_error_string_n, 'buf'
15142 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15144 * CONF library reworked to become more general. A new CONF
15145 configuration file reader "class" is implemented as well as a
15146 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15147 old `CONF_*` functions are still there, but are reimplemented to
15148 work in terms of the new functions. Also, a set of functions
15149 to handle the internal storage of the configuration data is
15150 provided to make it easier to write new configuration file
15151 reader "classes" (I can definitely see something reading a
15152 configuration file in XML format, for example), called `_CONF_*`,
15153 or "the configuration storage API"...
15155 The new configuration file reading functions are:
15157 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15158 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15160 NCONF_default, NCONF_WIN32
15162 NCONF_dump_fp, NCONF_dump_bio
15164 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15165 NCONF_new creates a new CONF object. This works in the same way
15166 as other interfaces in OpenSSL, like the BIO interface.
15167 `NCONF_dump_*` dump the internal storage of the configuration file,
15168 which is useful for debugging. All other functions take the same
15169 arguments as the old `CONF_*` functions with the exception of the
15170 first that must be a `CONF *` instead of a `LHASH *`.
15172 To make it easier to use the new classes with the old `CONF_*` functions,
15173 the function CONF_set_default_method is provided.
15177 * Add '-tls1' option to 'openssl ciphers', which was already
15178 mentioned in the documentation but had not been implemented.
15179 (This option is not yet really useful because even the additional
15180 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15184 * Initial DSO code added into libcrypto for letting OpenSSL (and
15185 OpenSSL-based applications) load shared libraries and bind to
15186 them in a portable way.
15188 *Geoff Thorpe, with contributions from Richard Levitte*
15190 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15192 * Make sure _lrotl and _lrotr are only used with MSVC.
15194 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15195 (the default implementation of RAND_status).
15197 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15198 to '-clrext' (= clear extensions), as intended and documented.
15199 *Bodo Moeller; inconsistency pointed out by Michael Attili
15200 <attili@amaxo.com>*
15202 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15203 was larger than the MD block size.
15205 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15207 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15208 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15209 using the passed key: if the passed key was a private key the result
15210 of X509_print(), for example, would be to print out all the private key
15215 * des_quad_cksum() byte order bug fix.
15216 *Ulf Möller, using the problem description in krb4-0.9.7, where
15217 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15219 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15222 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15224 * For easily testing in shell scripts whether some command
15225 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15226 returns with exit code 0 iff no command of the given name is available.
15227 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15228 the output goes to stdout and nothing is printed to stderr.
15229 Additional arguments are always ignored.
15231 Since for each cipher there is a command of the same name,
15232 the 'no-cipher' compilation switches can be tested this way.
15234 ('openssl no-XXX' is not able to detect pseudo-commands such
15235 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15239 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15243 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15244 is set; it will be thrown away anyway because each handshake creates
15246 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15247 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15248 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15249 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15253 * New s_client option -ign_eof: EOF at stdin is ignored, and
15254 'Q' and 'R' lose their special meanings (quit/renegotiate).
15255 This is part of what -quiet does; unlike -quiet, -ign_eof
15256 does not suppress any output.
15260 * Add compatibility options to the purpose and trust code. The
15261 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15262 accepts a certificate or CA, this was the previous behaviour,
15263 with all the associated security issues.
15265 X509_TRUST_COMPAT is the old trust behaviour: only and
15266 automatically trust self signed roots in certificate store. A
15267 new trust setting X509_TRUST_DEFAULT is used to specify that
15268 a purpose has no associated trust setting and it should instead
15269 use the value in the default purpose.
15273 * Fix the PKCS#8 DSA private key code so it decodes keys again
15274 and fix a memory leak.
15278 * In util/mkerr.pl (which implements 'make errors'), preserve
15279 reason strings from the previous version of the .c file, as
15280 the default to have only downcase letters (and digits) in
15281 automatically generated reasons codes is not always appropriate.
15285 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15286 using strerror. Previously, ERR_reason_error_string() returned
15287 library names as reason strings for SYSerr; but SYSerr is a special
15288 case where small numbers are errno values, not library numbers.
15292 * Add '-dsaparam' option to 'openssl dhparam' application. This
15293 converts DSA parameters into DH parameters. (When creating parameters,
15294 DSA_generate_parameters is used.)
15298 * Include 'length' (recommended exponent length) in C code generated
15299 by 'openssl dhparam -C'.
15303 * The second argument to set_label in perlasm was already being used
15304 so couldn't be used as a "file scope" flag. Moved to third argument
15309 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15310 instead of RAND_bytes for encryption IVs and salts.
15314 * Include RAND_status() into RAND_METHOD instead of implementing
15315 it only for md_rand.c Otherwise replacing the PRNG by calling
15316 RAND_set_rand_method would be impossible.
15320 * Don't let DSA_generate_key() enter an infinite loop if the random
15321 number generation fails.
15325 * New 'rand' application for creating pseudo-random output.
15329 * Added configuration support for Linux/IA64
15331 *Rolf Haberrecker <rolf@suse.de>*
15333 * Assembler module support for Mingw32.
15337 * Shared library support for HPUX (in shlib/).
15339 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15341 * Shared library support for Solaris gcc.
15343 *Lutz Behnke <behnke@trustcenter.de>*
15345 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15347 * PKCS7_encrypt() was adding text MIME headers twice because they
15348 were added manually and by SMIME_crlf_copy().
15352 * In bntest.c don't call BN_rand with zero bits argument.
15354 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15356 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15357 case was implemented. This caused BN_div_recp() to fail occasionally.
15361 * Add an optional second argument to the set_label() in the perl
15362 assembly language builder. If this argument exists and is set
15363 to 1 it signals that the assembler should use a symbol whose
15364 scope is the entire file, not just the current function. This
15365 is needed with MASM which uses the format label:: for this scope.
15367 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15369 * Change the ASN1 types so they are typedefs by default. Before
15370 almost all types were #define'd to ASN1_STRING which was causing
15371 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15376 * Change names of new functions to the new get1/get0 naming
15377 convention: After 'get1', the caller owns a reference count
15378 and has to call `..._free`; 'get0' returns a pointer to some
15379 data structure without incrementing reference counters.
15380 (Some of the existing 'get' functions increment a reference
15381 counter, some don't.)
15382 Similarly, 'set1' and 'add1' functions increase reference
15383 counters or duplicate objects.
15387 * Allow for the possibility of temp RSA key generation failure:
15388 the code used to assume it always worked and crashed on failure.
15392 * Fix potential buffer overrun problem in BIO_printf().
15393 *Ulf Möller, using public domain code by Patrick Powell; problem
15394 pointed out by David Sacerdote <das33@cornell.edu>*
15396 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15397 RAND_egd() and RAND_status(). In the command line application,
15398 the EGD socket can be specified like a seed file using RANDFILE
15403 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15404 Some CAs (e.g. Verisign) distribute certificates in this form.
15408 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15409 list to exclude them. This means that no special compilation option
15410 is needed to use anonymous DH: it just needs to be included in the
15415 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15416 EVP_MD_type. The old functionality is available in a new macro called
15417 EVP_MD_md(). Change code that uses it and update docs.
15421 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15422 where the `void *` argument is replaced by a function pointer argument.
15423 Previously `void *` was abused to point to functions, which works on
15424 many platforms, but is not correct. As these functions are usually
15425 called by macros defined in OpenSSL header files, most source code
15426 should work without changes.
15430 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15431 sections with information on -D... compiler switches used for
15432 compiling the library so that applications can see them. To enable
15433 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15434 must be defined. E.g.,
15435 #define OPENSSL_ALGORITHM_DEFINES
15436 #include <openssl/opensslconf.h>
15437 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15439 *Richard Levitte, Ulf and Bodo Möller*
15441 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15446 * Change the 'other' type in certificate aux info to a STACK_OF
15447 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15448 the required ASN1 format: arbitrary types determined by an OID.
15452 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15453 argument to 'req'. This is not because the function is newer or
15454 better than others it just uses the work 'NEW' in the certificate
15455 request header lines. Some software needs this.
15459 * Reorganise password command line arguments: now passwords can be
15460 obtained from various sources. Delete the PEM_cb function and make
15461 it the default behaviour: i.e. if the callback is NULL and the
15462 usrdata argument is not NULL interpret it as a null terminated pass
15463 phrase. If usrdata and the callback are NULL then the pass phrase
15464 is prompted for as usual.
15468 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15469 the support is automatically enabled. The resulting binaries will
15470 autodetect the card and use it if present.
15472 *Ben Laurie and Compaq Inc.*
15474 * Work around for Netscape hang bug. This sends certificate request
15475 and server done in one record. Since this is perfectly legal in the
15476 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15477 the bugs/SSLv3 entry for more info.
15481 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15485 * Add -rand argument to smime and pkcs12 applications and read/write
15490 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15494 * Add command line password options to the remaining applications.
15498 * Bug fix for BN_div_recp() for numerators with an even number of
15503 * More tests in bntest.c, and changed test_bn output.
15507 * ./config recognizes MacOS X now.
15511 * Bug fix for BN_div() when the first words of num and divisor are
15512 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15516 * Add support for various broken PKCS#8 formats, and command line
15517 options to produce them.
15521 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15522 get temporary BIGNUMs from a BN_CTX.
15526 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15531 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15532 include a #define from the old name to the new. The original intent
15533 was that statically linked binaries could for example just call
15534 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15535 link with digests. This never worked because SSLeay_add_all_digests()
15536 and SSLeay_add_all_ciphers() were in the same source file so calling
15537 one would link with the other. They are now in separate source files.
15541 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15545 * Use a less unusual form of the Miller-Rabin primality test (it used
15546 a binary algorithm for exponentiation integrated into the Miller-Rabin
15547 loop, our standard modexp algorithms are faster).
15551 * Support for the EBCDIC character set completed.
15553 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15555 * Source code cleanups: use const where appropriate, eliminate casts,
15556 use `void *` instead of `char *` in lhash.
15560 * Bugfix: ssl3_send_server_key_exchange was not restartable
15561 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15562 this the server could overwrite ephemeral keys that the client
15567 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15568 using 50 iterations of the Rabin-Miller test.
15570 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15571 iterations of the Rabin-Miller test as required by the appendix
15572 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15573 As BN_is_prime_fasttest includes trial division, DSA parameter
15574 generation becomes much faster.
