4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * The -cipher-commands and -digest-commands options of the command line
27 utility list has been deprecated.
28 Instead use the -cipher-algorithms and -digest-algorithms options.
32 * Deprecated all the libcrypto and libssl error string loading
33 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
34 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
35 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
36 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
37 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
38 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
39 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
40 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
41 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
42 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
44 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
45 now loads error strings automatically.
49 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
50 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
51 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
52 are to be used by servers requiring ephemeral DH keys. Instead applications
53 should consider using the built-in DH parameters that are available by
54 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
55 necessary then applications can use the alternative functions
56 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
57 replacement for the "callback" functions. The callback was originally useful
58 in order to have different parameters for export and non-export ciphersuites.
59 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
60 functions should be replaced by one of the other methods described above.
64 * The -crypt option to the passwd command line tool has been removed.
68 * The -C option to the x509, dhparam, dsaparam, and ecparam commands
73 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
75 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
76 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
77 The inverse ciphers use AES decryption for wrapping, and
78 AES encryption for unwrapping.
82 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
83 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
84 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
85 OpenSSL 3.0 these are replaced by the more generic functions
86 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
87 The old versions have been converted to deprecated macros that just call the
92 * The security callback, which can be customised by application code, supports
93 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
94 in the "other" parameter. In most places this is what is passed. All these
95 places occur server side. However there was one client side call of this
96 security operation and it passed a DH object instead. This is incorrect
97 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
98 of the other locations. Therefore this client side call has been changed to
99 pass an EVP_PKEY instead.
103 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
104 interface. Their functionality remains unchanged.
108 * Added new option for 'openssl list', '-providers', which will display the
109 list of loaded providers, their names, version and status. It optionally
110 displays their gettable parameters.
114 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
115 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
116 type is internally recognised so the workaround is no longer needed.
118 Functionality is still retained as it is, but will only work with
119 EVP_PKEYs with a legacy internal key.
123 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
124 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
128 * Changed all "STACK" functions to be macros instead of inline functions. Macro
129 parameters are still checked for type safety at compile time via helper
134 * Remove the RAND_DRBG API
136 The RAND_DRBG API did not fit well into the new provider concept as
137 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
138 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
139 and some of its API calls are rather low-level. This holds in particular
140 for the callback mechanism (RAND_DRBG_set_callbacks()).
142 Adding a compatibility layer to continue supporting the RAND_DRBG API as
143 a legacy API for a regular deprecation period turned out to come at the
144 price of complicating the new provider API unnecessarily. Since the
145 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
148 *Paul Dale and Matthias St. Pierre*
150 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
151 as well as actual hostnames.
155 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
156 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
157 conversely, silently ignore DTLS protocol version bounds when configuring
158 TLS-based contexts. The commands can be repeated to set bounds of both
159 types. The same applies with the corresponding "min_protocol" and
160 "max_protocol" command-line switches, in case some application uses both TLS
163 SSL_CTX instances that are created for a fixed protocol version (e.g.
164 TLSv1_server_method()) also silently ignore version bounds. Previously
165 attempts to apply bounds to these protocol versions would result in an
166 error. Now only the "version-flexible" SSL_CTX instances are subject to
167 limits in configuration files in command-line options.
171 * Deprecated the `ENGINE` API. Engines should be replaced with providers
176 * Reworked the recorded ERR codes to make better space for system errors.
177 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
178 given code is a system error (true) or an OpenSSL error (false).
182 * Reworked the test perl framework to better allow parallel testing.
184 *Nicola Tuveri and David von Oheimb*
186 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
187 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
191 * 'Configure' has been changed to figure out the configuration target if
192 none is given on the command line. Consequently, the 'config' script is
193 now only a mere wrapper. All documentation is changed to only mention
196 *Rich Salz and Richard Levitte*
198 * Added a library context `OSSL_LIB_CTX` that applications as well as
199 other libraries can use to form a separate context within which
200 libcrypto operations are performed.
202 There are two ways this can be used:
204 - Directly, by passing a library context to functions that take
205 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
207 - Indirectly, by creating a new library context and then assigning
208 it as the new default, with `OSSL_LIB_CTX_set0_default`.
210 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
211 apart from the functions directly related to `OSSL_LIB_CTX`, accept
212 NULL to indicate that the default library context should be used.
214 Library code that changes the default library context using
215 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
216 second call before returning to the caller.
218 _(Note: the library context was initially called `OPENSSL_CTX` and
219 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
223 * Handshake now fails if Extended Master Secret extension is dropped
228 * Dropped interactive mode from the 'openssl' program. From now on,
229 the `openssl` command without arguments is equivalent to `openssl
234 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
235 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
236 While the old function names have been retained for backward compatibility
237 they should not be used in new developments
238 because their return values are confusing: Unlike other `_cmp()` functions
239 they do not return 0 in case their arguments are equal.
243 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
244 EC_GROUP_get_field_type().
248 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
249 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
250 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
251 Applications should rely on the library automatically assigning a suitable
252 EC_METHOD internally upon EC_GROUP construction.
256 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
257 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
258 assigned internally without application intervention.
259 Users of EC_GROUP_new() should switch to a different suitable constructor.
263 * Add CAdES-BES signature verification support, mostly derived
264 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
266 *Filipe Raimundo da Silva*
268 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
272 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
273 functions are not widely used and now OpenSSL automatically perform this
274 conversion when needed.
278 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
279 EC_KEY_precompute_mult(). These functions are not widely used and
280 applications should instead switch to named curves which OpenSSL has
281 hardcoded lookup tables for.
285 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
286 should instead use the L<EC_POINT_mul(3)> function.
290 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
291 that are not applicable to the new provider model. Applications should
292 instead use EVP_default_properties_is_fips_enabled() and
293 EVP_default_properties_enable_fips().
297 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
298 is set, an unexpected EOF is ignored, it pretends a close notify was received
299 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
303 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
304 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
305 used and applications should instead use the
306 L<EC_POINT_set_affine_coordinates(3)> and
307 L<EC_POINT_get_affine_coordinates(3)> functions.
311 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
312 arrays to be more easily constructed via a series of utility functions.
313 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
314 the various push functions and finally convert to a passable OSSL_PARAM
315 array using OSSL_PARAM_BLD_to_param().
319 * The security strength of SHA1 and MD5 based signatures in TLS has been
320 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
321 working at the default security level of 1 and instead requires security
322 level 0. The security level can be changed either using the cipher string
323 with @SECLEVEL, or calling SSL_CTX_set_security_level().
327 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
328 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
329 internal keys, if they correspond to one of those built in types.
333 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
334 contain a provider side internal key.
338 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
339 They are old functions that we don't use, and that you could disable with
340 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
344 * Project text documents not yet having a proper file name extension
345 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
346 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
350 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
351 have been converted to Markdown with the goal to produce documents
352 which not only look pretty when viewed online in the browser, but
353 remain well readable inside a plain text editor.
355 To achieve this goal, a 'minimalistic' Markdown style has been applied
356 which avoids formatting elements that interfere too much with the
357 reading flow in the text file. For example, it
359 * avoids [ATX headings][] and uses [setext headings][] instead
360 (which works for `<h1>` and `<h2>` headings only).
361 * avoids [inline links][] and uses [reference links][] instead.
362 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
364 [ATX headings]: https://github.github.com/gfm/#atx-headings
365 [setext headings]: https://github.github.com/gfm/#setext-headings
366 [inline links]: https://github.github.com/gfm/#inline-link
367 [reference links]: https://github.github.com/gfm/#reference-link
368 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
369 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
371 *Matthias St. Pierre*
373 * The test suite is changed to preserve results of each test recipe.
374 A new directory test-runs/ with subdirectories named like the
375 test recipes are created in the build tree for this purpose.
379 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
380 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
381 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
383 *David von Oheimb, Martin Peylo*
385 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
386 The legacy OCSP-focused and only partly documented API is retained.
387 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
391 * Added `util/check-format.pl`, a tool for checking adherence to the
392 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
393 The checks performed are incomplete and yield some false positives.
394 Still the tool should be useful for detecting most typical glitches.
398 * BIO_do_connect and BIO_do_handshake have been extended:
399 If domain name resolution yields multiple IP addresses all of them are tried
400 after connect() failures.
404 * All of the low level RSA functions have been deprecated including:
406 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
407 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
408 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
409 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
410 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
411 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
412 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
413 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
414 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
415 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
416 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
417 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
418 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
419 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
420 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
421 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
422 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
423 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
424 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
425 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
426 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
427 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
428 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
429 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
430 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
431 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
432 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
433 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
435 Use of these low level functions has been informally discouraged for a long
436 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
437 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
438 L<EVP_PKEY_decrypt(3)>.
442 * X509 certificates signed using SHA1 are no longer allowed at security
444 In TLS/SSL the default security level is 1. It can be set either
445 using the cipher string with @SECLEVEL, or calling
446 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
447 a call to SSL_CTX_use_certificate() will fail if the security level is not
449 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
450 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
455 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
456 modified to use PKEY APIs. These commands are now in maintenance mode
457 and no new features will be added to them.
461 * The command line utility rsautl has been deprecated.
462 Instead use the pkeyutl program.
466 * The command line utilities genrsa and rsa have been modified to use PKEY
467 APIs. They now write PKCS#8 keys by default. These commands are now in
468 maintenance mode and no new features will be added to them.
472 * All of the low level DH functions have been deprecated including:
474 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
475 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
476 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
477 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
478 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
479 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
480 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
481 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
482 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
483 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
484 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
485 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
486 DH_meth_get_generate_params and DH_meth_set_generate_params.
488 Use of these low level functions has been informally discouraged for a long
489 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
490 and L<EVP_PKEY_derive(3)>.
492 Additionally functions that read and write DH objects such as d2i_DHparams,
493 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
494 functions have also been deprecated. Applications should instead use the
495 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
497 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
498 EVP_PKEY_assign_DH(), EVP_PKEY_get0_DH, EVP_PKEY_get1_DH, EVP_PKEY_set1_DH
499 are also deprecated. Applications should instead either read or write an
500 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs. Or load an
501 EVP_PKEY directly from DH data using EVP_PKEY_fromdata().
503 *Paul Dale and Matt Caswell*
505 * All of the low level DSA functions have been deprecated including:
507 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
508 DSA_get_default_method, DSA_set_method, DSA_get_method,
509 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
510 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
511 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
512 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
513 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
514 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
515 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
516 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
517 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
518 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
519 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
522 Use of these low level functions has been informally discouraged for a long
523 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
524 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
528 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
529 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
530 This means that applications don't have to look at the curve NID and
531 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
532 However, they still can, that EVP_PKEY_set_alias_type() call acts as
533 a no-op when the EVP_PKEY is already of the given type.
535 Parameter and key generation is also reworked to make it possible
536 to generate EVP_PKEY_SM2 parameters and keys without having to go
537 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
538 However, code that does the latter will still work as before.
542 * Deprecated low level ECDH and ECDSA functions. These include:
544 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
545 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
548 Use of these low level functions has been informally discouraged for a long
549 time. Instead applications should use the EVP_PKEY_derive(3),
550 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
554 * Deprecated the EC_KEY_METHOD functions. These include:
556 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
557 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
558 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
559 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
560 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
561 EC_KEY_METHOD_get_verify.
563 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
567 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
568 and EVP_PKEY_decrypt() instead.
569 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
570 and EVP_PKEY_encrypt() instead.
574 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
575 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
576 a new formulation to include all the things it can be used for,
577 as well as words of caution.
581 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
582 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
586 * All of the low level HMAC functions have been deprecated including:
588 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
589 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
592 Use of these low level functions has been informally discouraged for a long
593 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
594 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
595 and L<EVP_MAC_final(3)>.
599 * Over two thousand fixes were made to the documentation, including:
600 - Common options (such as -rand/-writerand, TLS version control, etc)
601 were refactored and point to newly-enhanced descriptions in openssl.pod.
602 - Added style conformance for all options (with help from Richard Levitte),
603 documented all reported missing options, added a CI build to check
604 that all options are documented and that no unimplemented options
606 - Documented some internals, such as all use of environment variables.
607 - Addressed all internal broken L<> references.
611 * All of the low level CMAC functions have been deprecated including:
613 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
614 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
616 Use of these low level functions has been informally discouraged for a long
617 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
618 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
619 and L<EVP_MAC_final(3)>.
623 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
624 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
627 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
628 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
629 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
630 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
631 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
632 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
633 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
634 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
635 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
636 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
638 Use of these low level functions has been informally discouraged
639 for a long time. Applications should use the EVP_DigestInit_ex(3),
640 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
644 * Corrected the documentation of the return values from the `EVP_DigestSign*`
645 set of functions. The documentation mentioned negative values for some
646 errors, but this was never the case, so the mention of negative values
649 Code that followed the documentation and thereby check with something
650 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
654 * All of the low level cipher functions have been deprecated including:
656 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
657 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
658 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
659 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
660 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
661 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
662 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
663 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
664 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
665 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
666 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
667 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
668 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
669 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
670 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
671 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
672 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
673 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
674 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
675 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
676 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
677 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
678 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
679 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
680 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
681 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
682 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
683 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
684 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
686 Use of these low level functions has been informally discouraged for
687 a long time. Applications should use the high level EVP APIs, e.g.
688 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
689 equivalently named decrypt functions instead.
691 *Matt Caswell and Paul Dale*
693 * Removed include/openssl/opensslconf.h.in and replaced it with
694 include/openssl/configuration.h.in, which differs in not including
695 <openssl/macros.h>. A short header include/openssl/opensslconf.h
696 was added to include both.
698 This allows internal hacks where one might need to modify the set
699 of configured macros, for example this if deprecated symbols are
700 still supposed to be available internally:
702 #include <openssl/configuration.h>
704 #undef OPENSSL_NO_DEPRECATED
705 #define OPENSSL_SUPPRESS_DEPRECATED
707 #include <openssl/macros.h>
709 This should not be used by applications that use the exported
710 symbols, as that will lead to linking errors.
714 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
715 used in exponentiation with 512-bit moduli. No EC algorithms are
716 affected. Analysis suggests that attacks against 2-prime RSA1024,
717 3-prime RSA1536, and DSA1024 as a result of this defect would be very
718 difficult to perform and are not believed likely. Attacks against DH512
719 are considered just feasible. However, for an attack the target would
720 have to re-use the DH512 private key, which is not recommended anyway.
721 Also applications directly using the low level API BN_mod_exp may be
722 affected if they use BN_FLG_CONSTTIME.
727 * Most memory-debug features have been deprecated, and the functionality
728 replaced with no-ops.
732 * Added documentation for the STACK API.
736 * Introduced a new method type and API, OSSL_ENCODER, to
737 represent generic encoders. An implementation is expected to
738 be able to encode an object associated with a given name (such
739 as an algorithm name for an asymmetric key) into forms given by
740 implementation properties.
742 Encoders are primarily used from inside libcrypto, through
743 calls to functions like EVP_PKEY_print_private(),
744 PEM_write_bio_PrivateKey() and similar.
746 Encoders are specified in such a way that they can be made to
747 directly handle the provider side portion of an object, if this
748 provider side part comes from the same provider as the encoder
749 itself, but can also be made to handle objects in parametrized
750 form (as an OSSL_PARAM array of data). This allows a provider to
751 offer generic encoders as a service for any other provider.
755 * Added a .pragma directive to the syntax of configuration files, to
756 allow varying behavior in a supported and predictable manner.
757 Currently added pragma:
761 This allows dollar signs to be a keyword character unless it's
762 followed by a opening brace or parenthesis. This is useful for
763 platforms where dollar signs are commonly used in names, such as
764 volume names and system directory names on VMS.
768 * Added functionality to create an EVP_PKEY from user data. This
769 is effectively the same as creating a RSA, DH or DSA object and
770 then assigning them to an EVP_PKEY, but directly using algorithm
771 agnostic EVP functions. A benefit is that this should be future
772 proof for public key algorithms to come.
776 * Change the interpretation of the '--api' configuration option to
777 mean that this is a desired API compatibility level with no
778 further meaning. The previous interpretation, that this would
779 also mean to remove all deprecated symbols up to and including
780 the given version, no requires that 'no-deprecated' is also used
781 in the configuration.
783 When building applications, the desired API compatibility level
784 can be set with the OPENSSL_API_COMPAT macro like before. For
785 API compatibility version below 3.0, the old style numerical
786 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
787 For version 3.0 and on, the value is expected to be the decimal
788 value calculated from the major and minor version like this:
790 MAJOR * 10000 + MINOR * 100
794 -DOPENSSL_API_COMPAT=30000 For 3.0
795 -DOPENSSL_API_COMPAT=30200 For 3.2
797 To hide declarations that are deprecated up to and including the
798 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
799 given when building the application as well.
803 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
804 access to certificate and CRL stores via URIs and OSSL_STORE
807 This adds the following functions:
809 - X509_LOOKUP_store()
810 - X509_STORE_load_file()
811 - X509_STORE_load_path()
812 - X509_STORE_load_store()
813 - SSL_add_store_cert_subjects_to_stack()
814 - SSL_CTX_set_default_verify_store()
815 - SSL_CTX_load_verify_file()
816 - SSL_CTX_load_verify_dir()
817 - SSL_CTX_load_verify_store()
821 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
822 The presence of this system service is determined at run-time.
826 * Added functionality to create an EVP_PKEY context based on data
827 for methods from providers. This takes an algorithm name and a
828 property query string and simply stores them, with the intent
829 that any operation that uses this context will use those strings
830 to fetch the needed methods implicitly, thereby making the port
831 of application written for pre-3.0 OpenSSL easier.
835 * The undocumented function NCONF_WIN32() has been deprecated; for
836 conversion details see the HISTORY section of doc/man5/config.pod
840 * Introduced the new functions EVP_DigestSignInit_ex() and
841 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
842 EVP_DigestVerifyUpdate() have been converted to functions. See the man
843 pages for further details.
847 * Over two thousand fixes were made to the documentation, including:
848 adding missing command flags, better style conformance, documentation
851 *Rich Salz, Richard Levitte*
853 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
854 X25519, X448, Ed25519 and Ed448.
858 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
863 * Deprecated the public definition of ERR_STATE as well as the function
864 ERR_get_state(). This is done in preparation of making ERR_STATE an
869 * Added ERR functionality to give callers access to the stored function
870 names that have replaced the older function code based functions.
872 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
873 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
874 ERR_peek_error_all() and ERR_peek_last_error_all().
876 These functions have become deprecated: ERR_get_error_line(),
877 ERR_get_error_line_data(), ERR_peek_error_line_data(),
878 ERR_peek_last_error_line_data() and ERR_func_error_string().
880 Users are recommended to use ERR_get_error_all(), or to pick information
881 with ERR_peek functions and finish off with getting the error code by using
886 * Extended testing to be verbose for failing tests only. The make variables
887 VERBOSE_FAILURE or VF can be used to enable this:
889 $ make VF=1 test # Unix
890 $ mms /macro=(VF=1) test ! OpenVMS
891 $ nmake VF=1 test # Windows
895 * Added several checks to X509_verify_cert() according to requirements in
896 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
897 (which may be done by using the CLI option `-x509_strict`):
898 * The basicConstraints of CA certificates must be marked critical.
899 * CA certificates must explicitly include the keyUsage extension.
900 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
901 * The issuer name of any certificate must not be empty.
902 * The subject name of CA certs, certs with keyUsage crlSign,
903 and certs without subjectAlternativeName must not be empty.
904 * If a subjectAlternativeName extension is given it must not be empty.
905 * The signatureAlgorithm field and the cert signature must be consistent.
906 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
907 must not be marked critical.
908 * The authorityKeyIdentifier must be given for X.509v3 certs
909 unless they are self-signed.
910 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
914 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
915 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
919 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
920 used even when parsing explicit parameters, when loading a encoded key
921 or calling `EC_GROUP_new_from_ecpkparameters()`/
922 `EC_GROUP_new_from_ecparameters()`.
923 This prevents bypass of security hardening and performance gains,
924 especially for curves with specialized EC_METHODs.
925 By default, if a key encoded with explicit parameters is loaded and later
926 encoded, the output is still encoded with explicit parameters, even if
927 internally a "named" EC_GROUP is used for computation.
931 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
932 this change, EC_GROUP_set_generator would accept order and/or cofactor as
933 NULL. After this change, only the cofactor parameter can be NULL. It also
934 does some minimal sanity checks on the passed order.
939 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
940 An attack is simple, if the first CMS_recipientInfo is valid but the
941 second CMS_recipientInfo is chosen ciphertext. If the second
942 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
943 encryption key will be replaced by garbage, and the message cannot be
944 decoded, but if the RSA decryption fails, the correct encryption key is
945 used and the recipient will not notice the attack.
946 As a work around for this potential attack the length of the decrypted
947 key must be equal to the cipher default key length, in case the
948 certifiate is not given and all recipientInfo are tried out.
949 The old behaviour can be re-enabled in the CMS code by setting the
950 CMS_DEBUG_DECRYPT flag.
954 * Early start up entropy quality from the DEVRANDOM seed source has been
955 improved for older Linux systems. The RAND subsystem will wait for
956 /dev/random to be producing output before seeding from /dev/urandom.
957 The seeded state is stored for future library initialisations using
958 a system global shared memory segment. The shared memory identifier
959 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
960 the desired value. The default identifier is 114.
964 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
965 when primes for RSA keys are computed.
966 Since we previously always generated primes == 2 (mod 3) for RSA keys,
967 the 2-prime and 3-prime RSA modules were easy to distinguish, since
968 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
969 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
970 This avoids possible fingerprinting of newly generated RSA modules.
974 * Correct the extended master secret constant on EBCDIC systems. Without this
975 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
976 negotiate EMS will fail. Unfortunately this also means that TLS connections
977 between EBCDIC systems with this fix, and EBCDIC systems without this
978 fix will fail if they negotiate EMS.
982 * Changed the library initialisation so that the config file is now loaded
983 by default. This was already the case for libssl. It now occurs for both
984 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
985 OPENSSL_init_crypto() to suppress automatic loading of a config file.
989 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
990 where the former acts as a replacement for ERR_put_error(), and the
991 latter replaces the combination ERR_put_error()+ERR_add_error_data().
992 ERR_raise_data() adds more flexibility by taking a format string and
993 an arbitrary number of arguments following it, to be processed with
998 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
999 to check if a named provider is loaded and available. When called, it
1000 will also activate all fallback providers if such are still present.
1004 * Enforce a minimum DH modulus size of 512 bits.
1008 * Changed DH parameters to generate the order q subgroup instead of 2q.
1009 Previously generated DH parameters are still accepted by DH_check
1010 but DH_generate_key works around that by clearing bit 0 of the
1011 private key for those. This avoids leaking bit 0 of the private key.
1015 * Significantly reduce secure memory usage by the randomness pools.
1019 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1024 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1025 algorithms. An implementation of a key exchange algorithm can be obtained
1026 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1027 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1028 the older EVP_PKEY_derive_init() function. See the man pages for the new
1029 functions for further details.
1033 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1037 * Removed the function names from error messages and deprecated the
1040 * Removed NextStep support and the macro OPENSSL_UNISTD
1044 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1045 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1046 Also removed "export var as function" capability; we do not export
1047 variables, only functions.
1051 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1052 an error and 1 indicating success. In previous versions of OpenSSL this
1053 was a void type. If a key was set longer than the maximum possible this
1058 * Support SM2 signing and verification schemes with X509 certificate.
1062 * Use SHA256 as the default digest for TS query in the ts app.
1066 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1067 This checks that the salt length is at least 128 bits, the derived key
1068 length is at least 112 bits, and that the iteration count is at least 1000.
1069 For backwards compatibility these checks are disabled by default in the
1070 default provider, but are enabled by default in the fips provider.
1071 To enable or disable these checks use the control
1072 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1076 * Default cipher lists/suites are now available via a function, the
1077 #defines are deprecated.
1081 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1082 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1083 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1087 * Join the directories crypto/x509 and crypto/x509v3
1091 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1092 This changes the size when using the genpkey app when no size is given. It
1093 fixes an omission in earlier changes that changed all RSA, DSA and DH
1094 generation apps to use 2048 bits by default.
1098 * Added command 'openssl kdf' that uses the EVP_KDF API.
1102 * Added command 'openssl mac' that uses the EVP_MAC API.
1106 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1107 as default directories. Also added the command 'openssl info'
1108 for scripting purposes.
1112 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1113 deprecated. These undocumented functions were never integrated into the EVP
1114 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1115 Bi-directional IGE mode. These modes were never formally standardised and
1116 usage of these functions is believed to be very small. In particular
1117 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1118 is ever used. The security implications are believed to be minimal, but
1119 this issue was never fixed for backwards compatibility reasons. New code
1120 should not use these modes.
1124 * Add prediction resistance to the DRBG reseeding process.
1128 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1129 mandated by IEEE Std 1619-2018.
1133 * Added newline escaping functionality to a filename when using openssl dgst.
1134 This output format is to replicate the output format found in the `*sum`
1135 checksum programs. This aims to preserve backward compatibility.
1137 *Matt Eaton, Richard Levitte, and Paul Dale*
1139 * Removed the heartbeat message in DTLS feature, as it has very
1140 little usage and doesn't seem to fulfill a valuable purpose.
1141 The configuration option is now deprecated.
1145 * Changed the output of 'openssl {digestname} < file' to display the
1146 digest name in its output.
1150 * Added a new generic trace API which provides support for enabling
1151 instrumentation through trace output. This feature is mainly intended
1152 as an aid for developers and is disabled by default. To utilize it,
1153 OpenSSL needs to be configured with the `enable-trace` option.
1155 If the tracing API is enabled, the application can activate trace output
1156 by registering BIOs as trace channels for a number of tracing and debugging
1159 The 'openssl' application has been expanded to enable any of the types
1160 available via environment variables defined by the user, and serves as
1161 one possible example on how to use this functionality.
1163 *Richard Levitte & Matthias St. Pierre*
1165 * Added build tests for C++. These are generated files that only do one
1166 thing, to include one public OpenSSL head file each. This tests that
1167 the public header files can be usefully included in a C++ application.
1169 This test isn't enabled by default. It can be enabled with the option
1170 'enable-buildtest-c++'.
1174 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1178 * Add KMAC to EVP_MAC.
1182 * Added property based algorithm implementation selection framework to
1187 * Added SCA hardening for modular field inversion in EC_GROUP through
1188 a new dedicated field_inv() pointer in EC_METHOD.
1189 This also addresses a leakage affecting conversions from projective
1190 to affine coordinates.
1192 *Billy Bob Brumley, Nicola Tuveri*
1194 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1195 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1196 those algorithms that were already supported through the EVP_PKEY API
1197 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1198 and scrypt are now wrappers that call EVP_KDF.
1202 * Build devcrypto engine as a dynamic engine.
1204 *Eneas U de Queiroz*
1206 * Add keyed BLAKE2 to EVP_MAC.
1210 * Fix a bug in the computation of the endpoint-pair shared secret used
1211 by DTLS over SCTP. This breaks interoperability with older versions
1212 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1213 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1214 interoperability with such broken implementations. However, enabling
1215 this switch breaks interoperability with correct implementations.
1217 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1218 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1222 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1226 * Change the license to the Apache License v2.0.
1230 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1232 - Major releases (indicated by incrementing the MAJOR release number)
1233 may introduce incompatible API/ABI changes.
1234 - Minor releases (indicated by incrementing the MINOR release number)
1235 may introduce new features but retain API/ABI compatibility.
1236 - Patch releases (indicated by incrementing the PATCH number)
1237 are intended for bug fixes and other improvements of existing
1238 features only (like improving performance or adding documentation)
1239 and retain API/ABI compatibility.
1243 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1247 * Remove the 'dist' target and add a tarball building script. The
1248 'dist' target has fallen out of use, and it shouldn't be
1249 necessary to configure just to create a source distribution.
1253 * Recreate the OS390-Unix config target. It no longer relies on a
1254 special script like it did for OpenSSL pre-1.1.0.
1258 * Instead of having the source directories listed in Configure, add
1259 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1264 * Add GMAC to EVP_MAC.
1268 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1272 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1273 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1274 to facilitate the continued use of MACs through raw private keys in
1275 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1279 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1280 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1284 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1285 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1286 are retained for backwards compatibility.
1290 * AES-XTS mode now enforces that its two keys are different to mitigate
1291 the attacked described in "Efficient Instantiations of Tweakable
1292 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1293 Details of this attack can be obtained from:
1294 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1298 * Rename the object files, i.e. give them other names than in previous
1299 versions. Their names now include the name of the final product, as
1300 well as its type mnemonic (bin, lib, shlib).
1304 * Added new option for 'openssl list', '-objects', which will display the
1305 list of built in objects, i.e. OIDs with names.
1309 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1310 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1315 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1316 improves application performance by removing data copies and providing
1317 applications with zero-copy system calls such as sendfile and splice.
1321 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1322 option is set, openssl cleanses (zeroize) plaintext bytes from
1323 internal buffers after delivering them to the application. Note,
1324 the application is still responsible for cleansing other copies
1325 (e.g.: data received by SSL_read(3)).
1329 * `PKCS12_parse` now maintains the order of the parsed certificates
1330 when outputting them via `*ca` (rather than reversing it).
1334 * Deprecated pthread fork support methods. These were unused so no
1335 replacement is required.
1337 - OPENSSL_fork_prepare()
1338 - OPENSSL_fork_parent()
1339 - OPENSSL_fork_child()
1346 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1348 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1349 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1350 If an attacker can control both items being compared then this could lead
1351 to a possible denial of service attack. OpenSSL itself uses the
1352 GENERAL_NAME_cmp function for two purposes:
1353 1) Comparing CRL distribution point names between an available CRL and a
1354 CRL distribution point embedded in an X509 certificate
1355 2) When verifying that a timestamp response token signer matches the
1356 timestamp authority name (exposed via the API functions
1357 TS_RESP_verify_response and TS_RESP_verify_token)
1362 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1364 * Certificates with explicit curve parameters are now disallowed in
1365 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1369 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1370 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1371 conversely, silently ignore DTLS protocol version bounds when configuring
1372 TLS-based contexts. The commands can be repeated to set bounds of both
1373 types. The same applies with the corresponding "min_protocol" and
1374 "max_protocol" command-line switches, in case some application uses both TLS
1377 SSL_CTX instances that are created for a fixed protocol version (e.g.
1378 TLSv1_server_method()) also silently ignore version bounds. Previously
1379 attempts to apply bounds to these protocol versions would result in an
1380 error. Now only the "version-flexible" SSL_CTX instances are subject to
1381 limits in configuration files in command-line options.
1385 * Handshake now fails if Extended Master Secret extension is dropped
1390 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1392 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1394 * Fixed segmentation fault in SSL_check_chain()
1395 Server or client applications that call the SSL_check_chain() function
1396 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1397 dereference as a result of incorrect handling of the
1398 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1399 or unrecognised signature algorithm is received from the peer. This could
1400 be exploited by a malicious peer in a Denial of Service attack.
1405 * Added AES consttime code for no-asm configurations
1406 an optional constant time support for AES was added
1407 when building openssl for no-asm.
1408 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1409 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1410 At this time this feature is by default disabled.
1411 It will be enabled by default in 3.0.
1415 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1417 * Revert the change of EOF detection while reading in libssl to avoid
1418 regressions in applications depending on the current way of reporting
1419 the EOF. As the existing method is not fully accurate the change to
1420 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1421 branch and will be present in the 3.0 release.
1425 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1426 when primes for RSA keys are computed.
1427 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1428 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1429 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1430 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1431 This avoids possible fingerprinting of newly generated RSA modules.
1435 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1437 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1438 while reading in libssl then we would report an error back to the
1439 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1440 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1441 therefore give a hint as to what went wrong.
1445 * Check that ed25519 and ed448 are allowed by the security level. Previously
1446 signature algorithms not using an MD were not being checked that they were
1447 allowed by the security level.
1451 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1452 was not quite right. The behaviour was not consistent between resumption
1453 and normal handshakes, and also not quite consistent with historical
1454 behaviour. The behaviour in various scenarios has been clarified and
1455 it has been updated to make it match historical behaviour as closely as
1460 * *[VMS only]* The header files that the VMS compilers include automatically,
1461 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1462 that the C++ compiler doesn't understand. This is a shortcoming in the
1463 compiler, but can be worked around with `__cplusplus` guards.
1465 C++ applications that use OpenSSL libraries must be compiled using the
1466 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1467 functions. Otherwise, only functions with symbols of less than 31
1468 characters can be used, as the linker will not be able to successfully
1469 resolve symbols with longer names.
1473 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1474 The presence of this system service is determined at run-time.
1478 * Added newline escaping functionality to a filename when using openssl dgst.
1479 This output format is to replicate the output format found in the `*sum`
1480 checksum programs. This aims to preserve backward compatibility.
1482 *Matt Eaton, Richard Levitte, and Paul Dale*
1484 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1489 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1491 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1492 number generator (RNG). This was intended to include protection in the
1493 event of a fork() system call in order to ensure that the parent and child
1494 processes did not share the same RNG state. However this protection was not
1495 being used in the default case.
1497 A partial mitigation for this issue is that the output from a high
1498 precision timer is mixed into the RNG state so the likelihood of a parent
1499 and child process sharing state is significantly reduced.
1501 If an application already calls OPENSSL_init_crypto() explicitly using
1502 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1505 *Matthias St. Pierre*
1507 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1508 used even when parsing explicit parameters, when loading a encoded key
1509 or calling `EC_GROUP_new_from_ecpkparameters()`/
1510 `EC_GROUP_new_from_ecparameters()`.
1511 This prevents bypass of security hardening and performance gains,
1512 especially for curves with specialized EC_METHODs.
1513 By default, if a key encoded with explicit parameters is loaded and later
1514 encoded, the output is still encoded with explicit parameters, even if
1515 internally a "named" EC_GROUP is used for computation.
1519 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1520 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1521 NULL. After this change, only the cofactor parameter can be NULL. It also
1522 does some minimal sanity checks on the passed order.
1527 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1528 An attack is simple, if the first CMS_recipientInfo is valid but the
1529 second CMS_recipientInfo is chosen ciphertext. If the second
1530 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1531 encryption key will be replaced by garbage, and the message cannot be
1532 decoded, but if the RSA decryption fails, the correct encryption key is
1533 used and the recipient will not notice the attack.
1534 As a work around for this potential attack the length of the decrypted
1535 key must be equal to the cipher default key length, in case the
1536 certifiate is not given and all recipientInfo are tried out.
1537 The old behaviour can be re-enabled in the CMS code by setting the
1538 CMS_DEBUG_DECRYPT flag.
1543 * Early start up entropy quality from the DEVRANDOM seed source has been
1544 improved for older Linux systems. The RAND subsystem will wait for
1545 /dev/random to be producing output before seeding from /dev/urandom.
1546 The seeded state is stored for future library initialisations using
1547 a system global shared memory segment. The shared memory identifier
1548 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1549 the desired value. The default identifier is 114.
1553 * Correct the extended master secret constant on EBCDIC systems. Without this
1554 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1555 negotiate EMS will fail. Unfortunately this also means that TLS connections
1556 between EBCDIC systems with this fix, and EBCDIC systems without this
1557 fix will fail if they negotiate EMS.
1561 * Use Windows installation paths in the mingw builds
1563 Mingw isn't a POSIX environment per se, which means that Windows
1564 paths should be used for installation.
1569 * Changed DH_check to accept parameters with order q and 2q subgroups.
1570 With order 2q subgroups the bit 0 of the private key is not secret
1571 but DH_generate_key works around that by clearing bit 0 of the
1572 private key for those. This avoids leaking bit 0 of the private key.
1576 * Significantly reduce secure memory usage by the randomness pools.
1580 * Revert the DEVRANDOM_WAIT feature for Linux systems
1582 The DEVRANDOM_WAIT feature added a select() call to wait for the
1583 /dev/random device to become readable before reading from the
1584 /dev/urandom device.
1586 It turned out that this change had negative side effects on
1587 performance which were not acceptable. After some discussion it
1588 was decided to revert this feature and leave it up to the OS
1589 resp. the platform maintainer to ensure a proper initialization
1590 during early boot time.
1592 *Matthias St. Pierre*
1594 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1596 * Add build tests for C++. These are generated files that only do one
1597 thing, to include one public OpenSSL head file each. This tests that
1598 the public header files can be usefully included in a C++ application.
1600 This test isn't enabled by default. It can be enabled with the option
1601 'enable-buildtest-c++'.
1605 * Enable SHA3 pre-hashing for ECDSA and DSA.
1609 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1610 This changes the size when using the genpkey app when no size is given. It
1611 fixes an omission in earlier changes that changed all RSA, DSA and DH
1612 generation apps to use 2048 bits by default.
1616 * Reorganize the manual pages to consistently have RETURN VALUES,
1617 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1618 util/fix-doc-nits accordingly.
1620 *Paul Yang, Joshua Lock*
1622 * Add the missing accessor EVP_PKEY_get0_engine()
1626 * Have apps like 's_client' and 's_server' output the signature scheme
1627 along with other cipher suite parameters when debugging.
1631 * Make OPENSSL_config() error agnostic again.
1635 * Do the error handling in RSA decryption constant time.
1639 * Prevent over long nonces in ChaCha20-Poly1305.
1641 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1642 for every encryption operation. RFC 7539 specifies that the nonce value
1643 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1644 and front pads the nonce with 0 bytes if it is less than 12
1645 bytes. However it also incorrectly allows a nonce to be set of up to 16
1646 bytes. In this case only the last 12 bytes are significant and any
1647 additional leading bytes are ignored.
1649 It is a requirement of using this cipher that nonce values are
1650 unique. Messages encrypted using a reused nonce value are susceptible to
1651 serious confidentiality and integrity attacks. If an application changes
1652 the default nonce length to be longer than 12 bytes and then makes a
1653 change to the leading bytes of the nonce expecting the new value to be a
1654 new unique nonce then such an application could inadvertently encrypt
1655 messages with a reused nonce.
1657 Additionally the ignored bytes in a long nonce are not covered by the
1658 integrity guarantee of this cipher. Any application that relies on the
1659 integrity of these ignored leading bytes of a long nonce may be further
1660 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1661 is safe because no such use sets such a long nonce value. However user
1662 applications that use this cipher directly and set a non-default nonce
1663 length to be longer than 12 bytes may be vulnerable.
1665 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1671 * Add DEVRANDOM_WAIT feature for Linux systems
1673 On older Linux systems where the getrandom() system call is not available,
1674 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1675 Contrary to getrandom(), the /dev/urandom device will not block during
1676 early boot when the kernel CSPRNG has not been seeded yet.
1678 To mitigate this known weakness, use select() to wait for /dev/random to
1679 become readable before reading from /dev/urandom.
1681 * Ensure that SM2 only uses SM3 as digest algorithm
1685 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1687 * Change the info callback signals for the start and end of a post-handshake
1688 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1689 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1690 confused by this and assume that a TLSv1.2 renegotiation has started. This
1691 can break KeyUpdate handling. Instead we no longer signal the start and end
1692 of a post handshake message exchange (although the messages themselves are
1693 still signalled). This could break some applications that were expecting
1694 the old signals. However without this KeyUpdate is not usable for many
1699 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1701 * Timing vulnerability in DSA signature generation
1703 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1704 timing side channel attack. An attacker could use variations in the signing
1705 algorithm to recover the private key.
1707 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1712 * Timing vulnerability in ECDSA signature generation
1714 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1715 timing side channel attack. An attacker could use variations in the signing
1716 algorithm to recover the private key.
1718 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1723 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1724 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1725 of two gigabytes and the error handling improved.
1727 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1728 categorized as a normal bug, not a security issue, because the DRBG reseeds
1729 automatically and is fully functional even without additional randomness
1730 provided by the application.