15576 This implies a change for the callback functions in DSA_is_prime
15577 and DSA_generate_parameters: The callback function is called once
15578 for each positive witness in the Rabin-Miller test, not just
15579 occasionally in the inner loop; and the parameters to the
15580 callback function now provide an iteration count for the outer
15581 loop rather than for the current invocation of the inner loop.
15582 DSA_generate_parameters additionally can call the callback
15583 function with an 'iteration count' of -1, meaning that a
15584 candidate has passed the trial division test (when q is generated
15585 from an application-provided seed, trial division is skipped).
15589 * New function BN_is_prime_fasttest that optionally does trial
15590 division before starting the Rabin-Miller test and has
15591 an additional BN_CTX * argument (whereas BN_is_prime always
15592 has to allocate at least one BN_CTX).
15593 'callback(1, -1, cb_arg)' is called when a number has passed the
15594 trial division stage.
15598 * Fix for bug in CRL encoding. The validity dates weren't being handled
15603 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15607 * New function BN_pseudo_rand().
15611 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15612 bignum version of BN_from_montgomery() with the working code from
15613 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15618 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15619 made it impossible to use the same SSL_SESSION data structure in
15620 SSL2 clients in multiple threads.
15624 * The return value of RAND_load_file() no longer counts bytes obtained
15625 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15626 to seed the PRNG (previously an explicit byte count was required).
15628 *Ulf Möller, Bodo Möller*
15630 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15631 used `char *` instead of `void *` and had casts all over the place.
15635 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15639 * Retain source code compatibility for BN_prime_checks macro:
15640 BN_is_prime(..., BN_prime_checks, ...) now uses
15641 BN_prime_checks_for_size to determine the appropriate number of
15642 Rabin-Miller iterations.
15646 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15647 DH_CHECK_P_NOT_SAFE_PRIME.
15648 (Check if this is true? OpenPGP calls them "strong".)
15652 * Merge the functionality of "dh" and "gendh" programs into a new program
15653 "dhparam". The old programs are retained for now but will handle DH keys
15654 (instead of parameters) in future.
15658 * Make the ciphers, s_server and s_client programs check the return values
15659 when a new cipher list is set.
15663 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15664 ciphers. Before when the 56bit ciphers were enabled the sorting was
15667 The syntax for the cipher sorting has been extended to support sorting by
15668 cipher-strength (using the strength_bits hard coded in the tables).
15669 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15671 Fix a bug in the cipher-command parser: when supplying a cipher command
15672 string with an "undefined" symbol (neither command nor alphanumeric
15673 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15674 an error is flagged.
15676 Due to the strength-sorting extension, the code of the
15677 ssl_create_cipher_list() function was completely rearranged. I hope that
15678 the readability was also increased :-)
15680 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15682 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15683 for the first serial number and places 2 in the serial number file. This
15684 avoids problems when the root CA is created with serial number zero and
15685 the first user certificate has the same issuer name and serial number
15690 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15691 the new code. Add documentation for this stuff.
15695 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15696 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15697 structures and behave in an analogous way to the X509v3 functions:
15698 they shouldn't be called directly but wrapper functions should be used
15701 So we also now have some wrapper functions that call the X509at functions
15702 when passed certificate requests. (TO DO: similar things can be done with
15703 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15704 things. Some of these need some d2i or i2d and print functionality
15705 because they handle more complex structures.)
15709 * Add missing #ifndefs that caused missing symbols when building libssl
15710 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15711 NO_RSA in `ssl/s2*.c`.
15713 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15715 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15716 has a return value which indicates the quality of the random data
15717 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15718 error queue. New function RAND_pseudo_bytes() generates output that is
15719 guaranteed to be unique but not unpredictable. RAND_add is like
15720 RAND_seed, but takes an extra argument for an entropy estimate
15721 (RAND_seed always assumes full entropy).
15725 * Do more iterations of Rabin-Miller probable prime test (specifically,
15726 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15727 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15728 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15729 false-positive rate of at most 2^-80 for random input.
15733 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15737 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15738 in the 0.9.5 release), this returns the chain
15739 from an X509_CTX structure with a dup of the stack and all
15740 the X509 reference counts upped: so the stack will exist
15741 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15744 Also make SSL_SESSION_print() print out the verify return
15749 * Add manpage for the pkcs12 command. Also change the default
15750 behaviour so MAC iteration counts are used unless the new
15751 -nomaciter option is used. This improves file security and
15752 only older versions of MSIE (4.0 for example) need it.
15756 * Honor the no-xxx Configure options when creating .DEF files.
15760 * Add PKCS#10 attributes to field table: challengePassword,
15761 unstructuredName and unstructuredAddress. These are taken from
15762 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15763 international characters are used.
15765 More changes to X509_ATTRIBUTE code: allow the setting of types
15766 based on strings. Remove the 'loc' parameter when adding
15767 attributes because these will be a SET OF encoding which is sorted
15772 * Initial changes to the 'req' utility to allow request generation
15773 automation. This will allow an application to just generate a template
15774 file containing all the field values and have req construct the
15777 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15778 used all over the place including certificate requests and PKCS#7
15779 structures. They are currently handled manually where necessary with
15780 some primitive wrappers for PKCS#7. The new functions behave in a
15781 manner analogous to the X509 extension functions: they allow
15782 attributes to be looked up by NID and added.
15784 Later something similar to the X509V3 code would be desirable to
15785 automatically handle the encoding, decoding and printing of the
15786 more complex types. The string types like challengePassword can
15787 be handled by the string table functions.
15789 Also modified the multi byte string table handling. Now there is
15790 a 'global mask' which masks out certain types. The table itself
15791 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15792 is useful when for example there is only one permissible type
15793 (as in countryName) and using the mask might result in no valid
15798 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15799 SSL_get_peer_finished to allow applications to obtain the latest
15800 Finished messages sent to the peer or expected from the peer,
15801 respectively. (SSL_get_peer_finished is usually the Finished message
15802 actually received from the peer, otherwise the protocol will be aborted.)
15804 As the Finished message are message digests of the complete handshake
15805 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15806 be used for external authentication procedures when the authentication
15807 provided by SSL/TLS is not desired or is not enough.
15811 * Enhanced support for Alpha Linux is added. Now ./config checks if
15812 the host supports BWX extension and if Compaq C is present on the
15813 $PATH. Just exploiting of the BWX extension results in 20-30%
15814 performance kick for some algorithms, e.g. DES and RC4 to mention
15815 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15820 * Add support for MS "fast SGC". This is arguably a violation of the
15821 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15822 weak crypto and after checking the certificate is SGC a second one
15823 with strong crypto. MS SGC stops the first handshake after receiving
15824 the server certificate message and sends a second client hello. Since
15825 a server will typically do all the time consuming operations before
15826 expecting any further messages from the client (server key exchange
15827 is the most expensive) there is little difference between the two.
15829 To get OpenSSL to support MS SGC we have to permit a second client
15830 hello message after we have sent server done. In addition we have to
15831 reset the MAC if we do get this second client hello.
15835 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15836 if a DER encoded private key is RSA or DSA traditional format. Changed
15837 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15838 format DER encoded private key. Newer code should use PKCS#8 format which
15839 has the key type encoded in the ASN1 structure. Added DER private key
15840 support to pkcs8 application.
15844 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15845 ciphersuites has been selected (as required by the SSL 3/TLS 1
15846 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15847 is set, we interpret this as a request to violate the specification
15848 (the worst that can happen is a handshake failure, and 'correct'
15849 behaviour would result in a handshake failure anyway).
15853 * In SSL_CTX_add_session, take into account that there might be multiple
15854 SSL_SESSION structures with the same session ID (e.g. when two threads
15855 concurrently obtain them from an external cache).
15856 The internal cache can handle only one SSL_SESSION with a given ID,
15857 so if there's a conflict, we now throw out the old one to achieve
15862 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15863 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15864 some routines that use cipher OIDs: some ciphers do not have OIDs
15865 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15870 * Simplify the trust setting structure and code. Now we just have
15871 two sequences of OIDs for trusted and rejected settings. These will
15872 typically have values the same as the extended key usage extension
15873 and any application specific purposes.
15875 The trust checking code now has a default behaviour: it will just
15876 check for an object with the same NID as the passed id. Functions can
15877 be provided to override either the default behaviour or the behaviour
15878 for a given id. SSL client, server and email already have functions
15879 in place for compatibility: they check the NID and also return "trusted"
15880 if the certificate is self signed.
15884 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15885 traditional format into an EVP_PKEY structure.
15889 * Add a password callback function PEM_cb() which either prompts for
15890 a password if usr_data is NULL or otherwise assumes it is a null
15891 terminated password. Allow passwords to be passed on command line
15892 environment or config files in a few more utilities.
15896 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
15897 keys. Add some short names for PKCS#8 PBE algorithms and allow them
15898 to be specified on the command line for the pkcs8 and pkcs12 utilities.
15899 Update documentation.
15903 * Support for ASN1 "NULL" type. This could be handled before by using
15904 ASN1_TYPE but there wasn't any function that would try to read a NULL
15905 and produce an error if it couldn't. For compatibility we also have
15906 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
15907 don't allocate anything because they don't need to.
15911 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
15914 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
15916 * Rebuild of the memory allocation routines used by OpenSSL code and
15917 possibly others as well. The purpose is to make an interface that
15918 provide hooks so anyone can build a separate set of allocation and
15919 deallocation routines to be used by OpenSSL, for example memory
15920 pool implementations, or something else, which was previously hard
15921 since Malloc(), Realloc() and Free() were defined as macros having
15922 the values malloc, realloc and free, respectively (except for Win32
15923 compilations). The same is provided for memory debugging code.
15924 OpenSSL already comes with functionality to find memory leaks, but
15925 this gives people a chance to debug other memory problems.
15927 With these changes, a new set of functions and macros have appeared:
15929 CRYPTO_set_mem_debug_functions() [F]
15930 CRYPTO_get_mem_debug_functions() [F]
15931 CRYPTO_dbg_set_options() [F]
15932 CRYPTO_dbg_get_options() [F]
15933 CRYPTO_malloc_debug_init() [M]
15935 The memory debug functions are NULL by default, unless the library
15936 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
15937 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
15938 gives the standard debugging functions that come with OpenSSL) or
15939 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
15940 provided by the library user) must be used. When the standard
15941 debugging functions are used, CRYPTO_dbg_set_options can be used to
15942 request additional information:
15943 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
15944 the CRYPTO_MDEBUG_xxx macro when compiling the library.