1732 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1734 * Add a new ClientHello callback. Provides a callback interface that gives
1735 the application the ability to adjust the nascent SSL object at the
1736 earliest stage of ClientHello processing, immediately after extensions have
1737 been collected but before they have been processed. In particular, this
1738 callback can adjust the supported TLS versions in response to the contents
1743 * Add SM2 base algorithm support.
1747 * s390x assembly pack: add (improved) hardware-support for the following
1748 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1749 aes-cfb/cfb8, aes-ecb.
1753 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1754 parameter is no longer accepted, as it leads to a corrupt table. NULL
1755 pem_str is reserved for alias entries only.
1759 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1760 step for prime curves. The new implementation is based on formulae from
1761 differential addition-and-doubling in homogeneous projective coordinates
1762 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1763 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1764 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1765 to work in projective coordinates.
1767 *Billy Bob Brumley, Nicola Tuveri*
1769 * Change generating and checking of primes so that the error rate of not
1770 being prime depends on the intended use based on the size of the input.
1771 For larger primes this will result in more rounds of Miller-Rabin.
1772 The maximal error rate for primes with more than 1080 bits is lowered
1775 *Kurt Roeckx, Annie Yousar*
1777 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1781 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1782 moving between systems, and to avoid confusion when a Windows build is
1783 done with mingw vs with MSVC. For POSIX installs, there's still a
1784 symlink or copy named 'tsget' to avoid that confusion as well.
1788 * Revert blinding in ECDSA sign and instead make problematic addition
1789 length-invariant. Switch even to fixed-length Montgomery multiplication.
1793 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1794 step for binary curves. The new implementation is based on formulae from
1795 differential addition-and-doubling in mixed Lopez-Dahab projective
1796 coordinates, modified to independently blind the operands.
1798 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1800 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1801 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1802 EC_METHODs to implement their own specialized "ladder step", to take
1803 advantage of more favorable coordinate systems or more efficient
1804 differential addition-and-doubling algorithms.
1806 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1808 * Modified the random device based seed sources to keep the relevant
1809 file descriptors open rather than reopening them on each access.
1810 This allows such sources to operate in a chroot() jail without
1811 the associated device nodes being available. This behaviour can be
1812 controlled using RAND_keep_random_devices_open().
1816 * Numerous side-channel attack mitigations have been applied. This may have
1817 performance impacts for some algorithms for the benefit of improved
1818 security. Specific changes are noted in this change log by their respective
1823 * AIX shared library support overhaul. Switch to AIX "natural" way of
1824 handling shared libraries, which means collecting shared objects of
1825 different versions and bitnesses in one common archive. This allows to
1826 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1827 doesn't affect the way 3rd party applications are linked, only how
1828 multi-version installation is managed.
1832 * Make ec_group_do_inverse_ord() more robust and available to other
1833 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1834 mitigations are applied to the fallback BN_mod_inverse().
1835 When using this function rather than BN_mod_inverse() directly, new
1836 EC cryptosystem implementations are then safer-by-default.
1840 * Add coordinate blinding for EC_POINT and implement projective
1841 coordinate blinding for generic prime curves as a countermeasure to
1842 chosen point SCA attacks.
1844 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1846 * Add blinding to ECDSA and DSA signatures to protect against side channel
1847 attacks discovered by Keegan Ryan (NCC Group).
1851 * Enforce checking in the pkeyutl command line app to ensure that the input
1852 length does not exceed the maximum supported digest length when performing
1853 a sign, verify or verifyrecover operation.
1857 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1858 I/O in combination with something like select() or poll() will hang. This
1859 can be turned off again using SSL_CTX_clear_mode().
1860 Many applications do not properly handle non-application data records, and
1861 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1862 around the problems in those applications, but can also break some.
1863 It's recommended to read the manpages about SSL_read(), SSL_write(),
1864 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1865 SSL_CTX_set_read_ahead() again.
1869 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1870 now allow empty (zero character) pass phrases.
1874 * Apply blinding to binary field modular inversion and remove patent
1875 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1879 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1880 binary and prime elliptic curves.
1884 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1885 constant time fixed point multiplication.
1889 * Revise elliptic curve scalar multiplication with timing attack
1890 defenses: ec_wNAF_mul redirects to a constant time implementation
1891 when computing fixed point and variable point multiplication (which
1892 in OpenSSL are mostly used with secret scalars in keygen, sign,
1893 ECDH derive operations).
1894 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1897 * Updated CONTRIBUTING
1901 * Updated DRBG / RAND to request nonce and additional low entropy
1902 randomness from the system.
1904 *Matthias St. Pierre*
1906 * Updated 'openssl rehash' to use OpenSSL consistent default.
1910 * Moved the load of the ssl_conf module to libcrypto, which helps
1911 loading engines that libssl uses before libssl is initialised.
1915 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1919 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1921 *Ingo Schwarze, Rich Salz*
1923 * Added output of accepting IP address and port for 'openssl s_server'
1927 * Added a new API for TLSv1.3 ciphersuites:
1928 SSL_CTX_set_ciphersuites()
1929 SSL_set_ciphersuites()
1933 * Memory allocation failures consistently add an error to the error
1938 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1939 in libcrypto when run as setuid/setgid.
1943 * Load any config file by default when libssl is used.
1947 * Added new public header file <openssl/rand_drbg.h> and documentation
1948 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1950 *Matthias St. Pierre*
1952 * QNX support removed (cannot find contributors to get their approval
1953 for the license change).
1957 * TLSv1.3 replay protection for early data has been implemented. See the
1958 SSL_read_early_data() man page for further details.
1962 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1963 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1964 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1965 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1966 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1967 configuration has been separated out. See the ciphers man page or the
1968 SSL_CTX_set_ciphersuites() man page for more information.
1972 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1973 in responder mode now supports the new "-multi" option, which
1974 spawns the specified number of child processes to handle OCSP
1975 requests. The "-timeout" option now also limits the OCSP
1976 responder's patience to wait to receive the full client request
1977 on a newly accepted connection. Child processes are respawned
1978 as needed, and the CA index file is automatically reloaded
1979 when changed. This makes it possible to run the "ocsp" responder
1980 as a long-running service, making the OpenSSL CA somewhat more
1981 feature-complete. In this mode, most diagnostic messages logged
1982 after entering the event loop are logged via syslog(3) rather than
1987 * Added support for X448 and Ed448. Heavily based on original work by
1992 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1993 objects loaded. This adds the functions OSSL_STORE_expect() and
1994 OSSL_STORE_find() as well as needed tools to construct searches and
1995 get the search data out of them.
1999 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2000 version of OpenSSL should review their configuration settings to ensure
2001 that they are still appropriate for TLSv1.3. For further information see:
2002 <https://wiki.openssl.org/index.php/TLS1.3>
2006 * Grand redesign of the OpenSSL random generator
2008 The default RAND method now utilizes an AES-CTR DRBG according to
2009 NIST standard SP 800-90Ar1. The new random generator is essentially
2010 a port of the default random generator from the OpenSSL FIPS 2.0
2011 object module. It is a hybrid deterministic random bit generator
2012 using an AES-CTR bit stream and which seeds and reseeds itself
2013 automatically using trusted system entropy sources.
2015 Some of its new features are:
2016 - Support for multiple DRBG instances with seed chaining.
2017 - The default RAND method makes use of a DRBG.
2018 - There is a public and private DRBG instance.
2019 - The DRBG instances are fork-safe.
2020 - Keep all global DRBG instances on the secure heap if it is enabled.
2021 - The public and private DRBG instance are per thread for lock free
2024 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2026 * Changed Configure so it only says what it does and doesn't dump
2027 so much data. Instead, ./configdata.pm should be used as a script
2028 to display all sorts of configuration data.
2032 * Added processing of "make variables" to Configure.
2036 * Added SHA512/224 and SHA512/256 algorithm support.
2040 * The last traces of Netware support, first removed in 1.1.0, have
2045 * Get rid of Makefile.shared, and in the process, make the processing
2046 of certain files (rc.obj, or the .def/.map/.opt files produced from
2047 the ordinal files) more visible and hopefully easier to trace and
2048 debug (or make silent).
2052 * Make it possible to have environment variable assignments as
2053 arguments to config / Configure.
2057 * Add multi-prime RSA (RFC 8017) support.
2061 * Add SM3 implemented according to GB/T 32905-2016
2062 *Jack Lloyd <jack.lloyd@ribose.com>,*
2063 *Ronald Tse <ronald.tse@ribose.com>,*
2064 *Erick Borsboom <erick.borsboom@ribose.com>*
2066 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2067 as documented in RFC6066.
2068 Based on a patch from Tomasz Moń
2070 *Filipe Raimundo da Silva*
2072 * Add SM4 implemented according to GB/T 32907-2016.
2073 *Jack Lloyd <jack.lloyd@ribose.com>,*
2074 *Ronald Tse <ronald.tse@ribose.com>,*
2075 *Erick Borsboom <erick.borsboom@ribose.com>*
2077 * Reimplement -newreq-nodes and ERR_error_string_n; the
2078 original author does not agree with the license change.
2082 * Add ARIA AEAD TLS support.
2086 * Some macro definitions to support VS6 have been removed. Visual
2087 Studio 6 has not worked since 1.1.0
2091 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2092 without clearing the errors.
2096 * Add "atfork" functions. If building on a system that without
2097 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2098 requirements. The RAND facility now uses/requires this.
2106 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2107 not possible to disable entirely. However, it's still possible to
2108 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2111 To disable, configure with 'no-ui-console'. 'no-ui' is still
2112 possible to use as an alias. Check at compile time with the
2113 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2114 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2118 * Add a STORE module, which implements a uniform and URI based reader of
2119 stores that can contain keys, certificates, CRLs and numerous other
2120 objects. The main API is loosely based on a few stdio functions,
2121 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2122 OSSL_STORE_error and OSSL_STORE_close.
2123 The implementation uses backends called "loaders" to implement arbitrary
2124 URI schemes. There is one built in "loader" for the 'file' scheme.
2128 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2129 then adjusted to work on FreeBSD 8.4 as well.
2130 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2131 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2135 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2136 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2137 error code calls like this:
2139 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2141 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2142 that can be encoded in C. For the foreseeable future, this will only
2145 *Richard Levitte and Tim Hudson*
2147 * Removed BSD cryptodev engine.
2151 * Add a build target 'build_all_generated', to build all generated files
2152 and only that. This can be used to prepare everything that requires
2153 things like perl for a system that lacks perl and then move everything
2154 to that system and do the rest of the build there.
2158 * In the UI interface, make it possible to duplicate the user data. This
2159 can be used by engines that need to retain the data for a longer time
2160 than just the call where this user data is passed.
2164 * Ignore the '-named_curve auto' value for compatibility of applications
2167 *Tomas Mraz <tmraz@fedoraproject.org>*
2169 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2170 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2171 alerts across multiple records (some of which could be empty). In practice
2172 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2173 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2174 support this at all. Supporting it adds significant complexity to the
2175 record layer, and its removal is unlikely to cause interoperability
2180 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2181 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2182 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2187 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2188 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2190 *Richard Levitte, Andy Polyakov*
2192 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2197 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2198 platform rather than 'mingw'.
2202 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2203 success if they are asked to add an object which already exists
2204 in the store. This change cascades to other functions which load
2205 certificates and CRLs.
2209 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2210 facilitate stack unwinding even from assembly subroutines.
2214 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2215 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2219 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2220 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2221 which is the minimum version we support.
2225 * Certificate time validation (X509_cmp_time) enforces stricter
2226 compliance with RFC 5280. Fractional seconds and timezone offsets
2227 are no longer allowed.
2231 * Add support for ARIA
2235 * s_client will now send the Server Name Indication (SNI) extension by
2236 default unless the new "-noservername" option is used. The server name is
2237 based on the host provided to the "-connect" option unless overridden by
2238 using "-servername".
2242 * Add support for SipHash
2246 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2247 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2248 prevent issues where no progress is being made and the peer continually
2249 sends unrecognised record types, using up resources processing them.
2253 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2254 using the algorithm defined in
2255 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2259 * Heartbeat support has been removed; the ABI is changed for now.
2261 *Richard Levitte, Rich Salz*
2263 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2267 * The RSA "null" method, which was partially supported to avoid patent
2268 issues, has been replaced to always returns NULL.
2275 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2277 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2278 used even when parsing explicit parameters, when loading a encoded key
2279 or calling `EC_GROUP_new_from_ecpkparameters()`/
2280 `EC_GROUP_new_from_ecparameters()`.
2281 This prevents bypass of security hardening and performance gains,
2282 especially for curves with specialized EC_METHODs.
2283 By default, if a key encoded with explicit parameters is loaded and later
2284 encoded, the output is still encoded with explicit parameters, even if
2285 internally a "named" EC_GROUP is used for computation.
2289 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2290 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2291 NULL. After this change, only the cofactor parameter can be NULL. It also
2292 does some minimal sanity checks on the passed order.
2297 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2298 An attack is simple, if the first CMS_recipientInfo is valid but the
2299 second CMS_recipientInfo is chosen ciphertext. If the second
2300 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2301 encryption key will be replaced by garbage, and the message cannot be
2302 decoded, but if the RSA decryption fails, the correct encryption key is
2303 used and the recipient will not notice the attack.
2304 As a work around for this potential attack the length of the decrypted
2305 key must be equal to the cipher default key length, in case the
2306 certifiate is not given and all recipientInfo are tried out.
2307 The old behaviour can be re-enabled in the CMS code by setting the
2308 CMS_DEBUG_DECRYPT flag.
2313 * Use Windows installation paths in the mingw builds
2315 Mingw isn't a POSIX environment per se, which means that Windows
2316 paths should be used for installation.
2321 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2323 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2324 This changes the size when using the genpkey app when no size is given. It
2325 fixes an omission in earlier changes that changed all RSA, DSA and DH
2326 generation apps to use 2048 bits by default.
2330 * Prevent over long nonces in ChaCha20-Poly1305.
2332 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2333 for every encryption operation. RFC 7539 specifies that the nonce value
2334 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2335 and front pads the nonce with 0 bytes if it is less than 12
2336 bytes. However it also incorrectly allows a nonce to be set of up to 16
2337 bytes. In this case only the last 12 bytes are significant and any
2338 additional leading bytes are ignored.
2340 It is a requirement of using this cipher that nonce values are
2341 unique. Messages encrypted using a reused nonce value are susceptible to
2342 serious confidentiality and integrity attacks. If an application changes
2343 the default nonce length to be longer than 12 bytes and then makes a
2344 change to the leading bytes of the nonce expecting the new value to be a
2345 new unique nonce then such an application could inadvertently encrypt
2346 messages with a reused nonce.
2348 Additionally the ignored bytes in a long nonce are not covered by the
2349 integrity guarantee of this cipher. Any application that relies on the
2350 integrity of these ignored leading bytes of a long nonce may be further
2351 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2352 is safe because no such use sets such a long nonce value. However user
2353 applications that use this cipher directly and set a non-default nonce
2354 length to be longer than 12 bytes may be vulnerable.
2356 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2362 * Added SCA hardening for modular field inversion in EC_GROUP through
2363 a new dedicated field_inv() pointer in EC_METHOD.
2364 This also addresses a leakage affecting conversions from projective
2365 to affine coordinates.
2367 *Billy Bob Brumley, Nicola Tuveri*
2369 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2370 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2374 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2378 * Remove the 'dist' target and add a tarball building script. The
2379 'dist' target has fallen out of use, and it shouldn't be
2380 necessary to configure just to create a source distribution.
2384 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2386 * Timing vulnerability in DSA signature generation
2388 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2389 timing side channel attack. An attacker could use variations in the signing
2390 algorithm to recover the private key.
2392 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2397 * Timing vulnerability in ECDSA signature generation
2399 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2400 timing side channel attack. An attacker could use variations in the signing
2401 algorithm to recover the private key.
2403 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2408 * Add coordinate blinding for EC_POINT and implement projective
2409 coordinate blinding for generic prime curves as a countermeasure to
2410 chosen point SCA attacks.
2412 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2414 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2416 * Client DoS due to large DH parameter
2418 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2419 malicious server can send a very large prime value to the client. This will
2420 cause the client to spend an unreasonably long period of time generating a
2421 key for this prime resulting in a hang until the client has finished. This
2422 could be exploited in a Denial Of Service attack.
2424 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2429 * Cache timing vulnerability in RSA Key Generation
2431 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2432 a cache timing side channel attack. An attacker with sufficient access to
2433 mount cache timing attacks during the RSA key generation process could
2434 recover the private key.
2436 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2437 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2442 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2443 parameter is no longer accepted, as it leads to a corrupt table. NULL
2444 pem_str is reserved for alias entries only.
2448 * Revert blinding in ECDSA sign and instead make problematic addition
2449 length-invariant. Switch even to fixed-length Montgomery multiplication.
2453 * Change generating and checking of primes so that the error rate of not
2454 being prime depends on the intended use based on the size of the input.
2455 For larger primes this will result in more rounds of Miller-Rabin.
2456 The maximal error rate for primes with more than 1080 bits is lowered
2459 *Kurt Roeckx, Annie Yousar*
2461 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2465 * Add blinding to ECDSA and DSA signatures to protect against side channel
2466 attacks discovered by Keegan Ryan (NCC Group).
2470 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2471 now allow empty (zero character) pass phrases.
2475 * Certificate time validation (X509_cmp_time) enforces stricter
2476 compliance with RFC 5280. Fractional seconds and timezone offsets
2477 are no longer allowed.
2481 * Fixed a text canonicalisation bug in CMS
2483 Where a CMS detached signature is used with text content the text goes
2484 through a canonicalisation process first prior to signing or verifying a
2485 signature. This process strips trailing space at the end of lines, converts
2486 line terminators to CRLF and removes additional trailing line terminators
2487 at the end of a file. A bug in the canonicalisation process meant that
2488 some characters, such as form-feed, were incorrectly treated as whitespace
2489 and removed. This is contrary to the specification (RFC5485). This fix
2490 could mean that detached text data signed with an earlier version of
2491 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2492 signed with a fixed OpenSSL may fail to verify with an earlier version of
2493 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2494 and use the "-binary" flag (for the "cms" command line application) or set
2495 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2499 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2501 * Constructed ASN.1 types with a recursive definition could exceed the stack
2503 Constructed ASN.1 types with a recursive definition (such as can be found
2504 in PKCS7) could eventually exceed the stack given malicious input with
2505 excessive recursion. This could result in a Denial Of Service attack. There
2506 are no such structures used within SSL/TLS that come from untrusted sources
2507 so this is considered safe.
2509 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2515 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2517 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2518 effectively reduced to only comparing the least significant bit of each
2519 byte. This allows an attacker to forge messages that would be considered as
2520 authenticated in an amount of tries lower than that guaranteed by the
2521 security claims of the scheme. The module can only be compiled by the
2522 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2524 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2530 * Add a build target 'build_all_generated', to build all generated files
2531 and only that. This can be used to prepare everything that requires
2532 things like perl for a system that lacks perl and then move everything
2533 to that system and do the rest of the build there.
2537 * Backport SSL_OP_NO_RENGOTIATION
2539 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2540 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2541 changes this is no longer possible in 1.1.0. Therefore the new
2542 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2543 1.1.0 to provide equivalent functionality.
2545 Note that if an application built against 1.1.0h headers (or above) is run
2546 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2547 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2551 * Removed the OS390-Unix config target. It relied on a script that doesn't
2556 * rsaz_1024_mul_avx2 overflow bug on x86_64
2558 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2559 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2560 Analysis suggests that attacks against RSA and DSA as a result of this
2561 defect would be very difficult to perform and are not believed likely.
2562 Attacks against DH1024 are considered just feasible, because most of the
2563 work necessary to deduce information about a private key may be performed
2564 offline. The amount of resources required for such an attack would be
2565 significant. However, for an attack on TLS to be meaningful, the server
2566 would have to share the DH1024 private key among multiple clients, which is
2567 no longer an option since CVE-2016-0701.
2569 This only affects processors that support the AVX2 but not ADX extensions
2570 like Intel Haswell (4th generation).
2572 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2573 was originally found via the OSS-Fuzz project.
2578 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2580 * bn_sqrx8x_internal carry bug on x86_64
2582 There is a carry propagating bug in the x86_64 Montgomery squaring
2583 procedure. No EC algorithms are affected. Analysis suggests that attacks
2584 against RSA and DSA as a result of this defect would be very difficult to
2585 perform and are not believed likely. Attacks against DH are considered just
2586 feasible (although very difficult) because most of the work necessary to
2587 deduce information about a private key may be performed offline. The amount
2588 of resources required for such an attack would be very significant and
2589 likely only accessible to a limited number of attackers. An attacker would
2590 additionally need online access to an unpatched system using the target
2591 private key in a scenario with persistent DH parameters and a private
2592 key that is shared between multiple clients.
2594 This only affects processors that support the BMI1, BMI2 and ADX extensions
2595 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2597 This issue was reported to OpenSSL by the OSS-Fuzz project.
2602 * Malformed X.509 IPAddressFamily could cause OOB read
2604 If an X.509 certificate has a malformed IPAddressFamily extension,
2605 OpenSSL could do a one-byte buffer overread. The most likely result
2606 would be an erroneous display of the certificate in text format.
2608 This issue was reported to OpenSSL by the OSS-Fuzz project.
2613 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2615 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2616 platform rather than 'mingw'.
2620 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2621 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2622 which is the minimum version we support.
2626 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2628 * Encrypt-Then-Mac renegotiation crash
2630 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2631 negotiated where it was not in the original handshake (or vice-versa) then
2632 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2633 and servers are affected.
2635 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2640 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2642 * Truncated packet could crash via OOB read
2644 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2645 cipher is being used, then a truncated packet can cause that host to
2646 perform an out-of-bounds read, usually resulting in a crash.
2648 This issue was reported to OpenSSL by Robert Święcki of Google.
2653 * Bad (EC)DHE parameters cause a client crash
2655 If a malicious server supplies bad parameters for a DHE or ECDHE key
2656 exchange then this can result in the client attempting to dereference a
2657 NULL pointer leading to a client crash. This could be exploited in a Denial
2660 This issue was reported to OpenSSL by Guido Vranken.
2665 * BN_mod_exp may produce incorrect results on x86_64
2667 There is a carry propagating bug in the x86_64 Montgomery squaring
2668 procedure. No EC algorithms are affected. Analysis suggests that attacks
2669 against RSA and DSA as a result of this defect would be very difficult to
2670 perform and are not believed likely. Attacks against DH are considered just
2671 feasible (although very difficult) because most of the work necessary to
2672 deduce information about a private key may be performed offline. The amount
2673 of resources required for such an attack would be very significant and
2674 likely only accessible to a limited number of attackers. An attacker would
2675 additionally need online access to an unpatched system using the target
2676 private key in a scenario with persistent DH parameters and a private
2677 key that is shared between multiple clients. For example this can occur by
2678 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2679 similar to CVE-2015-3193 but must be treated as a separate problem.
2681 This issue was reported to OpenSSL by the OSS-Fuzz project.
2686 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2688 * ChaCha20/Poly1305 heap-buffer-overflow
2690 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2691 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2692 crash. This issue is not considered to be exploitable beyond a DoS.
2694 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2699 * CMS Null dereference
2701 Applications parsing invalid CMS structures can crash with a NULL pointer
2702 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2703 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2704 structure callback if an attempt is made to free certain invalid encodings.
2705 Only CHOICE structures using a callback which do not handle NULL value are
2708 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2713 * Montgomery multiplication may produce incorrect results
2715 There is a carry propagating bug in the Broadwell-specific Montgomery
2716 multiplication procedure that handles input lengths divisible by, but
2717 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2718 and DH private keys are impossible. This is because the subroutine in
2719 question is not used in operations with the private key itself and an input
2720 of the attacker's direct choice. Otherwise the bug can manifest itself as
2721 transient authentication and key negotiation failures or reproducible
2722 erroneous outcome of public-key operations with specially crafted input.
2723 Among EC algorithms only Brainpool P-512 curves are affected and one
2724 presumably can attack ECDH key negotiation. Impact was not analyzed in
2725 detail, because pre-requisites for attack are considered unlikely. Namely
2726 multiple clients have to choose the curve in question and the server has to
2727 share the private key among them, neither of which is default behaviour.
2728 Even then only clients that chose the curve will be affected.
2730 This issue was publicly reported as transient failures and was not
2731 initially recognized as a security issue. Thanks to Richard Morgan for
2732 providing reproducible case.
2737 * Removed automatic addition of RPATH in shared libraries and executables,
2738 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2742 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2744 * Fix Use After Free for large message sizes
2746 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2747 message larger than approx 16k is received then the underlying buffer to
2748 store the incoming message is reallocated and moved. Unfortunately a
2749 dangling pointer to the old location is left which results in an attempt to
2750 write to the previously freed location. This is likely to result in a
2751 crash, however it could potentially lead to execution of arbitrary code.
2753 This issue only affects OpenSSL 1.1.0a.
2755 This issue was reported to OpenSSL by Robert Święcki.
2760 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2762 * OCSP Status Request extension unbounded memory growth
2764 A malicious client can send an excessively large OCSP Status Request
2765 extension. If that client continually requests renegotiation, sending a
2766 large OCSP Status Request extension each time, then there will be unbounded
2767 memory growth on the server. This will eventually lead to a Denial Of
2768 Service attack through memory exhaustion. Servers with a default
2769 configuration are vulnerable even if they do not support OCSP. Builds using
2770 the "no-ocsp" build time option are not affected.
2772 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2777 * SSL_peek() hang on empty record
2779 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2780 sends an empty record. This could be exploited by a malicious peer in a
2781 Denial Of Service attack.
2783 This issue was reported to OpenSSL by Alex Gaynor.
2788 * Excessive allocation of memory in tls_get_message_header() and
2789 dtls1_preprocess_fragment()
2791 A (D)TLS message includes 3 bytes for its length in the header for the
2792 message. This would allow for messages up to 16Mb in length. Messages of
2793 this length are excessive and OpenSSL includes a check to ensure that a
2794 peer is sending reasonably sized messages in order to avoid too much memory
2795 being consumed to service a connection. A flaw in the logic of version
2796 1.1.0 means that memory for the message is allocated too early, prior to
2797 the excessive message length check. Due to way memory is allocated in
2798 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2799 to service a connection. This could lead to a Denial of Service through
2800 memory exhaustion. However, the excessive message length check still takes
2801 place, and this would cause the connection to immediately fail. Assuming
2802 that the application calls SSL_free() on the failed connection in a timely
2803 manner then the 21Mb of allocated memory will then be immediately freed
2804 again. Therefore the excessive memory allocation will be transitory in
2805 nature. This then means that there is only a security impact if:
2807 1) The application does not call SSL_free() in a timely manner in the event
2808 that the connection fails
2810 2) The application is working in a constrained environment where there is
2811 very little free memory
2813 3) The attacker initiates multiple connection attempts such that there are
2814 multiple connections in a state where memory has been allocated for the
2815 connection; SSL_free() has not yet been called; and there is insufficient
2816 memory to service the multiple requests.
2818 Except in the instance of (1) above any Denial Of Service is likely to be
2819 transitory because as soon as the connection fails the memory is
2820 subsequently freed again in the SSL_free() call. However there is an
2821 increased risk during this period of application crashes due to the lack of
2822 memory - which would then mean a more serious Denial of Service.
2824 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2825 (CVE-2016-6307 and CVE-2016-6308)
2829 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2830 had to be removed. Primary reason is that vendor assembler can't
2831 assemble our modules with -KPIC flag. As result it, assembly
2832 support, was not even available as option. But its lack means
2833 lack of side-channel resistant code, which is incompatible with
2834 security by todays standards. Fortunately gcc is readily available
2835 prepackaged option, which we firmly point at...
2839 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2841 * Windows command-line tool supports UTF-8 opt-in option for arguments
2842 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2843 (to any value) allows Windows user to access PKCS#12 file generated
2844 with Windows CryptoAPI and protected with non-ASCII password, as well
2845 as files generated under UTF-8 locale on Linux also protected with
2850 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2851 have been disabled by default and removed from DEFAULT, just like RC4.
2852 See the RC4 item below to re-enable both.
2856 * The method for finding the storage location for the Windows RAND seed file
2857 has changed. First we check %RANDFILE%. If that is not set then we check
2858 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2859 all else fails we fall back to C:\.
2863 * The EVP_EncryptUpdate() function has had its return type changed from void
2864 to int. A return of 0 indicates and error while a return of 1 indicates
2869 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2870 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2871 off the constant time implementation for RSA, DSA and DH have been made
2872 no-ops and deprecated.
2876 * Windows RAND implementation was simplified to only get entropy by
2877 calling CryptGenRandom(). Various other RAND-related tickets
2880 *Joseph Wylie Yandle, Rich Salz*
2882 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2883 and `OPENSSL_LH_`, respectively. The old names are available
2884 with API compatibility. They new names are now completely documented.
2888 * Unify TYPE_up_ref(obj) methods signature.
2889 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2890 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2891 int (instead of void) like all others TYPE_up_ref() methods.
2892 So now these methods also check the return value of CRYPTO_atomic_add(),
2893 and the validity of object reference counter.
2895 *fdasilvayy@gmail.com*
2897 * With Windows Visual Studio builds, the .pdb files are installed
2898 alongside the installed libraries and executables. For a static
2899 library installation, ossl_static.pdb is the associate compiler
2900 generated .pdb file to be used when linking programs.
2904 * Remove openssl.spec. Packaging files belong with the packagers.
2908 * Automatic Darwin/OSX configuration has had a refresh, it will now
2909 recognise x86_64 architectures automatically. You can still decide
2910 to build for a different bitness with the environment variable
2911 KERNEL_BITS (can be 32 or 64), for example:
2913 KERNEL_BITS=32 ./config
2917 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2918 256 bit AES and HMAC with SHA256.
2922 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2926 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2930 * To enable users to have their own config files and build file templates,
2931 Configure looks in the directory indicated by the environment variable
2932 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2933 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2934 name and is used as is.
2938 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2939 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2940 X509_CERT_FILE_CTX was removed.
2944 * "shared" builds are now the default. To create only static libraries use
2945 the "no-shared" Configure option.
2949 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2950 All of these option have not worked for some while and are fundamental
2955 * Make various cleanup routines no-ops and mark them as deprecated. Most
2956 global cleanup functions are no longer required because they are handled
2957 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2958 Explicitly de-initing can cause problems (e.g. where a library that uses
2959 OpenSSL de-inits, but an application is still using it). The affected
2960 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2961 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2962 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2963 COMP_zlib_cleanup().
2967 * --strict-warnings no longer enables runtime debugging options
2968 such as REF_DEBUG. Instead, debug options are automatically
2969 enabled with '--debug' builds.
2971 *Andy Polyakov, Emilia Käsper*
2973 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2974 have been moved out of the public header files. New functions for managing
2975 these have been added.
2979 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2980 objects have been moved out of the public header files. New
2981 functions for managing these have been added.
2985 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2986 have been moved out of the public header files. New functions for managing
2987 these have been added.
2991 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2992 moved out of the public header files. New functions for managing these
2997 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3001 * Removed the mk1mf build scripts.
3005 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3006 it is always safe to #include a header now.
3010 * Removed the aged BC-32 config and all its supporting scripts
3014 * Removed support for Ultrix, Netware, and OS/2.
3018 * Add support for HKDF.
3020 *Alessandro Ghedini*
3022 * Add support for blake2b and blake2s
3026 * Added support for "pipelining". Ciphers that have the
3027 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3028 encryptions/decryptions simultaneously. There are currently no built-in
3029 ciphers with this property but the expectation is that engines will be able
3030 to offer it to significantly improve throughput. Support has been extended
3031 into libssl so that multiple records for a single connection can be
3032 processed in one go (for >=TLS 1.1).
3036 * Added the AFALG engine. This is an async capable engine which is able to
3037 offload work to the Linux kernel. In this initial version it only supports
3038 AES128-CBC. The kernel must be version 4.1.0 or greater.
3042 * OpenSSL now uses a new threading API. It is no longer necessary to
3043 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3044 are two supported threading models: pthreads and windows threads. It is
3045 also possible to configure OpenSSL at compile time for "no-threads". The
3046 old threading API should no longer be used. The functions have been
3047 replaced with "no-op" compatibility macros.
3049 *Alessandro Ghedini, Matt Caswell*
3051 * Modify behavior of ALPN to invoke callback after SNI/servername
3052 callback, such that updates to the SSL_CTX affect ALPN.
3056 * Add SSL_CIPHER queries for authentication and key-exchange.
3060 * Changes to the DEFAULT cipherlist:
3061 - Prefer (EC)DHE handshakes over plain RSA.
3062 - Prefer AEAD ciphers over legacy ciphers.
3063 - Prefer ECDSA over RSA when both certificates are available.
3064 - Prefer TLSv1.2 ciphers/PRF.
3065 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3070 * Change the ECC default curve list to be this, in order: x25519,
3071 secp256r1, secp521r1, secp384r1.
3075 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3076 disabled by default. They can be re-enabled using the
3077 enable-weak-ssl-ciphers option to Configure.
3081 * If the server has ALPN configured, but supports no protocols that the
3082 client advertises, send a fatal "no_application_protocol" alert.
3083 This behaviour is SHALL in RFC 7301, though it isn't universally
3084 implemented by other servers.
3088 * Add X25519 support.
3089 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3090 for public and private key encoding using the format documented in
3091 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3092 key generation and key derivation.
3094 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3099 * Deprecate SRP_VBASE_get_by_user.
3100 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3101 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3102 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3103 seed, even if the seed is configured.
3105 Users should use SRP_VBASE_get1_by_user instead. Note that in
3106 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3107 also that even though configuring the SRP seed attempts to hide
3108 invalid usernames by continuing the handshake with fake
3109 credentials, this behaviour is not constant time and no strong
3110 guarantees are made that the handshake is indistinguishable from
3111 that of a valid user.
3115 * Configuration change; it's now possible to build dynamic engines
3116 without having to build shared libraries and vice versa. This
3117 only applies to the engines in engines/, those in crypto/engine/
3118 will always be built into libcrypto (i.e. "static").
3120 Building dynamic engines is enabled by default; to disable, use
3121 the configuration option "disable-dynamic-engine".
3123 The only requirements for building dynamic engines are the
3124 presence of the DSO module and building with position independent
3125 code, so they will also automatically be disabled if configuring
3126 with "disable-dso" or "disable-pic".
3128 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3129 are also taken away from openssl/opensslconf.h, as they are
3134 * Configuration change; if there is a known flag to compile
3135 position independent code, it will always be applied on the
3136 libcrypto and libssl object files, and never on the application
3137 object files. This means other libraries that use routines from
3138 libcrypto / libssl can be made into shared libraries regardless
3139 of how OpenSSL was configured.
3141 If this isn't desirable, the configuration options "disable-pic"
3142 or "no-pic" can be used to disable the use of PIC. This will
3143 also disable building shared libraries and dynamic engines.
3147 * Removed JPAKE code. It was experimental and has no wide use.
3151 * The INSTALL_PREFIX Makefile variable has been renamed to
3152 DESTDIR. That makes for less confusion on what this variable
3153 is for. Also, the configuration option --install_prefix is
3158 * Heartbeat for TLS has been removed and is disabled by default
3159 for DTLS; configure with enable-heartbeats. Code that uses the
3160 old #define's might need to be updated.
3162 *Emilia Käsper, Rich Salz*
3164 * Rename REF_CHECK to REF_DEBUG.
3168 * New "unified" build system
3170 The "unified" build system is aimed to be a common system for all
3171 platforms we support. With it comes new support for VMS.
3173 This system builds supports building in a different directory tree
3174 than the source tree. It produces one Makefile (for unix family
3175 or lookalikes), or one descrip.mms (for VMS).
3177 The source of information to make the Makefile / descrip.mms is
3178 small files called 'build.info', holding the necessary
3179 information for each directory with source to compile, and a
3180 template in Configurations, like unix-Makefile.tmpl or
3183 With this change, the library names were also renamed on Windows
3184 and on VMS. They now have names that are closer to the standard
3185 on Unix, and include the major version number, and in certain
3186 cases, the architecture they are built for. See "Notes on shared
3187 libraries" in INSTALL.
3189 We rely heavily on the perl module Text::Template.
3193 * Added support for auto-initialisation and de-initialisation of the library.
3194 OpenSSL no longer requires explicit init or deinit routines to be called,
3195 except in certain circumstances. See the OPENSSL_init_crypto() and
3196 OPENSSL_init_ssl() man pages for further information.
3200 * The arguments to the DTLSv1_listen function have changed. Specifically the
3201 "peer" argument is now expected to be a BIO_ADDR object.
3203 * Rewrite of BIO networking library. The BIO library lacked consistent
3204 support of IPv6, and adding it required some more extensive
3205 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3206 which hold all types of addresses and chains of address information.
3207 It also introduces a new API, with functions like BIO_socket,
3208 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3209 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3210 have been adapted accordingly.
3214 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3219 * CRIME protection: disable compression by default, even if OpenSSL is
3220 compiled with zlib enabled. Applications can still enable compression
3221 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3222 using the SSL_CONF library to configure compression.
3226 * The signature of the session callback configured with
3227 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3228 was explicitly marked as `const unsigned char*` instead of
3233 * Always DPURIFY. Remove the use of uninitialized memory in the
3234 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3238 * Removed many obsolete configuration items, including
3239 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3240 MD2_CHAR, MD2_INT, MD2_LONG
3242 IDEA_SHORT, IDEA_LONG
3243 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3245 *Rich Salz, with advice from Andy Polyakov*
3247 * Many BN internals have been moved to an internal header file.
3249 *Rich Salz with help from Andy Polyakov*
3251 * Configuration and writing out the results from it has changed.
3252 Files such as Makefile include/openssl/opensslconf.h and are now
3253 produced through general templates, such as Makefile.in and
3254 crypto/opensslconf.h.in and some help from the perl module
3257 Also, the center of configuration information is no longer
3258 Makefile. Instead, Configure produces a perl module in
3259 configdata.pm which holds most of the config data (in the hash
3260 table %config), the target data that comes from the target
3261 configuration in one of the `Configurations/*.conf` files (in
3266 * To clarify their intended purposes, the Configure options
3267 --prefix and --openssldir change their semantics, and become more
3268 straightforward and less interdependent.
3270 --prefix shall be used exclusively to give the location INSTALLTOP
3271 where programs, scripts, libraries, include files and manuals are
3272 going to be installed. The default is now /usr/local.
3274 --openssldir shall be used exclusively to give the default
3275 location OPENSSLDIR where certificates, private keys, CRLs are
3276 managed. This is also where the default openssl.cnf gets
3278 If the directory given with this option is a relative path, the
3279 values of both the --prefix value and the --openssldir value will
3280 be combined to become OPENSSLDIR.
3281 The default for --openssldir is INSTALLTOP/ssl.
3283 Anyone who uses --openssldir to specify where OpenSSL is to be
3284 installed MUST change to use --prefix instead.
3288 * The GOST engine was out of date and therefore it has been removed. An up
3289 to date GOST engine is now being maintained in an external repository.
3290 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3291 support for GOST ciphersuites (these are only activated if a GOST engine
3296 * EGD is no longer supported by default; use enable-egd when
3299 *Ben Kaduk and Rich Salz*
3301 * The distribution now has Makefile.in files, which are used to
3302 create Makefile's when Configure is run. *Configure must be run
3303 before trying to build now.*
3307 * The return value for SSL_CIPHER_description() for error conditions
3312 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3314 Obtaining and performing DNSSEC validation of TLSA records is
3315 the application's responsibility. The application provides
3316 the TLSA records of its choice to OpenSSL, and these are then
3317 used to authenticate the peer.