15946 Also, things like CRYPTO_set_mem_functions will always give the
15947 expected result (the new set of functions is used for allocation
15948 and deallocation) at all times, regardless of platform and compiler
15951 To finish it up, some functions that were never use in any other
15952 way than through macros have a new API and new semantic:
15954 CRYPTO_dbg_malloc()
15955 CRYPTO_dbg_realloc()
15958 All macros of value have retained their old syntax.
15960 *Richard Levitte and Bodo Moeller*
15962 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
15963 ordering of SMIMECapabilities wasn't in "strength order" and there
15964 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
15969 * Some ASN1 types with illegal zero length encoding (INTEGER,
15970 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
15972 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
15974 * Merge in my S/MIME library for OpenSSL. This provides a simple
15975 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
15976 functionality to handle multipart/signed properly) and a utility
15977 called 'smime' to call all this stuff. This is based on code I
15978 originally wrote for Celo who have kindly allowed it to be
15979 included in OpenSSL.
15983 * Add variants des_set_key_checked and des_set_key_unchecked of
15984 des_set_key (aka des_key_sched). Global variable des_check_key
15985 decides which of these is called by des_set_key; this way
15986 des_check_key behaves as it always did, but applications and
15987 the library itself, which was buggy for des_check_key == 1,
15988 have a cleaner way to pick the version they need.
15992 * New function PKCS12_newpass() which changes the password of a
15997 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
15998 dynamic mix. In both cases the ids can be used as an index into the
15999 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16000 functions so they accept a list of the field values and the
16001 application doesn't need to directly manipulate the X509_TRUST
16006 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16011 * Modify the way the V3 extension code looks up extensions. This now
16012 works in a similar way to the object code: we have some "standard"
16013 extensions in a static table which is searched with OBJ_bsearch()
16014 and the application can add dynamic ones if needed. The file
16015 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16016 updated whenever a new extension is added to the core code and kept
16017 in ext_nid order. There is a simple program 'tabtest.c' which checks
16018 this. New extensions are not added too often so this file can readily
16019 be maintained manually.
16021 There are two big advantages in doing things this way. The extensions
16022 can be looked up immediately and no longer need to be "added" using
16023 X509V3_add_standard_extensions(): this function now does nothing.
16024 Side note: I get *lots* of email saying the extension code doesn't
16025 work because people forget to call this function.
16026 Also no dynamic allocation is done unless new extensions are added:
16027 so if we don't add custom extensions there is no need to call
16028 X509V3_EXT_cleanup().
16032 * Modify enc utility's salting as follows: make salting the default. Add a
16033 magic header, so unsalted files fail gracefully instead of just decrypting
16034 to garbage. This is because not salting is a big security hole, so people
16035 should be discouraged from doing it.
16039 * Fixes and enhancements to the 'x509' utility. It allowed a message
16040 digest to be passed on the command line but it only used this
16041 parameter when signing a certificate. Modified so all relevant
16042 operations are affected by the digest parameter including the
16043 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16044 DSA key was used because it didn't fix the digest.
16048 * Initial certificate chain verify code. Currently tests the untrusted
16049 certificates for consistency with the verify purpose (which is set
16050 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16052 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16053 this is because it will reject chains with invalid extensions whereas
16054 every previous version of OpenSSL and SSLeay made no checks at all.
16056 Trust code: checks the root CA for the relevant trust settings. Trust
16057 settings have an initial value consistent with the verify purpose: e.g.
16058 if the verify purpose is for SSL client use it expects the CA to be
16059 trusted for SSL client use. However the default value can be changed to
16060 permit custom trust settings: one example of this would be to only trust
16061 certificates from a specific "secure" set of CAs.
16063 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16064 which should be used for version portability: especially since the
16065 verify structure is likely to change more often now.
16067 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16068 to set them. If not set then assume SSL clients will verify SSL servers
16071 Two new options to the verify program: -untrusted allows a set of
16072 untrusted certificates to be passed in and -purpose which sets the
16073 intended purpose of the certificate. If a purpose is set then the
16074 new chain verify code is used to check extension consistency.
16078 * Support for the authority information access extension.
16082 * Modify RSA and DSA PEM read routines to transparently handle
16083 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16084 public keys in a format compatible with certificate
16085 SubjectPublicKeyInfo structures. Unfortunately there were already
16086 functions called *_PublicKey_* which used various odd formats so
16087 these are retained for compatibility: however the DSA variants were
16088 never in a public release so they have been deleted. Changed dsa/rsa
16089 utilities to handle the new format: note no releases ever handled public
16090 keys so we should be OK.
16092 The primary motivation for this change is to avoid the same fiasco
16093 that dogs private keys: there are several incompatible private key
16094 formats some of which are standard and some OpenSSL specific and
16095 require various evil hacks to allow partial transparent handling and
16096 even then it doesn't work with DER formats. Given the option anything
16097 other than PKCS#8 should be dumped: but the other formats have to
16098 stay in the name of compatibility.
16100 With public keys and the benefit of hindsight one standard format
16101 is used which works with EVP_PKEY, RSA or DSA structures: though
16102 it clearly returns an error if you try to read the wrong kind of key.
16104 Added a -pubkey option to the 'x509' utility to output the public key.
16105 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16106 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16107 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16108 that do the same as the `EVP_PKEY_assign_*()` except they up the
16109 reference count of the added key (they don't "swallow" the
16114 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16115 CRLs would fail if the file contained no certificates or no CRLs:
16116 added a new function to read in both types and return the number
16117 read: this means that if none are read it will be an error. The
16118 DER versions of the certificate and CRL reader would always fail
16119 because it isn't possible to mix certificates and CRLs in DER format
16120 without choking one or the other routine. Changed this to just read
16121 a certificate: this is the best we can do. Also modified the code
16122 in apps/verify.c to take notice of return codes: it was previously
16123 attempting to read in certificates from NULL pointers and ignoring
16124 any errors: this is one reason why the cert and CRL reader seemed
16125 to work. It doesn't check return codes from the default certificate
16126 routines: these may well fail if the certificates aren't installed.
16130 * Code to support otherName option in GeneralName.
16134 * First update to verify code. Change the verify utility
16135 so it warns if it is passed a self signed certificate:
16136 for consistency with the normal behaviour. X509_verify
16137 has been modified to it will now verify a self signed
16138 certificate if *exactly* the same certificate appears
16139 in the store: it was previously impossible to trust a
16140 single self signed certificate. This means that:
16141 openssl verify ss.pem
16142 now gives a warning about a self signed certificate but
16143 openssl verify -CAfile ss.pem ss.pem
16148 * For servers, store verify_result in SSL_SESSION data structure
16149 (and add it to external session representation).
16150 This is needed when client certificate verifications fails,
16151 but an application-provided verification callback (set by
16152 SSL_CTX_set_cert_verify_callback) allows accepting the session
16153 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16154 but returns 1): When the session is reused, we have to set
16155 ssl->verify_result to the appropriate error code to avoid
16158 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16160 * Fix a bug in the new PKCS#7 code: it didn't consider the
16161 case in PKCS7_dataInit() where the signed PKCS7 structure
16162 didn't contain any existing data because it was being created.
16164 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16166 * Add a salt to the key derivation routines in enc.c. This
16167 forms the first 8 bytes of the encrypted file. Also add a
16168 -S option to allow a salt to be input on the command line.
16172 * New function X509_cmp(). Oddly enough there wasn't a function
16173 to compare two certificates. We do this by working out the SHA1
16174 hash and comparing that. X509_cmp() will be needed by the trust
16179 * SSL_get1_session() is like SSL_get_session(), but increments
16180 the reference count in the SSL_SESSION returned.
16182 *Geoff Thorpe <geoff@eu.c2.net>*
16184 * Fix for 'req': it was adding a null to request attributes.
16185 Also change the X509_LOOKUP and X509_INFO code to handle
16186 certificate auxiliary information.
16190 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16195 * Add the possibility to add extra information to the memory leak
16196 detecting output, to form tracebacks, showing from where each
16197 allocation was originated: CRYPTO_push_info("constant string") adds
16198 the string plus current file name and line number to a per-thread
16199 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16200 is like calling CYRPTO_pop_info() until the stack is empty.
16201 Also updated memory leak detection code to be multi-thread-safe.
16205 * Add options -text and -noout to pkcs7 utility and delete the
16206 encryption options which never did anything. Update docs.
16210 * Add options to some of the utilities to allow the pass phrase
16211 to be included on either the command line (not recommended on
16212 OSes like Unix) or read from the environment. Update the
16213 manpages and fix a few bugs.
16217 * Add a few manpages for some of the openssl commands.
16221 * Fix the -revoke option in ca. It was freeing up memory twice,
16222 leaking and not finding already revoked certificates.
16226 * Extensive changes to support certificate auxiliary information.
16227 This involves the use of X509_CERT_AUX structure and X509_AUX
16228 functions. An X509_AUX function such as PEM_read_X509_AUX()
16229 can still read in a certificate file in the usual way but it
16230 will also read in any additional "auxiliary information". By
16231 doing things this way a fair degree of compatibility can be
16232 retained: existing certificates can have this information added
16233 using the new 'x509' options.
16235 Current auxiliary information includes an "alias" and some trust
16236 settings. The trust settings will ultimately be used in enhanced
16237 certificate chain verification routines: currently a certificate
16238 can only be trusted if it is self signed and then it is trusted
16243 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16244 The problem was that one of the replacement routines had not been working
16245 since SSLeay releases. For now the offending routine has been replaced
16246 with non-optimised assembler. Even so, this now gives around 95%
16247 performance improvement for 1024 bit RSA signs.
16251 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16252 handling. Most clients have the effective key size in bits equal to
16253 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16254 A few however don't do this and instead use the size of the decrypted key
16255 to determine the RC2 key length and the AlgorithmIdentifier to determine
16256 the effective key length. In this case the effective key length can still
16257 be 40 bits but the key length can be 168 bits for example. This is fixed
16258 by manually forcing an RC2 key into the EVP_PKEY structure because the
16259 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16260 the key length and effective key length are equal.
16264 * Add a bunch of functions that should simplify the creation of
16265 X509_NAME structures. Now you should be able to do:
16266 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16267 and have it automatically work out the correct field type and fill in
16268 the structures. The more adventurous can try:
16269 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16270 and it will (hopefully) work out the correct multibyte encoding.
16274 * Change the 'req' utility to use the new field handling and multibyte
16275 copy routines. Before the DN field creation was handled in an ad hoc
16276 way in req, ca, and x509 which was rather broken and didn't support
16277 BMPStrings or UTF8Strings. Since some software doesn't implement
16278 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16279 using the dirstring_type option. See the new comment in the default
16280 openssl.cnf for more info.
16284 * Make crypto/rand/md_rand.c more robust:
16285 - Assure unique random numbers after fork().
16286 - Make sure that concurrent threads access the global counter and
16287 md serializably so that we never lose entropy in them
16288 or use exactly the same state in multiple threads.
16289 Access to the large state is not always serializable because
16290 the additional locking could be a performance killer, and
16291 md should be large enough anyway.
16295 * New file apps/app_rand.c with commonly needed functionality
16296 for handling the random seed file.
16298 Use the random seed file in some applications that previously did not:
16300 dsaparam -genkey (which also ignored its '-rand' option),
16303 x509 (when signing).
16304 Except on systems with /dev/urandom, it is crucial to have a random
16305 seed file at least for key creation, DSA signing, and for DH exchanges;
16306 for RSA signatures we could do without one.
16308 gendh and gendsa (unlike genrsa) used to read only the first byte
16309 of each file listed in the '-rand' option. The function as previously
16310 found in genrsa is now in app_rand.c and is used by all programs
16311 that support '-rand'.
16315 * In RAND_write_file, use mode 0600 for creating files;
16316 don't just chmod when it may be too late.
16320 * Report an error from X509_STORE_load_locations
16321 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16325 * New function ASN1_mbstring_copy() this copies a string in either
16326 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16327 into an ASN1_STRING type. A mask of permissible types is passed
16328 and it chooses the "minimal" type to use or an error if not type
16333 * Add function equivalents to the various macros in asn1.h. The old
16334 macros are retained with an `M_` prefix. Code inside the library can
16335 use the `M_` macros. External code (including the openssl utility)
16336 should *NOT* in order to be "shared library friendly".
16340 * Add various functions that can check a certificate's extensions
16341 to see if it usable for various purposes such as SSL client,
16342 server or S/MIME and CAs of these types. This is currently
16343 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16344 verification. Also added a -purpose flag to x509 utility to
16345 print out all the purposes.
16349 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16354 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16355 for, obtain and decode and extension and obtain its critical flag.
16356 This allows all the necessary extension code to be handled in a
16357 single function call.
16361 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16362 platforms. See crypto/rc4/rc4_enc.c for further details.
16366 * New -noout option to asn1parse. This causes no output to be produced
16367 its main use is when combined with -strparse and -out to extract data
16368 from a file (which may not be in ASN.1 format).
16372 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16373 when producing the local key id.
16375 *Richard Levitte <levitte@stacken.kth.se>*
16377 * New option -dhparam in s_server. This allows a DH parameter file to be
16378 stated explicitly. If it is not stated then it tries the first server
16379 certificate file. The previous behaviour hard coded the filename
16384 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16385 a public key to be input or output. For example:
16386 openssl rsa -in key.pem -pubout -out pubkey.pem
16387 Also added necessary DSA public key functions to handle this.
16391 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16392 in the message. This was handled by allowing
16393 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16395 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16397 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16398 to the end of the strings whereas this didn't. This would cause problems
16399 if strings read with d2i_ASN1_bytes() were later modified.
16401 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16403 * Fix for base64 decode bug. When a base64 bio reads only one line of
16404 data and it contains EOF it will end up returning an error. This is
16405 caused by input 46 bytes long. The cause is due to the way base64
16406 BIOs find the start of base64 encoded data. They do this by trying a
16407 trial decode on each line until they find one that works. When they
16408 do a flag is set and it starts again knowing it can pass all the
16409 data directly through the decoder. Unfortunately it doesn't reset
16410 the context it uses. This means that if EOF is reached an attempt
16411 is made to pass two EOFs through the context and this causes the
16412 resulting error. This can also cause other problems as well. As is
16413 usual with these problems it takes *ages* to find and the fix is
16414 trivial: move one line.
16416 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16418 * Ugly workaround to get s_client and s_server working under Windows. The
16419 old code wouldn't work because it needed to select() on sockets and the
16420 tty (for keypresses and to see if data could be written). Win32 only
16421 supports select() on sockets so we select() with a 1s timeout on the
16422 sockets and then see if any characters are waiting to be read, if none
16423 are present then we retry, we also assume we can always write data to
16424 the tty. This isn't nice because the code then blocks until we've
16425 received a complete line of data and it is effectively polling the
16426 keyboard at 1s intervals: however it's quite a bit better than not
16427 working at all :-) A dedicated Windows application might handle this
16428 with an event loop for example.
16432 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16433 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16434 will be called when RSA_sign() and RSA_verify() are used. This is useful
16435 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16436 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16437 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16438 This necessitated the support of an extra signature type NID_md5_sha1
16439 for SSL signatures and modifications to the SSL library to use it instead
16440 of calling RSA_public_decrypt() and RSA_private_encrypt().
16444 * Add new -verify -CAfile and -CApath options to the crl program, these
16445 will lookup a CRL issuers certificate and verify the signature in a
16446 similar way to the verify program. Tidy up the crl program so it
16447 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16448 less strict. It will now permit CRL extensions even if it is not
16449 a V2 CRL: this will allow it to tolerate some broken CRLs.
16453 * Initialize all non-automatic variables each time one of the openssl
16454 sub-programs is started (this is necessary as they may be started
16455 multiple times from the "OpenSSL>" prompt).
16457 *Lennart Bang, Bodo Moeller*
16459 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16460 removing all other RSA functionality (this is what NO_RSA does). This
16461 is so (for example) those in the US can disable those operations covered
16462 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16467 * Non-copying interface to BIO pairs.
16468 (still largely untested)
16472 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16473 ASCII string. This was handled independently in various places before.
16477 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16478 UTF8 strings a character at a time.
16482 * Use client_version from client hello to select the protocol
16483 (s23_srvr.c) and for RSA client key exchange verification
16484 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16488 * Add various utility functions to handle SPKACs, these were previously
16489 handled by poking round in the structure internals. Added new function
16490 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16491 print, verify and generate SPKACs. Based on an original idea from
16492 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16496 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16500 * Allow the config file extension section to be overwritten on the
16501 command line. Based on an original idea from Massimiliano Pala
16502 <madwolf@comune.modena.it>. The new option is called -extensions
16503 and can be applied to ca, req and x509. Also -reqexts to override
16504 the request extensions in req and -crlexts to override the crl extensions
16509 * Add new feature to the SPKAC handling in ca. Now you can include
16510 the same field multiple times by preceding it by "XXXX." for example:
16513 this is the same syntax as used in the req config file.
16517 * Allow certificate extensions to be added to certificate requests. These
16518 are specified in a 'req_extensions' option of the req section of the
16519 config file. They can be printed out with the -text option to req but
16520 are otherwise ignored at present.
16524 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16525 data read consists of only the final block it would not decrypted because
16526 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16527 A misplaced 'break' also meant the decrypted final block might not be
16528 copied until the next read.
16532 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16533 a few extra parameters to the DH structure: these will be useful if
16534 for example we want the value of 'q' or implement X9.42 DH.
16538 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16539 provides hooks that allow the default DSA functions or functions on a
16540 "per key" basis to be replaced. This allows hardware acceleration and
16541 hardware key storage to be handled without major modification to the
16542 library. Also added low level modexp hooks and CRYPTO_EX structure and
16543 associated functions.
16547 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16548 as "read only": it can't be written to and the buffer it points to will
16549 not be freed. Reading from a read only BIO is much more efficient than
16550 a normal memory BIO. This was added because there are several times when
16551 an area of memory needs to be read from a BIO. The previous method was
16552 to create a memory BIO and write the data to it, this results in two
16553 copies of the data and an O(n^2) reading algorithm. There is a new
16554 function BIO_new_mem_buf() which creates a read only memory BIO from
16555 an area of memory. Also modified the PKCS#7 routines to use read only
16560 * Bugfix: ssl23_get_client_hello did not work properly when called in
16561 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16562 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16563 but a retry condition occurred while trying to read the rest.
16567 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16568 NID_pkcs7_encrypted by default: this was wrong since this should almost
16569 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16570 the encrypted data type: this is a more sensible place to put it and it
16571 allows the PKCS#12 code to be tidied up that duplicated this
16576 * Changed obj_dat.pl script so it takes its input and output files on
16577 the command line. This should avoid shell escape redirection problems
16582 * Initial support for certificate extension requests, these are included
16583 in things like Xenroll certificate requests. Included functions to allow
16584 extensions to be obtained and added.
16588 * -crlf option to s_client and s_server for sending newlines as
16589 CRLF (as required by many protocols).
16593 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16595 * Install libRSAglue.a when OpenSSL is built with RSAref.
16597 *Ralf S. Engelschall*
16599 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16601 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16603 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16608 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16609 DH parameters/keys (q is lost during that conversion, but the resulting
16610 DH parameters contain its length).
16612 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16613 much faster than DH_generate_parameters (which creates parameters
16614 where `p = 2*q + 1`), and also the smaller q makes DH computations
16615 much more efficient (160-bit exponentiation instead of 1024-bit
16616 exponentiation); so this provides a convenient way to support DHE
16617 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16618 utter importance to use
16619 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16621 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16622 when such DH parameters are used, because otherwise small subgroup
16623 attacks may become possible!
16627 * Avoid memory leak in i2d_DHparams.
16631 * Allow the -k option to be used more than once in the enc program:
16632 this allows the same encrypted message to be read by multiple recipients.
16636 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16637 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16638 it will always use the numerical form of the OID, even if it has a short
16643 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16644 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16645 otherwise bn_mod_exp was called. In the case of hardware keys for example
16646 no private key components need be present and it might store extra data
16647 in the RSA structure, which cannot be accessed from bn_mod_exp.
16648 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16649 private key operations.
16653 * Added support for SPARC Linux.
16657 * pem_password_cb function type incompatibly changed from
16658 typedef int pem_password_cb(char *buf, int size, int rwflag);
16660 ....(char *buf, int size, int rwflag, void *userdata);
16661 so that applications can pass data to their callbacks:
16662 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16663 additional void * argument, which is just handed through whenever
16664 the password callback is called.
16666 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16668 New function SSL_CTX_set_default_passwd_cb_userdata.
16670 Compatibility note: As many C implementations push function arguments
16671 onto the stack in reverse order, the new library version is likely to
16672 interoperate with programs that have been compiled with the old
16673 pem_password_cb definition (PEM_whatever takes some data that
16674 happens to be on the stack as its last argument, and the callback
16675 just ignores this garbage); but there is no guarantee whatsoever that
16678 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16679 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16680 problems not only on Windows, but also on some Unix platforms.