3319 The TLSA records need not even come from DNS. They can, for
3320 example, be used to implement local end-entity certificate or
3321 trust-anchor "pinning", where the "pin" data takes the form
3322 of TLSA records, which can augment or replace verification
3323 based on the usual WebPKI public certification authorities.
3327 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3328 continues to support deprecated interfaces in default builds.
3329 However, applications are strongly advised to compile their
3330 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3331 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3332 or the 1.1.0 releases.
3334 In environments in which all applications have been ported to
3335 not use any deprecated interfaces OpenSSL's Configure script
3336 should be used with the --api=1.1.0 option to entirely remove
3337 support for the deprecated features from the library and
3338 unconditionally disable them in the installed headers.
3339 Essentially the same effect can be achieved with the "no-deprecated"
3340 argument to Configure, except that this will always restrict
3341 the build to just the latest API, rather than a fixed API
3344 As applications are ported to future revisions of the API,
3345 they should update their compile-time OPENSSL_API_COMPAT define
3346 accordingly, but in most cases should be able to continue to
3347 compile with later releases.
3349 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3350 0x10000000L and 0x00908000L, respectively. However those
3351 versions did not support the OPENSSL_API_COMPAT feature, and
3352 so applications are not typically tested for explicit support
3353 of just the undeprecated features of either release.
3357 * Add support for setting the minimum and maximum supported protocol.
3358 It can bet set via the SSL_set_min_proto_version() and
3359 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3360 MaxProtocol. It's recommended to use the new APIs to disable
3361 protocols instead of disabling individual protocols using
3362 SSL_set_options() or SSL_CONF's Protocol. This change also
3363 removes support for disabling TLS 1.2 in the OpenSSL TLS
3364 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3368 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3372 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3373 and integrates ECDSA and ECDH functionality into EC. Implementations can
3374 now redirect key generation and no longer need to convert to or from
3377 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3378 include the ec.h header file instead.
3382 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3383 ciphers who are no longer supported and drops support the ephemeral RSA key
3384 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3388 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3389 opaque. For HMAC_CTX, the following constructors and destructors
3392 HMAC_CTX *HMAC_CTX_new(void);
3393 void HMAC_CTX_free(HMAC_CTX *ctx);
3395 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3396 destroy such methods has been added. See EVP_MD_meth_new(3) and
3397 EVP_CIPHER_meth_new(3) for documentation.
3400 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3401 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3402 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3403 an already created structure.
3404 2) For consistency with the majority of our object creators and
3405 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3406 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3407 for deprecated builds.
3411 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3412 cryptographic operations to be performed asynchronously as long as an
3413 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3414 further details. Libssl has also had this capability integrated with the
3415 introduction of the new mode SSL_MODE_ASYNC and associated error
3416 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3417 pages. This work was developed in partnership with Intel Corp.
3421 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3422 always enabled now. If you want to disable the support you should
3423 exclude it using the list of supported ciphers. This also means that the
3424 "-no_ecdhe" option has been removed from s_server.
3428 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3429 SSL_{CTX_}set1_curves() which can set a list.
3433 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3434 curve you want to support using SSL_{CTX_}set1_curves().
3438 * State machine rewrite. The state machine code has been significantly
3439 refactored in order to remove much duplication of code and solve issues
3440 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3441 further details). This change does have some associated API changes.
3442 Notably the SSL_state() function has been removed and replaced by
3443 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3444 SSL_set_state() has been removed altogether. The previous handshake states
3445 defined in ssl.h and ssl3.h have also been removed.
3449 * All instances of the string "ssleay" in the public API were replaced
3450 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3451 Some error codes related to internal RSA_eay API's were renamed.
3455 * The demo files in crypto/threads were moved to demo/threads.
3459 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3462 *Matt Caswell, Rich Salz*
3464 * New ASN.1 embed macro.
3466 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3467 structure is not allocated: it is part of the parent. That is instead of
3475 This reduces memory fragmentation and make it impossible to accidentally
3476 set a mandatory field to NULL.
3478 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3479 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3480 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3485 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3489 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3490 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3491 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3492 DES and RC4 ciphersuites.
3496 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3497 This changes the decoding behaviour for some invalid messages,
3498 though the change is mostly in the more lenient direction, and
3499 legacy behaviour is preserved as much as possible.
3503 * Fix no-stdio build.
3504 *David Woodhouse <David.Woodhouse@intel.com> and also*
3505 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3507 * New testing framework
3508 The testing framework has been largely rewritten and is now using
3509 perl and the perl modules Test::Harness and an extended variant of
3510 Test::More called OpenSSL::Test to do its work. All test scripts in
3511 test/ have been rewritten into test recipes, and all direct calls to
3512 executables in test/Makefile have become individual recipes using the
3513 simplified testing OpenSSL::Test::Simple.
3515 For documentation on our testing modules, do:
3517 perldoc test/testlib/OpenSSL/Test/Simple.pm
3518 perldoc test/testlib/OpenSSL/Test.pm
3522 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3523 are used; the latter aborts on memory leaks (usually checked on exit).
3524 Some undocumented "set malloc, etc., hooks" functions were removed
3525 and others were changed. All are now documented.
3529 * In DSA_generate_parameters_ex, if the provided seed is too short,
3532 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3534 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3535 from RFC4279, RFC4785, RFC5487, RFC5489.
3537 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3538 original RSA_PSK patch.
3542 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3543 era flag was never set throughout the codebase (only read). Also removed
3544 SSL3_FLAGS_POP_BUFFER which was only used if
3545 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3549 * Changed the default name options in the "ca", "crl", "req" and "x509"
3550 to be "oneline" instead of "compat".
3554 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3555 not aware of clients that still exhibit this bug, and the workaround
3556 hasn't been working properly for a while.
3560 * The return type of BIO_number_read() and BIO_number_written() as well as
3561 the corresponding num_read and num_write members in the BIO structure has
3562 changed from unsigned long to uint64_t. On platforms where an unsigned
3563 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3568 * Given the pervasive nature of TLS extensions it is inadvisable to run
3569 OpenSSL without support for them. It also means that maintaining
3570 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3571 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3575 * Removed support for the two export grade static DH ciphersuites
3576 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3577 were newly added (along with a number of other static DH ciphersuites) to
3578 1.0.2. However the two export ones have *never* worked since they were
3579 introduced. It seems strange in any case to be adding new export
3580 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3584 * Version negotiation has been rewritten. In particular SSLv23_method(),
3585 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3586 and turned into macros which simply call the new preferred function names
3587 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3588 should use the new names instead. Also as part of this change the ssl23.h
3589 header file has been removed.
3593 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3594 code and the associated standard is no longer considered fit-for-purpose.
3598 * RT2547 was closed. When generating a private key, try to make the
3599 output file readable only by the owner. This behavior change might
3600 be noticeable when interacting with other software.
3602 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3607 * Added HTTP GET support to the ocsp command.
3611 * Changed default digest for the dgst and enc commands from MD5 to
3616 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3620 * Added support for TLS extended master secret from
3621 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3622 initial patch which was a great help during development.
3626 * All libssl internal structures have been removed from the public header
3627 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3628 now redundant). Users should not attempt to access internal structures
3629 directly. Instead they should use the provided API functions.
3633 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3634 Access to deprecated functions can be re-enabled by running config with
3635 "enable-deprecated". In addition applications wishing to use deprecated
3636 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3637 will, by default, disable some transitive includes that previously existed
3638 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3642 * Added support for OCB mode. OpenSSL has been granted a patent license
3643 compatible with the OpenSSL license for use of OCB. Details are available
3644 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3645 for OCB can be removed by calling config with no-ocb.
3649 * SSLv2 support has been removed. It still supports receiving a SSLv2
3650 compatible client hello.
3654 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3655 done while fixing the error code for the key-too-small case.
3657 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3659 * CA.sh has been removed; use CA.pl instead.
3663 * Removed old DES API.
3667 * Remove various unsupported platforms:
3673 Sinix/ReliantUNIX RM400
3678 16-bit platforms such as WIN16
3682 * Clean up OPENSSL_NO_xxx #define's
3683 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3684 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3685 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3686 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3687 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3688 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3689 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3690 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3691 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3692 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3696 * Cleaned up dead code
3697 Remove all but one '#ifdef undef' which is to be looked at.
3701 * Clean up calling of xxx_free routines.
3702 Just like free(), fix most of the xxx_free routines to accept
3703 NULL. Remove the non-null checks from callers. Save much code.
3707 * Add secure heap for storage of private keys (when possible).
3708 Add BIO_s_secmem(), CBIGNUM, etc.
3709 Contributed by Akamai Technologies under our Corporate CLA.
3713 * Experimental support for a new, fast, unbiased prime candidate generator,
3714 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3716 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3718 * New output format NSS in the sess_id command line tool. This allows
3719 exporting the session id and the master key in NSS keylog format.
3721 *Martin Kaiser <martin@kaiser.cx>*
3723 * Harmonize version and its documentation. -f flag is used to display
3726 *mancha <mancha1@zoho.com>*
3728 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3729 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3731 *mancha <mancha1@zoho.com>*
3733 * Fix some double frees. These are not thought to be exploitable.
3735 *mancha <mancha1@zoho.com>*
3737 * A missing bounds check in the handling of the TLS heartbeat extension
3738 can be used to reveal up to 64k of memory to a connected client or
3741 Thanks for Neel Mehta of Google Security for discovering this bug and to
3742 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3743 preparing the fix ([CVE-2014-0160])
3745 *Adam Langley, Bodo Moeller*
3747 * Fix for the attack described in the paper "Recovering OpenSSL
3748 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3749 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3750 <http://eprint.iacr.org/2014/140>
3752 Thanks to Yuval Yarom and Naomi Benger for discovering this
3753 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3755 *Yuval Yarom and Naomi Benger*
3757 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3758 this fixes a limitation in previous versions of OpenSSL.
3762 * Experimental encrypt-then-mac support.
3764 Experimental support for encrypt then mac from
3765 draft-gutmann-tls-encrypt-then-mac-02.txt
3767 To enable it set the appropriate extension number (0x42 for the test
3768 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3770 For non-compliant peers (i.e. just about everything) this should have no
3773 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3777 * Add EVP support for key wrapping algorithms, to avoid problems with
3778 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3779 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3780 algorithms and include tests cases.
3784 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3789 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3790 MGF1 digest and OAEP label.
3794 * Make openssl verify return errors.
3796 *Chris Palmer <palmer@google.com> and Ben Laurie*
3798 * New function ASN1_TIME_diff to calculate the difference between two
3799 ASN1_TIME structures or one structure and the current time.
3803 * Update fips_test_suite to support multiple command line options. New
3804 test to induce all self test errors in sequence and check expected
3809 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3810 sign or verify all in one operation.
3814 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3815 test programs and fips_test_suite. Includes functionality to parse
3816 the minimal script output of fipsalgest.pl directly.
3820 * Add authorisation parameter to FIPS_module_mode_set().
3824 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3828 * Use separate DRBG fields for internal and external flags. New function
3829 FIPS_drbg_health_check() to perform on demand health checking. Add
3830 generation tests to fips_test_suite with reduced health check interval to
3831 demonstrate periodic health checking. Add "nodh" option to
3832 fips_test_suite to skip very slow DH test.
3836 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3841 * More extensive health check for DRBG checking many more failure modes.
3842 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3843 combination: call this in fips_test_suite.
3847 * Add support for canonical generation of DSA parameter 'g'. See
3850 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3851 POST to handle HMAC cases.
3855 * Add functions FIPS_module_version() and FIPS_module_version_text()
3856 to return numerical and string versions of the FIPS module number.
3860 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3861 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3862 outside the validated module in the FIPS capable OpenSSL.
3866 * Minor change to DRBG entropy callback semantics. In some cases
3867 there is no multiple of the block length between min_len and
3868 max_len. Allow the callback to return more than max_len bytes
3869 of entropy but discard any extra: it is the callback's responsibility
3870 to ensure that the extra data discarded does not impact the
3871 requested amount of entropy.
3875 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3876 information in FIPS186-3, SP800-57 and SP800-131A.
3880 * CCM support via EVP. Interface is very similar to GCM case except we
3881 must supply all data in one chunk (i.e. no update, final) and the
3882 message length must be supplied if AAD is used. Add algorithm test
3887 * Initial version of POST overhaul. Add POST callback to allow the status
3888 of POST to be monitored and/or failures induced. Modify fips_test_suite
3889 to use callback. Always run all selftests even if one fails.
3893 * XTS support including algorithm test driver in the fips_gcmtest program.
3894 Note: this does increase the maximum key length from 32 to 64 bytes but
3895 there should be no binary compatibility issues as existing applications
3896 will never use XTS mode.
3900 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3901 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3902 performs algorithm blocking for unapproved PRNG types. Also do not
3903 set PRNG type in FIPS_mode_set(): leave this to the application.
3904 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3905 the standard OpenSSL PRNG: set additional data to a date time vector.
3909 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3910 This shouldn't present any incompatibility problems because applications
3911 shouldn't be using these directly and any that are will need to rethink
3912 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3916 * Extensive self tests and health checking required by SP800-90 DRBG.
3917 Remove strength parameter from FIPS_drbg_instantiate and always
3918 instantiate at maximum supported strength.
3922 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3926 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3930 * New function DH_compute_key_padded() to compute a DH key and pad with
3931 leading zeroes if needed: this complies with SP800-56A et al.
3935 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3936 anything, incomplete, subject to change and largely untested at present.
3940 * Modify fipscanisteronly build option to only build the necessary object
3941 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3945 * Add experimental option FIPSSYMS to give all symbols in
3946 fipscanister.o and FIPS or fips prefix. This will avoid
3947 conflicts with future versions of OpenSSL. Add perl script
3948 util/fipsas.pl to preprocess assembly language source files
3949 and rename any affected symbols.
3953 * Add selftest checks and algorithm block of non-fips algorithms in
3954 FIPS mode. Remove DES2 from selftests.
3958 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3959 return internal method without any ENGINE dependencies. Add new
3960 tiny fips sign and verify functions.
3964 * New build option no-ec2m to disable characteristic 2 code.
3968 * New build option "fipscanisteronly". This only builds fipscanister.o
3969 and (currently) associated fips utilities. Uses the file Makefile.fips
3970 instead of Makefile.org as the prototype.
3974 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3975 Update fips_gcmtest to use IV generator.
3979 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3980 setting output buffer to NULL. The `*Final` function must be
3981 called although it will not retrieve any additional data. The tag
3982 can be set or retrieved with a ctrl. The IV length is by default 12
3983 bytes (96 bits) but can be set to an alternative value. If the IV
3984 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3989 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3990 underlying do_cipher function handles all cipher semantics itself
3991 including padding and finalisation. This is useful if (for example)
3992 an ENGINE cipher handles block padding itself. The behaviour of
3993 do_cipher is subtly changed if this flag is set: the return value
3994 is the number of characters written to the output buffer (zero is
3995 no longer an error code) or a negative error code. Also if the
3996 input buffer is NULL and length 0 finalisation should be performed.
4000 * If a candidate issuer certificate is already part of the constructed
4001 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4005 * Improve forward-security support: add functions
4007 void SSL_CTX_set_not_resumable_session_callback(
4008 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4009 void SSL_set_not_resumable_session_callback(
4010 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4012 for use by SSL/TLS servers; the callback function will be called whenever a
4013 new session is created, and gets to decide whether the session may be
4014 cached to make it resumable (return 0) or not (return 1). (As by the
4015 SSL/TLS protocol specifications, the session_id sent by the server will be
4016 empty to indicate that the session is not resumable; also, the server will
4017 not generate RFC 4507 (RFC 5077) session tickets.)
4019 A simple reasonable callback implementation is to return is_forward_secure.
4020 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4021 by the SSL/TLS server library, indicating whether it can provide forward
4024 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4026 * New -verify_name option in command line utilities to set verification
4031 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4032 Add CMAC pkey methods.
4036 * Experimental renegotiation in s_server -www mode. If the client
4037 browses /reneg connection is renegotiated. If /renegcert it is
4038 renegotiated requesting a certificate.
4042 * Add an "external" session cache for debugging purposes to s_server. This
4043 should help trace issues which normally are only apparent in deployed
4044 multi-process servers.
4048 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4049 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4050 BIO_set_cipher() and some obscure PEM functions were changed so they
4051 can now return an error. The RAND changes required a change to the
4052 RAND_METHOD structure.
4056 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4057 a gcc attribute to warn if the result of a function is ignored. This
4058 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4059 whose return value is often ignored.
4063 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4064 These allow SCTs (signed certificate timestamps) to be requested and
4065 validated when establishing a connection.
4067 *Rob Percival <robpercival@google.com>*
4072 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4074 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4075 used even when parsing explicit parameters, when loading a encoded key
4076 or calling `EC_GROUP_new_from_ecpkparameters()`/
4077 `EC_GROUP_new_from_ecparameters()`.
4078 This prevents bypass of security hardening and performance gains,
4079 especially for curves with specialized EC_METHODs.
4080 By default, if a key encoded with explicit parameters is loaded and later
4081 encoded, the output is still encoded with explicit parameters, even if
4082 internally a "named" EC_GROUP is used for computation.
4086 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4087 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4088 NULL. After this change, only the cofactor parameter can be NULL. It also
4089 does some minimal sanity checks on the passed order.
4094 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4095 An attack is simple, if the first CMS_recipientInfo is valid but the
4096 second CMS_recipientInfo is chosen ciphertext. If the second
4097 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4098 encryption key will be replaced by garbage, and the message cannot be
4099 decoded, but if the RSA decryption fails, the correct encryption key is
4100 used and the recipient will not notice the attack.
4101 As a work around for this potential attack the length of the decrypted
4102 key must be equal to the cipher default key length, in case the
4103 certifiate is not given and all recipientInfo are tried out.
4104 The old behaviour can be re-enabled in the CMS code by setting the
4105 CMS_DEBUG_DECRYPT flag.
4110 * Document issue with installation paths in diverse Windows builds
4112 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4113 binaries and run-time config file.
4118 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4120 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4121 This changes the size when using the genpkey app when no size is given. It
4122 fixes an omission in earlier changes that changed all RSA, DSA and DH
4123 generation apps to use 2048 bits by default.
4127 * Add FIPS support for Android Arm 64-bit
4129 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4130 Module in Version 2.0.10. For some reason, the corresponding target
4131 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4132 built with FIPS support on Android Arm 64-bit. This omission has been
4135 *Matthias St. Pierre*
4137 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4139 * 0-byte record padding oracle
4141 If an application encounters a fatal protocol error and then calls
4142 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4143 then OpenSSL can respond differently to the calling application if a 0 byte
4144 record is received with invalid padding compared to if a 0 byte record is
4145 received with an invalid MAC. If the application then behaves differently
4146 based on that in a way that is detectable to the remote peer, then this
4147 amounts to a padding oracle that could be used to decrypt data.
4149 In order for this to be exploitable "non-stitched" ciphersuites must be in
4150 use. Stitched ciphersuites are optimised implementations of certain
4151 commonly used ciphersuites. Also the application must call SSL_shutdown()
4152 twice even if a protocol error has occurred (applications should not do
4153 this but some do anyway).
4155 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4156 Aviram, with additional investigation by Steven Collison and Andrew
4157 Hourselt. It was reported to OpenSSL on 10th December 2018.
4162 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4166 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4168 * Microarchitecture timing vulnerability in ECC scalar multiplication
4170 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4171 shown to be vulnerable to a microarchitecture timing side channel attack.
4172 An attacker with sufficient access to mount local timing attacks during
4173 ECDSA signature generation could recover the private key.
4175 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4176 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4182 * Timing vulnerability in DSA signature generation
4184 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4185 timing side channel attack. An attacker could use variations in the signing
4186 algorithm to recover the private key.
4188 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4193 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4194 Module, accidentally introduced while backporting security fixes from the
4195 development branch and hindering the use of ECC in FIPS mode.
4199 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4201 * Client DoS due to large DH parameter
4203 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4204 malicious server can send a very large prime value to the client. This will
4205 cause the client to spend an unreasonably long period of time generating a
4206 key for this prime resulting in a hang until the client has finished. This
4207 could be exploited in a Denial Of Service attack.
4209 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4214 * Cache timing vulnerability in RSA Key Generation
4216 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4217 a cache timing side channel attack. An attacker with sufficient access to
4218 mount cache timing attacks during the RSA key generation process could
4219 recover the private key.
4221 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4222 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4227 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4228 parameter is no longer accepted, as it leads to a corrupt table. NULL
4229 pem_str is reserved for alias entries only.
4233 * Revert blinding in ECDSA sign and instead make problematic addition
4234 length-invariant. Switch even to fixed-length Montgomery multiplication.
4238 * Change generating and checking of primes so that the error rate of not
4239 being prime depends on the intended use based on the size of the input.
4240 For larger primes this will result in more rounds of Miller-Rabin.
4241 The maximal error rate for primes with more than 1080 bits is lowered
4244 *Kurt Roeckx, Annie Yousar*
4246 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4250 * Add blinding to ECDSA and DSA signatures to protect against side channel
4251 attacks discovered by Keegan Ryan (NCC Group).
4255 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4256 now allow empty (zero character) pass phrases.
4260 * Certificate time validation (X509_cmp_time) enforces stricter
4261 compliance with RFC 5280. Fractional seconds and timezone offsets
4262 are no longer allowed.
4266 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4268 * Constructed ASN.1 types with a recursive definition could exceed the stack
4270 Constructed ASN.1 types with a recursive definition (such as can be found
4271 in PKCS7) could eventually exceed the stack given malicious input with
4272 excessive recursion. This could result in a Denial Of Service attack. There
4273 are no such structures used within SSL/TLS that come from untrusted sources
4274 so this is considered safe.
4276 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4282 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4284 * Read/write after SSL object in error state
4286 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4287 mechanism. The intent was that if a fatal error occurred during a handshake
4288 then OpenSSL would move into the error state and would immediately fail if
4289 you attempted to continue the handshake. This works as designed for the
4290 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4291 SSL_connect()), however due to a bug it does not work correctly if
4292 SSL_read() or SSL_write() is called directly. In that scenario, if the
4293 handshake fails then a fatal error will be returned in the initial function
4294 call. If SSL_read()/SSL_write() is subsequently called by the application
4295 for the same SSL object then it will succeed and the data is passed without
4296 being decrypted/encrypted directly from the SSL/TLS record layer.
4298 In order to exploit this issue an application bug would have to be present
4299 that resulted in a call to SSL_read()/SSL_write() being issued after having
4300 already received a fatal error.
4302 This issue was reported to OpenSSL by David Benjamin (Google).
4307 * rsaz_1024_mul_avx2 overflow bug on x86_64
4309 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4310 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4311 Analysis suggests that attacks against RSA and DSA as a result of this
4312 defect would be very difficult to perform and are not believed likely.
4313 Attacks against DH1024 are considered just feasible, because most of the
4314 work necessary to deduce information about a private key may be performed
4315 offline. The amount of resources required for such an attack would be
4316 significant. However, for an attack on TLS to be meaningful, the server
4317 would have to share the DH1024 private key among multiple clients, which is
4318 no longer an option since CVE-2016-0701.
4320 This only affects processors that support the AVX2 but not ADX extensions
4321 like Intel Haswell (4th generation).
4323 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4324 was originally found via the OSS-Fuzz project.
4329 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4331 * bn_sqrx8x_internal carry bug on x86_64
4333 There is a carry propagating bug in the x86_64 Montgomery squaring
4334 procedure. No EC algorithms are affected. Analysis suggests that attacks
4335 against RSA and DSA as a result of this defect would be very difficult to
4336 perform and are not believed likely. Attacks against DH are considered just
4337 feasible (although very difficult) because most of the work necessary to
4338 deduce information about a private key may be performed offline. The amount
4339 of resources required for such an attack would be very significant and
4340 likely only accessible to a limited number of attackers. An attacker would
4341 additionally need online access to an unpatched system using the target
4342 private key in a scenario with persistent DH parameters and a private
4343 key that is shared between multiple clients.
4345 This only affects processors that support the BMI1, BMI2 and ADX extensions
4346 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4348 This issue was reported to OpenSSL by the OSS-Fuzz project.
4353 * Malformed X.509 IPAddressFamily could cause OOB read
4355 If an X.509 certificate has a malformed IPAddressFamily extension,
4356 OpenSSL could do a one-byte buffer overread. The most likely result
4357 would be an erroneous display of the certificate in text format.
4359 This issue was reported to OpenSSL by the OSS-Fuzz project.
4364 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4366 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4367 platform rather than 'mingw'.
4371 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4373 * Truncated packet could crash via OOB read
4375 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4376 cipher is being used, then a truncated packet can cause that host to
4377 perform an out-of-bounds read, usually resulting in a crash.
4379 This issue was reported to OpenSSL by Robert Święcki of Google.
4384 * BN_mod_exp may produce incorrect results on x86_64
4386 There is a carry propagating bug in the x86_64 Montgomery squaring
4387 procedure. No EC algorithms are affected. Analysis suggests that attacks
4388 against RSA and DSA as a result of this defect would be very difficult to
4389 perform and are not believed likely. Attacks against DH are considered just
4390 feasible (although very difficult) because most of the work necessary to
4391 deduce information about a private key may be performed offline. The amount
4392 of resources required for such an attack would be very significant and
4393 likely only accessible to a limited number of attackers. An attacker would
4394 additionally need online access to an unpatched system using the target
4395 private key in a scenario with persistent DH parameters and a private
4396 key that is shared between multiple clients. For example this can occur by
4397 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4398 similar to CVE-2015-3193 but must be treated as a separate problem.
4400 This issue was reported to OpenSSL by the OSS-Fuzz project.
4405 * Montgomery multiplication may produce incorrect results
4407 There is a carry propagating bug in the Broadwell-specific Montgomery
4408 multiplication procedure that handles input lengths divisible by, but
4409 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4410 and DH private keys are impossible. This is because the subroutine in
4411 question is not used in operations with the private key itself and an input
4412 of the attacker's direct choice. Otherwise the bug can manifest itself as
4413 transient authentication and key negotiation failures or reproducible
4414 erroneous outcome of public-key operations with specially crafted input.
4415 Among EC algorithms only Brainpool P-512 curves are affected and one
4416 presumably can attack ECDH key negotiation. Impact was not analyzed in
4417 detail, because pre-requisites for attack are considered unlikely. Namely
4418 multiple clients have to choose the curve in question and the server has to
4419 share the private key among them, neither of which is default behaviour.
4420 Even then only clients that chose the curve will be affected.
4422 This issue was publicly reported as transient failures and was not
4423 initially recognized as a security issue. Thanks to Richard Morgan for
4424 providing reproducible case.
4429 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4430 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4431 prevent issues where no progress is being made and the peer continually
4432 sends unrecognised record types, using up resources processing them.
4436 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4438 * Missing CRL sanity check
4440 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4441 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4442 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4444 This issue only affects the OpenSSL 1.0.2i
4449 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4451 * OCSP Status Request extension unbounded memory growth
4453 A malicious client can send an excessively large OCSP Status Request
4454 extension. If that client continually requests renegotiation, sending a
4455 large OCSP Status Request extension each time, then there will be unbounded
4456 memory growth on the server. This will eventually lead to a Denial Of
4457 Service attack through memory exhaustion. Servers with a default
4458 configuration are vulnerable even if they do not support OCSP. Builds using
4459 the "no-ocsp" build time option are not affected.
4461 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4466 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4469 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4475 * OOB write in MDC2_Update()
4477 An overflow can occur in MDC2_Update() either if called directly or
4478 through the EVP_DigestUpdate() function using MDC2. If an attacker
4479 is able to supply very large amounts of input data after a previous
4480 call to EVP_EncryptUpdate() with a partial block then a length check
4481 can overflow resulting in a heap corruption.
4483 The amount of data needed is comparable to SIZE_MAX which is impractical
4486 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4491 * Malformed SHA512 ticket DoS
4493 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4494 DoS attack where a malformed ticket will result in an OOB read which will
4497 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4498 a custom server callback and ticket lookup mechanism.
4500 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4505 * OOB write in BN_bn2dec()
4507 The function BN_bn2dec() does not check the return value of BN_div_word().
4508 This can cause an OOB write if an application uses this function with an
4509 overly large BIGNUM. This could be a problem if an overly large certificate
4510 or CRL is printed out from an untrusted source. TLS is not affected because
4511 record limits will reject an oversized certificate before it is parsed.
4513 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4518 * OOB read in TS_OBJ_print_bio()
4520 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4521 the total length the OID text representation would use and not the amount
4522 of data written. This will result in OOB reads when large OIDs are
4525 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4530 * Pointer arithmetic undefined behaviour
4532 Avoid some undefined pointer arithmetic
4534 A common idiom in the codebase is to check limits in the following manner:
4537 Where "p" points to some malloc'd data of SIZE bytes and
4540 "len" here could be from some externally supplied data (e.g. from a TLS
4543 The rules of C pointer arithmetic are such that "p + len" is only well
4544 defined where len <= SIZE. Therefore the above idiom is actually
4545 undefined behaviour.
4547 For example this could cause problems if some malloc implementation
4548 provides an address for "p" such that "p + len" actually overflows for
4549 values of len that are too big and therefore p + len < limit.
4551 This issue was reported to OpenSSL by Guido Vranken
4556 * Constant time flag not preserved in DSA signing
4558 Operations in the DSA signing algorithm should run in constant time in
4559 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4560 implementation means that a non-constant time codepath is followed for
4561 certain operations. This has been demonstrated through a cache-timing
4562 attack to be sufficient for an attacker to recover the private DSA key.
4564 This issue was reported by César Pereida (Aalto University), Billy Brumley
4565 (Tampere University of Technology), and Yuval Yarom (The University of
4566 Adelaide and NICTA).
4571 * DTLS buffered message DoS
4573 In a DTLS connection where handshake messages are delivered out-of-order
4574 those messages that OpenSSL is not yet ready to process will be buffered
4575 for later use. Under certain circumstances, a flaw in the logic means that
4576 those messages do not get removed from the buffer even though the handshake
4577 has been completed. An attacker could force up to approx. 15 messages to
4578 remain in the buffer when they are no longer required. These messages will
4579 be cleared when the DTLS connection is closed. The default maximum size for
4580 a message is 100k. Therefore the attacker could force an additional 1500k
4581 to be consumed per connection. By opening many simulataneous connections an
4582 attacker could cause a DoS attack through memory exhaustion.
4584 This issue was reported to OpenSSL by Quan Luo.
4589 * DTLS replay protection DoS
4591 A flaw in the DTLS replay attack protection mechanism means that records
4592 that arrive for future epochs update the replay protection "window" before
4593 the MAC for the record has been validated. This could be exploited by an
4594 attacker by sending a record for the next epoch (which does not have to
4595 decrypt or have a valid MAC), with a very large sequence number. This means
4596 that all subsequent legitimate packets are dropped causing a denial of
4597 service for a specific DTLS connection.
4599 This issue was reported to OpenSSL by the OCAP audit team.
4604 * Certificate message OOB reads
4606 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4607 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4608 theoretical DoS risk but this has not been observed in practice on common
4611 The messages affected are client certificate, client certificate request
4612 and server certificate. As a result the attack can only be performed
4613 against a client or a server which enables client authentication.
4615 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4620 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4622 * Prevent padding oracle in AES-NI CBC MAC check
4624 A MITM attacker can use a padding oracle attack to decrypt traffic
4625 when the connection uses an AES CBC cipher and the server support
4628 This issue was introduced as part of the fix for Lucky 13 padding
4629 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4630 constant time by making sure that always the same bytes are read and
4631 compared against either the MAC or padding bytes. But it no longer
4632 checked that there was enough data to have both the MAC and padding
4635 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4640 * Fix EVP_EncodeUpdate overflow
4642 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4643 Base64 encoding of binary data. If an attacker is able to supply very large
4644 amounts of input data then a length check can overflow resulting in a heap
4647 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4648 the `PEM_write_bio*` family of functions. These are mainly used within the
4649 OpenSSL command line applications, so any application which processes data
4650 from an untrusted source and outputs it as a PEM file should be considered
4651 vulnerable to this issue. User applications that call these APIs directly
4652 with large amounts of untrusted data may also be vulnerable.
4654 This issue was reported by Guido Vranken.
4659 * Fix EVP_EncryptUpdate overflow
4661 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4662 is able to supply very large amounts of input data after a previous call to
4663 EVP_EncryptUpdate() with a partial block then a length check can overflow
4664 resulting in a heap corruption. Following an analysis of all OpenSSL
4665 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4666 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4667 the first called function after an EVP_EncryptInit(), and therefore that
4668 specific call must be safe. The second form is where the length passed to
4669 EVP_EncryptUpdate() can be seen from the code to be some small value and
4670 therefore there is no possibility of an overflow. Since all instances are
4671 one of these two forms, it is believed that there can be no overflows in
4672 internal code due to this problem. It should be noted that
4673 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4674 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4675 of these calls have also been analysed too and it is believed there are no
4676 instances in internal usage where an overflow could occur.
4678 This issue was reported by Guido Vranken.
4683 * Prevent ASN.1 BIO excessive memory allocation
4685 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4686 a short invalid encoding can cause allocation of large amounts of memory
4687 potentially consuming excessive resources or exhausting memory.
4689 Any application parsing untrusted data through d2i BIO functions is
4690 affected. The memory based functions such as d2i_X509() are *not* affected.
4691 Since the memory based functions are used by the TLS library, TLS
4692 applications are not affected.
4694 This issue was reported by Brian Carpenter.
4701 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4702 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4703 in arbitrary stack data being returned in the buffer.
4705 This issue was reported by Guido Vranken.
4710 * Modify behavior of ALPN to invoke callback after SNI/servername
4711 callback, such that updates to the SSL_CTX affect ALPN.
4715 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4720 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4721 methods are enabled and ssl2 is disabled the methods return NULL.
4725 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4727 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4728 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4729 provide any "EXPORT" or "LOW" strength ciphers.
4733 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4734 is by default disabled at build-time. Builds that are not configured with
4735 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4736 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4737 will need to explicitly call either of:
4739 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4741 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4743 as appropriate. Even if either of those is used, or the application
4744 explicitly uses the version-specific SSLv2_method() or its client and
4745 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4746 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4747 ciphers, and SSLv2 56-bit DES are no longer available.
4752 * Fix a double-free in DSA code
4754 A double free bug was discovered when OpenSSL parses malformed DSA private
4755 keys and could lead to a DoS attack or memory corruption for applications
4756 that receive DSA private keys from untrusted sources. This scenario is
4759 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4765 * Disable SRP fake user seed to address a server memory leak.
4767 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4769 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4770 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4771 was changed to ignore the "fake user" SRP seed, even if the seed
4774 Users should use SRP_VBASE_get1_by_user instead. Note that in
4775 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4776 also that even though configuring the SRP seed attempts to hide
4777 invalid usernames by continuing the handshake with fake
4778 credentials, this behaviour is not constant time and no strong
4779 guarantees are made that the handshake is indistinguishable from
4780 that of a valid user.
4785 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4787 In the BN_hex2bn function the number of hex digits is calculated using an
4788 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4789 large values of `i` this can result in `bn_expand` not allocating any
4790 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4791 field as NULL leading to a subsequent NULL ptr deref. For very large values
4792 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4793 In this case memory is allocated to the internal BIGNUM data field, but it
4794 is insufficiently sized leading to heap corruption. A similar issue exists
4795 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4796 is ever called by user applications with very large untrusted hex/dec data.
4797 This is anticipated to be a rare occurrence.
4799 All OpenSSL internal usage of these functions use data that is not expected
4800 to be untrusted, e.g. config file data or application command line
4801 arguments. If user developed applications generate config file data based
4802 on untrusted data then it is possible that this could also lead to security
4803 consequences. This is also anticipated to be rare.
4805 This issue was reported to OpenSSL by Guido Vranken.
4810 * Fix memory issues in `BIO_*printf` functions
4812 The internal `fmtstr` function used in processing a "%s" format string in
4813 the `BIO_*printf` functions could overflow while calculating the length of a
4814 string and cause an OOB read when printing very long strings.
4816 Additionally the internal `doapr_outch` function can attempt to write to an
4817 OOB memory location (at an offset from the NULL pointer) in the event of a
4818 memory allocation failure. In 1.0.2 and below this could be caused where
4819 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4820 could be in processing a very long "%s" format string. Memory leaks can
4823 The first issue may mask the second issue dependent on compiler behaviour.
4824 These problems could enable attacks where large amounts of untrusted data
4825 is passed to the `BIO_*printf` functions. If applications use these functions
4826 in this way then they could be vulnerable. OpenSSL itself uses these
4827 functions when printing out human-readable dumps of ASN.1 data. Therefore
4828 applications that print this data could be vulnerable if the data is from
4829 untrusted sources. OpenSSL command line applications could also be
4830 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4831 as command line arguments.
4833 Libssl is not considered directly vulnerable. Additionally certificates etc
4834 received via remote connections via libssl are also unlikely to be able to
4835 trigger these issues because of message size limits enforced within libssl.
4837 This issue was reported to OpenSSL Guido Vranken.
4842 * Side channel attack on modular exponentiation
4844 A side-channel attack was found which makes use of cache-bank conflicts on
4845 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4846 of RSA keys. The ability to exploit this issue is limited as it relies on
4847 an attacker who has control of code in a thread running on the same
4848 hyper-threaded core as the victim thread which is performing decryptions.
4850 This issue was reported to OpenSSL by Yuval Yarom, The University of
4851 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4852 Nadia Heninger, University of Pennsylvania with more information at
4853 <http://cachebleed.info>.
4858 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4859 if no keysize is specified with default_bits. This fixes an
4860 omission in an earlier change that changed all RSA/DSA key generation
4861 apps to use 2048 bits by default.
4865 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4867 * DH small subgroups
4869 Historically OpenSSL only ever generated DH parameters based on "safe"
4870 primes. More recently (in version 1.0.2) support was provided for
4871 generating X9.42 style parameter files such as those required for RFC 5114
4872 support. The primes used in such files may not be "safe". Where an
4873 application is using DH configured with parameters based on primes that are
4874 not "safe" then an attacker could use this fact to find a peer's private
4875 DH exponent. This attack requires that the attacker complete multiple
4876 handshakes in which the peer uses the same private DH exponent. For example
4877 this could be used to discover a TLS server's private DH exponent if it's
4878 reusing the private DH exponent or it's using a static DH ciphersuite.
4880 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4881 TLS. It is not on by default. If the option is not set then the server
4882 reuses the same private DH exponent for the life of the server process and
4883 would be vulnerable to this attack. It is believed that many popular
4884 applications do set this option and would therefore not be at risk.
4886 The fix for this issue adds an additional check where a "q" parameter is
4887 available (as is the case in X9.42 based parameters). This detects the
4888 only known attack, and is the only possible defense for static DH
4889 ciphersuites. This could have some performance impact.
4891 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4892 default and cannot be disabled. This could have some performance impact.
4894 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4899 * SSLv2 doesn't block disabled ciphers
4901 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4902 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4903 been disabled, provided that the SSLv2 protocol was not also disabled via
4906 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4907 and Sebastian Schinzel.
4912 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4914 * BN_mod_exp may produce incorrect results on x86_64
4916 There is a carry propagating bug in the x86_64 Montgomery squaring
4917 procedure. No EC algorithms are affected. Analysis suggests that attacks
4918 against RSA and DSA as a result of this defect would be very difficult to
4919 perform and are not believed likely. Attacks against DH are considered just
4920 feasible (although very difficult) because most of the work necessary to
4921 deduce information about a private key may be performed offline. The amount
4922 of resources required for such an attack would be very significant and
4923 likely only accessible to a limited number of attackers. An attacker would
4924 additionally need online access to an unpatched system using the target
4925 private key in a scenario with persistent DH parameters and a private
4926 key that is shared between multiple clients. For example this can occur by
4927 default in OpenSSL DHE based SSL/TLS ciphersuites.