16681 To avoid problematic command lines, these definitions are now in an
16682 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16683 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16687 * MIPS III/IV assembler module is reimplemented.
16691 * More DES library cleanups: remove references to srand/rand and
16692 delete an unused file.
16696 * Add support for the free Netwide assembler (NASM) under Win32,
16697 since not many people have MASM (ml) and it can be hard to obtain.
16698 This is currently experimental but it seems to work OK and pass all
16699 the tests. Check out INSTALL.W32 for info.
16703 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16704 without temporary keys kept an extra copy of the server key,
16705 and connections with temporary keys did not free everything in case
16710 * New function RSA_check_key and new openssl rsa option -check
16711 for verifying the consistency of RSA keys.
16713 *Ulf Moeller, Bodo Moeller*
16715 * Various changes to make Win32 compile work:
16716 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16717 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16718 comparison" warnings.
16719 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16723 * Add a debugging option to PKCS#5 v2 key generation function: when
16724 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16725 derived keys are printed to stderr.
16729 * Copy the flags in ASN1_STRING_dup().
16731 *Roman E. Pavlov <pre@mo.msk.ru>*
16733 * The x509 application mishandled signing requests containing DSA
16734 keys when the signing key was also DSA and the parameters didn't match.
16736 It was supposed to omit the parameters when they matched the signing key:
16737 the verifying software was then supposed to automatically use the CA's
16738 parameters if they were absent from the end user certificate.
16740 Omitting parameters is no longer recommended. The test was also
16741 the wrong way round! This was probably due to unusual behaviour in
16742 EVP_cmp_parameters() which returns 1 if the parameters match.
16743 This meant that parameters were omitted when they *didn't* match and
16744 the certificate was useless. Certificates signed with 'ca' didn't have
16747 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16749 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16750 The interface is as follows:
16751 Applications can use
16752 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16753 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16754 "off" is now the default.
16755 The library internally uses
16756 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16757 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16758 to disable memory-checking temporarily.
16760 Some inconsistent states that previously were possible (and were
16761 even the default) are now avoided.
16763 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16764 with each memory chunk allocated; this is occasionally more helpful
16765 than just having a counter.
16767 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16769 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16774 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16775 which largely parallels "options", but is for changing API behaviour,
16776 whereas "options" are about protocol behaviour.
16777 Initial "mode" flags are:
16779 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16780 a single record has been written.
16781 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16782 retries use the same buffer location.
16783 (But all of the contents must be
16788 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16791 * Fix problems with no-hmac etc.
16793 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16795 * New functions RSA_get_default_method(), RSA_set_method() and
16796 RSA_get_method(). These allows replacement of RSA_METHODs without having
16797 to mess around with the internals of an RSA structure.
16801 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16802 Also really enable memory leak checks in openssl.c and in some
16805 *Chad C. Mulligan, Bodo Moeller*
16807 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16808 up the length of negative integers. This has now been simplified to just
16809 store the length when it is first determined and use it later, rather
16810 than trying to keep track of where data is copied and updating it to
16812 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16814 * Add a new function PKCS7_signatureVerify. This allows the verification
16815 of a PKCS#7 signature but with the signing certificate passed to the
16816 function itself. This contrasts with PKCS7_dataVerify which assumes the
16817 certificate is present in the PKCS#7 structure. This isn't always the
16818 case: certificates can be omitted from a PKCS#7 structure and be
16819 distributed by "out of band" means (such as a certificate database).
16823 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16824 function prototypes in pem.h, also change util/mkdef.pl to add the
16825 necessary function names.
16829 * mk1mf.pl (used by Windows builds) did not properly read the
16830 options set by Configure in the top level Makefile, and Configure
16831 was not even able to write more than one option correctly.
16832 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16836 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16837 file to be loaded from a BIO or FILE pointer. The BIO version will
16838 for example allow memory BIOs to contain config info.
16842 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16843 Whoever hopes to achieve shared-library compatibility across versions
16844 must use this, not the compile-time macro.
16845 (Exercise 0.9.4: Which is the minimum library version required by
16847 Note: All this applies only to multi-threaded programs, others don't
16852 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16853 through a BIO pair triggered the default case, i.e.
16854 SSLerr(...,SSL_R_UNKNOWN_STATE).
16858 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16859 can use the SSL library even if none of the specific BIOs is
16864 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16865 for the encoded length.
16867 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16869 * Add initial documentation of the X509V3 functions.
16873 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16874 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16875 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16876 secure PKCS#8 private key format with a high iteration count.
16880 * Fix determination of Perl interpreter: A perl or perl5
16881 *directory* in $PATH was also accepted as the interpreter.
16883 *Ralf S. Engelschall*
16885 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16886 wrong with it but it was very old and did things like calling
16887 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
16888 unusual formatting.
16892 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
16893 to use the new extension code.
16897 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
16898 with macros. This should make it easier to change their form, add extra
16899 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
16904 * Add to configuration table a new entry that can specify an alternative
16905 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
16906 according to Mark Crispin <MRC@Panda.COM>.
16910 * DES CBC did not update the IV. Weird.
16914 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
16915 Changing the behaviour of the former might break existing programs --
16916 where IV updating is needed, des_ncbc_encrypt can be used.
16919 * When bntest is run from "make test" it drives bc to check its
16920 calculations, as well as internally checking them. If an internal check
16921 fails, it needs to cause bc to give a non-zero result or make test carries
16922 on without noticing the failure. Fixed.
16926 * DES library cleanups.
16930 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
16931 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
16932 ciphers. NOTE: although the key derivation function has been verified
16933 against some published test vectors it has not been extensively tested
16934 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
16939 * Instead of "mkdir -p", which is not fully portable, use new
16940 Perl script "util/mkdir-p.pl".
16944 * Rewrite the way password based encryption (PBE) is handled. It used to
16945 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
16946 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
16947 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
16948 the 'parameter' field of the AlgorithmIdentifier is passed to the
16949 underlying key generation function so it must do its own ASN1 parsing.
16950 This has also changed the EVP_PBE_CipherInit() function which now has a
16951 'parameter' argument instead of literal salt and iteration count values
16952 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
16956 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
16957 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
16958 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
16959 KEY" because this clashed with PKCS#8 unencrypted string. Since this
16960 value was just used as a "magic string" and not used directly its
16961 value doesn't matter.
16965 * Introduce some semblance of const correctness to BN. Shame C doesn't
16970 * "linux-sparc64" configuration (ultrapenguin).
16972 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
16973 "linux-sparc" configuration.
16975 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
16977 * config now generates no-xxx options for missing ciphers.
16981 * Support the EBCDIC character set (work in progress).
16982 File ebcdic.c not yet included because it has a different license.
16984 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16986 * Support BS2000/OSD-POSIX.
16988 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
16990 * Make callbacks for key generation use `void *` instead of `char *`.
16994 * Make S/MIME samples compile (not yet tested).
16998 * Additional typesafe stacks.
17002 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17006 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17008 * New configuration variant "sco5-gcc".
17010 * Updated some demos.
17012 *Sean O Riordain, Wade Scholine*
17014 * Add missing BIO_free at exit of pkcs12 application.
17018 * Fix memory leak in conf.c.
17022 * Updates for Win32 to assembler version of MD5.
17026 * Set #! path to perl in apps/der_chop to where we found it
17027 instead of using a fixed path.
17031 * SHA library changes for irix64-mips4-cc.
17035 * Improvements for VMS support.
17039 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17041 * Bignum library bug fix. IRIX 6 passes "make test" now!
17042 This also avoids the problems with SC4.2 and unpatched SC5.
17044 *Andy Polyakov <appro@fy.chalmers.se>*
17046 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17047 These are required because of the typesafe stack would otherwise break
17048 existing code. If old code used a structure member which used to be STACK
17049 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17050 sk_num or sk_value it would produce an error because the num, data members
17051 are not present in STACK_OF. Now it just produces a warning. sk_set
17052 replaces the old method of assigning a value to sk_value
17053 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17054 that does this will no longer work (and should use sk_set instead) but
17055 this could be regarded as a "questionable" behaviour anyway.
17059 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17060 correctly handle encrypted S/MIME data.
17064 * Change type of various DES function arguments from des_cblock
17065 (which means, in function argument declarations, pointer to char)
17066 to des_cblock * (meaning pointer to array with 8 char elements),
17067 which allows the compiler to do more typechecking; it was like
17068 that back in SSLeay, but with lots of ugly casts.
17070 Introduce new type const_des_cblock.
17074 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17075 problems: find RecipientInfo structure that matches recipient certificate
17076 and initialise the ASN1 structures properly based on passed cipher.
17080 * Belatedly make the BN tests actually check the results.
17084 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17085 to and from BNs: it was completely broken. New compilation option
17086 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17087 key elements as negative integers.
17091 * Reorganize and speed up MD5.
17093 *Andy Polyakov <appro@fy.chalmers.se>*
17097 *Richard Levitte <richard@levitte.org>*
17099 * New option -out to asn1parse to allow the parsed structure to be
17100 output to a file. This is most useful when combined with the -strparse
17101 option to examine the output of things like OCTET STRINGS.
17105 * Make SSL library a little more fool-proof by not requiring any longer
17106 that `SSL_set_{accept,connect}_state` be called before
17107 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17108 in many applications because usually everything *appeared* to work as
17109 intended anyway -- now it really works as intended).
17113 * Move openssl.cnf out of lib/.
17117 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17118 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17119 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17121 *Ralf S. Engelschall*
17123 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17124 handle PKCS#7 enveloped data properly.
17126 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17128 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17129 copying pointers. The cert_st handling is changed by this in
17130 various ways (and thus what used to be known as ctx->default_cert
17131 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17132 any longer when s->cert does not give us what we need).
17133 ssl_cert_instantiate becomes obsolete by this change.
17134 As soon as we've got the new code right (possibly it already is?),
17135 we have solved a couple of bugs of the earlier code where s->cert
17136 was used as if it could not have been shared with other SSL structures.
17138 Note that using the SSL API in certain dirty ways now will result
17139 in different behaviour than observed with earlier library versions:
17140 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17141 does not influence s as it used to.
17143 In order to clean up things more thoroughly, inside SSL_SESSION
17144 we don't use CERT any longer, but a new structure SESS_CERT
17145 that holds per-session data (if available); currently, this is
17146 the peer's certificate chain and, for clients, the server's certificate
17147 and temporary key. CERT holds only those values that can have
17148 meaningful defaults in an SSL_CTX.
17152 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17153 from the internal representation. Various PKCS#7 fixes: remove some
17154 evil casts and set the enc_dig_alg field properly based on the signing
17159 * Allow PKCS#12 password to be set from the command line or the
17160 environment. Let 'ca' get its config file name from the environment
17161 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17166 * Allow certificate policies extension to use an IA5STRING for the
17167 organization field. This is contrary to the PKIX definition but
17168 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17173 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17174 without disallowing inline assembler and the like for non-pedantic builds.
17178 * Support Borland C++ builder.
17180 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17186 * SHA-1 cleanups and performance enhancements.
17188 *Andy Polyakov <appro@fy.chalmers.se>*
17190 * Sparc v8plus assembler for the bignum library.
17192 *Andy Polyakov <appro@fy.chalmers.se>*
17194 * Accept any -xxx and +xxx compiler options in Configure.
17198 * Update HPUX configuration.
17202 * Add missing `sk_<type>_unshift()` function to safestack.h
17204 *Ralf S. Engelschall*
17206 * New function SSL_CTX_use_certificate_chain_file that sets the
17207 "extra_cert"s in addition to the certificate. (This makes sense
17208 only for "PEM" format files, as chains as a whole are not
17213 * Support verify_depth from the SSL API.
17214 x509_vfy.c had what can be considered an off-by-one-error:
17215 Its depth (which was not part of the external interface)
17216 was actually counting the number of certificates in a chain;
17217 now it really counts the depth.
17221 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17222 instead of X509err, which often resulted in confusing error
17223 messages since the error codes are not globally unique
17224 (e.g. an alleged error in ssl3_accept when a certificate
17225 didn't match the private key).
17227 * New function SSL_CTX_set_session_id_context that allows to set a default
17228 value (so that you don't need SSL_set_session_id_context for each
17229 connection using the SSL_CTX).
17233 * OAEP decoding bug fix.
17237 * Support INSTALL_PREFIX for package builders, as proposed by
17242 * New Configure options "threads" and "no-threads". For systems
17243 where the proper compiler options are known (currently Solaris
17244 and Linux), "threads" is the default.
17248 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17252 * Install various scripts to $(OPENSSLDIR)/misc, not to
17253 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17254 such as /usr/local/bin.
17258 * "make linux-shared" to build shared libraries.
17260 *Niels Poppe <niels@netbox.org>*
17262 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17266 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17267 extension adding in x509 utility.
17271 * Remove NOPROTO sections and error code comments.
17275 * Partial rewrite of the DEF file generator to now parse the ANSI
17280 * New Configure options --prefix=DIR and --openssldir=DIR.
17284 * Complete rewrite of the error code script(s). It is all now handled
17285 by one script at the top level which handles error code gathering,
17286 header rewriting and C source file generation. It should be much better
17287 than the old method: it now uses a modified version of Ulf's parser to
17288 read the ANSI prototypes in all header files (thus the old K&R definitions
17289 aren't needed for error creation any more) and do a better job of
17290 translating function codes into names. The old 'ASN1 error code embedded
17291 in a comment' is no longer necessary and it doesn't use .err files which
17292 have now been deleted. Also the error code call doesn't have to appear all
17293 on one line (which resulted in some large lines...).
17297 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17301 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17302 0 (which usually indicates a closed connection), but continue reading.
17306 * Fix some race conditions.
17310 * Add support for CRL distribution points extension. Add Certificate
17311 Policies and CRL distribution points documentation.
17315 * Move the autogenerated header file parts to crypto/opensslconf.h.
17319 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17320 8 of keying material. Merlin has also confirmed interop with this fix
17321 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17323 *Merlin Hughes <merlin@baltimore.ie>*
17325 * Fix lots of warnings.
17327 *Richard Levitte <levitte@stacken.kth.se>*
17329 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17330 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17332 *Richard Levitte <levitte@stacken.kth.se>*
17334 * Fix problems with sizeof(long) == 8.
17336 *Andy Polyakov <appro@fy.chalmers.se>*
17338 * Change functions to ANSI C.
17342 * Fix typos in error codes.
17344 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17346 * Remove defunct assembler files from Configure.
17350 * SPARC v8 assembler BIGNUM implementation.
17352 *Andy Polyakov <appro@fy.chalmers.se>*
17354 * Support for Certificate Policies extension: both print and set.
17355 Various additions to support the r2i method this uses.
17359 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17360 return a const string when you are expecting an allocated buffer.
17364 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17365 types DirectoryString and DisplayText.
17369 * Add code to allow r2i extensions to access the configuration database,
17370 add an LHASH database driver and add several ctx helper functions.
17374 * Fix an evil bug in bn_expand2() which caused various BN functions to
17375 fail when they extended the size of a BIGNUM.
17379 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17380 support typesafe stack.
17384 * Fix typo in SSL_[gs]et_options().
17386 *Nils Frostberg <nils@medcom.se>*
17388 * Delete various functions and files that belonged to the (now obsolete)
17389 old X509V3 handling code.
17393 * New Configure option "rsaref".
17397 * Don't auto-generate pem.h.
17401 * Introduce type-safe ASN.1 SETs.
17405 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17407 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17409 * Introduce type-safe STACKs. This will almost certainly break lots of code
17410 that links with OpenSSL (well at least cause lots of warnings), but fear
17411 not: the conversion is trivial, and it eliminates loads of evil casts. A
17412 few STACKed things have been converted already. Feel free to convert more.
17413 In the fullness of time, I'll do away with the STACK type altogether.
17417 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17418 specified in `<certfile>` by updating the entry in the index.txt file.
17419 This way one no longer has to edit the index.txt file manually for
17420 revoking a certificate. The -revoke option does the gory details now.
17422 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17424 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17425 `-text` option at all and this way the `-noout -text` combination was
17426 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17428 *Ralf S. Engelschall*
17430 * Make sure a corresponding plain text error message exists for the
17431 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17432 verify callback function determined that a certificate was revoked.
17434 *Ralf S. Engelschall*
17436 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17437 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17438 all available ciphers including rc5, which was forgotten until now.
17439 In order to let the testing shell script know which algorithms
17440 are available, a new (up to now undocumented) command
17441 `openssl list-cipher-commands` is used.
17445 * Bugfix: s_client occasionally would sleep in select() when
17446 it should have checked SSL_pending() first.
17450 * New functions DSA_do_sign and DSA_do_verify to provide access to
17451 the raw DSA values prior to ASN.1 encoding.
17455 * Tweaks to Configure
17457 *Niels Poppe <niels@netbox.org>*
17459 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17464 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17468 * New config option to avoid instructions that are illegal on the 80386.
17469 The default code is faster, but requires at least a 486.
17473 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17474 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17475 same as SSL2_VERSION anyway.
17479 * New "-showcerts" option for s_client.
17483 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17484 application. Various cleanups and fixes.
17488 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17489 modify error routines to work internally. Add error codes and PBE init
17490 to library startup routines.
17494 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17495 packing functions to asn1 and evp. Changed function names and error
17496 codes along the way.
17500 * PKCS12 integration: and so it begins... First of several patches to
17501 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17502 objects to objects.h
17506 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17507 and display support for Thawte strong extranet extension.
17511 * Add LinuxPPC support.
17513 *Jeff Dubrule <igor@pobox.org>*
17515 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17516 bn_div_words in alpha.s.
17518 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17520 * Make sure the RSA OAEP test is skipped under -DRSAref because
17521 OAEP isn't supported when OpenSSL is built with RSAref.
17523 *Ulf Moeller <ulf@fitug.de>*
17525 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17526 so they no longer are missing under -DNOPROTO.
17528 *Soren S. Jorvang <soren@t.dk>*
17530 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17532 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17533 doesn't work when the session is reused. Coming soon!
17537 * Fix a security hole, that allows sessions to be reused in the wrong
17538 context thus bypassing client cert protection! All software that uses
17539 client certs and session caches in multiple contexts NEEDS PATCHING to
17540 allow session reuse! A fuller solution is in the works.
17542 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17544 * Some more source tree cleanups (removed obsolete files
17545 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17546 permission on "config" script to be executable) and a fix for the INSTALL
17549 *Ulf Moeller <ulf@fitug.de>*
17551 * Remove some legacy and erroneous uses of malloc, free instead of
17554 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17556 * Make rsa_oaep_test return non-zero on error.
17558 *Ulf Moeller <ulf@fitug.de>*
17560 * Add support for native Solaris shared libraries. Configure
17561 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17562 if someone would make that last step automatic.
17564 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17566 * ctx_size was not built with the right compiler during "make links". Fixed.
17570 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17571 except NULL ciphers". This means the default cipher list will no longer
17572 enable NULL ciphers. They need to be specifically enabled e.g. with
17573 the string "DEFAULT:eNULL".
17577 * Fix to RSA private encryption routines: if p < q then it would
17578 occasionally produce an invalid result. This will only happen with
17579 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17583 * Be less restrictive and allow also `perl util/perlpath.pl
17584 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17585 because this way one can also use an interpreter named `perl5` (which is
17586 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17587 installed as `perl`).
17589 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17591 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17593 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17595 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17596 advapi32.lib to Win32 build and change the pem test comparison
17597 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17598 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17599 and crypto/des/ede_cbcm_enc.c.
17603 * DES quad checksum was broken on big-endian architectures. Fixed.
17607 * Comment out two functions in bio.h that aren't implemented. Fix up the
17608 Win32 test batch file so it (might) work again. The Win32 test batch file
17609 is horrible: I feel ill....
17613 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17614 in e_os.h. Audit of header files to check ANSI and non ANSI
17615 sections: 10 functions were absent from non ANSI section and not exported
17616 from Windows DLLs. Fixed up libeay.num for new functions.
17620 * Make `openssl version` output lines consistent.
17622 *Ralf S. Engelschall*
17624 * Fix Win32 symbol export lists for BIO functions: Added
17625 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17626 to ms/libeay{16,32}.def.
17628 *Ralf S. Engelschall*
17630 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17631 fine under Unix and passes some trivial tests I've now added. But the
17632 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17633 added to make sure no one expects that this stuff really works in the
17634 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17635 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17638 *Ralf S. Engelschall*
17640 * Fix the generation of two part addresses in perl.
17642 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17644 * Add config entry for Linux on MIPS.
17646 *John Tobey <jtobey@channel1.com>*
17648 * Make links whenever Configure is run, unless we are on Windoze.