4929 This issue was reported to OpenSSL by Hanno Böck.
4934 * Certificate verify crash with missing PSS parameter
4936 The signature verification routines will crash with a NULL pointer
4937 dereference if presented with an ASN.1 signature using the RSA PSS
4938 algorithm and absent mask generation function parameter. Since these
4939 routines are used to verify certificate signature algorithms this can be
4940 used to crash any certificate verification operation and exploited in a
4941 DoS attack. Any application which performs certificate verification is
4942 vulnerable including OpenSSL clients and servers which enable client
4945 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4950 * X509_ATTRIBUTE memory leak
4952 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4953 memory. This structure is used by the PKCS#7 and CMS routines so any
4954 application which reads PKCS#7 or CMS data from untrusted sources is
4955 affected. SSL/TLS is not affected.
4957 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4963 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4964 This changes the decoding behaviour for some invalid messages,
4965 though the change is mostly in the more lenient direction, and
4966 legacy behaviour is preserved as much as possible.
4970 * In DSA_generate_parameters_ex, if the provided seed is too short,
4973 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4975 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4977 * Alternate chains certificate forgery
4979 During certificate verification, OpenSSL will attempt to find an
4980 alternative certificate chain if the first attempt to build such a chain
4981 fails. An error in the implementation of this logic can mean that an
4982 attacker could cause certain checks on untrusted certificates to be
4983 bypassed, such as the CA flag, enabling them to use a valid leaf
4984 certificate to act as a CA and "issue" an invalid certificate.
4986 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4991 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4993 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4994 incompatibility in the handling of HMAC. The previous ABI has now been
4999 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5001 * Malformed ECParameters causes infinite loop
5003 When processing an ECParameters structure OpenSSL enters an infinite loop
5004 if the curve specified is over a specially malformed binary polynomial
5007 This can be used to perform denial of service against any
5008 system which processes public keys, certificate requests or
5009 certificates. This includes TLS clients and TLS servers with
5010 client authentication enabled.
5012 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5017 * Exploitable out-of-bounds read in X509_cmp_time
5019 X509_cmp_time does not properly check the length of the ASN1_TIME
5020 string and can read a few bytes out of bounds. In addition,
5021 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5024 An attacker can use this to craft malformed certificates and CRLs of
5025 various sizes and potentially cause a segmentation fault, resulting in
5026 a DoS on applications that verify certificates or CRLs. TLS clients
5027 that verify CRLs are affected. TLS clients and servers with client
5028 authentication enabled may be affected if they use custom verification
5031 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5032 independently by Hanno Böck.
5037 * PKCS7 crash with missing EnvelopedContent
5039 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5040 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5041 with missing content and trigger a NULL pointer dereference on parsing.
5043 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5044 structures from untrusted sources are affected. OpenSSL clients and
5045 servers are not affected.
5047 This issue was reported to OpenSSL by Michal Zalewski (Google).
5052 * CMS verify infinite loop with unknown hash function
5054 When verifying a signedData message the CMS code can enter an infinite loop
5055 if presented with an unknown hash function OID. This can be used to perform
5056 denial of service against any system which verifies signedData messages using
5058 This issue was reported to OpenSSL by Johannes Bauer.
5063 * Race condition handling NewSessionTicket
5065 If a NewSessionTicket is received by a multi-threaded client when attempting to
5066 reuse a previous ticket then a race condition can occur potentially leading to
5067 a double free of the ticket data.
5072 * Only support 256-bit or stronger elliptic curves with the
5073 'ecdh_auto' setting (server) or by default (client). Of supported
5074 curves, prefer P-256 (both).
5078 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5080 * ClientHello sigalgs DoS fix
5082 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5083 invalid signature algorithms extension a NULL pointer dereference will
5084 occur. This can be exploited in a DoS attack against the server.
5086 This issue was was reported to OpenSSL by David Ramos of Stanford
5090 *Stephen Henson and Matt Caswell*
5092 * Multiblock corrupted pointer fix
5094 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5095 feature only applies on 64 bit x86 architecture platforms that support AES
5096 NI instructions. A defect in the implementation of "multiblock" can cause
5097 OpenSSL's internal write buffer to become incorrectly set to NULL when
5098 using non-blocking IO. Typically, when the user application is using a
5099 socket BIO for writing, this will only result in a failed connection.
5100 However if some other BIO is used then it is likely that a segmentation
5101 fault will be triggered, thus enabling a potential DoS attack.
5103 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5108 * Segmentation fault in DTLSv1_listen fix
5110 The DTLSv1_listen function is intended to be stateless and processes the
5111 initial ClientHello from many peers. It is common for user code to loop
5112 over the call to DTLSv1_listen until a valid ClientHello is received with
5113 an associated cookie. A defect in the implementation of DTLSv1_listen means
5114 that state is preserved in the SSL object from one invocation to the next
5115 that can lead to a segmentation fault. Errors processing the initial
5116 ClientHello can trigger this scenario. An example of such an error could be
5117 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5120 This issue was reported to OpenSSL by Per Allansson.
5125 * Segmentation fault in ASN1_TYPE_cmp fix
5127 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5128 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5129 certificate signature algorithm consistency this can be used to crash any
5130 certificate verification operation and exploited in a DoS attack. Any
5131 application which performs certificate verification is vulnerable including
5132 OpenSSL clients and servers which enable client authentication.
5137 * Segmentation fault for invalid PSS parameters fix
5139 The signature verification routines will crash with a NULL pointer
5140 dereference if presented with an ASN.1 signature using the RSA PSS
5141 algorithm and invalid parameters. Since these routines are used to verify
5142 certificate signature algorithms this can be used to crash any
5143 certificate verification operation and exploited in a DoS attack. Any
5144 application which performs certificate verification is vulnerable including
5145 OpenSSL clients and servers which enable client authentication.
5147 This issue was was reported to OpenSSL by Brian Carpenter.
5152 * ASN.1 structure reuse memory corruption fix
5154 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5155 memory corruption via an invalid write. Such reuse is and has been
5156 strongly discouraged and is believed to be rare.
5158 Applications that parse structures containing CHOICE or ANY DEFINED BY
5159 components may be affected. Certificate parsing (d2i_X509 and related
5160 functions) are however not affected. OpenSSL clients and servers are
5166 * PKCS7 NULL pointer dereferences fix
5168 The PKCS#7 parsing code does not handle missing outer ContentInfo
5169 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5170 missing content and trigger a NULL pointer dereference on parsing.
5172 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5173 otherwise parse PKCS#7 structures from untrusted sources are
5174 affected. OpenSSL clients and servers are not affected.
5176 This issue was reported to OpenSSL by Michal Zalewski (Google).
5181 * DoS via reachable assert in SSLv2 servers fix
5183 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5184 servers that both support SSLv2 and enable export cipher suites by sending
5185 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5187 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5188 (OpenSSL development team).
5193 * Empty CKE with client auth and DHE fix
5195 If client auth is used then a server can seg fault in the event of a DHE
5196 ciphersuite being selected and a zero length ClientKeyExchange message
5197 being sent by the client. This could be exploited in a DoS attack.
5202 * Handshake with unseeded PRNG fix
5204 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5205 with an unseeded PRNG. The conditions are:
5206 - The client is on a platform where the PRNG has not been seeded
5207 automatically, and the user has not seeded manually
5208 - A protocol specific client method version has been used (i.e. not
5209 SSL_client_methodv23)
5210 - A ciphersuite is used that does not require additional random data from
5211 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5213 If the handshake succeeds then the client random that has been used will
5214 have been generated from a PRNG with insufficient entropy and therefore the
5215 output may be predictable.
5217 For example using the following command with an unseeded openssl will
5218 succeed on an unpatched platform:
5220 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5225 * Use After Free following d2i_ECPrivatekey error fix
5227 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5228 could cause a use after free condition. This, in turn, could cause a double
5229 free in several private key parsing functions (such as d2i_PrivateKey
5230 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5231 for applications that receive EC private keys from untrusted
5232 sources. This scenario is considered rare.
5234 This issue was discovered by the BoringSSL project and fixed in their
5240 * X509_to_X509_REQ NULL pointer deref fix
5242 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5243 the certificate key is invalid. This function is rarely used in practice.
5245 This issue was discovered by Brian Carpenter.
5250 * Removed the export ciphers from the DEFAULT ciphers
5254 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5256 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5257 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5258 So far those who have to target multiple platforms would compromise
5259 and argue that binary targeting say ARMv5 would still execute on
5260 ARMv8. "Universal" build resolves this compromise by providing
5261 near-optimal performance even on newer platforms.
5265 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5266 (other platforms pending).
5268 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5270 * Add support for the SignedCertificateTimestampList certificate and
5271 OCSP response extensions from RFC6962.
5275 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5276 for corner cases. (Certain input points at infinity could lead to
5277 bogus results, with non-infinity inputs mapped to infinity too.)
5281 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5282 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5283 common cases are optimized and there still is room for further
5284 improvements. Vector Permutation AES for Altivec is also added.
5288 * Add support for little-endian ppc64 Linux target.
5290 *Marcelo Cerri (IBM)*
5292 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5293 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5294 are optimized and there still is room for further improvements.
5295 Both 32- and 64-bit modes are supported.
5297 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5299 * Improved ARMv7 NEON support.
5303 * Support for SPARC Architecture 2011 crypto extensions, first
5304 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5305 SHA256/512, MD5, GHASH and modular exponentiation.
5307 *Andy Polyakov, David Miller*
5309 * Accelerated modular exponentiation for Intel processors, a.k.a.
5312 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5314 * Support for new and upcoming Intel processors, including AVX2,
5315 BMI and SHA ISA extensions. This includes additional "stitched"
5316 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5319 This work was sponsored by Intel Corp.
5323 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5324 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5325 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5329 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5330 this fixes a limitation in previous versions of OpenSSL.
5334 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5335 MGF1 digest and OAEP label.
5339 * Add EVP support for key wrapping algorithms, to avoid problems with
5340 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5341 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5342 algorithms and include tests cases.
5346 * Add functions to allocate and set the fields of an ECDSA_METHOD
5349 *Douglas E. Engert, Steve Henson*
5351 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5352 difference in days and seconds between two tm or ASN1_TIME structures.
5356 * Add -rev test option to s_server to just reverse order of characters
5357 received by client and send back to server. Also prints an abbreviated
5358 summary of the connection parameters.
5362 * New option -brief for s_client and s_server to print out a brief summary
5363 of connection parameters.
5367 * Add callbacks for arbitrary TLS extensions.
5369 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5371 * New option -crl_download in several openssl utilities to download CRLs
5372 from CRLDP extension in certificates.
5376 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5380 * New function X509_CRL_diff to generate a delta CRL from the difference
5381 of two full CRLs. Add support to "crl" utility.
5385 * New functions to set lookup_crls function and to retrieve
5386 X509_STORE from X509_STORE_CTX.
5390 * Print out deprecated issuer and subject unique ID fields in
5395 * Extend OCSP I/O functions so they can be used for simple general purpose
5396 HTTP as well as OCSP. New wrapper function which can be used to download
5397 CRLs using the OCSP API.
5401 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5405 * `SSL_CONF*` functions. These provide a common framework for application
5406 configuration using configuration files or command lines.
5410 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5411 message callback and prints the results. Needs compile time option
5412 "enable-ssl-trace". New options to s_client and s_server to enable
5417 * New ctrl and macro to retrieve supported points extensions.
5418 Print out extension in s_server and s_client.
5422 * New functions to retrieve certificate signature and signature
5427 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5432 * New Suite B modes for TLS code. These use and enforce the requirements
5433 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5434 only use Suite B curves. The Suite B modes can be set by using the
5435 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5439 * New chain verification flags for Suite B levels of security. Check
5440 algorithms are acceptable when flags are set in X509_verify_cert.
5444 * Make tls1_check_chain return a set of flags indicating checks passed
5445 by a certificate chain. Add additional tests to handle client
5446 certificates: checks for matching certificate type and issuer name
5451 * If an attempt is made to use a signature algorithm not in the peer
5452 preference list abort the handshake. If client has no suitable
5453 signature algorithms in response to a certificate request do not
5454 use the certificate.
5458 * If server EC tmp key is not in client preference list abort handshake.
5462 * Add support for certificate stores in CERT structure. This makes it
5463 possible to have different stores per SSL structure or one store in
5464 the parent SSL_CTX. Include distinct stores for certificate chain
5465 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5466 to build and store a certificate chain in CERT structure: returning
5467 an error if the chain cannot be built: this will allow applications
5468 to test if a chain is correctly configured.
5470 Note: if the CERT based stores are not set then the parent SSL_CTX
5471 store is used to retain compatibility with existing behaviour.
5475 * New function ssl_set_client_disabled to set a ciphersuite disabled
5476 mask based on the current session, check mask when sending client
5477 hello and checking the requested ciphersuite.
5481 * New ctrls to retrieve and set certificate types in a certificate
5482 request message. Print out received values in s_client. If certificate
5483 types is not set with custom values set sensible values based on
5484 supported signature algorithms.
5488 * Support for distinct client and server supported signature algorithms.
5492 * Add certificate callback. If set this is called whenever a certificate
5493 is required by client or server. An application can decide which
5494 certificate chain to present based on arbitrary criteria: for example
5495 supported signature algorithms. Add very simple example to s_server.
5496 This fixes many of the problems and restrictions of the existing client
5497 certificate callback: for example you can now clear an existing
5498 certificate and specify the whole chain.
5502 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5503 the certificate can be used for (if anything). Set valid_flags field
5504 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5505 to have similar checks in it.
5507 Add new "cert_flags" field to CERT structure and include a "strict mode".
5508 This enforces some TLS certificate requirements (such as only permitting
5509 certificate signature algorithms contained in the supported algorithms
5510 extension) which some implementations ignore: this option should be used
5511 with caution as it could cause interoperability issues.
5515 * Update and tidy signature algorithm extension processing. Work out
5516 shared signature algorithms based on preferences and peer algorithms
5517 and print them out in s_client and s_server. Abort handshake if no
5518 shared signature algorithms.
5522 * Add new functions to allow customised supported signature algorithms
5523 for SSL and SSL_CTX structures. Add options to s_client and s_server
5528 * New function SSL_certs_clear() to delete all references to certificates
5529 from an SSL structure. Before this once a certificate had been added
5530 it couldn't be removed.
5534 * Integrate hostname, email address and IP address checking with certificate
5535 verification. New verify options supporting checking in openssl utility.
5539 * Fixes and wildcard matching support to hostname and email checking
5540 functions. Add manual page.
5542 *Florian Weimer (Red Hat Product Security Team)*
5544 * New functions to check a hostname email or IP address against a
5545 certificate. Add options x509 utility to print results of checks against
5550 * Fix OCSP checking.
5552 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5554 * Initial experimental support for explicitly trusted non-root CAs.
5555 OpenSSL still tries to build a complete chain to a root but if an
5556 intermediate CA has a trust setting included that is used. The first
5557 setting is used: whether to trust (e.g., -addtrust option to the x509
5562 * Add -trusted_first option which attempts to find certificates in the
5563 trusted store even if an untrusted chain is also supplied.
5567 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5568 platform support for Linux and Android.
5572 * Support for linux-x32, ILP32 environment in x86_64 framework.
5576 * Experimental multi-implementation support for FIPS capable OpenSSL.
5577 When in FIPS mode the approved implementations are used as normal,
5578 when not in FIPS mode the internal unapproved versions are used instead.
5579 This means that the FIPS capable OpenSSL isn't forced to use the
5580 (often lower performance) FIPS implementations outside FIPS mode.
5584 * Transparently support X9.42 DH parameters when calling
5585 PEM_read_bio_DHparameters. This means existing applications can handle
5586 the new parameter format automatically.
5590 * Initial experimental support for X9.42 DH parameter format: mainly
5591 to support use of 'q' parameter for RFC5114 parameters.
5595 * Add DH parameters from RFC5114 including test data to dhtest.
5599 * Support for automatic EC temporary key parameter selection. If enabled
5600 the most preferred EC parameters are automatically used instead of
5601 hardcoded fixed parameters. Now a server just has to call:
5602 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5603 support ECDH and use the most appropriate parameters.
5607 * Enhance and tidy EC curve and point format TLS extension code. Use
5608 static structures instead of allocation if default values are used.
5609 New ctrls to set curves we wish to support and to retrieve shared curves.
5610 Print out shared curves in s_server. New options to s_server and s_client
5611 to set list of supported curves.
5615 * New ctrls to retrieve supported signature algorithms and
5616 supported curve values as an array of NIDs. Extend openssl utility
5617 to print out received values.
5621 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5622 between NIDs and the more common NIST names such as "P-256". Enhance
5623 ecparam utility and ECC method to recognise the NIST names for curves.
5627 * Enhance SSL/TLS certificate chain handling to support different
5628 chains for each certificate instead of one chain in the parent SSL_CTX.
5632 * Support for fixed DH ciphersuite client authentication: where both
5633 server and client use DH certificates with common parameters.
5637 * Support for fixed DH ciphersuites: those requiring DH server
5642 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5644 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5645 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5646 X509_CINF_get_signature were reverted post internal team review.
5651 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5653 * OCSP Status Request extension unbounded memory growth
5655 A malicious client can send an excessively large OCSP Status Request
5656 extension. If that client continually requests renegotiation, sending a
5657 large OCSP Status Request extension each time, then there will be unbounded
5658 memory growth on the server. This will eventually lead to a Denial Of
5659 Service attack through memory exhaustion. Servers with a default
5660 configuration are vulnerable even if they do not support OCSP. Builds using
5661 the "no-ocsp" build time option are not affected.
5663 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5668 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5671 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5677 * OOB write in MDC2_Update()
5679 An overflow can occur in MDC2_Update() either if called directly or
5680 through the EVP_DigestUpdate() function using MDC2. If an attacker
5681 is able to supply very large amounts of input data after a previous
5682 call to EVP_EncryptUpdate() with a partial block then a length check
5683 can overflow resulting in a heap corruption.
5685 The amount of data needed is comparable to SIZE_MAX which is impractical
5688 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5693 * Malformed SHA512 ticket DoS
5695 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5696 DoS attack where a malformed ticket will result in an OOB read which will
5699 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5700 a custom server callback and ticket lookup mechanism.
5702 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5707 * OOB write in BN_bn2dec()
5709 The function BN_bn2dec() does not check the return value of BN_div_word().
5710 This can cause an OOB write if an application uses this function with an
5711 overly large BIGNUM. This could be a problem if an overly large certificate
5712 or CRL is printed out from an untrusted source. TLS is not affected because
5713 record limits will reject an oversized certificate before it is parsed.
5715 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5720 * OOB read in TS_OBJ_print_bio()
5722 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5723 the total length the OID text representation would use and not the amount
5724 of data written. This will result in OOB reads when large OIDs are
5727 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5732 * Pointer arithmetic undefined behaviour
5734 Avoid some undefined pointer arithmetic
5736 A common idiom in the codebase is to check limits in the following manner:
5739 Where "p" points to some malloc'd data of SIZE bytes and
5742 "len" here could be from some externally supplied data (e.g. from a TLS
5745 The rules of C pointer arithmetic are such that "p + len" is only well
5746 defined where len <= SIZE. Therefore the above idiom is actually
5747 undefined behaviour.
5749 For example this could cause problems if some malloc implementation
5750 provides an address for "p" such that "p + len" actually overflows for
5751 values of len that are too big and therefore p + len < limit.
5753 This issue was reported to OpenSSL by Guido Vranken
5758 * Constant time flag not preserved in DSA signing
5760 Operations in the DSA signing algorithm should run in constant time in
5761 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5762 implementation means that a non-constant time codepath is followed for
5763 certain operations. This has been demonstrated through a cache-timing
5764 attack to be sufficient for an attacker to recover the private DSA key.
5766 This issue was reported by César Pereida (Aalto University), Billy Brumley
5767 (Tampere University of Technology), and Yuval Yarom (The University of
5768 Adelaide and NICTA).
5773 * DTLS buffered message DoS
5775 In a DTLS connection where handshake messages are delivered out-of-order
5776 those messages that OpenSSL is not yet ready to process will be buffered
5777 for later use. Under certain circumstances, a flaw in the logic means that
5778 those messages do not get removed from the buffer even though the handshake
5779 has been completed. An attacker could force up to approx. 15 messages to
5780 remain in the buffer when they are no longer required. These messages will
5781 be cleared when the DTLS connection is closed. The default maximum size for
5782 a message is 100k. Therefore the attacker could force an additional 1500k
5783 to be consumed per connection. By opening many simulataneous connections an
5784 attacker could cause a DoS attack through memory exhaustion.
5786 This issue was reported to OpenSSL by Quan Luo.
5791 * DTLS replay protection DoS
5793 A flaw in the DTLS replay attack protection mechanism means that records
5794 that arrive for future epochs update the replay protection "window" before
5795 the MAC for the record has been validated. This could be exploited by an
5796 attacker by sending a record for the next epoch (which does not have to
5797 decrypt or have a valid MAC), with a very large sequence number. This means
5798 that all subsequent legitimate packets are dropped causing a denial of
5799 service for a specific DTLS connection.
5801 This issue was reported to OpenSSL by the OCAP audit team.
5806 * Certificate message OOB reads
5808 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5809 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5810 theoretical DoS risk but this has not been observed in practice on common
5813 The messages affected are client certificate, client certificate request
5814 and server certificate. As a result the attack can only be performed
5815 against a client or a server which enables client authentication.
5817 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5822 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5824 * Prevent padding oracle in AES-NI CBC MAC check
5826 A MITM attacker can use a padding oracle attack to decrypt traffic
5827 when the connection uses an AES CBC cipher and the server support
5830 This issue was introduced as part of the fix for Lucky 13 padding
5831 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5832 constant time by making sure that always the same bytes are read and
5833 compared against either the MAC or padding bytes. But it no longer
5834 checked that there was enough data to have both the MAC and padding
5837 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5842 * Fix EVP_EncodeUpdate overflow
5844 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5845 Base64 encoding of binary data. If an attacker is able to supply very large
5846 amounts of input data then a length check can overflow resulting in a heap
5849 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5850 the `PEM_write_bio*` family of functions. These are mainly used within the
5851 OpenSSL command line applications, so any application which processes data
5852 from an untrusted source and outputs it as a PEM file should be considered
5853 vulnerable to this issue. User applications that call these APIs directly
5854 with large amounts of untrusted data may also be vulnerable.
5856 This issue was reported by Guido Vranken.
5861 * Fix EVP_EncryptUpdate overflow
5863 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5864 is able to supply very large amounts of input data after a previous call to
5865 EVP_EncryptUpdate() with a partial block then a length check can overflow
5866 resulting in a heap corruption. Following an analysis of all OpenSSL
5867 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5868 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5869 the first called function after an EVP_EncryptInit(), and therefore that
5870 specific call must be safe. The second form is where the length passed to
5871 EVP_EncryptUpdate() can be seen from the code to be some small value and
5872 therefore there is no possibility of an overflow. Since all instances are
5873 one of these two forms, it is believed that there can be no overflows in
5874 internal code due to this problem. It should be noted that
5875 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5876 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5877 of these calls have also been analysed too and it is believed there are no
5878 instances in internal usage where an overflow could occur.
5880 This issue was reported by Guido Vranken.
5885 * Prevent ASN.1 BIO excessive memory allocation
5887 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5888 a short invalid encoding can casuse allocation of large amounts of memory
5889 potentially consuming excessive resources or exhausting memory.
5891 Any application parsing untrusted data through d2i BIO functions is
5892 affected. The memory based functions such as d2i_X509() are *not* affected.
5893 Since the memory based functions are used by the TLS library, TLS
5894 applications are not affected.
5896 This issue was reported by Brian Carpenter.
5903 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5904 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5905 in arbitrary stack data being returned in the buffer.
5907 This issue was reported by Guido Vranken.
5912 * Modify behavior of ALPN to invoke callback after SNI/servername
5913 callback, such that updates to the SSL_CTX affect ALPN.
5917 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5922 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5923 methods are enabled and ssl2 is disabled the methods return NULL.
5927 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5929 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5930 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5931 provide any "EXPORT" or "LOW" strength ciphers.
5935 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5936 is by default disabled at build-time. Builds that are not configured with
5937 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5938 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5939 will need to explicitly call either of:
5941 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5943 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5945 as appropriate. Even if either of those is used, or the application
5946 explicitly uses the version-specific SSLv2_method() or its client and
5947 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5948 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5949 ciphers, and SSLv2 56-bit DES are no longer available.
5954 * Fix a double-free in DSA code
5956 A double free bug was discovered when OpenSSL parses malformed DSA private
5957 keys and could lead to a DoS attack or memory corruption for applications
5958 that receive DSA private keys from untrusted sources. This scenario is
5961 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5967 * Disable SRP fake user seed to address a server memory leak.
5969 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5971 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5972 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5973 was changed to ignore the "fake user" SRP seed, even if the seed
5976 Users should use SRP_VBASE_get1_by_user instead. Note that in
5977 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5978 also that even though configuring the SRP seed attempts to hide
5979 invalid usernames by continuing the handshake with fake
5980 credentials, this behaviour is not constant time and no strong
5981 guarantees are made that the handshake is indistinguishable from
5982 that of a valid user.
5987 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5989 In the BN_hex2bn function the number of hex digits is calculated using an
5990 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5991 large values of `i` this can result in `bn_expand` not allocating any
5992 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5993 field as NULL leading to a subsequent NULL ptr deref. For very large values
5994 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5995 In this case memory is allocated to the internal BIGNUM data field, but it
5996 is insufficiently sized leading to heap corruption. A similar issue exists
5997 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5998 is ever called by user applications with very large untrusted hex/dec data.
5999 This is anticipated to be a rare occurrence.
6001 All OpenSSL internal usage of these functions use data that is not expected
6002 to be untrusted, e.g. config file data or application command line
6003 arguments. If user developed applications generate config file data based
6004 on untrusted data then it is possible that this could also lead to security
6005 consequences. This is also anticipated to be rare.
6007 This issue was reported to OpenSSL by Guido Vranken.
6012 * Fix memory issues in `BIO_*printf` functions
6014 The internal `fmtstr` function used in processing a "%s" format string in
6015 the `BIO_*printf` functions could overflow while calculating the length of a
6016 string and cause an OOB read when printing very long strings.
6018 Additionally the internal `doapr_outch` function can attempt to write to an
6019 OOB memory location (at an offset from the NULL pointer) in the event of a
6020 memory allocation failure. In 1.0.2 and below this could be caused where
6021 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6022 could be in processing a very long "%s" format string. Memory leaks can
6025 The first issue may mask the second issue dependent on compiler behaviour.
6026 These problems could enable attacks where large amounts of untrusted data
6027 is passed to the `BIO_*printf` functions. If applications use these functions
6028 in this way then they could be vulnerable. OpenSSL itself uses these
6029 functions when printing out human-readable dumps of ASN.1 data. Therefore
6030 applications that print this data could be vulnerable if the data is from
6031 untrusted sources. OpenSSL command line applications could also be
6032 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6033 as command line arguments.
6035 Libssl is not considered directly vulnerable. Additionally certificates etc
6036 received via remote connections via libssl are also unlikely to be able to
6037 trigger these issues because of message size limits enforced within libssl.
6039 This issue was reported to OpenSSL Guido Vranken.
6044 * Side channel attack on modular exponentiation
6046 A side-channel attack was found which makes use of cache-bank conflicts on
6047 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6048 of RSA keys. The ability to exploit this issue is limited as it relies on
6049 an attacker who has control of code in a thread running on the same
6050 hyper-threaded core as the victim thread which is performing decryptions.
6052 This issue was reported to OpenSSL by Yuval Yarom, The University of
6053 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6054 Nadia Heninger, University of Pennsylvania with more information at
6055 <http://cachebleed.info>.
6060 * Change the req app to generate a 2048-bit RSA/DSA key by default,
6061 if no keysize is specified with default_bits. This fixes an
6062 omission in an earlier change that changed all RSA/DSA key generation
6063 apps to use 2048 bits by default.
6067 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6069 * Protection for DH small subgroup attacks
6071 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6072 switched on by default and cannot be disabled. This could have some
6077 * SSLv2 doesn't block disabled ciphers
6079 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6080 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6081 been disabled, provided that the SSLv2 protocol was not also disabled via
6084 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6085 and Sebastian Schinzel.
6090 * Reject DH handshakes with parameters shorter than 1024 bits.
6094 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6096 * Certificate verify crash with missing PSS parameter
6098 The signature verification routines will crash with a NULL pointer
6099 dereference if presented with an ASN.1 signature using the RSA PSS
6100 algorithm and absent mask generation function parameter. Since these
6101 routines are used to verify certificate signature algorithms this can be
6102 used to crash any certificate verification operation and exploited in a
6103 DoS attack. Any application which performs certificate verification is
6104 vulnerable including OpenSSL clients and servers which enable client
6107 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6112 * X509_ATTRIBUTE memory leak
6114 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6115 memory. This structure is used by the PKCS#7 and CMS routines so any
6116 application which reads PKCS#7 or CMS data from untrusted sources is
6117 affected. SSL/TLS is not affected.
6119 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6125 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6126 This changes the decoding behaviour for some invalid messages,
6127 though the change is mostly in the more lenient direction, and
6128 legacy behaviour is preserved as much as possible.
6132 * In DSA_generate_parameters_ex, if the provided seed is too short,
6133 use a random seed, as already documented.
6135 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6137 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6139 * Alternate chains certificate forgery
6141 During certificate verfification, OpenSSL will attempt to find an
6142 alternative certificate chain if the first attempt to build such a chain
6143 fails. An error in the implementation of this logic can mean that an
6144 attacker could cause certain checks on untrusted certificates to be
6145 bypassed, such as the CA flag, enabling them to use a valid leaf
6146 certificate to act as a CA and "issue" an invalid certificate.
6148 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6154 * Race condition handling PSK identify hint
6156 If PSK identity hints are received by a multi-threaded client then
6157 the values are wrongly updated in the parent SSL_CTX structure. This can
6158 result in a race condition potentially leading to a double free of the
6164 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6166 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6167 incompatibility in the handling of HMAC. The previous ABI has now been
6170 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6172 * Malformed ECParameters causes infinite loop
6174 When processing an ECParameters structure OpenSSL enters an infinite loop
6175 if the curve specified is over a specially malformed binary polynomial
6178 This can be used to perform denial of service against any
6179 system which processes public keys, certificate requests or
6180 certificates. This includes TLS clients and TLS servers with
6181 client authentication enabled.
6183 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6188 * Exploitable out-of-bounds read in X509_cmp_time
6190 X509_cmp_time does not properly check the length of the ASN1_TIME
6191 string and can read a few bytes out of bounds. In addition,
6192 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6195 An attacker can use this to craft malformed certificates and CRLs of
6196 various sizes and potentially cause a segmentation fault, resulting in
6197 a DoS on applications that verify certificates or CRLs. TLS clients
6198 that verify CRLs are affected. TLS clients and servers with client
6199 authentication enabled may be affected if they use custom verification
6202 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6203 independently by Hanno Böck.
6208 * PKCS7 crash with missing EnvelopedContent
6210 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6211 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6212 with missing content and trigger a NULL pointer dereference on parsing.
6214 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6215 structures from untrusted sources are affected. OpenSSL clients and
6216 servers are not affected.
6218 This issue was reported to OpenSSL by Michal Zalewski (Google).
6223 * CMS verify infinite loop with unknown hash function
6225 When verifying a signedData message the CMS code can enter an infinite loop
6226 if presented with an unknown hash function OID. This can be used to perform
6227 denial of service against any system which verifies signedData messages using
6229 This issue was reported to OpenSSL by Johannes Bauer.
6234 * Race condition handling NewSessionTicket
6236 If a NewSessionTicket is received by a multi-threaded client when attempting to
6237 reuse a previous ticket then a race condition can occur potentially leading to
6238 a double free of the ticket data.
6243 * Reject DH handshakes with parameters shorter than 768 bits.
6245 *Kurt Roeckx and Emilia Kasper*
6247 * dhparam: generate 2048-bit parameters by default.
6249 *Kurt Roeckx and Emilia Kasper*
6251 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6253 * Segmentation fault in ASN1_TYPE_cmp fix
6255 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6256 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6257 certificate signature algorithm consistency this can be used to crash any
6258 certificate verification operation and exploited in a DoS attack. Any
6259 application which performs certificate verification is vulnerable including
6260 OpenSSL clients and servers which enable client authentication.
6265 * ASN.1 structure reuse memory corruption fix
6267 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6268 memory corruption via an invalid write. Such reuse is and has been
6269 strongly discouraged and is believed to be rare.
6271 Applications that parse structures containing CHOICE or ANY DEFINED BY
6272 components may be affected. Certificate parsing (d2i_X509 and related
6273 functions) are however not affected. OpenSSL clients and servers are
6279 * PKCS7 NULL pointer dereferences fix
6281 The PKCS#7 parsing code does not handle missing outer ContentInfo
6282 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6283 missing content and trigger a NULL pointer dereference on parsing.
6285 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6286 otherwise parse PKCS#7 structures from untrusted sources are
6287 affected. OpenSSL clients and servers are not affected.
6289 This issue was reported to OpenSSL by Michal Zalewski (Google).
6294 * DoS via reachable assert in SSLv2 servers fix
6296 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6297 servers that both support SSLv2 and enable export cipher suites by sending
6298 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6300 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6301 (OpenSSL development team).
6306 * Use After Free following d2i_ECPrivatekey error fix
6308 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6309 could cause a use after free condition. This, in turn, could cause a double
6310 free in several private key parsing functions (such as d2i_PrivateKey
6311 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6312 for applications that receive EC private keys from untrusted
6313 sources. This scenario is considered rare.
6315 This issue was discovered by the BoringSSL project and fixed in their
6321 * X509_to_X509_REQ NULL pointer deref fix
6323 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6324 the certificate key is invalid. This function is rarely used in practice.
6326 This issue was discovered by Brian Carpenter.
6331 * Removed the export ciphers from the DEFAULT ciphers
6335 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6337 * Build fixes for the Windows and OpenVMS platforms
6339 *Matt Caswell and Richard Levitte*
6341 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6343 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6344 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6345 dereference. This could lead to a Denial Of Service attack. Thanks to
6346 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6351 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6352 dtls1_buffer_record function under certain conditions. In particular this
6353 could occur if an attacker sent repeated DTLS records with the same
6354 sequence number but for the next epoch. The memory leak could be exploited
6355 by an attacker in a Denial of Service attack through memory exhaustion.
6356 Thanks to Chris Mueller for reporting this issue.
6361 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6362 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6363 method would be set to NULL which could later result in a NULL pointer
6364 dereference. Thanks to Frank Schmirler for reporting this issue.
6369 * Abort handshake if server key exchange message is omitted for ephemeral
6372 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6373 reporting this issue.
6378 * Remove non-export ephemeral RSA code on client and server. This code
6379 violated the TLS standard by allowing the use of temporary RSA keys in
6380 non-export ciphersuites and could be used by a server to effectively
6381 downgrade the RSA key length used to a value smaller than the server
6382 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6383 INRIA or reporting this issue.
6388 * Fixed issue where DH client certificates are accepted without verification.
6389 An OpenSSL server will accept a DH certificate for client authentication
6390 without the certificate verify message. This effectively allows a client to
6391 authenticate without the use of a private key. This only affects servers
6392 which trust a client certificate authority which issues certificates
6393 containing DH keys: these are extremely rare and hardly ever encountered.
6394 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6400 * Ensure that the session ID context of an SSL is updated when its
6401 SSL_CTX is updated via SSL_set_SSL_CTX.
6403 The session ID context is typically set from the parent SSL_CTX,
6404 and can vary with the CTX.
6408 * Fix various certificate fingerprint issues.
6410 By using non-DER or invalid encodings outside the signed portion of a
6411 certificate the fingerprint can be changed without breaking the signature.
6412 Although no details of the signed portion of the certificate can be changed
6413 this can cause problems with some applications: e.g. those using the
6414 certificate fingerprint for blacklists.
6416 1. Reject signatures with non zero unused bits.
6418 If the BIT STRING containing the signature has non zero unused bits reject
6419 the signature. All current signature algorithms require zero unused bits.
6421 2. Check certificate algorithm consistency.
6423 Check the AlgorithmIdentifier inside TBS matches the one in the
6424 certificate signature. NB: this will result in signature failure
6425 errors for some broken certificates.
6427 Thanks to Konrad Kraszewski from Google for reporting this issue.
6429 3. Check DSA/ECDSA signatures use DER.
6431 Re-encode DSA/ECDSA signatures and compare with the original received
6432 signature. Return an error if there is a mismatch.
6434 This will reject various cases including garbage after signature
6435 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6436 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6437 (negative or with leading zeroes).
6439 Further analysis was conducted and fixes were developed by Stephen Henson
6440 of the OpenSSL core team.
6446 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6447 results on some platforms, including x86_64. This bug occurs at random
6448 with a very low probability, and is not known to be exploitable in any
6449 way, though its exact impact is difficult to determine. Thanks to Pieter
6450 Wuille (Blockstream) who reported this issue and also suggested an initial
6451 fix. Further analysis was conducted by the OpenSSL development team and
6452 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6453 the OpenSSL core team.
6458 * Do not resume sessions on the server if the negotiated protocol
6459 version does not match the session's version. Resuming with a different
6460 version, while not strictly forbidden by the RFC, is of questionable
6461 sanity and breaks all known clients.
6463 *David Benjamin, Emilia Käsper*
6465 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6466 early CCS messages during renegotiation. (Note that because
6467 renegotiation is encrypted, this early CCS was not exploitable.)
6471 * Tighten client-side session ticket handling during renegotiation:
6472 ensure that the client only accepts a session ticket if the server sends
6473 the extension anew in the ServerHello. Previously, a TLS client would
6474 reuse the old extension state and thus accept a session ticket if one was
6475 announced in the initial ServerHello.
6477 Similarly, ensure that the client requires a session ticket if one
6478 was advertised in the ServerHello. Previously, a TLS client would
6479 ignore a missing NewSessionTicket message.
6483 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6487 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6488 sends a carefully crafted handshake message, to cause OpenSSL to fail
6489 to free up to 64k of memory causing a memory leak. This could be
6490 exploited in a Denial Of Service attack. This issue affects OpenSSL
6491 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6492 whether SRTP is used or configured. Implementations of OpenSSL that
6493 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6495 The fix was developed by the OpenSSL team.
6500 * Session Ticket Memory Leak.
6502 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6503 integrity of that ticket is first verified. In the event of a session
6504 ticket integrity check failing, OpenSSL will fail to free memory
6505 causing a memory leak. By sending a large number of invalid session
6506 tickets an attacker could exploit this issue in a Denial Of Service
6512 * Build option no-ssl3 is incomplete.
6514 When OpenSSL is configured with "no-ssl3" as a build option, servers
6515 could accept and complete a SSL 3.0 handshake, and clients could be
6516 configured to send them.
6519 *Akamai and the OpenSSL team*
6521 * Add support for TLS_FALLBACK_SCSV.
6522 Client applications doing fallback retries should call
6523 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6526 *Adam Langley, Bodo Moeller*
6528 * Add additional DigestInfo checks.
6530 Re-encode DigestInto in DER and check against the original when
6531 verifying RSA signature: this will reject any improperly encoded
6532 DigestInfo structures.
6534 Note: this is a precautionary measure and no attacks are currently known.
6538 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6540 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6541 SRP code can be overrun an internal buffer. Add sanity check that
6542 g, A, B < N to SRP code.
6544 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6545 Group for discovering this issue.
6550 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6551 TLS 1.0 instead of higher protocol versions when the ClientHello message
6552 is badly fragmented. This allows a man-in-the-middle attacker to force a
6553 downgrade to TLS 1.0 even if both the server and the client support a
6554 higher protocol version, by modifying the client's TLS records.