17652 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17653 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17658 * Add a useful kludge to allow package maintainers to specify compiler and
17659 other platforms details on the command line without having to patch the
17660 Configure script every time: One now can use
17661 `perl Configure <id>:<details>`,
17662 i.e. platform ids are allowed to have details appended
17663 to them (separated by colons). This is treated as there would be a static
17664 pre-configured entry in Configure's %table under key `<id>` with value
17665 `<details>` and `perl Configure <id>` is called. So, when you want to
17666 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17667 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17668 now, which overrides the FreeBSD-elf entry on-the-fly.
17670 *Ralf S. Engelschall*
17672 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17676 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17677 on the `perl Configure ...` command line. This way one can compile
17678 OpenSSL libraries with Position Independent Code (PIC) which is needed
17679 for linking it into DSOs.
17681 *Ralf S. Engelschall*
17683 * Remarkably, export ciphers were totally broken and no-one had noticed!
17688 * Cleaned up the LICENSE document: The official contact for any license
17689 questions now is the OpenSSL core team under openssl-core@openssl.org.
17690 And add a paragraph about the dual-license situation to make sure people
17691 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17692 to the OpenSSL toolkit.
17694 *Ralf S. Engelschall*
17696 * General source tree makefile cleanups: Made `making xxx in yyy...`
17697 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17698 Additionally cleaned up the `make links` target: Remove unnecessary
17699 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17700 to speed processing and no longer clutter the display with confusing
17701 stuff. Instead only the actually done links are displayed.
17703 *Ralf S. Engelschall*
17705 * Permit null encryption ciphersuites, used for authentication only. It used
17706 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17707 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17712 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17713 signed attributes when verifying signatures (this would break them),
17714 the detached data encoding was wrong and public keys obtained using
17715 X509_get_pubkey() weren't freed.
17719 * Add text documentation for the BUFFER functions. Also added a work around
17720 to a Win95 console bug. This was triggered by the password read stuff: the
17721 last character typed gets carried over to the next fread(). If you were
17722 generating a new cert request using 'req' for example then the last
17723 character of the passphrase would be CR which would then enter the first
17728 * Added the new 'Includes OpenSSL Cryptography Software' button as
17729 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17730 button and can be used by applications based on OpenSSL to show the
17731 relationship to the OpenSSL project.
17733 *Ralf S. Engelschall*
17735 * Remove confusing variables in function signatures in files
17736 ssl/ssl_lib.c and ssl/ssl.h.
17738 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17740 * Don't install bss_file.c under PREFIX/include/
17742 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17744 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17745 functions that return function pointers and has support for NT specific
17746 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17747 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17748 unsigned to signed types: this was killing the Win32 compile.
17752 * Add new certificate file to stack functions,
17753 SSL_add_dir_cert_subjects_to_stack() and
17754 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17755 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17756 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17757 This means that Apache-SSL and similar packages don't have to mess around
17758 to add as many CAs as they want to the preferred list.
17762 * Experiment with doxygen documentation. Currently only partially applied to
17764 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17765 openssl.doxy as the configuration file.
17769 * Get rid of remaining C++-style comments which strict C compilers hate.
17771 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17773 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17774 compiled in by default: it has problems with large keys.
17778 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17779 DH private keys and/or callback functions which directly correspond to
17780 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17781 is needed for applications which have to configure certificates on a
17782 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17784 For the RSA certificate situation is makes no difference, but
17785 for the DSA certificate situation this fixes the "no shared cipher"
17786 problem where the OpenSSL cipher selection procedure failed because the
17787 temporary keys were not overtaken from the context and the API provided
17788 no way to reconfigure them.
17789 The new functions now let applications reconfigure the stuff and they
17790 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17791 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17792 non-public-API function ssl_cert_instantiate() is used as a helper
17793 function and also to reduce code redundancy inside ssl_rsa.c.
17795 *Ralf S. Engelschall*
17797 * Move s_server -dcert and -dkey options out of the undocumented feature
17798 area because they are useful for the DSA situation and should be
17799 recognized by the users.
17801 *Ralf S. Engelschall*
17803 * Fix the cipher decision scheme for export ciphers: the export bits are
17804 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17805 SSL_EXP_MASK. So, the original variable has to be used instead of the
17806 already masked variable.
17808 *Richard Levitte <levitte@stacken.kth.se>*
17810 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17812 *Richard Levitte <levitte@stacken.kth.se>*
17814 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17815 from `int` to `unsigned int` because it is a length and initialized by
17816 EVP_DigestFinal() which expects an `unsigned int *`.
17818 *Richard Levitte <levitte@stacken.kth.se>*
17820 * Don't hard-code path to Perl interpreter on shebang line of Configure
17821 script. Instead use the usual Shell->Perl transition trick.
17823 *Ralf S. Engelschall*
17825 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17826 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17827 -noout -modulus` as it's already the case for `openssl rsa -noout
17828 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17829 currently the public key is printed (a decision which was already done by
17830 `openssl dsa -modulus` in the past) which serves a similar purpose.
17831 Additionally the NO_RSA no longer completely removes the whole -modulus
17832 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17835 *Ralf S. Engelschall*
17837 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17838 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17840 *Arne Ansper <arne@ats.cyber.ee>*
17842 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17843 to be added. Now both 'req' and 'ca' can use new objects defined in the
17848 * Add cool BIO that does syslog (or event log on NT).
17850 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17852 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17853 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17854 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17855 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17859 * Add preliminary config info for new extension code.
17863 * Make RSA_NO_PADDING really use no padding.
17865 *Ulf Moeller <ulf@fitug.de>*
17867 * Generate errors when private/public key check is done.
17871 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17872 for some CRL extensions and new objects added.
17876 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17877 key usage extension and fuller support for authority key id.
17881 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17882 padding method for RSA, which is recommended for new applications in PKCS
17883 #1 v2.0 (RFC 2437, October 1998).
17884 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17885 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17886 against Bleichbacher's attack on RSA.
17887 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
17890 * Updates to the new SSL compression code
17892 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17894 * Fix so that the version number in the master secret, when passed
17895 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
17896 (because the server will not accept higher), that the version number
17897 is 0x03,0x01, not 0x03,0x00
17899 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
17901 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
17902 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
17903 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
17907 * Support for RAW extensions where an arbitrary extension can be
17908 created by including its DER encoding. See apps/openssl.cnf for
17913 * Make sure latest Perl versions don't interpret some generated C array
17914 code as Perl array code in the crypto/err/err_genc.pl script.
17916 *Lars Weber <3weber@informatik.uni-hamburg.de>*
17918 * Modify ms/do_ms.bat to not generate assembly language makefiles since
17919 not many people have the assembler. Various Win32 compilation fixes and
17920 update to the INSTALL.W32 file with (hopefully) more accurate Win32
17921 build instructions.
17925 * Modify configure script 'Configure' to automatically create crypto/date.h
17926 file under Win32 and also build pem.h from pem.org. New script
17927 util/mkfiles.pl to create the MINFO file on environments that can't do a
17928 'make files': perl util/mkfiles.pl >MINFO should work.
17932 * Major rework of DES function declarations, in the pursuit of correctness
17933 and purity. As a result, many evil casts evaporated, and some weirdness,
17934 too. You may find this causes warnings in your code. Zapping your evil
17935 casts will probably fix them. Mostly.
17939 * Fix for a typo in asn1.h. Bug fix to object creation script
17940 obj_dat.pl. It considered a zero in an object definition to mean
17941 "end of object": none of the objects in objects.h have any zeros
17942 so it wasn't spotted.
17944 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
17946 * Add support for Triple DES Cipher Block Chaining with Output Feedback
17947 Masking (CBCM). In the absence of test vectors, the best I have been able
17948 to do is check that the decrypt undoes the encrypt, so far. Send me test
17949 vectors if you have them.
17953 * Correct calculation of key length for export ciphers (too much space was
17954 allocated for null ciphers). This has not been tested!
17958 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
17959 message is now correct (it understands "crypto" and "ssl" on its
17960 command line). There is also now an "update" option. This will update
17961 the util/ssleay.num and util/libeay.num files with any new functions.
17963 perl util/mkdef.pl crypto ssl update
17964 it will update them.
17968 * Overhauled the Perl interface:
17969 - ported BN stuff to OpenSSL's different BN library
17970 - made the perl/ source tree CVS-aware
17971 - renamed the package from SSLeay to OpenSSL (the files still contain
17972 their history because I've copied them in the repository)
17973 - removed obsolete files (the test scripts will be replaced
17974 by better Test::Harness variants in the future)
17976 *Ralf S. Engelschall*
17978 * First cut for a very conservative source tree cleanup:
17979 1. merge various obsolete readme texts into doc/ssleay.txt
17980 where we collect the old documents and readme texts.
17981 2. remove the first part of files where I'm already sure that we no
17982 longer need them because of three reasons: either they are just temporary
17983 files which were left by Eric or they are preserved original files where
17984 I've verified that the diff is also available in the CVS via "cvs diff
17985 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
17986 the crypto/md/ stuff).
17988 *Ralf S. Engelschall*
17990 * More extension code. Incomplete support for subject and issuer alt
17991 name, issuer and authority key id. Change the i2v function parameters
17992 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
17993 what that's for :-) Fix to ASN1 macro which messed up
17994 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
17998 * Preliminary support for ENUMERATED type. This is largely copied from the
18003 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18005 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18007 * Make sure `make rehash` target really finds the `openssl` program.
18009 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18011 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18012 like to hear about it if this slows down other processors.
18016 * Add CygWin32 platform information to Configure script.
18018 *Alan Batie <batie@aahz.jf.intel.com>*
18020 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18022 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18024 * New program nseq to manipulate netscape certificate sequences
18028 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18033 * Fixes to BN code. Previously the default was to define BN_RECURSION
18034 but the BN code had some problems that would cause failures when
18035 doing certificate verification and some other functions.
18037 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18039 * Add ASN1 and PEM code to support netscape certificate sequences.
18043 * Add ASN1 and PEM code to support netscape certificate sequences.
18047 * Add several PKIX and private extended key usage OIDs.
18051 * Modify the 'ca' program to handle the new extension code. Modify
18052 openssl.cnf for new extension format, add comments.
18056 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18057 and add a sample to openssl.cnf so req -x509 now adds appropriate
18062 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18063 error code, add initial support to X509_print() and x509 application.
18067 * Takes a deep breath and start adding X509 V3 extension support code. Add
18068 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18069 stuff is currently isolated and isn't even compiled yet.
18073 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18074 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18075 Removed the versions check from X509 routines when loading extensions:
18076 this allows certain broken certificates that don't set the version
18077 properly to be processed.
18081 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18082 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18083 can still be regenerated with "make depend".
18087 * Spelling mistake in C version of CAST-128.
18089 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18091 * Changes to the error generation code. The perl script err-code.pl
18092 now reads in the old error codes and retains the old numbers, only
18093 adding new ones if necessary. It also only changes the .err files if new
18094 codes are added. The makefiles have been modified to only insert errors
18095 when needed (to avoid needlessly modifying header files). This is done
18096 by only inserting errors if the .err file is newer than the auto generated
18097 C file. To rebuild all the error codes from scratch (the old behaviour)
18098 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18099 or delete all the .err files.
18103 * CAST-128 was incorrectly implemented for short keys. The C version has
18104 been fixed, but is untested. The assembler versions are also fixed, but
18105 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18106 to regenerate it if needed.
18107 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18108 Hagino <itojun@kame.net>*
18110 * File was opened incorrectly in randfile.c.
18112 *Ulf Möller <ulf@fitug.de>*
18114 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18115 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18116 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18117 al: it's just almost always a UTCTime. Note this patch adds new error
18118 codes so do a "make errors" if there are problems.
18122 * Correct Linux 1 recognition in config.
18124 *Ulf Möller <ulf@fitug.de>*
18126 * Remove pointless MD5 hash when using DSA keys in ca.
18128 *Anonymous <nobody@replay.com>*
18130 * Generate an error if given an empty string as a cert directory. Also
18131 generate an error if handed NULL (previously returned 0 to indicate an
18132 error, but didn't set one).
18134 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18136 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18140 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18141 parameters. This was causing a warning which killed off the Win32 compile.
18145 * Remove C++ style comments from crypto/bn/bn_local.h.
18147 *Neil Costigan <neil.costigan@celocom.com>*
18149 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18150 based on a text string, looking up short and long names and finally
18151 "dot" format. The "dot" format stuff didn't work. Added new function
18152 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18153 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18154 OID is not part of the table.
18158 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18159 X509_LOOKUP_by_alias().
18163 * Sort openssl functions by name.
18167 * Get the gendsa program working (hopefully) and add it to app list. Remove
18168 encryption from sample DSA keys (in case anyone is interested the password
18173 * Make *all* `*_free` functions accept a NULL pointer.
18175 *Frans Heymans <fheymans@isaserver.be>*
18177 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18180 *Anonymous <nobody@replay.com>*
18182 * s_server should send the CAfile as acceptable CAs, not its own cert.
18184 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18186 * Don't blow it for numeric -newkey arguments to apps/req.
18188 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18190 * Temp key "for export" tests were wrong in s3_srvr.c.
18192 *Anonymous <nobody@replay.com>*
18194 * Add prototype for temp key callback functions
18195 SSL_CTX_set_tmp_{rsa,dh}_callback().
18199 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18200 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18204 * X509_name_add_entry() freed the wrong thing after an error.
18206 *Arne Ansper <arne@ats.cyber.ee>*
18208 * rsa_eay.c would attempt to free a NULL context.
18210 *Arne Ansper <arne@ats.cyber.ee>*
18212 * BIO_s_socket() had a broken should_retry() on Windoze.
18214 *Arne Ansper <arne@ats.cyber.ee>*
18216 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18218 *Arne Ansper <arne@ats.cyber.ee>*
18220 * Make sure the already existing X509_STORE->depth variable is initialized
18221 in X509_STORE_new(), but document the fact that this variable is still
18222 unused in the certificate verification process.
18224 *Ralf S. Engelschall*
18226 * Fix the various library and apps files to free up pkeys obtained from
18227 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18231 * Fix reference counting in X509_PUBKEY_get(). This makes
18232 demos/maurice/example2.c work, amongst others, probably.
18234 *Steve Henson and Ben Laurie*
18236 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18237 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18238 are no longer created. This way we have a single and consistent command
18239 line interface `openssl <command>`, similar to `cvs <command>`.
18241 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18243 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18244 BIT STRING wrapper always have zero unused bits.
18248 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18252 * Make the top-level INSTALL documentation easier to understand.
18256 * Makefiles updated to exit if an error occurs in a sub-directory
18257 make (including if user presses ^C) [Paul Sutton]
18259 * Make Montgomery context stuff explicit in RSA data structure.
18263 * Fix build order of pem and err to allow for generated pem.h.
18267 * Fix renumbering bug in X509_NAME_delete_entry().
18271 * Enhanced the err-ins.pl script so it makes the error library number
18272 global and can add a library name. This is needed for external ASN1 and
18273 other error libraries.
18277 * Fixed sk_insert which never worked properly.
18281 * Fix ASN1 macros so they can handle indefinite length constructed
18282 EXPLICIT tags. Some non standard certificates use these: they can now
18287 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18288 into a single doc/ssleay.txt bundle. This way the information is still
18289 preserved but no longer messes up this directory. Now it's new room for
18290 the new set of documentation files.
18292 *Ralf S. Engelschall*
18294 * SETs were incorrectly DER encoded. This was a major pain, because they
18295 shared code with SEQUENCEs, which aren't coded the same. This means that
18296 almost everything to do with SETs or SEQUENCEs has either changed name or
18297 number of arguments.
18299 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18301 * Fix test data to work with the above.
18305 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18306 was already fixed by Eric for 0.9.1 it seems.
18308 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18310 * Autodetect FreeBSD3.
18314 * Fix various bugs in Configure. This affects the following platforms:
18318 unixware-2.0-pentium
18323 * Eliminate generated files from CVS. Reorder tests to regenerate files
18324 before they are needed.
18328 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18332 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18334 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18335 changed SSLeay to OpenSSL in version strings.
18337 *Ralf S. Engelschall*
18339 * Some fixups to the top-level documents.
18343 * Fixed the nasty bug where rsaref.h was not found under compile-time
18344 because the symlink to include/ was missing.
18346 *Ralf S. Engelschall*
18348 * Incorporated the popular no-RSA/DSA-only patches
18349 which allow to compile a RSA-free SSLeay.
18351 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18353 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18354 when "ssleay" is still not found.
18356 *Ralf S. Engelschall*
18358 * Added more platforms to Configure: Cray T3E, HPUX 11,
18360 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18362 * Updated the README file.
18364 *Ralf S. Engelschall*
18366 * Added various .cvsignore files in the CVS repository subdirs
18367 to make a "cvs update" really silent.
18369 *Ralf S. Engelschall*
18371 * Recompiled the error-definition header files and added
18372 missing symbols to the Win32 linker tables.
18374 *Ralf S. Engelschall*
18376 * Cleaned up the top-level documents;
18377 o new files: CHANGES and LICENSE
18378 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18379 o merged COPYRIGHT into LICENSE
18380 o removed obsolete TODO file
18381 o renamed MICROSOFT to INSTALL.W32
18383 *Ralf S. Engelschall*
18385 * Removed dummy files from the 0.9.1b source tree:
18386 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18387 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18388 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18389 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18390 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18392 *Ralf S. Engelschall*
18394 * Added various platform portability fixes.
18398 * The Genesis of the OpenSSL rpject:
18399 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18400 Young and Tim J. Hudson created while they were working for C2Net until
18403 *The OpenSSL Project*
18405 ### Changes between 0.9.0b and 0.9.1b [not released]
18407 * Updated a few CA certificates under certs/
18411 * Changed some BIGNUM api stuff.
18415 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18416 DGUX x86, Linux Alpha, etc.
18420 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18421 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18426 * Add -strparse option to asn1pars program which parses nested
18429 *Dr Stephen Henson <shenson@bigfoot.com>*
18431 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18435 * DSA fix for "ca" program.
18439 * Added "-genkey" option to "dsaparam" program.
18443 * Added RIPE MD160 (rmd160) message digest.
18447 * Added -a (all) option to "ssleay version" command.
18451 * Added PLATFORM define which is the id given to Configure.
18455 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18459 * Extended the ASN.1 parser routines.
18463 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18467 * Added a BN_CTX to the BN library.
18471 * Fixed the weak key values in DES library
18475 * Changed API in EVP library for cipher aliases.
18479 * Added support for RC2/64bit cipher.
18483 * Converted the lhash library to the crypto/mem.c functions.
18487 * Added more recognized ASN.1 object ids.
18491 * Added more RSA padding checks for SSL/TLS.
18495 * Added BIO proxy/filter functionality.
18499 * Added extra_certs to SSL_CTX which can be used
18500 send extra CA certificates to the client in the CA cert chain sending
18501 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18505 * Now Fortezza is denied in the authentication phase because
18506 this is key exchange mechanism is not supported by SSLeay at all.
18510 * Additional PKCS1 checks.
18514 * Support the string "TLSv1" for all TLS v1 ciphers.
18518 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18519 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18523 * Fixed a few memory leaks.
18527 * Fixed various code and comment typos.
18531 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18532 bytes sent in the client random.
18534 *Edward Bishop <ebishop@spyglass.com>*
18538 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18539 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18540 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18541 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18542 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18543 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18544 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18545 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18546 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18547 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18548 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18549 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18550 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18551 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18552 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18553 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18554 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18555 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18556 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18557 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18558 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18559 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18560 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18561 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18562 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18563 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18564 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18565 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18566 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18567 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18568 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18569 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18570 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18571 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18572 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18573 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18574 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18575 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18576 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18577 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18578 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18579 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18580 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18581 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18582 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18583 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18584 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18585 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18586 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18587 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18588 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18589 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18590 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18591 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18592 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18593 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18594 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18595 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18596 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18597 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18598 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18599 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18600 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18601 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18602 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18603 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18604 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18605 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18606 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18607 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18608 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18609 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18610 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18611 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18612 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18613 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18614 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18615 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18616 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18617 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18618 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18619 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18620 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18621 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18622 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18623 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18624 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18625 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18626 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18627 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18628 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18629 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18630 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18631 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18632 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18633 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18634 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18635 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18636 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18637 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18638 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18639 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18640 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18641 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18642 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18643 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18644 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18645 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18646 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18647 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18648 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18649 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18650 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18651 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18652 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18653 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18654 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18655 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18656 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18657 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18658 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18659 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18660 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18661 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18662 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18663 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18664 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18665 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18666 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18667 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18668 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18669 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18670 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18671 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18672 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18673 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18674 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18675 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18676 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18677 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18678 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18679 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18680 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18681 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18682 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18683 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18684 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18685 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18686 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18687 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18688 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18689 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18690 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18691 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18692 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18693 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18694 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18695 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18696 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18697 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18698 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655