6556 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6557 researching this issue.
6562 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6563 to a denial of service attack. A malicious server can crash the client
6564 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6565 ciphersuite and sending carefully crafted handshake messages.
6567 Thanks to Felix Gröbert (Google) for discovering and researching this
6573 * By sending carefully crafted DTLS packets an attacker could cause openssl
6574 to leak memory. This can be exploited through a Denial of Service attack.
6575 Thanks to Adam Langley for discovering and researching this issue.
6580 * An attacker can force openssl to consume large amounts of memory whilst
6581 processing DTLS handshake messages. This can be exploited through a
6582 Denial of Service attack.
6583 Thanks to Adam Langley for discovering and researching this issue.
6588 * An attacker can force an error condition which causes openssl to crash
6589 whilst processing DTLS packets due to memory being freed twice. This
6590 can be exploited through a Denial of Service attack.
6591 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6597 * If a multithreaded client connects to a malicious server using a resumed
6598 session and the server sends an ec point format extension it could write
6599 up to 255 bytes to freed memory.
6601 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6607 * A malicious server can crash an OpenSSL client with a null pointer
6608 dereference (read) by specifying an SRP ciphersuite even though it was not
6609 properly negotiated with the client. This can be exploited through a
6610 Denial of Service attack.
6612 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6613 discovering and researching this issue.
6618 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6619 X509_name_oneline, X509_name_print_ex et al. to leak some information
6620 from the stack. Applications may be affected if they echo pretty printing
6621 output to the attacker.
6623 Thanks to Ivan Fratric (Google) for discovering this issue.
6626 *Emilia Käsper, and Steve Henson*
6628 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6629 for corner cases. (Certain input points at infinity could lead to
6630 bogus results, with non-infinity inputs mapped to infinity too.)
6634 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6636 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6637 handshake can force the use of weak keying material in OpenSSL
6638 SSL/TLS clients and servers.
6640 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6641 researching this issue. ([CVE-2014-0224])
6643 *KIKUCHI Masashi, Steve Henson*
6645 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6646 OpenSSL DTLS client the code can be made to recurse eventually crashing
6649 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6652 *Imre Rad, Steve Henson*
6654 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6655 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6656 client or server. This is potentially exploitable to run arbitrary
6657 code on a vulnerable client or server.
6659 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6661 *Jüri Aedla, Steve Henson*
6663 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6664 are subject to a denial of service attack.
6666 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6667 this issue. ([CVE-2014-3470])
6669 *Felix Gröbert, Ivan Fratric, Steve Henson*
6671 * Harmonize version and its documentation. -f flag is used to display
6674 *mancha <mancha1@zoho.com>*
6676 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6677 in i2d_ECPrivateKey.
6679 *mancha <mancha1@zoho.com>*
6681 * Fix some double frees. These are not thought to be exploitable.
6683 *mancha <mancha1@zoho.com>*
6685 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6687 * A missing bounds check in the handling of the TLS heartbeat extension
6688 can be used to reveal up to 64k of memory to a connected client or
6691 Thanks for Neel Mehta of Google Security for discovering this bug and to
6692 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6693 preparing the fix ([CVE-2014-0160])
6695 *Adam Langley, Bodo Moeller*
6697 * Fix for the attack described in the paper "Recovering OpenSSL
6698 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6699 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6700 <http://eprint.iacr.org/2014/140>
6702 Thanks to Yuval Yarom and Naomi Benger for discovering this
6703 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6705 *Yuval Yarom and Naomi Benger*
6707 * TLS pad extension: draft-agl-tls-padding-03
6709 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6710 TLS client Hello record length value would otherwise be > 255 and
6711 less that 512 pad with a dummy extension containing zeroes so it
6712 is at least 512 bytes long.
6714 *Adam Langley, Steve Henson*
6716 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6718 * Fix for TLS record tampering bug. A carefully crafted invalid
6719 handshake could crash OpenSSL with a NULL pointer exception.
6720 Thanks to Anton Johansson for reporting this issues.
6723 * Keep original DTLS digest and encryption contexts in retransmission
6724 structures so we can use the previous session parameters if they need
6725 to be resent. ([CVE-2013-6450])
6729 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6730 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6731 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6732 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6733 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6734 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6736 *Rob Stradling, Adam Langley*
6738 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6740 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6741 supporting platforms or when small records were transferred.
6743 *Andy Polyakov, Steve Henson*
6745 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6747 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6749 This addresses the flaw in CBC record processing discovered by
6750 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6751 at: <http://www.isg.rhul.ac.uk/tls/>
6753 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6754 Security Group at Royal Holloway, University of London
6755 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6756 Emilia Käsper for the initial patch.
6759 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6761 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6762 ciphersuites which can be exploited in a denial of service attack.
6763 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6764 and detecting this bug and to Wolfgang Ettlinger
6765 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6770 * Return an error when checking OCSP signatures when key is NULL.
6771 This fixes a DoS attack. ([CVE-2013-0166])
6775 * Make openssl verify return errors.
6777 *Chris Palmer <palmer@google.com> and Ben Laurie*
6779 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6780 the right response is stapled. Also change SSL_get_certificate()
6781 so it returns the certificate actually sent.
6782 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6784 *Rob Stradling <rob.stradling@comodo.com>*
6786 * Fix possible deadlock when decoding public keys.
6790 * Don't use TLS 1.0 record version number in initial client hello
6795 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6797 * Sanity check record length before skipping explicit IV in TLS
6798 1.2, 1.1 and DTLS to fix DoS attack.
6800 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6801 fuzzing as a service testing platform.
6806 * Initialise tkeylen properly when encrypting CMS messages.
6807 Thanks to Solar Designer of Openwall for reporting this issue.
6811 * In FIPS mode don't try to use composite ciphers as they are not
6816 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6818 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6819 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6820 mean any application compiled against OpenSSL 1.0.0 headers setting
6821 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6822 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6823 0x10000000L Any application which was previously compiled against
6824 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6825 will need to be recompiled as a result. Letting be results in
6826 inability to disable specifically TLS 1.1 and in client context,
6827 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6831 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6832 disable just protocol X, but all protocols above X *if* there are
6833 protocols *below* X still enabled. In more practical terms it means
6834 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6835 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6836 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6841 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6843 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6844 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6845 in CRYPTO_realloc_clean.
6847 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6848 issue and to Adam Langley <agl@chromium.org> for fixing it.
6851 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6853 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6857 * Workarounds for some broken servers that "hang" if a client hello
6858 record length exceeds 255 bytes.
6860 1. Do not use record version number > TLS 1.0 in initial client
6861 hello: some (but not all) hanging servers will now work.
6862 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6863 the number of ciphers sent in the client hello. This should be
6864 set to an even number, such as 50, for example by passing:
6865 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6866 Most broken servers should now work.
6867 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6868 TLS 1.2 client support entirely.
6872 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6876 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6878 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6879 STRING form instead of a DigestInfo.
6883 * The format used for MDC2 RSA signatures is inconsistent between EVP
6884 and the RSA_sign/RSA_verify functions. This was made more apparent when
6885 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6886 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6887 the correct format in RSA_verify so both forms transparently work.
6891 * Some servers which support TLS 1.0 can choke if we initially indicate
6892 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6893 encrypted premaster secret. As a workaround use the maximum permitted
6894 client version in client hello, this should keep such servers happy
6895 and still work with previous versions of OpenSSL.
6899 * Add support for TLS/DTLS heartbeats.
6901 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6903 * Add support for SCTP.
6905 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6907 * Improved PRNG seeding for VOS.
6909 *Paul Green <Paul.Green@stratus.com>*
6911 * Extensive assembler packs updates, most notably:
6913 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6914 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6915 - x86_64: bit-sliced AES implementation;
6916 - ARM: NEON support, contemporary platforms optimizations;
6917 - s390x: z196 support;
6918 - `*`: GHASH and GF(2^m) multiplication implementations;
6922 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6923 (removal of unnecessary code)
6925 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6927 * Add TLS key material exporter from RFC 5705.
6931 * Add DTLS-SRTP negotiation from RFC 5764.
6935 * Add Next Protocol Negotiation,
6936 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6937 disabled with a no-npn flag to config or Configure. Code donated
6940 *Adam Langley <agl@google.com> and Ben Laurie*
6942 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6943 NIST-P256, NIST-P521, with constant-time single point multiplication on
6944 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6945 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6946 Code made available under Apache License version 2.0.
6948 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6949 line to include this in your build of OpenSSL, and run "make depend" (or
6950 "make update"). This enables the following EC_METHODs:
6952 EC_GFp_nistp224_method()
6953 EC_GFp_nistp256_method()
6954 EC_GFp_nistp521_method()
6956 EC_GROUP_new_by_curve_name() will automatically use these (while
6957 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6960 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6962 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6963 all platforms. Move ssize_t definition from e_os.h to the public
6964 header file e_os2.h as it now appears in public header file cms.h
6968 * New -sigopt option to the ca, req and x509 utilities. Additional
6969 signature parameters can be passed using this option and in
6974 * Add RSA PSS signing function. This will generate and set the
6975 appropriate AlgorithmIdentifiers for PSS based on those in the
6976 corresponding EVP_MD_CTX structure. No application support yet.
6980 * Support for companion algorithm specific ASN1 signing routines.
6981 New function ASN1_item_sign_ctx() signs a pre-initialised
6982 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6983 the appropriate parameters.
6987 * Add new algorithm specific ASN1 verification initialisation function
6988 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6989 handling will be the same no matter what EVP_PKEY_METHOD is used.
6990 Add a PSS handler to support verification of PSS signatures: checked
6991 against a number of sample certificates.
6995 * Add signature printing for PSS. Add PSS OIDs.
6997 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6999 * Add algorithm specific signature printing. An individual ASN1 method
7000 can now print out signatures instead of the standard hex dump.
7002 More complex signatures (e.g. PSS) can print out more meaningful
7003 information. Include DSA version that prints out the signature
7008 * Password based recipient info support for CMS library: implementing
7013 * Split password based encryption into PBES2 and PBKDF2 functions. This
7014 neatly separates the code into cipher and PBE sections and is required
7015 for some algorithms that split PBES2 into separate pieces (such as
7016 password based CMS).
7020 * Session-handling fixes:
7021 - Fix handling of connections that are resuming with a session ID,
7022 but also support Session Tickets.
7023 - Fix a bug that suppressed issuing of a new ticket if the client
7024 presented a ticket with an expired session.
7025 - Try to set the ticket lifetime hint to something reasonable.
7026 - Make tickets shorter by excluding irrelevant information.
7027 - On the client side, don't ignore renewed tickets.
7029 *Adam Langley, Bodo Moeller (Google)*
7031 * Fix PSK session representation.
7035 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7037 This work was sponsored by Intel.
7041 * Add GCM support to TLS library. Some custom code is needed to split
7042 the IV between the fixed (from PRF) and explicit (from TLS record)
7043 portions. This adds all GCM ciphersuites supported by RFC5288 and
7044 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7045 add a special AESGCM string for GCM only.
7049 * Expand range of ctrls for AES GCM. Permit setting invocation
7050 field on decrypt and retrieval of invocation field only on encrypt.
7054 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7055 As required by RFC5289 these ciphersuites cannot be used if for
7056 versions of TLS earlier than 1.2.
7060 * For FIPS capable OpenSSL interpret a NULL default public key method
7061 as unset and return the appropriate default but do *not* set the default.
7062 This means we can return the appropriate method in applications that
7063 switch between FIPS and non-FIPS modes.
7067 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7068 ENGINE is used then we cannot handle that in the FIPS module so we
7069 keep original code iff non-FIPS operations are allowed.
7073 * Add -attime option to openssl utilities.
7075 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7077 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7081 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7082 FIPS EC methods unconditionally for now.
7086 * New build option no-ec2m to disable characteristic 2 code.
7090 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7091 all cases can be covered as some introduce binary incompatibilities.
7095 * Redirect RSA operations to FIPS module including keygen,
7096 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7100 * Add similar low level API blocking to ciphers.
7104 * Low level digest APIs are not approved in FIPS mode: any attempt
7105 to use these will cause a fatal error. Applications that *really* want
7106 to use them can use the `private_*` version instead.
7110 * Redirect cipher operations to FIPS module for FIPS builds.
7114 * Redirect digest operations to FIPS module for FIPS builds.
7118 * Update build system to add "fips" flag which will link in fipscanister.o
7119 for static and shared library builds embedding a signature if needed.
7123 * Output TLS supported curves in preference order instead of numerical
7124 order. This is currently hardcoded for the highest order curves first.
7125 This should be configurable so applications can judge speed vs strength.
7129 * Add TLS v1.2 server support for client authentication.
7133 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7138 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7139 FIPS modules versions.
7143 * Add TLS v1.2 client side support for client authentication. Keep cache
7144 of handshake records longer as we don't know the hash algorithm to use
7145 until after the certificate request message is received.
7149 * Initial TLS v1.2 client support. Add a default signature algorithms
7150 extension including all the algorithms we support. Parse new signature
7151 format in client key exchange. Relax some ECC signing restrictions for
7152 TLS v1.2 as indicated in RFC5246.
7156 * Add server support for TLS v1.2 signature algorithms extension. Switch
7157 to new signature format when needed using client digest preference.
7158 All server ciphersuites should now work correctly in TLS v1.2. No client
7159 support yet and no support for client certificates.
7163 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7164 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7165 ciphersuites. At present only RSA key exchange ciphersuites work with
7166 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7167 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7168 and version checking.
7172 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7173 with this defined it will not be affected by any changes to ssl internal
7174 structures. Add several utility functions to allow openssl application
7175 to work with OPENSSL_NO_SSL_INTERN defined.
7179 * A long standing patch to add support for SRP from EdelWeb (Peter
7180 Sylvester and Christophe Renou) was integrated.
7181 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7182 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7185 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7189 * Permit abbreviated handshakes when renegotiating using the function
7190 SSL_renegotiate_abbreviated().
7192 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7194 * Add call to ENGINE_register_all_complete() to
7195 ENGINE_load_builtin_engines(), so some implementations get used
7196 automatically instead of needing explicit application support.
7200 * Add support for TLS key exporter as described in RFC5705.
7202 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7204 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7205 a few changes are required:
7207 Add SSL_OP_NO_TLSv1_1 flag.
7208 Add TLSv1_1 methods.
7209 Update version checking logic to handle version 1.1.
7210 Add explicit IV handling (ported from DTLS code).
7211 Add command line options to s_client/s_server.
7218 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7220 * X509_ATTRIBUTE memory leak
7222 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7223 memory. This structure is used by the PKCS#7 and CMS routines so any
7224 application which reads PKCS#7 or CMS data from untrusted sources is
7225 affected. SSL/TLS is not affected.
7227 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7233 * Race condition handling PSK identify hint
7235 If PSK identity hints are received by a multi-threaded client then
7236 the values are wrongly updated in the parent SSL_CTX structure. This can
7237 result in a race condition potentially leading to a double free of the
7243 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7245 * Malformed ECParameters causes infinite loop
7247 When processing an ECParameters structure OpenSSL enters an infinite loop
7248 if the curve specified is over a specially malformed binary polynomial
7251 This can be used to perform denial of service against any
7252 system which processes public keys, certificate requests or
7253 certificates. This includes TLS clients and TLS servers with
7254 client authentication enabled.
7256 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7261 * Exploitable out-of-bounds read in X509_cmp_time
7263 X509_cmp_time does not properly check the length of the ASN1_TIME
7264 string and can read a few bytes out of bounds. In addition,
7265 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7268 An attacker can use this to craft malformed certificates and CRLs of
7269 various sizes and potentially cause a segmentation fault, resulting in
7270 a DoS on applications that verify certificates or CRLs. TLS clients
7271 that verify CRLs are affected. TLS clients and servers with client
7272 authentication enabled may be affected if they use custom verification
7275 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7276 independently by Hanno Böck.
7281 * PKCS7 crash with missing EnvelopedContent
7283 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7284 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7285 with missing content and trigger a NULL pointer dereference on parsing.
7287 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7288 structures from untrusted sources are affected. OpenSSL clients and
7289 servers are not affected.
7291 This issue was reported to OpenSSL by Michal Zalewski (Google).
7296 * CMS verify infinite loop with unknown hash function
7298 When verifying a signedData message the CMS code can enter an infinite loop
7299 if presented with an unknown hash function OID. This can be used to perform
7300 denial of service against any system which verifies signedData messages using
7302 This issue was reported to OpenSSL by Johannes Bauer.
7307 * Race condition handling NewSessionTicket
7309 If a NewSessionTicket is received by a multi-threaded client when attempting to
7310 reuse a previous ticket then a race condition can occur potentially leading to
7311 a double free of the ticket data.
7316 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7318 * Segmentation fault in ASN1_TYPE_cmp fix
7320 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7321 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7322 certificate signature algorithm consistency this can be used to crash any
7323 certificate verification operation and exploited in a DoS attack. Any
7324 application which performs certificate verification is vulnerable including
7325 OpenSSL clients and servers which enable client authentication.
7330 * ASN.1 structure reuse memory corruption fix
7332 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7333 memory corruption via an invalid write. Such reuse is and has been
7334 strongly discouraged and is believed to be rare.
7336 Applications that parse structures containing CHOICE or ANY DEFINED BY
7337 components may be affected. Certificate parsing (d2i_X509 and related
7338 functions) are however not affected. OpenSSL clients and servers are
7344 * PKCS7 NULL pointer dereferences fix
7346 The PKCS#7 parsing code does not handle missing outer ContentInfo
7347 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7348 missing content and trigger a NULL pointer dereference on parsing.
7350 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7351 otherwise parse PKCS#7 structures from untrusted sources are
7352 affected. OpenSSL clients and servers are not affected.
7354 This issue was reported to OpenSSL by Michal Zalewski (Google).
7359 * DoS via reachable assert in SSLv2 servers fix
7361 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7362 servers that both support SSLv2 and enable export cipher suites by sending
7363 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7365 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7366 (OpenSSL development team).
7371 * Use After Free following d2i_ECPrivatekey error fix
7373 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7374 could cause a use after free condition. This, in turn, could cause a double
7375 free in several private key parsing functions (such as d2i_PrivateKey
7376 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7377 for applications that receive EC private keys from untrusted
7378 sources. This scenario is considered rare.
7380 This issue was discovered by the BoringSSL project and fixed in their
7386 * X509_to_X509_REQ NULL pointer deref fix
7388 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7389 the certificate key is invalid. This function is rarely used in practice.
7391 This issue was discovered by Brian Carpenter.
7396 * Removed the export ciphers from the DEFAULT ciphers
7400 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7402 * Build fixes for the Windows and OpenVMS platforms
7404 *Matt Caswell and Richard Levitte*
7406 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7408 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7409 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7410 dereference. This could lead to a Denial Of Service attack. Thanks to
7411 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7416 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7417 dtls1_buffer_record function under certain conditions. In particular this
7418 could occur if an attacker sent repeated DTLS records with the same
7419 sequence number but for the next epoch. The memory leak could be exploited
7420 by an attacker in a Denial of Service attack through memory exhaustion.
7421 Thanks to Chris Mueller for reporting this issue.
7426 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7427 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7428 method would be set to NULL which could later result in a NULL pointer
7429 dereference. Thanks to Frank Schmirler for reporting this issue.
7434 * Abort handshake if server key exchange message is omitted for ephemeral
7437 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7438 reporting this issue.
7443 * Remove non-export ephemeral RSA code on client and server. This code
7444 violated the TLS standard by allowing the use of temporary RSA keys in
7445 non-export ciphersuites and could be used by a server to effectively
7446 downgrade the RSA key length used to a value smaller than the server
7447 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7448 INRIA or reporting this issue.
7453 * Fixed issue where DH client certificates are accepted without verification.
7454 An OpenSSL server will accept a DH certificate for client authentication
7455 without the certificate verify message. This effectively allows a client to
7456 authenticate without the use of a private key. This only affects servers
7457 which trust a client certificate authority which issues certificates
7458 containing DH keys: these are extremely rare and hardly ever encountered.
7459 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7465 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7466 results on some platforms, including x86_64. This bug occurs at random
7467 with a very low probability, and is not known to be exploitable in any
7468 way, though its exact impact is difficult to determine. Thanks to Pieter
7469 Wuille (Blockstream) who reported this issue and also suggested an initial
7470 fix. Further analysis was conducted by the OpenSSL development team and
7471 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7472 the OpenSSL core team.
7477 * Fix various certificate fingerprint issues.
7479 By using non-DER or invalid encodings outside the signed portion of a
7480 certificate the fingerprint can be changed without breaking the signature.
7481 Although no details of the signed portion of the certificate can be changed
7482 this can cause problems with some applications: e.g. those using the
7483 certificate fingerprint for blacklists.
7485 1. Reject signatures with non zero unused bits.
7487 If the BIT STRING containing the signature has non zero unused bits reject
7488 the signature. All current signature algorithms require zero unused bits.
7490 2. Check certificate algorithm consistency.
7492 Check the AlgorithmIdentifier inside TBS matches the one in the
7493 certificate signature. NB: this will result in signature failure
7494 errors for some broken certificates.
7496 Thanks to Konrad Kraszewski from Google for reporting this issue.
7498 3. Check DSA/ECDSA signatures use DER.
7500 Reencode DSA/ECDSA signatures and compare with the original received
7501 signature. Return an error if there is a mismatch.
7503 This will reject various cases including garbage after signature
7504 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7505 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7506 (negative or with leading zeroes).
7508 Further analysis was conducted and fixes were developed by Stephen Henson
7509 of the OpenSSL core team.
7515 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7517 * Session Ticket Memory Leak.
7519 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7520 integrity of that ticket is first verified. In the event of a session
7521 ticket integrity check failing, OpenSSL will fail to free memory
7522 causing a memory leak. By sending a large number of invalid session
7523 tickets an attacker could exploit this issue in a Denial Of Service
7529 * Build option no-ssl3 is incomplete.
7531 When OpenSSL is configured with "no-ssl3" as a build option, servers
7532 could accept and complete a SSL 3.0 handshake, and clients could be
7533 configured to send them.
7536 *Akamai and the OpenSSL team*
7538 * Add support for TLS_FALLBACK_SCSV.
7539 Client applications doing fallback retries should call
7540 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7543 *Adam Langley, Bodo Moeller*
7545 * Add additional DigestInfo checks.
7547 Reencode DigestInto in DER and check against the original when
7548 verifying RSA signature: this will reject any improperly encoded
7549 DigestInfo structures.
7551 Note: this is a precautionary measure and no attacks are currently known.
7555 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7557 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7558 to a denial of service attack. A malicious server can crash the client
7559 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7560 ciphersuite and sending carefully crafted handshake messages.
7562 Thanks to Felix Gröbert (Google) for discovering and researching this
7568 * By sending carefully crafted DTLS packets an attacker could cause openssl
7569 to leak memory. This can be exploited through a Denial of Service attack.
7570 Thanks to Adam Langley for discovering and researching this issue.
7575 * An attacker can force openssl to consume large amounts of memory whilst
7576 processing DTLS handshake messages. This can be exploited through a
7577 Denial of Service attack.
7578 Thanks to Adam Langley for discovering and researching this issue.
7583 * An attacker can force an error condition which causes openssl to crash
7584 whilst processing DTLS packets due to memory being freed twice. This
7585 can be exploited through a Denial of Service attack.
7586 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7592 * If a multithreaded client connects to a malicious server using a resumed
7593 session and the server sends an ec point format extension it could write
7594 up to 255 bytes to freed memory.
7596 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7602 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7603 X509_name_oneline, X509_name_print_ex et al. to leak some information
7604 from the stack. Applications may be affected if they echo pretty printing
7605 output to the attacker.
7607 Thanks to Ivan Fratric (Google) for discovering this issue.
7610 *Emilia Käsper, and Steve Henson*
7612 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7613 for corner cases. (Certain input points at infinity could lead to
7614 bogus results, with non-infinity inputs mapped to infinity too.)
7618 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7620 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7621 handshake can force the use of weak keying material in OpenSSL
7622 SSL/TLS clients and servers.
7624 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7625 researching this issue. ([CVE-2014-0224])
7627 *KIKUCHI Masashi, Steve Henson*
7629 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7630 OpenSSL DTLS client the code can be made to recurse eventually crashing
7633 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7636 *Imre Rad, Steve Henson*
7638 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7639 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7640 client or server. This is potentially exploitable to run arbitrary
7641 code on a vulnerable client or server.
7643 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7645 *Jüri Aedla, Steve Henson*
7647 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7648 are subject to a denial of service attack.
7650 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7651 this issue. ([CVE-2014-3470])
7653 *Felix Gröbert, Ivan Fratric, Steve Henson*
7655 * Harmonize version and its documentation. -f flag is used to display
7658 *mancha <mancha1@zoho.com>*
7660 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7661 in i2d_ECPrivateKey.
7663 *mancha <mancha1@zoho.com>*
7665 * Fix some double frees. These are not thought to be exploitable.
7667 *mancha <mancha1@zoho.com>*
7669 * Fix for the attack described in the paper "Recovering OpenSSL
7670 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7671 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7672 <http://eprint.iacr.org/2014/140>
7674 Thanks to Yuval Yarom and Naomi Benger for discovering this
7675 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7677 *Yuval Yarom and Naomi Benger*
7679 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7681 * Keep original DTLS digest and encryption contexts in retransmission
7682 structures so we can use the previous session parameters if they need
7683 to be resent. ([CVE-2013-6450])
7687 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7688 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7689 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7690 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7691 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7692 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7694 *Rob Stradling, Adam Langley*
7696 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7698 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7700 This addresses the flaw in CBC record processing discovered by
7701 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7702 at: <http://www.isg.rhul.ac.uk/tls/>
7704 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7705 Security Group at Royal Holloway, University of London
7706 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7707 Emilia Käsper for the initial patch.
7710 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7712 * Return an error when checking OCSP signatures when key is NULL.
7713 This fixes a DoS attack. ([CVE-2013-0166])
7717 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7718 the right response is stapled. Also change SSL_get_certificate()
7719 so it returns the certificate actually sent.
7720 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7721 (This is a backport)
7723 *Rob Stradling <rob.stradling@comodo.com>*
7725 * Fix possible deadlock when decoding public keys.
7729 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7731 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7734 * Sanity check record length before skipping explicit IV in DTLS
7737 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7738 fuzzing as a service testing platform.
7743 * Initialise tkeylen properly when encrypting CMS messages.
7744 Thanks to Solar Designer of Openwall for reporting this issue.
7748 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7750 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7751 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7752 in CRYPTO_realloc_clean.
7754 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7755 issue and to Adam Langley <agl@chromium.org> for fixing it.
7758 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7760 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7762 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7763 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7764 content decryption and always return the same error. Note: this attack
7765 needs on average 2^20 messages so it only affects automated senders. The
7766 old behaviour can be re-enabled in the CMS code by setting the
7767 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7768 an MMA defence is not necessary.
7769 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7770 this issue. ([CVE-2012-0884])
7774 * Fix CVE-2011-4619: make sure we really are receiving a
7775 client hello before rejecting multiple SGC restarts. Thanks to
7776 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7780 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7782 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7783 Thanks to Antonio Martin, Enterprise Secure Access Research and
7784 Development, Cisco Systems, Inc. for discovering this bug and
7785 preparing a fix. ([CVE-2012-0050])
7789 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7791 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7792 of the Vaudenay padding oracle attack on CBC mode encryption
7793 which enables an efficient plaintext recovery attack against
7794 the OpenSSL implementation of DTLS. Their attack exploits timing
7795 differences arising during decryption processing. A research
7796 paper describing this attack can be found at:
7797 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7798 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7799 Security Group at Royal Holloway, University of London
7800 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7801 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7802 for preparing the fix. ([CVE-2011-4108])
7804 *Robin Seggelmann, Michael Tuexen*
7806 * Clear bytes used for block padding of SSL 3.0 records.
7809 *Adam Langley (Google)*
7811 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7812 Kadianakis <desnacked@gmail.com> for discovering this issue and
7813 Adam Langley for preparing the fix. ([CVE-2011-4619])
7815 *Adam Langley (Google)*
7817 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7819 *Andrey Kulikov <amdeich@gmail.com>*
7821 * Prevent malformed RFC3779 data triggering an assertion failure.
7822 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7823 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7825 *Rob Austein <sra@hactrn.net>*
7827 * Improved PRNG seeding for VOS.
7829 *Paul Green <Paul.Green@stratus.com>*
7831 * Fix ssl_ciph.c set-up race.
7833 *Adam Langley (Google)*
7835 * Fix spurious failures in ecdsatest.c.
7837 *Emilia Käsper (Google)*
7839 * Fix the BIO_f_buffer() implementation (which was mixing different
7840 interpretations of the `..._len` fields).
7842 *Adam Langley (Google)*
7844 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7845 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7846 threads won't reuse the same blinding coefficients.
7848 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7849 lock to call BN_BLINDING_invert_ex, and avoids one use of
7850 BN_BLINDING_update for each BN_BLINDING structure (previously,
7851 the last update always remained unused).
7853 *Emilia Käsper (Google)*
7855 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7857 *Bob Buckholz (Google)*
7859 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7861 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7862 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7864 *Kaspar Brand <ossl@velox.ch>*
7866 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7867 for multi-threaded use of ECDH. ([CVE-2011-3210])
7869 *Adam Langley (Google)*
7871 * Fix x509_name_ex_d2i memory leak on bad inputs.
7875 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7876 signature public key algorithm by using OID xref utilities instead.
7877 Before this you could only use some ECC ciphersuites with SHA1 only.
7881 * Add protection against ECDSA timing attacks as mentioned in the paper
7882 by Billy Bob Brumley and Nicola Tuveri, see:
7883 <http://eprint.iacr.org/2011/232.pdf>
7885 *Billy Bob Brumley and Nicola Tuveri*
7887 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7889 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7891 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7893 * Fix bug in string printing code: if *any* escaping is enabled we must
7894 escape the escape character (backslash) or the resulting string is
7899 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7901 * Disable code workaround for ancient and obsolete Netscape browsers
7902 and servers: an attacker can use it in a ciphersuite downgrade attack.
7903 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7907 * Fixed J-PAKE implementation error, originally discovered by
7908 Sebastien Martini, further info and confirmation from Stefan
7909 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7913 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7915 * Fix extension code to avoid race conditions which can result in a buffer
7916 overrun vulnerability: resumed sessions must not be modified as they can
7917 be shared by multiple threads. CVE-2010-3864
7921 * Fix WIN32 build system to correctly link an ENGINE directory into
7926 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7928 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7931 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7933 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7935 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7936 context. The operation can be customised via the ctrl mechanism in
7937 case ENGINEs want to include additional functionality.
7941 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7945 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7946 output hashes compatible with older versions of OpenSSL.
7948 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7950 * Fix compression algorithm handling: if resuming a session use the
7951 compression algorithm of the resumed session instead of determining
7952 it from client hello again. Don't allow server to change algorithm.
7956 * Add load_crls() function to apps tidying load_certs() too. Add option
7957 to verify utility to allow additional CRLs to be included.
7961 * Update OCSP request code to permit adding custom headers to the request:
7962 some responders need this.
7966 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7969 *Julia Lawall <julia@diku.dk>*
7971 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7972 needlessly dereferenced structures, used obsolete functions and
7973 didn't handle all updated verify codes correctly.
7977 * Disable MD2 in the default configuration.
7981 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7982 indicate the initial BIO being pushed or popped. This makes it possible
7983 to determine whether the BIO is the one explicitly called or as a result
7984 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7985 it handles reference counts correctly and doesn't zero out the I/O bio
7986 when it is not being explicitly popped. WARNING: applications which
7987 included workarounds for the old buggy behaviour will need to be modified
7988 or they could free up already freed BIOs.
7992 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7993 renaming to all platforms (within the 0.9.8 branch, this was
7994 done conditionally on Netware platforms to avoid a name clash).
7996 *Guenter <lists@gknw.net>*
7998 * Add ECDHE and PSK support to DTLS.
8000 *Michael Tuexen <tuexen@fh-muenster.de>*
8002 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8007 * Add "missing" function EVP_MD_flags() (without this the only way to
8008 retrieve a digest flags is by accessing the structure directly. Update
8009 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8010 or cipher is registered as in the "from" argument. Print out all
8011 registered digests in the dgst usage message instead of manually
8012 attempting to work them out.
8016 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8017 this allows the use of compression and extensions. Change default cipher
8018 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8019 by default unless an application cipher string requests it.
8023 * Alter match criteria in PKCS12_parse(). It used to try to use local
8024 key ids to find matching certificates and keys but some PKCS#12 files
8025 don't follow the (somewhat unwritten) rules and this strategy fails.
8026 Now just gather all certificates together and the first private key
8027 then look for the first certificate that matches the key.
8031 * Support use of registered digest and cipher names for dgst and cipher
8032 commands instead of having to add each one as a special case. So now
8039 openssl dgst -sha256 foo
8041 and this works for ENGINE based algorithms too.
8045 * Update Gost ENGINE to support parameter files.
8047 *Victor B. Wagner <vitus@cryptocom.ru>*
8049 * Support GeneralizedTime in ca utility.
8051 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8053 * Enhance the hash format used for certificate directory links. The new
8054 form uses the canonical encoding (meaning equivalent names will work
8055 even if they aren't identical) and uses SHA1 instead of MD5. This form
8056 is incompatible with the older format and as a result c_rehash should
8057 be used to rebuild symbolic links.
8061 * Make PKCS#8 the default write format for private keys, replacing the
8062 traditional format. This form is standardised, more secure and doesn't
8063 include an implicit MD5 dependency.
8067 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8068 committed to OpenSSL should pass this lot as a minimum.
8072 * Add session ticket override functionality for use by EAP-FAST.
8074 *Jouni Malinen <j@w1.fi>*
8076 * Modify HMAC functions to return a value. Since these can be implemented
8077 in an ENGINE errors can occur.
8081 * Type-checked OBJ_bsearch_ex.
8085 * Type-checked OBJ_bsearch. Also some constification necessitated
8086 by type-checking. Still to come: TXT_DB, bsearch(?),
8087 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8092 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8093 seconds to a tm structure directly, instead of going through OS
8094 specific date routines. This avoids any issues with OS routines such
8095 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8096 and X509_time_adj_ex() to cover the extended range. The existing
8097 X509_time_adj() is still usable and will no longer have any date issues.
8101 * Delta CRL support. New use deltas option which will attempt to locate
8102 and search any appropriate delta CRLs available.
8104 This work was sponsored by Google.
8108 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8109 code and add additional score elements. Validate alternate CRL paths
8110 as part of the CRL checking and indicate a new error "CRL path validation
8111 error" in this case. Applications wanting additional details can use
8112 the verify callback and check the new "parent" field. If this is not
8113 NULL CRL path validation is taking place. Existing applications won't
8114 see this because it requires extended CRL support which is off by
8117 This work was sponsored by Google.
8121 * Support for freshest CRL extension.
8123 This work was sponsored by Google.
8127 * Initial indirect CRL support. Currently only supported in the CRLs
8128 passed directly and not via lookup. Process certificate issuer
8129 CRL entry extension and lookup CRL entries by bother issuer name
8130 and serial number. Check and process CRL issuer entry in IDP extension.
8132 This work was sponsored by Google.
8136 * Add support for distinct certificate and CRL paths. The CRL issuer
8137 certificate is validated separately in this case. Only enabled if
8138 an extended CRL support flag is set: this flag will enable additional
8139 CRL functionality in future.
8141 This work was sponsored by Google.
8145 * Add support for policy mappings extension.
8147 This work was sponsored by Google.
8151 * Fixes to pathlength constraint, self issued certificate handling,
8152 policy processing to align with RFC3280 and PKITS tests.
8154 This work was sponsored by Google.
8158 * Support for name constraints certificate extension. DN, email, DNS
8159 and URI types are currently supported.
8161 This work was sponsored by Google.
8165 * To cater for systems that provide a pointer-based thread ID rather
8166 than numeric, deprecate the current numeric thread ID mechanism and
8167 replace it with a structure and associated callback type. This
8168 mechanism allows a numeric "hash" to be extracted from a thread ID in
8169 either case, and on platforms where pointers are larger than 'long',
8170 mixing is done to help ensure the numeric 'hash' is usable even if it
8171 can't be guaranteed unique. The default mechanism is to use "&errno"
8172 as a pointer-based thread ID to distinguish between threads.
8174 Applications that want to provide their own thread IDs should now use
8175 CRYPTO_THREADID_set_callback() to register a callback that will call
8176 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8178 Note that ERR_remove_state() is now deprecated, because it is tied
8179 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8180 to free the current thread's error state should be replaced by
8181 ERR_remove_thread_state(NULL).
8183 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8184 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8185 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8186 application was previously providing a numeric thread callback that
8187 was inappropriate for distinguishing threads, then uniqueness might
8188 have been obtained with &errno that happened immediately in the
8189 intermediate development versions of OpenSSL; this is no longer the
8190 case, the numeric thread callback will now override the automatic use
8193 *Geoff Thorpe, with help from Bodo Moeller*
8195 * Initial support for different CRL issuing certificates. This covers a
8196 simple case where the self issued certificates in the chain exist and
8197 the real CRL issuer is higher in the existing chain.
8199 This work was sponsored by Google.
8203 * Removed effectively defunct crypto/store from the build.
8207 * Revamp of STACK to provide stronger type-checking. Still to come:
8208 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8209 ASN1_STRING, CONF_VALUE.
8213 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8214 RAM on SSL connections. This option can save about 34k per idle SSL.
8218 * Revamp of LHASH to provide stronger type-checking. Still to come:
8219 STACK, TXT_DB, bsearch, qsort.
8223 * Initial support for Cryptographic Message Syntax (aka CMS) based
8224 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8225 support for data, signedData, compressedData, digestedData and
8226 encryptedData, envelopedData types included. Scripts to check against
8227 RFC4134 examples draft and interop and consistency checks of many
8228 content types and variants.
8232 * Add options to enc utility to support use of zlib compression BIO.
8236 * Extend mk1mf to support importing of options and assembly language
8237 files from Configure script, currently only included in VC-WIN32.
8238 The assembly language rules can now optionally generate the source
8239 files from the associated perl scripts.
8243 * Implement remaining functionality needed to support GOST ciphersuites.
8244 Interop testing has been performed using CryptoPro implementations.
8246 *Victor B. Wagner <vitus@cryptocom.ru>*
8248 * s390x assembler pack.
8252 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8257 * Implement Opaque PRF Input TLS extension as specified in
8258 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8259 official specification yet and no extension type assignment by
8260 IANA exists, this extension (for now) will have to be explicitly
8261 enabled when building OpenSSL by providing the extension number
8262 to use. For example, specify an option
8264 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8266 to the "config" or "Configure" script to enable the extension,
8267 assuming extension number 0x9527 (which is a completely arbitrary
8268 and unofficial assignment based on the MD5 hash of the Internet
8269 Draft). Note that by doing so, you potentially lose
8270 interoperability with other TLS implementations since these might
8271 be using the same extension number for other purposes.
8273 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8274 opaque PRF input value to use in the handshake. This will create
8275 an internal copy of the length-'len' string at 'src', and will
8276 return non-zero for success.
8278 To get more control and flexibility, provide a callback function
8281 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8282 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8286 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8289 Callback function 'cb' will be called in handshakes, and is
8290 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8291 Argument 'arg' is for application purposes (the value as given to
8292 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8293 be provided to the callback function). The callback function
8294 has to return non-zero to report success: usually 1 to use opaque
8295 PRF input just if possible, or 2 to enforce use of the opaque PRF
8296 input. In the latter case, the library will abort the handshake
8297 if opaque PRF input is not successfully negotiated.
8299 Arguments 'peerinput' and 'len' given to the callback function
8300 will always be NULL and 0 in the case of a client. A server will
8301 see the client's opaque PRF input through these variables if
8302 available (NULL and 0 otherwise). Note that if the server
8303 provides an opaque PRF input, the length must be the same as the
8304 length of the client's opaque PRF input.
8306 Note that the callback function will only be called when creating
8307 a new session (session resumption can resume whatever was
8308 previously negotiated), and will not be called in SSL 2.0
8309 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8310 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8311 for applications that need to enforce opaque PRF input.
8315 * Update ssl code to support digests other than SHA1+MD5 for handshake
8318 *Victor B. Wagner <vitus@cryptocom.ru>*
8320 * Add RFC4507 support to OpenSSL. This includes the corrections in
8321 RFC4507bis. The encrypted ticket format is an encrypted encoded
8322 SSL_SESSION structure, that way new session features are automatically
8325 If a client application caches session in an SSL_SESSION structure
8326 support is transparent because tickets are now stored in the encoded
8329 The SSL_CTX structure automatically generates keys for ticket
8330 protection in servers so again support should be possible
8331 with no application modification.
8333 If a client or server wishes to disable RFC4507 support then the option
8334 SSL_OP_NO_TICKET can be set.
8336 Add a TLS extension debugging callback to allow the contents of any client
8337 or server extensions to be examined.
8339 This work was sponsored by Google.
8343 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8344 OpenSSL should now compile cleanly on gcc 4.2
8346 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8348 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8349 support including streaming MAC support: this is required for GOST
8350 ciphersuite support.
8352 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8354 * Add option -stream to use PKCS#7 streaming in smime utility. New
8355 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8356 to output in BER and PEM format.
8360 * Experimental support for use of HMAC via EVP_PKEY interface. This
8361 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8362 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8363 ENGINE support for HMAC keys which are unextractable. New -mac and
8364 -macopt options to dgst utility.
8368 * New option -sigopt to dgst utility. Update dgst to use
8369 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8370 alternative signing parameters such as X9.31 or PSS in the dgst
8375 * Change ssl_cipher_apply_rule(), the internal function that does
8376 the work each time a ciphersuite string requests enabling
8377 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8378 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8379 the order of disabled ciphersuites such that those ciphersuites
8380 that most recently went from enabled to disabled not only stay
8381 in order with respect to each other, but also have higher priority
8382 than other disabled ciphersuites the next time ciphersuites are
8385 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8386 the same ciphersuites as with "HIGH" alone, but in a specific
8387 order where the PSK ciphersuites come first (since they are the
8388 most recently disabled ciphersuites when "HIGH" is parsed).
8390 Also, change ssl_create_cipher_list() (using this new
8391 functionality) such that between otherwise identical
8392 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8397 * Change ssl_create_cipher_list() so that it automatically
8398 arranges the ciphersuites in reasonable order before starting
8399 to process the rule string. Thus, the definition for "DEFAULT"
8400 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8401 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8402 This makes it much easier to arrive at a reasonable default order
8403 in applications for which anonymous ciphers are OK (meaning
8404 that you can't actually use DEFAULT).
8406 *Bodo Moeller; suggested by Victor Duchovni*
8408 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8409 processing) into multiple integers instead of setting
8410 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8411 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8412 (These masks as well as the individual bit definitions are hidden
8413 away into the non-exported interface ssl/ssl_locl.h, so this
8414 change to the definition of the SSL_CIPHER structure shouldn't
8415 affect applications.) This give us more bits for each of these
8416 categories, so there is no longer a need to coagulate AES128 and
8417 AES256 into a single algorithm bit, and to coagulate Camellia128
8418 and Camellia256 into a single algorithm bit, which has led to all
8421 Thus, among other things, the kludge introduced in 0.9.7m and
8422 0.9.8e for masking out AES256 independently of AES128 or masking
8423 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8425 With the change, we also introduce new ciphersuite aliases that
8426 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8431 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8432 Use the leftmost N bytes of the signature input if the input is
8433 larger than the prime q (with N being the size in bytes of q).
8437 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8438 it yet and it is largely untested.
8442 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8446 * Initial incomplete changes to avoid need for function casts in OpenSSL
8447 some compilers (gcc 4.2 and later) reject their use. Safestack is
8448 reimplemented. Update ASN1 to avoid use of legacy functions.
8452 * Win32/64 targets are linked with Winsock2.
8456 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8457 to external functions. This can be used to increase CRL handling
8458 efficiency especially when CRLs are very large by (for example) storing
8459 the CRL revoked certificates in a database.
8463 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8464 new CRLs added to a directory can be used. New command line option
8465 -verify_return_error to s_client and s_server. This causes real errors
8466 to be returned by the verify callback instead of carrying on no matter
8467 what. This reflects the way a "real world" verify callback would behave.
8471 * GOST engine, supporting several GOST algorithms and public key formats.
8472 Kindly donated by Cryptocom.
8476 * Partial support for Issuing Distribution Point CRL extension. CRLs
8477 partitioned by DP are handled but no indirect CRL or reason partitioning
8478 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8479 selected via a scoring technique which handles IDP and AKID in CRLs.
8483 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8484 will ultimately be used for all verify operations: this will remove the
8485 X509_STORE dependency on certificate verification and allow alternative
8486 lookup methods. X509_STORE based implementations of these two callbacks.
8490 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8491 Modify get_crl() to find a valid (unexpired) CRL if possible.
8495 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8496 this would be called X509_CRL_cmp() but that name is already used by
8497 a function that just compares CRL issuer names. Cache several CRL
8498 extensions in X509_CRL structure and cache CRLDP in X509.
8502 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8503 this maps equivalent X509_NAME structures into a consistent structure.
8504 Name comparison can then be performed rapidly using memcmp().
8508 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8513 * Allow digests to supply their own micalg string for S/MIME type using
8514 the ctrl EVP_MD_CTRL_MICALG.
8518 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8519 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8520 ctrl. It can then customise the structure before and/or after signing
8525 * New function OBJ_add_sigid() to allow application defined signature OIDs
8526 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8527 to free up any added signature OIDs.
8531 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8532 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8533 digest and cipher tables. New options added to openssl utility:
8534 list-message-digest-algorithms and list-cipher-algorithms.
8538 * Change the array representation of binary polynomials: the list
8539 of degrees of non-zero coefficients is now terminated with -1.
8540 Previously it was terminated with 0, which was also part of the
8541 value; thus, the array representation was not applicable to
8542 polynomials where t^0 has coefficient zero. This change makes
8543 the array representation useful in a more general context.
8547 * Various modifications and fixes to SSL/TLS cipher string
8548 handling. For ECC, the code now distinguishes between fixed ECDH
8549 with RSA certificates on the one hand and with ECDSA certificates
8550 on the other hand, since these are separate ciphersuites. The
8551 unused code for Fortezza ciphersuites has been removed.
8553 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8554 (not "ECDHE"). For consistency with the code for DH
8555 certificates, use of ECDH certificates is now considered ECDH
8556 authentication, not RSA or ECDSA authentication (the latter is
8557 merely the CA's signing algorithm and not actively used in the
8560 The temporary ciphersuite alias "ECCdraft" is no longer
8561 available, and ECC ciphersuites are no longer excluded from "ALL"
8562 and "DEFAULT". The following aliases now exist for RFC 4492
8563 ciphersuites, most of these by analogy with the DH case:
8565 kECDHr - ECDH cert, signed with RSA
8566 kECDHe - ECDH cert, signed with ECDSA
8567 kECDH - ECDH cert (signed with either RSA or ECDSA)
8568 kEECDH - ephemeral ECDH
8569 ECDH - ECDH cert or ephemeral ECDH
8575 AECDH - anonymous ECDH
8576 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8580 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8581 Use correct micalg parameters depending on digest(s) in signed message.
8585 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8586 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8590 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8591 an engine to register a method. Add ENGINE lookups for methods and
8592 functional reference processing.
8596 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8597 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8602 * New -resign option to smime utility. This adds one or more signers
8603 to an existing PKCS#7 signedData structure. Also -md option to use an
8604 alternative message digest algorithm for signing.
8608 * Tidy up PKCS#7 routines and add new functions to make it easier to
8609 create PKCS7 structures containing multiple signers. Update smime
8610 application to support multiple signers.
8614 * New -macalg option to pkcs12 utility to allow setting of an alternative
8619 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8620 Reorganize PBE internals to lookup from a static table using NIDs,
8621 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8622 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8623 PRF which will be automatically used with PBES2.
8627 * Replace the algorithm specific calls to generate keys in "req" with the
8632 * Update PKCS#7 enveloped data routines to use new API. This is now
8633 supported by any public key method supporting the encrypt operation. A
8634 ctrl is added to allow the public key algorithm to examine or modify
8635 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8640 * Add a ctrl to asn1 method to allow a public key algorithm to express
8641 a default digest type to use. In most cases this will be SHA1 but some
8642 algorithms (such as GOST) need to specify an alternative digest. The
8643 return value indicates how strong the preference is 1 means optional and
8644 2 is mandatory (that is it is the only supported type). Modify
8645 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8646 use the default md. Update openssl utilities to use the default digest
8647 type for signing if it is not explicitly indicated.
8651 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8652 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8653 signing method from the key type. This effectively removes the link
8654 between digests and public key types.
8658 * Add an OID cross reference table and utility functions. Its purpose is to
8659 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8660 rsaEncryption. This will allow some of the algorithm specific hackery
8661 needed to use the correct OID to be removed.
8665 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8666 structures for PKCS7_sign(). They are now set up by the relevant public
8671 * Add provisional EC pkey method with support for ECDSA and ECDH.
8675 * Add support for key derivation (agreement) in the API, DH method and
8680 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8681 public and private key formats. As a side effect these add additional
8682 command line functionality not previously available: DSA signatures can be
8683 generated and verified using pkeyutl and DH key support and generation in
8690 *Oliver Tappe <zooey@hirschkaefer.de>*
8692 * New make target "install_html_docs" installs HTML renditions of the
8695 *Oliver Tappe <zooey@hirschkaefer.de>*
8697 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8698 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8699 support key and parameter generation and add initial key generation
8700 functionality for RSA.
8704 * Add functions for main EVP_PKEY_method operations. The undocumented
8705 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8706 `EVP_PKEY_{encrypt,decrypt}_old`.
8710 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8711 key API, doesn't do much yet.
8715 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8716 public key algorithms. New option to openssl utility:
8717 "list-public-key-algorithms" to print out info.
8721 * Implement the Supported Elliptic Curves Extension for
8722 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8726 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8727 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8731 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8732 utilities such as rsa, dsa, dsaparam etc except they process any key
8737 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8738 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8739 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8744 * Initial support for pluggable public key ASN1.
8745 De-spaghettify the public key ASN1 handling. Move public and private
8746 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8747 algorithm specific handling to a single module within the relevant
8748 algorithm directory. Add functions to allow (near) opaque processing
8749 of public and private key structures.
8753 * Implement the Supported Point Formats Extension for
8754 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8758 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8759 for the psk identity [hint] and the psk callback functions to the
8760 SSL_SESSION, SSL and SSL_CTX structure.
8763 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8767 SSL_CTX_use_psk_identity_hint
8768 SSL_get_psk_identity_hint
8769 SSL_get_psk_identity
8770 SSL_use_psk_identity_hint
8772 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8774 * Add RFC 3161 compliant time stamp request creation, response generation
8775 and response verification functionality.
8777 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8779 * Add initial support for TLS extensions, specifically for the server_name
8780 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8781 have new members for a host name. The SSL data structure has an
8782 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8783 stored in that context to allow for session resumption, even after the
8784 SSL has been switched to a new SSL_CTX in reaction to a client's
8785 server_name extension.
8787 New functions (subject to change):
8789 SSL_get_servername()
8790 SSL_get_servername_type()
8793 New CTRL codes and macros (subject to change):
8795 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8796 - SSL_CTX_set_tlsext_servername_callback()
8797 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8798 - SSL_CTX_set_tlsext_servername_arg()
8799 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8801 openssl s_client has a new '-servername ...' option.
8803 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8804 '-key2 ...', '-servername_fatal' (subject to change). This allows
8805 testing the HostName extension for a specific single host name ('-cert'
8806 and '-key' remain fallbacks for handshakes without HostName
8807 negotiation). If the unrecognized_name alert has to be sent, this by
8808 default is a warning; it becomes fatal with the '-servername_fatal'
8811 *Peter Sylvester, Remy Allais, Christophe Renou*
8813 * Whirlpool hash implementation is added.
8817 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8818 bn(64,32). Because of instruction set limitations it doesn't have
8819 any negative impact on performance. This was done mostly in order
8820 to make it possible to share assembler modules, such as bn_mul_mont
8821 implementations, between 32- and 64-bit builds without hassle.
8825 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8826 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8831 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8832 dedicated Montgomery multiplication procedure, is introduced.
8833 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8834 "64-bit" performance on certain 32-bit targets.
8838 * New option SSL_OP_NO_COMP to disable use of compression selectively
8839 in SSL structures. New SSL ctrl to set maximum send fragment size.
8840 Save memory by setting the I/O buffer sizes dynamically instead of
8841 using the maximum available value.
8845 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8846 in addition to the text details.
8850 * Very, very preliminary EXPERIMENTAL support for printing of general
8851 ASN1 structures. This currently produces rather ugly output and doesn't
8852 handle several customised structures at all.
8856 * Integrated support for PVK file format and some related formats such
8857 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8858 these in the 'rsa' and 'dsa' utilities.
8862 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8866 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8867 place for the (very old) "NETSCAPE" format certificates which are now
8868 handled using new ASN1 code equivalents.
8872 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8873 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8874 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8878 * Modify CRL distribution points extension code to print out previously
8879 unsupported fields. Enhance extension setting code to allow setting of
8884 * Add print and set support for Issuing Distribution Point CRL extension.
8888 * Change 'Configure' script to enable Camellia by default.
8895 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8897 * When rejecting SSL/TLS records due to an incorrect version number, never
8898 update s->server with a new major version number. As of
8899 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8900 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8901 the previous behavior could result in a read attempt at NULL when
8902 receiving specific incorrect SSL/TLS records once record payload
8903 protection is active. ([CVE-2010-0740])
8905 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8907 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8908 could be crashed if the relevant tables were not present (e.g. chrooted).
8910 *Tomas Hoger <thoger@redhat.com>*
8912 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8914 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8916 *Martin Olsson, Neel Mehta*
8918 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8919 accommodate for stack sorting, always a write lock!).
8923 * On some versions of WIN32 Heap32Next is very slow. This can cause
8924 excessive delays in the RAND_poll(): over a minute. As a workaround
8925 include a time check in the inner Heap32Next loop too.
8929 * The code that handled flushing of data in SSL/TLS originally used the
8930 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8931 the problem outlined in PR#1949. The fix suggested there however can
8932 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8933 of Apache). So instead simplify the code to flush unconditionally.
8934 This should be fine since flushing with no data to flush is a no op.
8938 * Handle TLS versions 2.0 and later properly and correctly use the
8939 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8940 off ancient servers have a habit of sticking around for a while...
8944 * Modify compression code so it frees up structures without using the
8945 ex_data callbacks. This works around a problem where some applications
8946 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8947 restarting) then use compression (e.g. SSL with compression) later.
8948 This results in significant per-connection memory leaks and
8949 has caused some security issues including CVE-2008-1678 and
8954 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8955 change when encrypting or decrypting.
8959 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8960 connect and renegotiate with servers which do not support RI.
8961 Until RI is more widely deployed this option is enabled by default.
8965 * Add "missing" ssl ctrls to clear options and mode.
8969 * If client attempts to renegotiate and doesn't support RI respond with
8970 a no_renegotiation alert as required by RFC5746. Some renegotiating
8971 TLS clients will continue a connection gracefully when they receive
8972 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8973 waiting for a server hello which it will never receive. Now we treat a
8974 received no_renegotiation alert as a fatal error. This is because
8975 applications requesting a renegotiation might well expect it to succeed
8976 and would have no code in place to handle the server denying it so the
8977 only safe thing to do is to terminate the connection.
8981 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8982 peer supports secure renegotiation and 0 otherwise. Print out peer
8983 renegotiation support in s_client/s_server.
8987 * Replace the highly broken and deprecated SPKAC certification method with
8988 the updated NID creation version. This should correctly handle UTF8.
8992 * Implement RFC5746. Re-enable renegotiation but require the extension
8993 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8994 turns out to be a bad idea. It has been replaced by
8995 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8996 SSL_CTX_set_options(). This is really not recommended unless you
8997 know what you are doing.
8999 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9001 * Fixes to stateless session resumption handling. Use initial_ctx when
9002 issuing and attempting to decrypt tickets in case it has changed during
9003 servername handling. Use a non-zero length session ID when attempting
9004 stateless session resumption: this makes it possible to determine if
9005 a resumption has occurred immediately after receiving server hello
9006 (several places in OpenSSL subtly assume this) instead of later in
9011 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9012 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9013 fixes for a few places where the return code is not checked
9016 *Julia Lawall <julia@diku.dk>*
9018 * Add --strict-warnings option to Configure script to include devteam
9019 warnings in other configurations.
9023 * Add support for --libdir option and LIBDIR variable in makefiles. This
9024 makes it possible to install openssl libraries in locations which
9025 have names other than "lib", for example "/usr/lib64" which some
9028 *Steve Henson, based on patch from Jeremy Utley*
9030 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9031 X690 8.9.12 and can produce some misleading textual output of OIDs.
9033 *Steve Henson, reported by Dan Kaminsky*
9035 * Delete MD2 from algorithm tables. This follows the recommendation in
9036 several standards that it is not used in new applications due to
9037 several cryptographic weaknesses. For binary compatibility reasons
9038 the MD2 API is still compiled in by default.
9042 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9047 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9048 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9051 *Guenter <lists@gknw.net>*
9053 * Fix the server certificate chain building code to use X509_verify_cert(),
9054 it used to have an ad-hoc builder which was unable to cope with anything
9055 other than a simple chain.
9057 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9059 * Don't check self signed certificate signatures in X509_verify_cert()
9060 by default (a flag can override this): it just wastes time without
9061 adding any security. As a useful side effect self signed root CAs
9062 with non-FIPS digests are now usable in FIPS mode.
9066 * In dtls1_process_out_of_seq_message() the check if the current message
9067 is already buffered was missing. For every new message was memory
9068 allocated, allowing an attacker to perform an denial of service attack
9069 with sending out of seq handshake messages until there is no memory
9070 left. Additionally every future message was buffered, even if the
9071 sequence number made no sense and would be part of another handshake.
9072 So only messages with sequence numbers less than 10 in advance will be
9073 buffered. ([CVE-2009-1378])
9075 *Robin Seggelmann, discovered by Daniel Mentz*
9077 * Records are buffered if they arrive with a future epoch to be
9078 processed after finishing the corresponding handshake. There is
9079 currently no limitation to this buffer allowing an attacker to perform
9080 a DOS attack with sending records with future epochs until there is no
9081 memory left. This patch adds the pqueue_size() function to determine
9082 the size of a buffer and limits the record buffer to 100 entries.
9085 *Robin Seggelmann, discovered by Daniel Mentz*
9087 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9088 parent structure is freed. ([CVE-2009-1379])
9092 * Handle non-blocking I/O properly in SSL_shutdown() call.
9094 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9096 * Add `2.5.4.*` OIDs
9098 *Ilya O. <vrghost@gmail.com>*
9100 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9102 * Disable renegotiation completely - this fixes a severe security
9103 problem ([CVE-2009-3555]) at the cost of breaking all
9104 renegotiation. Renegotiation can be re-enabled by setting
9105 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9106 run-time. This is really not recommended unless you know what
9111 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9113 * Don't set val to NULL when freeing up structures, it is freed up by
9114 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9115 zeroing past the valid field. ([CVE-2009-0789])
9117 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9119 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9120 checked correctly. This would allow some invalid signed attributes to
9121 appear to verify correctly. ([CVE-2009-0591])
9123 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9125 * Reject UniversalString and BMPString types with invalid lengths. This
9126 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9127 a legal length. ([CVE-2009-0590])
9131 * Set S/MIME signing as the default purpose rather than setting it
9132 unconditionally. This allows applications to override it at the store
9137 * Permit restricted recursion of ASN1 strings. This is needed in practice
9138 to handle some structures.
9142 * Improve efficiency of mem_gets: don't search whole buffer each time
9145 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9147 * New -hex option for openssl rand.
9151 * Print out UTF8String and NumericString when parsing ASN1.
9155 * Support NumericString type for name components.
9159 * Allow CC in the environment to override the automatically chosen
9160 compiler. Note that nothing is done to ensure flags work with the
9165 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9167 * Properly check EVP_VerifyFinal() and similar return values
9170 *Ben Laurie, Bodo Moeller, Google Security Team*
9172 * Enable TLS extensions by default.
9176 * Allow the CHIL engine to be loaded, whether the application is
9177 multithreaded or not. (This does not release the developer from the
9178 obligation to set up the dynamic locking callbacks.)
9180 *Sander Temme <sander@temme.net>*
9182 * Use correct exit code if there is an error in dgst command.
9184 *Steve Henson; problem pointed out by Roland Dirlewanger*
9186 * Tweak Configure so that you need to say "experimental-jpake" to enable
9187 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9191 * Add experimental JPAKE support, including demo authentication in
9192 s_client and s_server.
9196 * Set the comparison function in v3_addr_canonize().
9198 *Rob Austein <sra@hactrn.net>*
9200 * Add support for XMPP STARTTLS in s_client.
9202 *Philip Paeps <philip@freebsd.org>*
9204 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9205 to ensure that even with this option, only ciphersuites in the
9206 server's preference list will be accepted. (Note that the option
9207 applies only when resuming a session, so the earlier behavior was
9208 just about the algorithm choice for symmetric cryptography.)
9212 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9214 * Fix NULL pointer dereference if a DTLS server received
9215 ChangeCipherSpec as first record ([CVE-2009-1386]).
9219 * Fix a state transition in s3_srvr.c and d1_srvr.c
9220 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9224 * The fix in 0.9.8c that supposedly got rid of unsafe
9225 double-checked locking was incomplete for RSA blinding,
9226 addressing just one layer of what turns out to have been
9227 doubly unsafe triple-checked locking.
9229 So now fix this for real by retiring the MONT_HELPER macro
9230 in crypto/rsa/rsa_eay.c.
9232 *Bodo Moeller; problem pointed out by Marius Schilder*
9234 * Various precautionary measures:
9236 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9238 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9239 (NB: This would require knowledge of the secret session ticket key
9240 to exploit, in which case you'd be SOL either way.)
9242 - Change bn_nist.c so that it will properly handle input BIGNUMs
9243 outside the expected range.
9245 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9248 *Neel Mehta, Bodo Moeller*
9250 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9251 the load fails. Useful for distros.
9253 *Ben Laurie and the FreeBSD team*
9255 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9259 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9263 * Expand ENGINE to support engine supplied SSL client certificate functions.
9265 This work was sponsored by Logica.
9269 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9270 keystores. Support for SSL/TLS client authentication too.
9271 Not compiled unless enable-capieng specified to Configure.
9273 This work was sponsored by Logica.
9277 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9278 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9279 attribute creation routines such as certificate requests and PKCS#12
9284 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9286 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9287 handshake which could lead to a client crash as found using the
9288 Codenomicon TLS test suite ([CVE-2008-1672])
9290 *Steve Henson, Mark Cox*
9292 * Fix double free in TLS server name extensions which could lead to
9293 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9297 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9299 Clear the error queue to ensure that error entries left from
9300 older function calls do not interfere with the correct operation.
9302 *Lutz Jaenicke, Erik de Castro Lopo*
9304 * Remove root CA certificates of commercial CAs:
9306 The OpenSSL project does not recommend any specific CA and does not
9307 have any policy with respect to including or excluding any CA.
9308 Therefore it does not make any sense to ship an arbitrary selection
9309 of root CA certificates with the OpenSSL software.
9313 * RSA OAEP patches to fix two separate invalid memory reads.
9314 The first one involves inputs when 'lzero' is greater than
9315 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9316 before the beginning of from). The second one involves inputs where
9317 the 'db' section contains nothing but zeroes (there is a one-byte
9318 invalid read after the end of 'db').
9320 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9322 * Partial backport from 0.9.9-dev:
9324 Introduce bn_mul_mont (dedicated Montgomery multiplication
9325 procedure) as a candidate for BIGNUM assembler implementation.
9326 While 0.9.9-dev uses assembler for various architectures, only
9327 x86_64 is available by default here in the 0.9.8 branch, and
9328 32-bit x86 is available through a compile-time setting.
9330 To try the 32-bit x86 assembler implementation, use Configure
9331 option "enable-montasm" (which exists only for this backport).
9333 As "enable-montasm" for 32-bit x86 disclaims code stability
9334 anyway, in this constellation we activate additional code
9335 backported from 0.9.9-dev for further performance improvements,
9336 namely BN_from_montgomery_word. (To enable this otherwise,
9337 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9339 *Andy Polyakov (backport partially by Bodo Moeller)*
9341 * Add TLS session ticket callback. This allows an application to set
9342 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9343 values. This is useful for key rollover for example where several key
9344 sets may exist with different names.
9348 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9349 This was broken until now in 0.9.8 releases, such that the only way
9350 a registered ENGINE could be used (assuming it initialises
9351 successfully on the host) was to explicitly set it as the default
9352 for the relevant algorithms. This is in contradiction with 0.9.7
9353 behaviour and the documentation. With this fix, when an ENGINE is
9354 registered into a given algorithm's table of implementations, the
9355 'uptodate' flag is reset so that auto-discovery will be used next
9356 time a new context for that algorithm attempts to select an
9359 *Ian Lister (tweaked by Geoff Thorpe)*
9361 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9362 implementation in the following ways:
9364 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9367 Lack of BER streaming support means one pass streaming processing is
9368 only supported if data is detached: setting the streaming flag is
9369 ignored for embedded content.
9371 CMS support is disabled by default and must be explicitly enabled
9372 with the enable-cms configuration option.
9376 * Update the GMP engine glue to do direct copies between BIGNUM and
9377 mpz_t when openssl and GMP use the same limb size. Otherwise the
9378 existing "conversion via a text string export" trick is still used.
9380 *Paul Sheer <paulsheer@gmail.com>*
9382 * Zlib compression BIO. This is a filter BIO which compressed and
9383 uncompresses any data passed through it.
9387 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9388 RFC3394 compatible AES key wrapping.
9392 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9393 sets string data without copying. X509_ALGOR_set0() and
9394 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9395 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9396 from an X509_ATTRIBUTE structure optionally checking it occurs only
9397 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9402 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9403 to get the expected BN_FLG_CONSTTIME behavior.
9405 *Bodo Moeller (Google)*
9409 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9410 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9411 - added some more tests to do_tests.pl
9412 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9413 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9414 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9415 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9416 - various changes to netware.pl to enable gcc-cross builds on Win32
9418 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9419 - various changes to fix missing prototype warnings
9420 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9421 - added AES, WHIRLPOOL and CPUID assembler code to build files
9422 - added missing AES assembler make rules to mk1mf.pl
9423 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9425 *Guenter Knauf <eflash@gmx.net>*
9427 * Implement certificate status request TLS extension defined in RFC3546.
9428 A client can set the appropriate parameters and receive the encoded
9429 OCSP response via a callback. A server can query the supplied parameters
9430 and set the encoded OCSP response in the callback. Add simplified examples
9431 to s_client and s_server.
9435 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9438 + Binary incompatibility of ssl_ctx_st structure
9439 + DTLS interoperation with non-compliant servers
9440 + Don't call get_session_cb() without proposed session
9441 + Fix ia64 assembler code
9443 *Andy Polyakov, Steve Henson*
9445 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9447 * DTLS Handshake overhaul. There were longstanding issues with
9448 OpenSSL DTLS implementation, which were making it impossible for
9449 RFC 4347 compliant client to communicate with OpenSSL server.
9450 Unfortunately just fixing these incompatibilities would "cut off"
9451 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9452 server keeps tolerating non RFC compliant syntax. The opposite is
9453 not true, 0.9.8f client can not communicate with earlier server.
9454 This update even addresses CVE-2007-4995.
9458 * Changes to avoid need for function casts in OpenSSL: some compilers
9459 (gcc 4.2 and later) reject their use.
9460 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9463 * Add RFC4507 support to OpenSSL. This includes the corrections in
9464 RFC4507bis. The encrypted ticket format is an encrypted encoded
9465 SSL_SESSION structure, that way new session features are automatically
9468 If a client application caches session in an SSL_SESSION structure
9469 support is transparent because tickets are now stored in the encoded
9472 The SSL_CTX structure automatically generates keys for ticket
9473 protection in servers so again support should be possible
9474 with no application modification.
9476 If a client or server wishes to disable RFC4507 support then the option
9477 SSL_OP_NO_TICKET can be set.
9479 Add a TLS extension debugging callback to allow the contents of any client
9480 or server extensions to be examined.
9482 This work was sponsored by Google.
9486 * Add initial support for TLS extensions, specifically for the server_name
9487 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9488 have new members for a host name. The SSL data structure has an
9489 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9490 stored in that context to allow for session resumption, even after the
9491 SSL has been switched to a new SSL_CTX in reaction to a client's
9492 server_name extension.
9494 New functions (subject to change):
9496 SSL_get_servername()
9497 SSL_get_servername_type()
9500 New CTRL codes and macros (subject to change):
9502 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9503 - SSL_CTX_set_tlsext_servername_callback()
9504 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9505 - SSL_CTX_set_tlsext_servername_arg()
9506 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9508 openssl s_client has a new '-servername ...' option.
9510 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9511 '-key2 ...', '-servername_fatal' (subject to change). This allows
9512 testing the HostName extension for a specific single host name ('-cert'
9513 and '-key' remain fallbacks for handshakes without HostName
9514 negotiation). If the unrecognized_name alert has to be sent, this by
9515 default is a warning; it becomes fatal with the '-servername_fatal'
9518 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9520 * Add AES and SSE2 assembly language support to VC++ build.
9524 * Mitigate attack on final subtraction in Montgomery reduction.
9528 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9529 (which previously caused an internal error).
9533 * Squeeze another 10% out of IGE mode when in != out.
9537 * AES IGE mode speedup.
9539 *Dean Gaudet (Google)*
9541 * Add the Korean symmetric 128-bit cipher SEED (see
9542 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9543 add SEED ciphersuites from RFC 4162:
9545 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9546 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9547 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9548 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9550 To minimize changes between patchlevels in the OpenSSL 0.9.8
9551 series, SEED remains excluded from compilation unless OpenSSL
9552 is configured with 'enable-seed'.
9554 *KISA, Bodo Moeller*
9556 * Mitigate branch prediction attacks, which can be practical if a
9557 single processor is shared, allowing a spy process to extract
9558 information. For detailed background information, see
9559 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9560 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9561 and Necessary Software Countermeasures"). The core of the change
9562 are new versions BN_div_no_branch() and
9563 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9564 respectively, which are slower, but avoid the security-relevant
9565 conditional branches. These are automatically called by BN_div()
9566 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9567 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9568 remove a conditional branch.
9570 BN_FLG_CONSTTIME is the new name for the previous
9571 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9572 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9573 in the exponent causes BN_mod_exp_mont() to use the alternative
9574 implementation in BN_mod_exp_mont_consttime().) The old name
9575 remains as a deprecated alias.
9577 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9578 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9579 constant-time implementations for more than just exponentiation.
9580 Here too the old name is kept as a deprecated alias.
9582 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9583 the BN_BLINDING structure gets an independent copy of the
9584 modulus. This means that the previous `BIGNUM *m` argument to
9585 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9586 essentially becomes `const BIGNUM *m`, although we can't actually
9587 change this in the header file before 0.9.9. It allows
9588 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9589 enable BN_FLG_CONSTTIME.
9591 *Matthew D Wood (Intel Corp)*
9593 * In the SSL/TLS server implementation, be strict about session ID
9594 context matching (which matters if an application uses a single
9595 external cache for different purposes). Previously,
9596 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9597 set. This did ensure strict client verification, but meant that,
9598 with applications using a single external cache for quite
9599 different requirements, clients could circumvent ciphersuite
9600 restrictions for a given session ID context by starting a session
9601 in a different context.
9605 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9606 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9607 authentication-only ciphersuites.
9611 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9612 not complete and could lead to a possible single byte overflow
9613 ([CVE-2007-5135]) [Ben Laurie]
9615 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9617 * Since AES128 and AES256 (and similarly Camellia128 and
9618 Camellia256) share a single mask bit in the logic of
9619 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9620 kludge to work properly if AES128 is available and AES256 isn't
9621 (or if Camellia128 is available and Camellia256 isn't).
9625 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9626 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9627 When a point or a seed is encoded in a BIT STRING, we need to
9628 prevent the removal of trailing zero bits to get the proper DER
9629 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9630 of a NamedBitList, for which trailing 0 bits need to be removed.)
9634 * Have SSL/TLS server implementation tolerate "mismatched" record
9635 protocol version while receiving ClientHello even if the
9636 ClientHello is fragmented. (The server can't insist on the
9637 particular protocol version it has chosen before the ServerHello
9638 message has informed the client about his choice.)
9642 * Add RFC 3779 support.
9644 *Rob Austein for ARIN, Ben Laurie*
9646 * Load error codes if they are not already present instead of using a
9647 static variable. This allows them to be cleanly unloaded and reloaded.
9648 Improve header file function name parsing.
9652 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9653 or CAPABILITY handshake as required by RFCs.
9657 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9659 * Introduce limits to prevent malicious keys being able to
9660 cause a denial of service. ([CVE-2006-2940])
9662 *Steve Henson, Bodo Moeller*
9664 * Fix ASN.1 parsing of certain invalid structures that can result
9665 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9667 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9668 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9670 * Fix SSL client code which could crash if connecting to a
9671 malicious SSLv2 server. ([CVE-2006-4343])
9673 *Tavis Ormandy and Will Drewry, Google Security Team*
9675 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9676 match only those. Before that, "AES256-SHA" would be interpreted
9677 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9678 the same strength classification in 0.9.7h) as we currently only
9679 have a single AES bit in the ciphersuite description bitmap.
9680 That change, however, also applied to ciphersuite strings such as
9681 "RC4-MD5" that intentionally matched multiple ciphersuites --
9682 namely, SSL 2.0 ciphersuites in addition to the more common ones
9683 from SSL 3.0/TLS 1.0.
9685 So we change the selection algorithm again: Naming an explicit
9686 ciphersuite selects this one ciphersuite, and any other similar
9687 ciphersuite (same bitmap) from *other* protocol versions.
9688 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9689 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9691 Since SSL 2.0 does not have any ciphersuites for which the
9692 128/256 bit distinction would be relevant, this works for now.
9693 The proper fix will be to use different bits for AES128 and
9694 AES256, which would have avoided the problems from the beginning;
9695 however, bits are scarce, so we can only do this in a new release
9696 (not just a patchlevel) when we can change the SSL_CIPHER
9697 definition to split the single 'unsigned long mask' bitmap into
9698 multiple values to extend the available space.
9702 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9704 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9705 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9707 * Add AES IGE and biIGE modes.
9711 * Change the Unix randomness entropy gathering to use poll() when
9712 possible instead of select(), since the latter has some
9713 undesirable limitations.
9715 *Darryl Miles via Richard Levitte and Bodo Moeller*
9717 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9718 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9719 cannot be implicitly activated as part of, e.g., the "AES" alias.
9720 However, please upgrade to OpenSSL 0.9.9[-dev] for
9721 non-experimental use of the ECC ciphersuites to get TLS extension
9722 support, which is required for curve and point format negotiation
9723 to avoid potential handshake problems.
9727 * Disable rogue ciphersuites:
9729 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9730 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9731 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9733 The latter two were purportedly from
9734 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9737 Also deactivate the remaining ciphersuites from
9738 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9739 unofficial, and the ID has long expired.
9743 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9744 dual-core machines) and other potential thread-safety issues.
9748 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9749 versions), which is now available for royalty-free use
9750 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9751 Also, add Camellia TLS ciphersuites from RFC 4132.
9753 To minimize changes between patchlevels in the OpenSSL 0.9.8
9754 series, Camellia remains excluded from compilation unless OpenSSL
9755 is configured with 'enable-camellia'.
9759 * Disable the padding bug check when compression is in use. The padding
9760 bug check assumes the first packet is of even length, this is not
9761 necessarily true if compression is enabled and can result in false
9762 positives causing handshake failure. The actual bug test is ancient
9763 code so it is hoped that implementations will either have fixed it by
9764 now or any which still have the bug do not support compression.
9768 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9770 * When applying a cipher rule check to see if string match is an explicit
9771 cipher suite and only match that one cipher suite if it is.
9775 * Link in manifests for VC++ if needed.
9777 *Austin Ziegler <halostatue@gmail.com>*
9779 * Update support for ECC-based TLS ciphersuites according to
9780 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9781 TLS extensions, which are supported starting with the 0.9.9
9782 branch, not in the OpenSSL 0.9.8 branch).
9786 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9787 opaque EVP_CIPHER_CTX handling.
9791 * Fixes and enhancements to zlib compression code. We now only use
9792 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9793 to conform with the standards mentioned here:
9794 <http://www.zlib.net/DLL_FAQ.txt>
9795 Static zlib linking now works on Windows and the new --with-zlib-include
9796 --with-zlib-lib options to Configure can be used to supply the location
9797 of the headers and library. Gracefully handle case where zlib library
9802 * Several fixes and enhancements to the OID generation code. The old code
9803 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9804 handle numbers larger than ULONG_MAX, truncated printing and had a
9805 non standard OBJ_obj2txt() behaviour.
9809 * Add support for building of engines under engine/ as shared libraries
9810 under VC++ build system.
9814 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9815 Hopefully, we will not see any false combination of paths any more.
9819 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9821 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9822 (part of SSL_OP_ALL). This option used to disable the
9823 countermeasure against man-in-the-middle protocol-version
9824 rollback in the SSL 2.0 server implementation, which is a bad
9825 idea. ([CVE-2005-2969])
9827 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9828 for Information Security, National Institute of Advanced Industrial
9829 Science and Technology [AIST], Japan)*
9831 * Add two function to clear and return the verify parameter flags.
9835 * Keep cipherlists sorted in the source instead of sorting them at
9836 runtime, thus removing the need for a lock.
9840 * Avoid some small subgroup attacks in Diffie-Hellman.
9842 *Nick Mathewson and Ben Laurie*
9844 * Add functions for well-known primes.
9848 * Extended Windows CE support.
9850 *Satoshi Nakamura and Andy Polyakov*
9852 * Initialize SSL_METHOD structures at compile time instead of during
9853 runtime, thus removing the need for a lock.
9857 * Make PKCS7_decrypt() work even if no certificate is supplied by
9858 attempting to decrypt each encrypted key in turn. Add support to
9863 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9865 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9868 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9872 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9873 key into the same file any more.
9877 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9881 * Add -utf8 command line and config file option to 'ca'.
9883 *Stefan <stf@udoma.org*
9885 * Removed the macro des_crypt(), as it seems to conflict with some
9886 libraries. Use DES_crypt().
9890 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9891 involves renaming the source and generated shared-libs for
9892 both. The engines will accept the corrected or legacy ids
9893 ('ncipher' and '4758_cca' respectively) when binding. NB,
9894 this only applies when building 'shared'.
9896 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9898 * Add attribute functions to EVP_PKEY structure. Modify
9899 PKCS12_create() to recognize a CSP name attribute and
9900 use it. Make -CSP option work again in pkcs12 utility.
9904 * Add new functionality to the bn blinding code:
9905 - automatic re-creation of the BN_BLINDING parameters after
9906 a fixed number of uses (currently 32)
9907 - add new function for parameter creation
9908 - introduce flags to control the update behaviour of the
9909 BN_BLINDING parameters
9910 - hide BN_BLINDING structure
9911 Add a second BN_BLINDING slot to the RSA structure to improve
9912 performance when a single RSA object is shared among several
9917 * Add support for DTLS.
9919 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9921 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9922 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9926 * Remove buggy and incomplete DH cert support from
9927 ssl/ssl_rsa.c and ssl/s3_both.c
9931 * Use SHA-1 instead of MD5 as the default digest algorithm for
9932 the apps/openssl applications.
9936 * Compile clean with "-Wall -Wmissing-prototypes
9937 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9938 DEBUG_SAFESTACK must also be set.
9942 * Change ./Configure so that certain algorithms can be disabled by default.
9943 The new counterpiece to "no-xxx" is "enable-xxx".
9945 The patented RC5 and MDC2 algorithms will now be disabled unless
9946 "enable-rc5" and "enable-mdc2", respectively, are specified.
9948 (IDEA remains enabled despite being patented. This is because IDEA
9949 is frequently required for interoperability, and there is no license
9950 fee for non-commercial use. As before, "no-idea" can be used to
9951 avoid this algorithm.)
9955 * Add processing of proxy certificates (see RFC 3820). This work was
9956 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9957 EGEE (Enabling Grids for E-science in Europe).
9961 * RC4 performance overhaul on modern architectures/implementations, such
9962 as Intel P4, IA-64 and AMD64.
9966 * New utility extract-section.pl. This can be used specify an alternative
9967 section number in a pod file instead of having to treat each file as
9968 a separate case in Makefile. This can be done by adding two lines to the
9971 =for comment openssl_section:XXX
9973 The blank line is mandatory.
9977 * New arguments -certform, -keyform and -pass for s_client and s_server
9978 to allow alternative format key and certificate files and passphrase
9983 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9984 update associated structures and add various utility functions.
9986 Add new policy related verify parameters, include policy checking in
9987 standard verify code. Enhance 'smime' application with extra parameters
9988 to support policy checking and print out.
9992 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9993 Nehemiah processors. These extensions support AES encryption in hardware
9994 as well as RNG (though RNG support is currently disabled).
9996 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9998 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10002 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10004 *Andy Polyakov and a number of other people*
10006 * Improved PowerPC platform support. Most notably BIGNUM assembler
10007 implementation contributed by IBM.
10009 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10011 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10012 exponent rather than 'unsigned long'. There is a corresponding change to
10013 the new 'rsa_keygen' element of the RSA_METHOD structure.
10015 *Jelte Jansen, Geoff Thorpe*
10017 * Functionality for creating the initial serial number file is now
10018 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10020 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10021 number file to 1, which is bound to cause problems. To avoid
10022 the problems while respecting compatibility between different 0.9.7
10023 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10024 CA.pl for serial number initialization. With the new release 0.9.8,
10025 we can fix the problem directly in the 'ca' utility.)
10029 * Reduced header interdependencies by declaring more opaque objects in
10030 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10031 give fewer recursive includes, which could break lazy source code - so
10032 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10033 developers should define this symbol when building and using openssl to
10034 ensure they track the recommended behaviour, interfaces, [etc], but
10035 backwards-compatible behaviour prevails when this isn't defined.
10039 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10043 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10044 This will generate a random key of the appropriate length based on the
10045 cipher context. The EVP_CIPHER can provide its own random key generation
10046 routine to support keys of a specific form. This is used in the des and
10047 3des routines to generate a key of the correct parity. Update S/MIME
10048 code to use new functions and hence generate correct parity DES keys.
10049 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10050 valid (weak or incorrect parity).
10054 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10055 as looking them up. This is useful when the verified structure may contain
10056 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10057 present unless the new PKCS7_NO_CRL flag is asserted.
10061 * Extend ASN1 oid configuration module. It now additionally accepts the
10064 shortName = some long name, 1.2.3.4
10068 * Reimplemented the BN_CTX implementation. There is now no more static
10069 limitation on the number of variables it can handle nor the depth of the
10070 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10071 information can now expand as required, and rather than having a single
10072 static array of bignums, BN_CTX now uses a linked-list of such arrays
10073 allowing it to expand on demand whilst maintaining the usefulness of
10074 BN_CTX's "bundling".
10078 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10079 to allow all RSA operations to function using a single BN_CTX.
10083 * Preliminary support for certificate policy evaluation and checking. This
10084 is initially intended to pass the tests outlined in "Conformance Testing
10085 of Relying Party Client Certificate Path Processing Logic" v1.07.
10089 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10090 remained unused and not that useful. A variety of other little bignum
10091 tweaks and fixes have also been made continuing on from the audit (see
10096 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10097 associated ASN1, EVP and SSL functions and old ASN1 macros.
10101 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10102 and this should never fail. So the return value from the use of
10103 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10104 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10108 * BN_CTX_get() should return zero-valued bignums, providing the same
10109 initialised value as BN_new().
10111 *Geoff Thorpe, suggested by Ulf Möller*
10113 * Support for inhibitAnyPolicy certificate extension.
10117 * An audit of the BIGNUM code is underway, for which debugging code is
10118 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10119 is considered valid when processing BIGNUMs, and causes execution to
10120 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10121 further steps are taken to deliberately pollute unused data in BIGNUM
10122 structures to try and expose faulty code further on. For now, openssl will
10123 (in its default mode of operation) continue to tolerate the inconsistent
10124 forms that it has tolerated in the past, but authors and packagers should
10125 consider trying openssl and their own applications when compiled with
10126 these debugging symbols defined. It will help highlight potential bugs in
10127 their own code, and will improve the test coverage for OpenSSL itself. At
10128 some point, these tighter rules will become openssl's default to improve
10129 maintainability, though the assert()s and other overheads will remain only
10130 in debugging configurations. See bn.h for more details.
10132 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10134 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10135 that can only be obtained through BN_CTX_new() (which implicitly
10136 initialises it). The presence of this function only made it possible
10137 to overwrite an existing structure (and cause memory leaks).
10141 * Because of the callback-based approach for implementing LHASH as a
10142 template type, lh_insert() adds opaque objects to hash-tables and
10143 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10144 to clean up those corresponding objects before destroying the hash table
10145 (and losing the object pointers). So some over-zealous constifications in
10146 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10147 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10148 prototyped to have "const" restrictions on the object pointers they are
10149 given (and so aren't required to cast them away any more).
10153 * The tmdiff.h API was so ugly and minimal that our own timing utility
10154 (speed) prefers to use its own implementation. The two implementations
10155 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10156 its object type properly exposed (MS_TM) instead of casting to/from
10157 `char *`. This may still change yet if someone realises MS_TM and
10159 aren't necessarily the greatest nomenclatures - but this is what was used
10160 internally to the implementation so I've used that for now.
10164 * Ensure that deprecated functions do not get compiled when
10165 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10166 the self-tests were still using deprecated key-generation functions so
10167 these have been updated also.
10171 * Reorganise PKCS#7 code to separate the digest location functionality
10172 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10173 New function PKCS7_set_digest() to set the digest type for PKCS#7
10174 digestedData type. Add additional code to correctly generate the
10175 digestedData type and add support for this type in PKCS7 initialization
10180 * New function PKCS7_set0_type_other() this initializes a PKCS7
10181 structure of type "other".
10185 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10186 sure the loop does correctly stop and breaking ("division by zero")
10187 modulus operations are not performed. The (pre-generated) prime
10188 table crypto/bn/bn_prime.h was already correct, but it could not be
10189 re-generated on some platforms because of the "division by zero"
10190 situation in the script.
10192 *Ralf S. Engelschall*
10194 * Update support for ECC-based TLS ciphersuites according to
10195 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10196 SHA-1 now is only used for "small" curves (where the
10197 representation of a field element takes up to 24 bytes); for
10198 larger curves, the field element resulting from ECDH is directly
10199 used as premaster secret.
10201 *Douglas Stebila (Sun Microsystems Laboratories)*
10203 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10204 curve secp160r1 to the tests.
10206 *Douglas Stebila (Sun Microsystems Laboratories)*
10208 * Add the possibility to load symbols globally with DSO.
10210 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10212 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10213 control of the error stack.
10217 * Add support for STORE in ENGINE.
10221 * Add the STORE type. The intention is to provide a common interface
10222 to certificate and key stores, be they simple file-based stores, or
10223 HSM-type store, or LDAP stores, or...
10224 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10228 * Add a generic structure called OPENSSL_ITEM. This can be used to
10229 pass a list of arguments to any function as well as provide a way
10230 for a function to pass data back to the caller.
10234 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10235 works like BUF_strdup() but can be used to duplicate a portion of
10236 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10241 * Add the function sk_find_ex() which works like sk_find(), but will
10242 return an index to an element even if an exact match couldn't be
10243 found. The index is guaranteed to point at the element where the
10244 searched-for key would be inserted to preserve sorting order.
10248 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10249 takes an extra flags argument for optional functionality. Currently,
10250 the following flags are defined:
10252 OBJ_BSEARCH_VALUE_ON_NOMATCH
10253 This one gets OBJ_bsearch_ex() to return a pointer to the first
10254 element where the comparing function returns a negative or zero
10257 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10258 This one gets OBJ_bsearch_ex() to return a pointer to the first
10259 element where the comparing function returns zero. This is useful
10260 if there are more than one element where the comparing function
10265 * Make it possible to create self-signed certificates with 'openssl ca'
10266 in such a way that the self-signed certificate becomes part of the
10267 CA database and uses the same mechanisms for serial number generation
10268 as all other certificate signing. The new flag '-selfsign' enables
10269 this functionality. Adapt CA.sh and CA.pl.in.
10273 * Add functionality to check the public key of a certificate request
10274 against a given private. This is useful to check that a certificate
10275 request can be signed by that key (self-signing).
10279 * Make it possible to have multiple active certificates with the same
10280 subject in the CA index file. This is done only if the keyword
10281 'unique_subject' is set to 'no' in the main CA section (default
10282 if 'CA_default') of the configuration file. The value is saved
10283 with the database itself in a separate index attribute file,
10284 named like the index file with '.attr' appended to the name.
10288 * Generate multi-valued AVAs using '+' notation in config files for
10293 * Support for nameConstraints certificate extension.
10297 * Support for policyConstraints certificate extension.
10301 * Support for policyMappings certificate extension.
10305 * Make sure the default DSA_METHOD implementation only uses its
10306 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10307 and change its own handlers to be NULL so as to remove unnecessary
10308 indirection. This lets alternative implementations fallback to the
10309 default implementation more easily.
10313 * Support for directoryName in GeneralName related extensions
10318 * Make it possible to link applications using Makefile.shared.
10319 Make that possible even when linking against static libraries!
10323 * Support for single pass processing for S/MIME signing. This now
10324 means that S/MIME signing can be done from a pipe, in addition
10325 cleartext signing (multipart/signed type) is effectively streaming
10326 and the signed data does not need to be all held in memory.
10328 This is done with a new flag PKCS7_STREAM. When this flag is set
10329 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10330 is done after the data is output (and digests calculated) in
10331 SMIME_write_PKCS7().
10335 * Add full support for -rpath/-R, both in shared libraries and
10336 applications, at least on the platforms where it's known how
10341 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10342 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10343 will now compute a table of multiples of the generator that
10344 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10345 faster (notably in the case of a single point multiplication,
10346 scalar * generator).
10348 *Nils Larsch, Bodo Moeller*
10350 * IPv6 support for certificate extensions. The various extensions
10351 which use the IP:a.b.c.d can now take IPv6 addresses using the
10352 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10357 * Added an ENGINE that implements RSA by performing private key
10358 exponentiations with the GMP library. The conversions to and from
10359 GMP's mpz_t format aren't optimised nor are any montgomery forms
10360 cached, and on x86 it appears OpenSSL's own performance has caught up.
10361 However there are likely to be other architectures where GMP could
10362 provide a boost. This ENGINE is not built in by default, but it can be
10363 specified at Configure time and should be accompanied by the necessary
10364 linker additions, eg;
10365 ./config -DOPENSSL_USE_GMP -lgmp
10369 * "openssl engine" will not display ENGINE/DSO load failure errors when
10370 testing availability of engines with "-t" - the old behaviour is
10371 produced by increasing the feature's verbosity with "-tt".
10375 * ECDSA routines: under certain error conditions uninitialized BN objects
10376 could be freed. Solution: make sure initialization is performed early
10377 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10382 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10383 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10384 software implementations. For DSA and DH, parameter generation can
10385 also be overridden by providing the appropriate method callbacks.
10389 * Change the "progress" mechanism used in key-generation and
10390 primality testing to functions that take a new BN_GENCB pointer in
10391 place of callback/argument pairs. The new API functions have `_ex`
10392 postfixes and the older functions are reimplemented as wrappers for
10393 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10394 declarations of the old functions to help (graceful) attempts to
10395 migrate to the new functions. Also, the new key-generation API
10396 functions operate on a caller-supplied key-structure and return
10397 success/failure rather than returning a key or NULL - this is to
10398 help make "keygen" another member function of RSA_METHOD etc.
10400 Example for using the new callback interface:
10402 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10403 void *my_arg = ...;
10406 BN_GENCB_set(&my_cb, my_callback, my_arg);
10408 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10409 /* For the meaning of a, b in calls to my_callback(), see the
10410 * documentation of the function that calls the callback.
10411 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10412 * my_callback should return 1 if it wants BN_is_prime_ex()
10413 * to continue, or 0 to stop.
10418 * Change the ZLIB compression method to be stateful, and make it
10419 available to TLS with the number defined in
10420 draft-ietf-tls-compression-04.txt.
10424 * Add the ASN.1 structures and functions for CertificatePair, which
10425 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10427 CertificatePair ::= SEQUENCE {
10428 forward [0] Certificate OPTIONAL,
10429 reverse [1] Certificate OPTIONAL,
10430 -- at least one of the pair shall be present -- }
10432 Also implement the PEM functions to read and write certificate
10433 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10435 This needed to be defined, mostly for the sake of the LDAP
10436 attribute crossCertificatePair, but may prove useful elsewhere as
10441 * Make it possible to inhibit symlinking of shared libraries in
10442 Makefile.shared, for Cygwin's sake.
10446 * Extend the BIGNUM API by creating a function
10447 void BN_set_negative(BIGNUM *a, int neg);
10448 and a macro that behave like
10449 int BN_is_negative(const BIGNUM *a);
10451 to avoid the need to access 'a->neg' directly in applications.
10455 * Implement fast modular reduction for pseudo-Mersenne primes
10456 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10457 EC_GROUP_new_curve_GFp() will now automatically use this
10460 *Nils Larsch <nla@trustcenter.de>*
10462 * Add new lock type (CRYPTO_LOCK_BN).
10466 * Change the ENGINE framework to automatically load engines
10467 dynamically from specific directories unless they could be
10468 found to already be built in or loaded. Move all the
10469 current engines except for the cryptodev one to a new
10470 directory engines/.
10471 The engines in engines/ are built as shared libraries if
10472 the "shared" options was given to ./Configure or ./config.
10473 Otherwise, they are inserted in libcrypto.a.
10474 /usr/local/ssl/engines is the default directory for dynamic
10475 engines, but that can be overridden at configure time through
10476 the usual use of --prefix and/or --openssldir, and at run
10477 time with the environment variable OPENSSL_ENGINES.
10479 *Geoff Thorpe and Richard Levitte*
10481 * Add Makefile.shared, a helper makefile to build shared
10482 libraries. Adapt Makefile.org.
10486 * Add version info to Win32 DLLs.
10488 *Peter 'Luna' Runestig" <peter@runestig.com>*
10490 * Add new 'medium level' PKCS#12 API. Certificates and keys
10491 can be added using this API to created arbitrary PKCS#12
10492 files while avoiding the low level API.
10494 New options to PKCS12_create(), key or cert can be NULL and
10495 will then be omitted from the output file. The encryption
10496 algorithm NIDs can be set to -1 for no encryption, the mac
10497 iteration count can be set to 0 to omit the mac.
10499 Enhance pkcs12 utility by making the -nokeys and -nocerts
10500 options work when creating a PKCS#12 file. New option -nomac
10501 to omit the mac, NONE can be set for an encryption algorithm.
10502 New code is modified to use the enhanced PKCS12_create()
10503 instead of the low level API.
10507 * Extend ASN1 encoder to support indefinite length constructed
10508 encoding. This can output sequences tags and octet strings in
10509 this form. Modify pk7_asn1.c to support indefinite length
10510 encoding. This is experimental and needs additional code to
10511 be useful, such as an ASN1 bio and some enhanced streaming
10514 Extend template encode functionality so that tagging is passed
10515 down to the template encoder.
10519 * Let 'openssl req' fail if an argument to '-newkey' is not
10520 recognized instead of using RSA as a default.
10524 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10525 As these are not official, they are not included in "ALL";
10526 the "ECCdraft" ciphersuite group alias can be used to select them.
10528 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10530 * Add ECDH engine support.
10532 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10534 * Add ECDH in new directory crypto/ecdh/.
10536 *Douglas Stebila (Sun Microsystems Laboratories)*
10538 * Let BN_rand_range() abort with an error after 100 iterations
10539 without success (which indicates a broken PRNG).
10543 * Change BN_mod_sqrt() so that it verifies that the input value
10544 is really the square of the return value. (Previously,
10545 BN_mod_sqrt would show GIGO behaviour.)
10549 * Add named elliptic curves over binary fields from X9.62, SECG,
10550 and WAP/WTLS; add OIDs that were still missing.
10552 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10554 * Extend the EC library for elliptic curves over binary fields
10555 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10558 EC_GF2m_simple_method
10562 EC_GROUP_new_curve_GF2m
10563 EC_GROUP_set_curve_GF2m
10564 EC_GROUP_get_curve_GF2m
10565 EC_POINT_set_affine_coordinates_GF2m
10566 EC_POINT_get_affine_coordinates_GF2m
10567 EC_POINT_set_compressed_coordinates_GF2m
10569 Point compression for binary fields is disabled by default for
10570 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10573 As binary polynomials are represented as BIGNUMs, various members
10574 of the EC_GROUP and EC_POINT data structures can be shared
10575 between the implementations for prime fields and binary fields;
10576 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10577 are essentially identical to their `..._GFp` counterparts.
10578 (For simplicity, the `..._GFp` prefix has been dropped from
10579 various internal method names.)
10581 An internal 'field_div' method (similar to 'field_mul' and
10582 'field_sqr') has been added; this is used only for binary fields.
10584 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10586 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10587 through methods ('mul', 'precompute_mult').
10589 The generic implementations (now internally called 'ec_wNAF_mul'
10590 and 'ec_wNAF_precomputed_mult') remain the default if these
10591 methods are undefined.
10593 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10595 * New function EC_GROUP_get_degree, which is defined through
10596 EC_METHOD. For curves over prime fields, this returns the bit
10597 length of the modulus.
10599 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10601 * New functions EC_GROUP_dup, EC_POINT_dup.
10602 (These simply call ..._new and ..._copy).
10604 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10606 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10607 Polynomials are represented as BIGNUMs (where the sign bit is not
10608 used) in the following functions [macros]:
10611 BN_GF2m_sub [= BN_GF2m_add]
10612 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10613 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10614 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10616 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10617 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10618 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10619 BN_GF2m_cmp [= BN_ucmp]
10621 (Note that only the 'mod' functions are actually for fields GF(2^m).
10622 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10624 For some functions, an the irreducible polynomial defining a
10625 field can be given as an 'unsigned int[]' with strictly
10626 decreasing elements giving the indices of those bits that are set;
10627 i.e., p[] represents the polynomial
10628 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10630 p[0] > p[1] > ... > p[k] = 0.
10631 This applies to the following functions:
10634 BN_GF2m_mod_mul_arr
10635 BN_GF2m_mod_sqr_arr
10636 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10637 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10638 BN_GF2m_mod_exp_arr
10639 BN_GF2m_mod_sqrt_arr
10640 BN_GF2m_mod_solve_quad_arr
10644 Conversion can be performed by the following functions:
10649 bntest.c has additional tests for binary polynomial arithmetic.
10651 Two implementations for BN_GF2m_mod_div() are available.
10652 The default algorithm simply uses BN_GF2m_mod_inv() and
10653 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10654 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10655 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10657 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10659 * Add new error code 'ERR_R_DISABLED' that can be used when some
10660 functionality is disabled at compile-time.
10662 *Douglas Stebila <douglas.stebila@sun.com>*
10664 * Change default behaviour of 'openssl asn1parse' so that more
10665 information is visible when viewing, e.g., a certificate:
10667 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10668 mode the content of non-printable OCTET STRINGs is output in a
10669 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10670 avoid the appearance of a printable string.
10672 *Nils Larsch <nla@trustcenter.de>*
10674 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10676 EC_GROUP_set_asn1_flag()
10677 EC_GROUP_get_asn1_flag()
10678 EC_GROUP_set_point_conversion_form()
10679 EC_GROUP_get_point_conversion_form()
10680 These control ASN1 encoding details:
10681 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10682 has been set to OPENSSL_EC_NAMED_CURVE.
10683 - Points are encoded in uncompressed form by default; options for
10684 asn1_for are as for point2oct, namely
10685 POINT_CONVERSION_COMPRESSED
10686 POINT_CONVERSION_UNCOMPRESSED
10687 POINT_CONVERSION_HYBRID
10689 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10691 EC_GROUP_set_seed()
10692 EC_GROUP_get0_seed()
10693 EC_GROUP_get_seed_len()
10694 This is used only for ASN1 purposes (so far).
10696 *Nils Larsch <nla@trustcenter.de>*
10698 * Add 'field_type' member to EC_METHOD, which holds the NID
10699 of the appropriate field type OID. The new function
10700 EC_METHOD_get_field_type() returns this value.
10702 *Nils Larsch <nla@trustcenter.de>*
10705 EC_POINT_point2bn()
10706 EC_POINT_bn2point()
10707 EC_POINT_point2hex()
10708 EC_POINT_hex2point()
10709 providing useful interfaces to EC_POINT_point2oct() and
10710 EC_POINT_oct2point().
10712 *Nils Larsch <nla@trustcenter.de>*
10714 * Change internals of the EC library so that the functions
10715 EC_GROUP_set_generator()
10716 EC_GROUP_get_generator()
10717 EC_GROUP_get_order()
10718 EC_GROUP_get_cofactor()
10719 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10720 to methods, which would lead to unnecessary code duplication when
10721 adding different types of curves.
10723 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10725 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10726 arithmetic, and such that modified wNAFs are generated
10727 (which avoid length expansion in many cases).
10731 * Add a function EC_GROUP_check_discriminant() (defined via
10732 EC_METHOD) that verifies that the curve discriminant is non-zero.
10734 Add a function EC_GROUP_check() that makes some sanity tests
10735 on a EC_GROUP, its generator and order. This includes
10736 EC_GROUP_check_discriminant().
10738 *Nils Larsch <nla@trustcenter.de>*
10740 * Add ECDSA in new directory crypto/ecdsa/.
10742 Add applications 'openssl ecparam' and 'openssl ecdsa'
10743 (these are based on 'openssl dsaparam' and 'openssl dsa').
10745 ECDSA support is also included in various other files across the
10746 library. Most notably,
10747 - 'openssl req' now has a '-newkey ecdsa:file' option;
10748 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10749 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10750 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10751 them suitable for ECDSA where domain parameters must be
10752 extracted before the specific public key;
10753 - ECDSA engine support has been added.
10755 *Nils Larsch <nla@trustcenter.de>*
10757 * Include some named elliptic curves, and add OIDs from X9.62,
10758 SECG, and WAP/WTLS. Each curve can be obtained from the new
10760 EC_GROUP_new_by_curve_name(),
10761 and the list of available named curves can be obtained with
10762 EC_get_builtin_curves().
10763 Also add a 'curve_name' member to EC_GROUP objects, which can be
10765 EC_GROUP_set_curve_name()
10766 EC_GROUP_get_curve_name()
10768 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10770 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10771 was actually never needed) and in BN_mul(). The removal in BN_mul()
10772 required a small change in bn_mul_part_recursive() and the addition
10773 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10774 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10775 bn_sub_words() and bn_add_words() except they take arrays with
10780 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10782 * Cleanse PEM buffers before freeing them since they may contain
10785 *Benjamin Bennett <ben@psc.edu>*
10787 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10788 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10789 authentication-only ciphersuites.
10793 * Since AES128 and AES256 share a single mask bit in the logic of
10794 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10795 kludge to work properly if AES128 is available and AES256 isn't.
10799 * Expand security boundary to match 1.1.1 module.
10803 * Remove redundant features: hash file source, editing of test vectors
10804 modify fipsld to use external fips_premain.c signature.
10808 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10809 run algorithm test programs.
10813 * Make algorithm test programs more tolerant of whitespace.
10817 * Have SSL/TLS server implementation tolerate "mismatched" record
10818 protocol version while receiving ClientHello even if the
10819 ClientHello is fragmented. (The server can't insist on the
10820 particular protocol version it has chosen before the ServerHello
10821 message has informed the client about his choice.)
10825 * Load error codes if they are not already present instead of using a
10826 static variable. This allows them to be cleanly unloaded and reloaded.
10830 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10832 * Introduce limits to prevent malicious keys being able to
10833 cause a denial of service. ([CVE-2006-2940])
10835 *Steve Henson, Bodo Moeller*
10837 * Fix ASN.1 parsing of certain invalid structures that can result
10838 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10840 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10841 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10843 * Fix SSL client code which could crash if connecting to a
10844 malicious SSLv2 server. ([CVE-2006-4343])
10846 *Tavis Ormandy and Will Drewry, Google Security Team*
10848 * Change ciphersuite string processing so that an explicit
10849 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10850 will no longer include "AES128-SHA"), and any other similar
10851 ciphersuite (same bitmap) from *other* protocol versions (so that
10852 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10853 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10854 changes from 0.9.8b and 0.9.8d.
10858 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10860 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10861 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10863 * Change the Unix randomness entropy gathering to use poll() when
10864 possible instead of select(), since the latter has some
10865 undesirable limitations.
10867 *Darryl Miles via Richard Levitte and Bodo Moeller*
10869 * Disable rogue ciphersuites:
10871 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10872 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10873 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10875 The latter two were purportedly from
10876 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10879 Also deactivate the remaining ciphersuites from
10880 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10881 unofficial, and the ID has long expired.
10885 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10886 dual-core machines) and other potential thread-safety issues.
10890 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10892 * Adapt fipsld and the build system to link against the validated FIPS
10893 module in FIPS mode.
10897 * Fixes for VC++ 2005 build under Windows.
10901 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10902 from a Windows bash shell such as MSYS. It is autodetected from the
10903 "config" script when run from a VC++ environment. Modify standard VC++
10904 build to use fipscanister.o from the GNU make build.
10908 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10910 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10911 The value now differs depending on if you build for FIPS or not.
10912 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10913 safely run with a non-FIPSed libcrypto, as it may crash because of
10914 the difference induced by this change.
10918 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10920 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10921 (part of SSL_OP_ALL). This option used to disable the
10922 countermeasure against man-in-the-middle protocol-version
10923 rollback in the SSL 2.0 server implementation, which is a bad
10924 idea. ([CVE-2005-2969])
10926 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10927 for Information Security, National Institute of Advanced Industrial
10928 Science and Technology [AIST, Japan)]*
10930 * Minimal support for X9.31 signatures and PSS padding modes. This is
10931 mainly for FIPS compliance and not fully integrated at this stage.
10935 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10936 the exponentiation using a fixed-length exponent. (Otherwise,
10937 the information leaked through timing could expose the secret key
10938 after many signatures; cf. Bleichenbacher's attack on DSA with
10943 * Make a new fixed-window mod_exp implementation the default for
10944 RSA, DSA, and DH private-key operations so that the sequence of
10945 squares and multiplies and the memory access pattern are
10946 independent of the particular secret key. This will mitigate
10947 cache-timing and potential related attacks.
10949 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10950 and this is automatically used by BN_mod_exp_mont() if the new flag
10951 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10952 will use this BN flag for private exponents unless the flag
10953 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10954 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10956 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10958 * Change the client implementation for SSLv23_method() and
10959 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10960 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10961 (Previously, the SSL 2.0 backwards compatible Client Hello
10962 message format would be used even with SSL_OP_NO_SSLv2.)
10966 * Add support for smime-type MIME parameter in S/MIME messages which some
10971 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10972 a threadsafe manner. Modify rsa code to use new function and add calls
10973 to dsa and dh code (which had race conditions before).
10977 * Include the fixed error library code in the C error file definitions
10978 instead of fixing them up at runtime. This keeps the error code
10979 structures constant.
10983 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10985 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10988 * Fixes for newer kerberos headers. NB: the casts are needed because
10989 the 'length' field is signed on one version and unsigned on another
10990 with no (?) obvious way to tell the difference, without these VC++
10991 complains. Also the "definition" of FAR (blank) is no longer included
10992 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10993 some needed definitions.
10997 * Undo Cygwin change.
11001 * Added support for proxy certificates according to RFC 3820.
11002 Because they may be a security thread to unaware applications,
11003 they must be explicitly allowed in run-time. See
11004 docs/HOWTO/proxy_certificates.txt for further information.
11008 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11010 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11011 server and client random values. Previously
11012 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11013 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11015 This change has negligible security impact because:
11017 1. Server and client random values still have 24 bytes of pseudo random
11020 2. Server and client random values are sent in the clear in the initial
11023 3. The master secret is derived using the premaster secret (48 bytes in
11024 size for static RSA ciphersuites) as well as client server and random
11027 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11030 *Stephen Henson, reported by UK NISCC*
11032 * Use Windows randomness collection on Cygwin.
11036 * Fix hang in EGD/PRNGD query when communication socket is closed
11037 prematurely by EGD/PRNGD.
11039 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11041 * Prompt for pass phrases when appropriate for PKCS12 input format.
11045 * Back-port of selected performance improvements from development
11046 branch, as well as improved support for PowerPC platforms.
11050 * Add lots of checks for memory allocation failure, error codes to indicate
11051 failure and freeing up memory if a failure occurs.
11053 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11055 * Add new -passin argument to dgst.
11059 * Perform some character comparisons of different types in X509_NAME_cmp:
11060 this is needed for some certificates that re-encode DNs into UTF8Strings
11061 (in violation of RFC3280) and can't or won't issue name rollover
11066 * Make an explicit check during certificate validation to see that
11067 the CA setting in each certificate on the chain is correct. As a
11068 side effect always do the following basic checks on extensions,
11069 not just when there's an associated purpose to the check:
11071 - if there is an unhandled critical extension (unless the user
11072 has chosen to ignore this fault)
11073 - if the path length has been exceeded (if one is set at all)
11074 - that certain extensions fit the associated purpose (if one has
11079 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11081 * Avoid a race condition when CRLs are checked in a multi threaded
11082 environment. This would happen due to the reordering of the revoked
11083 entries during signature checking and serial number lookup. Now the
11084 encoding is cached and the serial number sort performed under a lock.
11085 Add new STACK function sk_is_sorted().
11089 * Add Delta CRL to the extension code.
11093 * Various fixes to s3_pkt.c so alerts are sent properly.
11095 *David Holmes <d.holmes@f5.com>*
11097 * Reduce the chances of duplicate issuer name and serial numbers (in
11098 violation of RFC3280) using the OpenSSL certificate creation utilities.
11099 This is done by creating a random 64 bit value for the initial serial
11100 number when a serial number file is created or when a self signed
11101 certificate is created using 'openssl req -x509'. The initial serial
11102 number file is created using 'openssl x509 -next_serial' in CA.pl
11103 rather than being initialized to 1.
11107 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11109 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11110 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11112 *Joe Orton, Steve Henson*
11114 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11117 *Joe Orton, Steve Henson*
11119 * Make it possible to have multiple active certificates with the same
11120 subject in the CA index file. This is done only if the keyword
11121 'unique_subject' is set to 'no' in the main CA section (default
11122 if 'CA_default') of the configuration file. The value is saved
11123 with the database itself in a separate index attribute file,
11124 named like the index file with '.attr' appended to the name.
11128 * X509 verify fixes. Disable broken certificate workarounds when
11129 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11130 keyUsage extension present. Don't accept CRLs with unhandled critical
11131 extensions: since verify currently doesn't process CRL extensions this
11132 rejects a CRL with *any* critical extensions. Add new verify error codes
11137 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11138 A clarification of RFC2560 will require the use of OCTET STRINGs and
11139 some implementations cannot handle the current raw format. Since OpenSSL
11140 copies and compares OCSP nonces as opaque blobs without any attempt at
11141 parsing them this should not create any compatibility issues.
11145 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11146 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11147 this HMAC (and other) operations are several times slower than OpenSSL
11152 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11154 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11156 * Use the correct content when signing type "other".
11160 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11162 * Fix various bugs revealed by running the NISCC test suite:
11164 Stop out of bounds reads in the ASN1 code when presented with
11165 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11167 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11169 If verify callback ignores invalid public key errors don't try to check
11170 certificate signature with the NULL public key.
11174 * New -ignore_err option in ocsp application to stop the server
11175 exiting on the first error in a request.
11179 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11180 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11185 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11186 extra data after the compression methods not only for TLS 1.0
11187 but also for SSL 3.0 (as required by the specification).
11189 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11191 * Change X509_certificate_type() to mark the key as exported/exportable
11192 when it's 512 *bits* long, not 512 bytes.
11196 * Change AES_cbc_encrypt() so it outputs exact multiple of
11197 blocks during encryption.
11201 * Various fixes to base64 BIO and non blocking I/O. On write
11202 flushes were not handled properly if the BIO retried. On read
11203 data was not being buffered properly and had various logic bugs.
11204 This also affects blocking I/O when the data being decoded is a
11209 * Various S/MIME bugfixes and compatibility changes:
11210 output correct application/pkcs7 MIME type if
11211 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11212 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11213 of files as .eml work). Correctly handle very long lines in MIME
11218 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11220 * Countermeasure against the Klima-Pokorny-Rosa extension of
11221 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11222 a protocol version number mismatch like a decryption error
11223 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11227 * Turn on RSA blinding by default in the default implementation
11228 to avoid a timing attack. Applications that don't want it can call
11229 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11230 They would be ill-advised to do so in most cases.
11232 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11234 * Change RSA blinding code so that it works when the PRNG is not
11235 seeded (in this case, the secret RSA exponent is abused as
11236 an unpredictable seed -- if it is not unpredictable, there
11237 is no point in blinding anyway). Make RSA blinding thread-safe
11238 by remembering the creator's thread ID in rsa->blinding and
11239 having all other threads use local one-time blinding factors
11240 (this requires more computation than sharing rsa->blinding, but
11241 avoids excessive locking; and if an RSA object is not shared
11242 between threads, blinding will still be very fast).
11246 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11247 ENGINE as defaults for all supported algorithms irrespective of
11248 the 'flags' parameter. 'flags' is now honoured, so applications
11249 should make sure they are passing it correctly.
11253 * Target "mingw" now allows native Windows code to be generated in
11254 the Cygwin environment as well as with the MinGW compiler.
11258 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11260 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11261 via timing by performing a MAC computation even if incorrect
11262 block cipher padding has been found. This is a countermeasure
11263 against active attacks where the attacker has to distinguish
11264 between bad padding and a MAC verification error. ([CVE-2003-0078])
11266 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11267 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11268 Martin Vuagnoux (EPFL, Ilion)*
11270 * Make the no-err option work as intended. The intention with no-err
11271 is not to have the whole error stack handling routines removed from
11272 libcrypto, it's only intended to remove all the function name and
11273 reason texts, thereby removing some of the footprint that may not
11274 be interesting if those errors aren't displayed anyway.
11276 NOTE: it's still possible for any application or module to have its
11277 own set of error texts inserted. The routines are there, just not
11278 used by default when no-err is given.
11282 * Add support for FreeBSD on IA64.
11284 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11286 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11287 Kerberos function mit_des_cbc_cksum(). Before this change,
11288 the value returned by DES_cbc_cksum() was like the one from
11289 mit_des_cbc_cksum(), except the bytes were swapped.
11291 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11293 * Allow an application to disable the automatic SSL chain building.
11294 Before this a rather primitive chain build was always performed in
11295 ssl3_output_cert_chain(): an application had no way to send the
11296 correct chain if the automatic operation produced an incorrect result.
11298 Now the chain builder is disabled if either:
11300 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11302 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11304 The reasoning behind this is that an application would not want the
11305 auto chain building to take place if extra chain certificates are
11306 present and it might also want a means of sending no additional
11307 certificates (for example the chain has two certificates and the
11312 * Add the possibility to build without the ENGINE framework.
11314 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11316 * Under Win32 gmtime() can return NULL: check return value in
11317 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11321 * DSA routines: under certain error conditions uninitialized BN objects
11322 could be freed. Solution: make sure initialization is performed early
11323 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11324 Nils Larsch <nla@trustcenter.de> via PR#459)
11328 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11329 checked on reconnect on the client side, therefore session resumption
11330 could still fail with a "ssl session id is different" error. This
11331 behaviour is masked when SSL_OP_ALL is used due to
11332 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11333 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11334 followup to PR #377.
11338 * IA-32 assembler support enhancements: unified ELF targets, support
11339 for SCO/Caldera platforms, fix for Cygwin shared build.
11343 * Add support for FreeBSD on sparc64. As a consequence, support for
11344 FreeBSD on non-x86 processors is separate from x86 processors on
11345 the config script, much like the NetBSD support.
11347 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11349 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11351 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11354 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11355 code (06) was taken as the first octet of the session ID and the last
11356 octet was ignored consequently. As a result SSLv2 client side session
11357 caching could not have worked due to the session ID mismatch between
11359 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11364 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11365 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11370 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11371 seems that in spite of existing for more than a year, many application
11372 author have done nothing to provide the necessary callbacks, which
11373 means that this particular engine will not work properly anywhere.
11374 This is a very unfortunate situation which forces us, in the name
11375 of usability, to give the hw_ncipher.c a static lock, which is part
11377 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11378 appear in 0.9.8 or later. We EXPECT application authors to have
11379 dealt properly with this when 0.9.8 is released (unless we actually
11380 make such changes in the libcrypto locking code that changes will
11381 have to be made anyway).
11385 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11386 octets have been read, EOF or an error occurs. Without this change
11387 some truncated ASN1 structures will not produce an error.
11391 * Disable Heimdal support, since it hasn't been fully implemented.
11392 Still give the possibility to force the use of Heimdal, but with
11393 warnings and a request that patches get sent to openssl-dev.
11397 * Add the VC-CE target, introduce the WINCE sysname, and add
11398 INSTALL.WCE and appropriate conditionals to make it build.
11400 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11402 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11403 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11404 edit numbers of the version.
11406 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11408 * Introduce safe string copy and catenation functions
11409 (BUF_strlcpy() and BUF_strlcat()).
11411 *Ben Laurie (CHATS) and Richard Levitte*
11413 * Avoid using fixed-size buffers for one-line DNs.
11415 *Ben Laurie (CHATS)*
11417 * Add BUF_MEM_grow_clean() to avoid information leakage when
11418 resizing buffers containing secrets, and use where appropriate.
11420 *Ben Laurie (CHATS)*
11422 * Avoid using fixed size buffers for configuration file location.
11424 *Ben Laurie (CHATS)*
11426 * Avoid filename truncation for various CA files.
11428 *Ben Laurie (CHATS)*
11430 * Use sizeof in preference to magic numbers.
11432 *Ben Laurie (CHATS)*
11434 * Avoid filename truncation in cert requests.
11436 *Ben Laurie (CHATS)*
11438 * Add assertions to check for (supposedly impossible) buffer
11441 *Ben Laurie (CHATS)*
11443 * Don't cache truncated DNS entries in the local cache (this could
11444 potentially lead to a spoofing attack).
11446 *Ben Laurie (CHATS)*
11448 * Fix various buffers to be large enough for hex/decimal
11449 representations in a platform independent manner.
11451 *Ben Laurie (CHATS)*
11453 * Add CRYPTO_realloc_clean() to avoid information leakage when
11454 resizing buffers containing secrets, and use where appropriate.
11456 *Ben Laurie (CHATS)*
11458 * Add BIO_indent() to avoid much slightly worrying code to do
11461 *Ben Laurie (CHATS)*
11463 * Convert sprintf()/BIO_puts() to BIO_printf().
11465 *Ben Laurie (CHATS)*
11467 * buffer_gets() could terminate with the buffer only half
11470 *Ben Laurie (CHATS)*
11472 * Add assertions to prevent user-supplied crypto functions from
11473 overflowing internal buffers by having large block sizes, etc.
11475 *Ben Laurie (CHATS)*
11477 * New OPENSSL_assert() macro (similar to assert(), but enabled
11480 *Ben Laurie (CHATS)*
11482 * Eliminate unused copy of key in RC4.
11484 *Ben Laurie (CHATS)*
11486 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11488 *Ben Laurie (CHATS)*
11490 * Fix off-by-one error in EGD path.
11492 *Ben Laurie (CHATS)*
11494 * If RANDFILE path is too long, ignore instead of truncating.
11496 *Ben Laurie (CHATS)*
11498 * Eliminate unused and incorrectly sized X.509 structure
11501 *Ben Laurie (CHATS)*
11503 * Eliminate unused and dangerous function knumber().
11505 *Ben Laurie (CHATS)*
11507 * Eliminate unused and dangerous structure, KSSL_ERR.
11509 *Ben Laurie (CHATS)*
11511 * Protect against overlong session ID context length in an encoded
11512 session object. Since these are local, this does not appear to be
11515 *Ben Laurie (CHATS)*
11517 * Change from security patch (see 0.9.6e below) that did not affect
11518 the 0.9.6 release series:
11520 Remote buffer overflow in SSL3 protocol - an attacker could
11521 supply an oversized master key in Kerberos-enabled versions.
11524 *Ben Laurie (CHATS)*
11526 * Change the SSL kerb5 codes to match RFC 2712.
11530 * Make -nameopt work fully for req and add -reqopt switch.
11532 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11534 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11536 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11538 * Make sure tests can be performed even if the corresponding algorithms
11539 have been removed entirely. This was also the last step to make
11540 OpenSSL compilable with DJGPP under all reasonable conditions.
11542 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11544 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11545 to allow version independent disabling of normally unselected ciphers,
11546 which may be activated as a side-effect of selecting a single cipher.
11548 (E.g., cipher list string "RSA" enables ciphersuites that are left
11549 out of "ALL" because they do not provide symmetric encryption.
11550 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11552 *Lutz Jaenicke, Bodo Moeller*
11554 * Add appropriate support for separate platform-dependent build
11555 directories. The recommended way to make a platform-dependent
11556 build directory is the following (tested on Linux), maybe with
11559 # Place yourself outside of the OpenSSL source tree. In
11560 # this example, the environment variable OPENSSL_SOURCE
11561 # is assumed to contain the absolute OpenSSL source directory.
11562 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11563 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11564 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11565 mkdir -p `dirname $F`
11566 ln -s $OPENSSL_SOURCE/$F $F
11569 To be absolutely sure not to disturb the source tree, a "make clean"
11570 is a good thing. If it isn't successful, don't worry about it,
11571 it probably means the source directory is very clean.
11575 * Make sure any ENGINE control commands make local copies of string
11576 pointers passed to them whenever necessary. Otherwise it is possible
11577 the caller may have overwritten (or deallocated) the original string
11578 data when a later ENGINE operation tries to use the stored values.
11580 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11582 * Improve diagnostics in file reading and command-line digests.
11584 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11586 * Add AES modes CFB and OFB to the object database. Correct an
11587 error in AES-CFB decryption.
11591 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11592 allows existing EVP_CIPHER_CTX structures to be reused after
11593 calling `EVP_*Final()`. This behaviour is used by encryption
11594 BIOs and some applications. This has the side effect that
11595 applications must explicitly clean up cipher contexts with
11596 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11600 * Check the values of dna and dnb in bn_mul_recursive before calling
11601 bn_mul_comba (a non zero value means the a or b arrays do not contain
11602 n2 elements) and fallback to bn_mul_normal if either is not zero.
11606 * Fix escaping of non-ASCII characters when using the -subj option
11607 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11611 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11612 form for "surname", serialNumber has no short form.
11613 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11614 therefore remove "mail" short name for "internet 7".
11615 The OID for unique identifiers in X509 certificates is
11616 x500UniqueIdentifier, not uniqueIdentifier.
11617 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11621 * Add an "init" command to the ENGINE config module and auto initialize
11622 ENGINEs. Without any "init" command the ENGINE will be initialized
11623 after all ctrl commands have been executed on it. If init=1 the
11624 ENGINE is initialized at that point (ctrls before that point are run
11625 on the uninitialized ENGINE and after on the initialized one). If
11626 init=0 then the ENGINE will not be initialized at all.
11630 * Fix the 'app_verify_callback' interface so that the user-defined
11631 argument is actually passed to the callback: In the
11632 SSL_CTX_set_cert_verify_callback() prototype, the callback
11633 declaration has been changed from
11636 int (*cb)(X509_STORE_CTX *,void *);
11637 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11638 i=s->ctx->app_verify_callback(&ctx)
11639 has been changed into
11640 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11642 To update applications using SSL_CTX_set_cert_verify_callback(),
11643 a dummy argument can be added to their callback functions.
11645 *D. K. Smetters <smetters@parc.xerox.com>*
11647 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11649 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11651 * Add and OPENSSL_LOAD_CONF define which will cause
11652 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11653 This allows older applications to transparently support certain
11654 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11655 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11656 load the config file and OPENSSL_add_all_algorithms_conf() which will
11657 always load it have also been added.
11661 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11662 Adjust NIDs and EVP layer.
11664 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11666 * Config modules support in openssl utility.
11668 Most commands now load modules from the config file,
11669 though in a few (such as version) this isn't done
11670 because it couldn't be used for anything.
11672 In the case of ca and req the config file used is
11673 the same as the utility itself: that is the -config
11674 command line option can be used to specify an
11679 * Move default behaviour from OPENSSL_config(). If appname is NULL
11680 use "openssl_conf" if filename is NULL use default openssl config file.
11684 * Add an argument to OPENSSL_config() to allow the use of an alternative
11685 config section name. Add a new flag to tolerate a missing config file
11686 and move code to CONF_modules_load_file().
11690 * Support for crypto accelerator cards from Accelerated Encryption
11691 Processing, www.aep.ie. (Use engine 'aep')
11692 The support was copied from 0.9.6c [engine] and adapted/corrected
11693 to work with the new engine framework.
11695 *AEP Inc. and Richard Levitte*
11697 * Support for SureWare crypto accelerator cards from Baltimore
11698 Technologies. (Use engine 'sureware')
11699 The support was copied from 0.9.6c [engine] and adapted
11700 to work with the new engine framework.
11704 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11705 make the newer ENGINE framework commands for the CHIL engine work.
11707 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11709 * Make it possible to produce shared libraries on ReliantUNIX.
11711 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11713 * Add the configuration target debug-linux-ppro.
11714 Make 'openssl rsa' use the general key loading routines
11715 implemented in apps.c, and make those routines able to
11716 handle the key format FORMAT_NETSCAPE and the variant
11719 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11721 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11723 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11725 * Add -keyform to rsautl, and document -engine.
11727 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11729 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11730 BIO_R_NO_SUCH_FILE error code rather than the generic
11731 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11735 * Add new functions
11736 ERR_peek_last_error
11737 ERR_peek_last_error_line
11738 ERR_peek_last_error_line_data.
11739 These are similar to
11741 ERR_peek_error_line
11742 ERR_peek_error_line_data,
11743 but report on the latest error recorded rather than the first one
11744 still in the error queue.
11746 *Ben Laurie, Bodo Moeller*
11748 * default_algorithms option in ENGINE config module. This allows things
11750 default_algorithms = ALL
11751 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11755 * Preliminary ENGINE config module.
11759 * New experimental application configuration code.
11763 * Change the AES code to follow the same name structure as all other
11764 symmetric ciphers, and behave the same way. Move everything to
11765 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11767 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11769 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11771 *Ben Laurie and Theo de Raadt*
11773 * Add option to output public keys in req command.
11775 *Massimiliano Pala madwolf@openca.org*
11777 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11778 (up to about 10% better than before for P-192 and P-224).
11782 * New functions/macros
11784 SSL_CTX_set_msg_callback(ctx, cb)
11785 SSL_CTX_set_msg_callback_arg(ctx, arg)
11786 SSL_set_msg_callback(ssl, cb)
11787 SSL_set_msg_callback_arg(ssl, arg)
11789 to request calling a callback function
11791 void cb(int write_p, int version, int content_type,
11792 const void *buf, size_t len, SSL *ssl, void *arg)
11794 whenever a protocol message has been completely received
11795 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11796 protocol version according to which the SSL library interprets
11797 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11798 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11799 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11800 specification (change_cipher_spec(20), alert(21), handshake(22)).
11801 'buf' and 'len' point to the actual message, 'ssl' to the
11802 SSL object, and 'arg' is the application-defined value set by
11803 SSL[_CTX]_set_msg_callback_arg().
11805 'openssl s_client' and 'openssl s_server' have new '-msg' options
11806 to enable a callback that displays all protocol messages.
11810 * Change the shared library support so shared libraries are built as
11811 soon as the corresponding static library is finished, and thereby get
11812 openssl and the test programs linked against the shared library.
11813 This still only happens when the keyword "shard" has been given to
11814 the configuration scripts.
11816 NOTE: shared library support is still an experimental thing, and
11817 backward binary compatibility is still not guaranteed.
11819 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11821 * Add support for Subject Information Access extension.
11823 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11825 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11826 additional bytes when new memory had to be allocated, not just
11827 when reusing an existing buffer.
11831 * New command line and configuration option 'utf8' for the req command.
11832 This allows field values to be specified as UTF8 strings.
11836 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11837 runs for the former and machine-readable output for the latter.
11841 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11842 of the e-mail address in the DN (i.e., it will go into a certificate
11843 extension only). The new configuration file option 'email_in_dn = no'
11844 has the same effect.
11846 *Massimiliano Pala madwolf@openca.org*
11848 * Change all functions with names starting with `des_` to be starting
11849 with `DES_` instead. Add wrappers that are compatible with libdes,
11850 but are named `_ossl_old_des_*`. Finally, add macros that map the
11851 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11852 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11853 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11856 Since we provide two compatibility mappings, the user needs to
11857 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11858 compatibility is desired. The default (i.e., when that macro
11859 isn't defined) is OpenSSL 0.9.6c compatibility.
11861 There are also macros that enable and disable the support of old
11862 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11863 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11864 are defined, the default will apply: to support the old des routines.
11866 In either case, one must include openssl/des.h to get the correct
11867 definitions. Do not try to just include openssl/des_old.h, that
11870 NOTE: This is a major break of an old API into a new one. Software
11871 authors are encouraged to switch to the `DES_` style functions. Some
11872 time in the future, des_old.h and the libdes compatibility functions
11873 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11874 default), and then completely removed.
11878 * Test for certificates which contain unsupported critical extensions.
11879 If such a certificate is found during a verify operation it is
11880 rejected by default: this behaviour can be overridden by either
11881 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11882 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11883 X509_supported_extension() has also been added which returns 1 if a
11884 particular extension is supported.
11888 * Modify the behaviour of EVP cipher functions in similar way to digests
11889 to retain compatibility with existing code.
11893 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11894 compatibility with existing code. In particular the 'ctx' parameter does
11895 not have to be to be initialized before the call to EVP_DigestInit() and
11896 it is tidied up after a call to EVP_DigestFinal(). New function
11897 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11898 EVP_MD_CTX_copy() changed to not require the destination to be
11899 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11900 requires the destination to be valid.
11902 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11903 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11907 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11908 so that complete 'Handshake' protocol structures are kept in memory
11909 instead of overwriting 'msg_type' and 'length' with 'body' data.
11913 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11915 *Massimo Santin via Richard Levitte*
11917 * Major restructuring to the underlying ENGINE code. This includes
11918 reduction of linker bloat, separation of pure "ENGINE" manipulation
11919 (initialisation, etc) from functionality dealing with implementations
11920 of specific crypto interfaces. This change also introduces integrated
11921 support for symmetric ciphers and digest implementations - so ENGINEs
11922 can now accelerate these by providing EVP_CIPHER and EVP_MD
11923 implementations of their own. This is detailed in
11924 [crypto/engine/README.md](crypto/engine/README.md)
11925 as it couldn't be adequately described here. However, there are a few
11926 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11927 were changed in the original introduction of ENGINE code have now
11928 reverted back - the hooking from this code to ENGINE is now a good
11929 deal more passive and at run-time, operations deal directly with
11930 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11931 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11932 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11933 they were not being used by the framework as there is no concept of a
11934 BIGNUM_METHOD and they could not be generalised to the new
11935 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11936 ENGINE_cpy() has been removed as it cannot be consistently defined in
11941 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11945 * Change mkdef.pl to sort symbols that get the same entry number,
11946 and make sure the automatically generated functions `ERR_load_*`
11947 become part of libeay.num as well.
11951 * New function SSL_renegotiate_pending(). This returns true once
11952 renegotiation has been requested (either SSL_renegotiate() call
11953 or HelloRequest/ClientHello received from the peer) and becomes
11954 false once a handshake has been completed.
11955 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11956 sends a HelloRequest, but does not ensure that a handshake takes
11957 place. SSL_renegotiate_pending() is useful for checking if the
11958 client has followed the request.)
11962 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11963 By default, clients may request session resumption even during
11964 renegotiation (if session ID contexts permit); with this option,
11965 session resumption is possible only in the first handshake.
11967 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11968 more bits available for options that should not be part of
11969 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11973 * Add some demos for certificate and certificate request creation.
11977 * Make maximum certificate chain size accepted from the peer application
11978 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11979 "Douglas E. Engert" <deengert@anl.gov>.
11983 * Add support for shared libraries for Unixware-7
11984 (Boyd Lynn Gerber <gerberb@zenez.com>).
11988 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11989 be done prior to destruction. Use this to unload error strings from
11990 ENGINEs that load their own error strings. NB: This adds two new API
11991 functions to "get" and "set" this destroy handler in an ENGINE.
11995 * Alter all existing ENGINE implementations (except "openssl" and
11996 "openbsd") to dynamically instantiate their own error strings. This
11997 makes them more flexible to be built both as statically-linked ENGINEs
11998 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11999 Also, add stub code to each that makes building them as self-contained
12000 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12004 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12005 implementations into applications that are completely implemented in
12006 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12007 commands that can be used to configure what shared-library to load and
12008 to control aspects of the way it is handled. Also, made an update to
12009 the [README-Engine.md](README-Engine.md) file
12010 that brings its information up-to-date and
12011 provides some information and instructions on the "dynamic" ENGINE
12012 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12016 * Make it possible to unload ranges of ERR strings with a new
12017 "ERR_unload_strings" function.
12021 * Add a copy() function to EVP_MD.
12025 * Make EVP_MD routines take a context pointer instead of just the
12026 md_data void pointer.
12030 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12031 that the digest can only process a single chunk of data
12032 (typically because it is provided by a piece of
12033 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12034 is only going to provide a single chunk of data, and hence the
12035 framework needn't accumulate the data for oneshot drivers.
12039 * As with "ERR", make it possible to replace the underlying "ex_data"
12040 functions. This change also alters the storage and management of global
12041 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12042 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12043 index counters. The API functions that use this state have been changed
12044 to take a "class_index" rather than pointers to the class's local STACK
12045 and counter, and there is now an API function to dynamically create new
12046 classes. This centralisation allows us to (a) plug a lot of the
12047 thread-safety problems that existed, and (b) makes it possible to clean
12048 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12049 such data would previously have always leaked in application code and
12050 workarounds were in place to make the memory debugging turn a blind eye
12051 to it. Application code that doesn't use this new function will still
12052 leak as before, but their memory debugging output will announce it now
12053 rather than letting it slide.
12055 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12056 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12057 has a return value to indicate success or failure.
12061 * Make it possible to replace the underlying "ERR" functions such that the
12062 global state (2 LHASH tables and 2 locks) is only used by the "default"
12063 implementation. This change also adds two functions to "get" and "set"
12064 the implementation prior to it being automatically set the first time
12065 any other ERR function takes place. Ie. an application can call "get",
12066 pass the return value to a module it has just loaded, and that module
12067 can call its own "set" function using that value. This means the
12068 module's "ERR" operations will use (and modify) the error state in the
12069 application and not in its own statically linked copy of OpenSSL code.
12073 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12074 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12075 the operation, and provides a more encapsulated way for external code
12076 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12077 to use these functions rather than manually incrementing the counts.
12079 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12083 * Add EVP test program.
12087 * Add symmetric cipher support to ENGINE. Expect the API to change!
12091 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12092 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12093 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12094 These allow a CRL to be built without having to access X509_CRL fields
12095 directly. Modify 'ca' application to use new functions.
12099 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12100 bug workarounds. Rollback attack detection is a security feature.
12101 The problem will only arise on OpenSSL servers when TLSv1 is not
12102 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12103 Software authors not wanting to support TLSv1 will have special reasons
12104 for their choice and can explicitly enable this option.
12106 *Bodo Moeller, Lutz Jaenicke*
12108 * Rationalise EVP so it can be extended: don't include a union of
12109 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12110 (similar to those existing for EVP_CIPHER_CTX).
12115 EVP_MD_CTX_init(&md); /* new function call */
12116 EVP_DigestInit(&md, EVP_sha1());
12117 EVP_DigestUpdate(&md, in, len);
12118 EVP_DigestFinal(&md, out, NULL);
12119 EVP_MD_CTX_cleanup(&md); /* new function call */
12123 * Make DES key schedule conform to the usual scheme, as well as
12124 correcting its structure. This means that calls to DES functions
12125 now have to pass a pointer to a des_key_schedule instead of a
12126 plain des_key_schedule (which was actually always a pointer
12129 des_key_schedule ks;
12131 des_set_key_checked(..., &ks);
12132 des_ncbc_encrypt(..., &ks, ...);
12134 (Note that a later change renames 'des_...' into 'DES_...'.)
12138 * Initial reduction of linker bloat: the use of some functions, such as
12139 PEM causes large amounts of unused functions to be linked in due to
12140 poor organisation. For example pem_all.c contains every PEM function
12141 which has a knock on effect of linking in large amounts of (unused)
12142 ASN1 code. Grouping together similar functions and splitting unrelated
12143 functions prevents this.
12147 * Cleanup of EVP macros.
12151 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12152 correct `_ecb suffix`.
12156 * Add initial OCSP responder support to ocsp application. The
12157 revocation information is handled using the text based index
12158 use by the ca application. The responder can either handle
12159 requests generated internally, supplied in files (for example
12160 via a CGI script) or using an internal minimal server.
12164 * Add configuration choices to get zlib compression for TLS.
12168 * Changes to Kerberos SSL for RFC 2712 compliance:
12169 1. Implemented real KerberosWrapper, instead of just using
12170 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12171 2. Implemented optional authenticator field of KerberosWrapper.
12173 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12174 and authenticator structs; see crypto/krb5/.
12176 Generalized Kerberos calls to support multiple Kerberos libraries.
12177 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12178 via Richard Levitte*
12180 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12181 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12182 values for each of the key sizes rather than having just
12183 parameters (and 'speed' generating keys each time).
12187 * Speed up EVP routines.
12190 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12191 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12192 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12193 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12195 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12196 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12197 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12200 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12202 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12206 * Added the OS2-EMX target.
12208 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12210 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12211 to support NCONF routines in extension code. New function CONF_set_nconf()
12212 to allow functions which take an NCONF to also handle the old LHASH
12213 structure: this means that the old CONF compatible routines can be
12214 retained (in particular wrt extensions) without having to duplicate the
12215 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12219 * Enhance the general user interface with mechanisms for inner control
12220 and with possibilities to have yes/no kind of prompts.
12224 * Change all calls to low level digest routines in the library and
12225 applications to use EVP. Add missing calls to HMAC_cleanup() and
12226 don't assume HMAC_CTX can be copied using memcpy().
12228 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12230 * Add the possibility to control engines through control names but with
12231 arbitrary arguments instead of just a string.
12232 Change the key loaders to take a UI_METHOD instead of a callback
12233 function pointer. NOTE: this breaks binary compatibility with earlier
12234 versions of OpenSSL [engine].
12235 Adapt the nCipher code for these new conditions and add a card insertion
12240 * Enhance the general user interface with mechanisms to better support
12241 dialog box interfaces, application-defined prompts, the possibility
12242 to use defaults (for example default passwords from somewhere else)
12243 and interrupts/cancellations.
12247 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12248 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12252 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12253 tidy up some unnecessarily weird code in 'sk_new()').
12255 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12257 * Change the key loading routines for ENGINEs to use the same kind
12258 callback (pem_password_cb) as all other routines that need this
12263 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12264 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12265 than this minimum value is recommended.
12269 * New random seeder for OpenVMS, using the system process statistics
12270 that are easily reachable.
12274 * Windows apparently can't transparently handle global
12275 variables defined in DLLs. Initialisations such as:
12277 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12279 won't compile. This is used by the any applications that need to
12280 declare their own ASN1 modules. This was fixed by adding the option
12281 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12282 needed for static libraries under Win32.
12286 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12287 setting of purpose and trust fields. New X509_STORE trust and
12288 purpose functions and tidy up setting in other SSL functions.
12292 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12293 structure. These are inherited by X509_STORE_CTX when it is
12294 initialised. This allows various defaults to be set in the
12295 X509_STORE structure (such as flags for CRL checking and custom
12296 purpose or trust settings) for functions which only use X509_STORE_CTX
12297 internally such as S/MIME.
12299 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12300 trust settings if they are not set in X509_STORE. This allows X509_STORE
12301 purposes and trust (in S/MIME for example) to override any set by default.
12303 Add command line options for CRL checking to smime, s_client and s_server
12308 * Initial CRL based revocation checking. If the CRL checking flag(s)
12309 are set then the CRL is looked up in the X509_STORE structure and
12310 its validity and signature checked, then if the certificate is found
12311 in the CRL the verify fails with a revoked error.
12313 Various new CRL related callbacks added to X509_STORE_CTX structure.
12315 Command line options added to 'verify' application to support this.
12317 This needs some additional work, such as being able to handle multiple
12318 CRLs with different times, extension based lookup (rather than just
12319 by subject name) and ultimately more complete V2 CRL extension
12324 * Add a general user interface API (crypto/ui/). This is designed
12325 to replace things like des_read_password and friends (backward
12326 compatibility functions using this new API are provided).
12327 The purpose is to remove prompting functions from the DES code
12328 section as well as provide for prompting through dialog boxes in
12329 a window system and the like.
12333 * Add "ex_data" support to ENGINE so implementations can add state at a
12334 per-structure level rather than having to store it globally.
12338 * Make it possible for ENGINE structures to be copied when retrieved by
12339 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12340 This causes the "original" ENGINE structure to act like a template,
12341 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12342 operational state can be localised to each ENGINE structure, despite the
12343 fact they all share the same "methods". New ENGINE structures returned in
12344 this case have no functional references and the return value is the single
12345 structural reference. This matches the single structural reference returned
12346 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12351 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12352 needs to match any other type at all we need to manually clear the
12357 * Changes to the "openssl engine" utility to include;
12358 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12359 about an ENGINE's available control commands.
12360 - executing control commands from command line arguments using the
12361 '-pre' and '-post' switches. '-post' is only used if '-t' is
12362 specified and the ENGINE is successfully initialised. The syntax for
12363 the individual commands are colon-separated, for example;
12364 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12368 * New dynamic control command support for ENGINEs. ENGINEs can now
12369 declare their own commands (numbers), names (strings), descriptions,
12370 and input types for run-time discovery by calling applications. A
12371 subset of these commands are implicitly classed as "executable"
12372 depending on their input type, and only these can be invoked through
12373 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12374 can be based on user input, config files, etc). The distinction is
12375 that "executable" commands cannot return anything other than a boolean
12376 result and can only support numeric or string input, whereas some
12377 discoverable commands may only be for direct use through
12378 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12379 pointers, or other custom uses. The "executable" commands are to
12380 support parameterisations of ENGINE behaviour that can be
12381 unambiguously defined by ENGINEs and used consistently across any
12382 OpenSSL-based application. Commands have been added to all the
12383 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12384 control over shared-library paths without source code alterations.
12388 * Changed all ENGINE implementations to dynamically allocate their
12389 ENGINEs rather than declaring them statically. Apart from this being
12390 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12391 this also allows the implementations to compile without using the
12392 internal engine_int.h header.
12396 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12397 'const' value. Any code that should be able to modify a RAND_METHOD
12398 should already have non-const pointers to it (ie. they should only
12399 modify their own ones).
12403 * Made a variety of little tweaks to the ENGINE code.
12404 - "atalla" and "ubsec" string definitions were moved from header files
12405 to C code. "nuron" string definitions were placed in variables
12406 rather than hard-coded - allowing parameterisation of these values
12407 later on via ctrl() commands.
12408 - Removed unused "#if 0"'d code.
12409 - Fixed engine list iteration code so it uses ENGINE_free() to release
12410 structural references.
12411 - Constified the RAND_METHOD element of ENGINE structures.
12412 - Constified various get/set functions as appropriate and added
12413 missing functions (including a catch-all ENGINE_cpy that duplicates
12414 all ENGINE values onto a new ENGINE except reference counts/state).
12415 - Removed NULL parameter checks in get/set functions. Setting a method
12416 or function to NULL is a way of cancelling out a previously set
12417 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12418 and doesn't justify the extra error symbols and code.
12419 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12420 flags from engine_int.h to engine.h.
12421 - Changed prototypes for ENGINE handler functions (init(), finish(),
12422 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12426 * Implement binary inversion algorithm for BN_mod_inverse in addition
12427 to the algorithm using long division. The binary algorithm can be
12428 used only if the modulus is odd. On 32-bit systems, it is faster
12429 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12430 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12431 up to 450 bits. In 64-bit environments, the binary algorithm
12432 appears to be advantageous for much longer moduli; here we use it
12433 for moduli up to 2048 bits.
12437 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12438 could not support the combine flag in choice fields.
12442 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12443 extensions from a certificate request to the certificate.
12447 * Allow multiple 'certopt' and 'nameopt' options to be separated
12448 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12449 file: this allows the display of the certificate about to be
12450 signed to be customised, to allow certain fields to be included
12451 or excluded and extension details. The old system didn't display
12452 multicharacter strings properly, omitted fields not in the policy
12453 and couldn't display additional details such as extensions.
12457 * Function EC_POINTs_mul for multiple scalar multiplication
12458 of an arbitrary number of elliptic curve points
12459 \sum scalars[i]*points[i],
12460 optionally including the generator defined for the EC_GROUP:
12461 scalar*generator + \sum scalars[i]*points[i].
12463 EC_POINT_mul is a simple wrapper function for the typical case
12464 that the point list has just one item (besides the optional
12469 * First EC_METHODs for curves over GF(p):
12471 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12472 operations and provides various method functions that can also
12473 operate with faster implementations of modular arithmetic.
12475 EC_GFp_mont_method() reuses most functions that are part of
12476 EC_GFp_simple_method, but uses Montgomery arithmetic.
12478 *Bodo Moeller; point addition and point doubling
12479 implementation directly derived from source code provided by
12480 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12482 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12483 crypto/ec/ec_lib.c):
12485 Curves are EC_GROUP objects (with an optional group generator)
12486 based on EC_METHODs that are built into the library.
12488 Points are EC_POINT objects based on EC_GROUP objects.
12490 Most of the framework would be able to handle curves over arbitrary
12491 finite fields, but as there are no obvious types for fields other
12492 than GF(p), some functions are limited to that for now.
12496 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12497 that the file contains a complete HTTP response.
12501 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12502 change the def and num file printf format specifier from "%-40sXXX"
12503 to "%-39s XXX". The latter will always guarantee a space after the
12504 field while the former will cause them to run together if the field
12505 is 40 of more characters long.
12509 * Constify the cipher and digest 'method' functions and structures
12510 and modify related functions to take constant EVP_MD and EVP_CIPHER
12515 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12516 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12520 * Modify `EVP_Digest*()` routines so they now return values. Although the
12521 internal software routines can never fail additional hardware versions
12526 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12528 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12529 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12532 ERR_R_NESTED_ASN1_ERROR
12534 ERR_R_MISSING_ASN1_EOS
12535 were 4 .. 9, conflicting with
12536 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12538 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12539 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12541 Add new error code 'ERR_R_INTERNAL_ERROR'.
12545 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12550 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12551 sets the subject name for a new request or supersedes the
12552 subject name in a given request. Formats that can be parsed are
12553 'CN=Some Name, OU=myOU, C=IT'
12555 'CN=Some Name/OU=myOU/C=IT'.
12557 Add options '-batch' and '-verbose' to 'openssl req'.
12559 *Massimiliano Pala <madwolf@hackmasters.net>*
12561 * Introduce the possibility to access global variables through
12562 functions on platform were that's the best way to handle exporting
12563 global variables in shared libraries. To enable this functionality,
12564 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12565 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12566 is normally done by Configure or something similar).
12568 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12569 in the source file (foo.c) like this:
12571 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12572 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12574 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12575 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12577 OPENSSL_DECLARE_GLOBAL(int,foo);
12578 #define foo OPENSSL_GLOBAL_REF(foo)
12579 OPENSSL_DECLARE_GLOBAL(double,bar);
12580 #define bar OPENSSL_GLOBAL_REF(bar)
12582 The #defines are very important, and therefore so is including the
12583 header file everywhere where the defined globals are used.
12585 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12586 of ASN.1 items, but that structure is a bit different.
12588 The largest change is in util/mkdef.pl which has been enhanced with
12589 better and easier to understand logic to choose which symbols should
12590 go into the Windows .def files as well as a number of fixes and code
12591 cleanup (among others, algorithm keywords are now sorted
12592 lexicographically to avoid constant rewrites).
12596 * In BN_div() keep a copy of the sign of 'num' before writing the
12597 result to 'rm' because if rm==num the value will be overwritten
12598 and produce the wrong result if 'num' is negative: this caused
12599 problems with BN_mod() and BN_nnmod().
12603 * Function OCSP_request_verify(). This checks the signature on an
12604 OCSP request and verifies the signer certificate. The signer
12605 certificate is just checked for a generic purpose and OCSP request
12610 * Add OCSP_check_validity() function to check the validity of OCSP
12611 responses. OCSP responses are prepared in real time and may only
12612 be a few seconds old. Simply checking that the current time lies
12613 between thisUpdate and nextUpdate max reject otherwise valid responses
12614 caused by either OCSP responder or client clock inaccuracy. Instead
12615 we allow thisUpdate and nextUpdate to fall within a certain period of
12616 the current time. The age of the response can also optionally be
12617 checked. Two new options -validity_period and -status_age added to
12622 * If signature or public key algorithm is unrecognized print out its
12623 OID rather that just UNKNOWN.
12627 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12628 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12629 ID to be generated from the issuer certificate alone which can then be
12630 passed to OCSP_id_issuer_cmp().
12634 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12635 ASN1 modules to export functions returning ASN1_ITEM pointers
12636 instead of the ASN1_ITEM structures themselves. This adds several
12637 new macros which allow the underlying ASN1 function/structure to
12638 be accessed transparently. As a result code should not use ASN1_ITEM
12639 references directly (such as &X509_it) but instead use the relevant
12640 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12641 use of the new ASN1 code on platforms where exporting structures
12642 is problematical (for example in shared libraries) but exporting
12643 functions returning pointers to structures is not.
12647 * Add support for overriding the generation of SSL/TLS session IDs.
12648 These callbacks can be registered either in an SSL_CTX or per SSL.
12649 The purpose of this is to allow applications to control, if they wish,
12650 the arbitrary values chosen for use as session IDs, particularly as it
12651 can be useful for session caching in multiple-server environments. A
12652 command-line switch for testing this (and any client code that wishes
12653 to use such a feature) has been added to "s_server".
12655 *Geoff Thorpe, Lutz Jaenicke*
12657 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12658 of the form `#if defined(...) || defined(...) || ...` and
12659 `#if !defined(...) && !defined(...) && ...`. This also avoids
12660 the growing number of special cases it was previously handling.
12664 * Make all configuration macros available for application by making
12665 sure they are available in opensslconf.h, by giving them names starting
12666 with `OPENSSL_` to avoid conflicts with other packages and by making
12667 sure e_os2.h will cover all platform-specific cases together with
12669 Additionally, it is now possible to define configuration/platform-
12670 specific names (called "system identities"). In the C code, these
12671 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12672 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12673 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12678 * New option -set_serial to 'req' and 'x509' this allows the serial
12679 number to use to be specified on the command line. Previously self
12680 signed certificates were hard coded with serial number 0 and the
12681 CA options of 'x509' had to use a serial number in a file which was
12686 * New options to 'ca' utility to support V2 CRL entry extensions.
12687 Currently CRL reason, invalidity date and hold instruction are
12688 supported. Add new CRL extensions to V3 code and some new objects.
12692 * New function EVP_CIPHER_CTX_set_padding() this is used to
12693 disable standard block padding (aka PKCS#5 padding) in the EVP
12694 API, which was previously mandatory. This means that the data is
12695 not padded in any way and so the total length much be a multiple
12696 of the block size, otherwise an error occurs.
12700 * Initial (incomplete) OCSP SSL support.
12704 * New function OCSP_parse_url(). This splits up a URL into its host,
12705 port and path components: primarily to parse OCSP URLs. New -url
12706 option to ocsp utility.
12710 * New nonce behavior. The return value of OCSP_check_nonce() now
12711 reflects the various checks performed. Applications can decide
12712 whether to tolerate certain situations such as an absent nonce
12713 in a response when one was present in a request: the ocsp application
12714 just prints out a warning. New function OCSP_add1_basic_nonce()
12715 this is to allow responders to include a nonce in a response even if
12716 the request is nonce-less.
12720 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12721 skipped when using openssl x509 multiple times on a single input file,
12722 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12726 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12727 set string type: to handle setting ASN1_TIME structures. Fix ca
12728 utility to correctly initialize revocation date of CRLs.
12732 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12733 the clients preferred ciphersuites and rather use its own preferences.
12734 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12735 Internet Explorer by ensuring unchanged hash method during stepup.
12736 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12740 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12741 to aes and add a new 'exist' option to print out symbols that don't
12746 * Additional options to ocsp utility to allow flags to be set and
12747 additional certificates supplied.
12751 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12752 OCSP client a number of certificate to only verify the response
12757 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12758 handle the new API. Currently only ECB, CBC modes supported. Add new
12761 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12762 Encryption Standard (AES) Ciphersuites for Transport Layer
12763 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12764 not enabled by default and were not part of the "ALL" ciphersuite
12765 alias because they were not yet official; they could be
12766 explicitly requested by specifying the "AESdraft" ciphersuite
12767 group alias. In the final release of OpenSSL 0.9.7, the group
12768 alias is called "AES" and is part of "ALL".)
12770 *Ben Laurie, Steve Henson, Bodo Moeller*
12772 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12773 request to response.
12777 * Functions for OCSP responders. OCSP_request_onereq_count(),
12778 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12779 extract information from a certificate request. OCSP_response_create()
12780 creates a response and optionally adds a basic response structure.
12781 OCSP_basic_add1_status() adds a complete single response to a basic
12782 response and returns the OCSP_SINGLERESP structure just added (to allow
12783 extensions to be included for example). OCSP_basic_add1_cert() adds a
12784 certificate to a basic response and OCSP_basic_sign() signs a basic
12785 response with various flags. New helper functions ASN1_TIME_check()
12786 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12787 (converts ASN1_TIME to GeneralizedTime).
12791 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12792 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12793 structure from a certificate. X509_pubkey_digest() digests the public_key
12794 contents: this is used in various key identifiers.
12798 * Make sk_sort() tolerate a NULL argument.
12800 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12802 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12803 passed by the function are trusted implicitly. If any of them signed the
12804 response then it is assumed to be valid and is not verified.
12808 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12809 to data. This was previously part of the PKCS7 ASN1 code. This
12810 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12811 *Steve Henson, reported by Kenneth R. Robinette
12812 <support@securenetterm.com>*
12814 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12815 routines: without these tracing memory leaks is very painful.
12816 Fix leaks in PKCS12 and PKCS7 routines.
12820 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12821 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12822 effectively meant GeneralizedTime would never be used. Now it
12823 is initialised to -1 but X509_time_adj() now has to check the value
12824 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12825 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12826 *Steve Henson, reported by Kenneth R. Robinette
12827 <support@securenetterm.com>*
12829 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12830 result in a zero length in the ASN1_INTEGER structure which was
12831 not consistent with the structure when d2i_ASN1_INTEGER() was used
12832 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12833 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12834 where it did not print out a minus for negative ASN1_INTEGER.
12838 * Add summary printout to ocsp utility. The various functions which
12839 convert status values to strings have been renamed to:
12840 OCSP_response_status_str(), OCSP_cert_status_str() and
12841 OCSP_crl_reason_str() and are no longer static. New options
12842 to verify nonce values and to disable verification. OCSP response
12843 printout format cleaned up.
12847 * Add additional OCSP certificate checks. These are those specified
12848 in RFC2560. This consists of two separate checks: the CA of the
12849 certificate being checked must either be the OCSP signer certificate
12850 or the issuer of the OCSP signer certificate. In the latter case the
12851 OCSP signer certificate must contain the OCSP signing extended key
12852 usage. This check is performed by attempting to match the OCSP
12853 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12854 in the OCSP_CERTID structures of the response.
12858 * Initial OCSP certificate verification added to OCSP_basic_verify()
12859 and related routines. This uses the standard OpenSSL certificate
12860 verify routines to perform initial checks (just CA validity) and
12861 to obtain the certificate chain. Then additional checks will be
12862 performed on the chain. Currently the root CA is checked to see
12863 if it is explicitly trusted for OCSP signing. This is used to set
12864 a root CA as a global signing root: that is any certificate that
12865 chains to that CA is an acceptable OCSP signing certificate.
12869 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12870 extensions from a separate configuration file.
12871 As when reading extensions from the main configuration file,
12872 the '-extensions ...' option may be used for specifying the
12875 *Massimiliano Pala <madwolf@comune.modena.it>*
12877 * New OCSP utility. Allows OCSP requests to be generated or
12878 read. The request can be sent to a responder and the output
12879 parsed, outputted or printed in text form. Not complete yet:
12880 still needs to check the OCSP response validity.
12884 * New subcommands for 'openssl ca':
12885 `openssl ca -status <serial>` prints the status of the cert with
12886 the given serial number (according to the index file).
12887 `openssl ca -updatedb` updates the expiry status of certificates
12890 *Massimiliano Pala <madwolf@comune.modena.it>*
12892 * New '-newreq-nodes' command option to CA.pl. This is like
12893 '-newreq', but calls 'openssl req' with the '-nodes' option
12894 so that the resulting key is not encrypted.
12896 *Damien Miller <djm@mindrot.org>*
12898 * New configuration for the GNU Hurd.
12900 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12902 * Initial code to implement OCSP basic response verify. This
12903 is currently incomplete. Currently just finds the signer's
12904 certificate and verifies the signature on the response.
12908 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12909 value of OPENSSLDIR. This is available via the new '-d' option
12910 to 'openssl version', and is also included in 'openssl version -a'.
12914 * Allowing defining memory allocation callbacks that will be given
12915 file name and line number information in additional arguments
12916 (a `const char*` and an int). The basic functionality remains, as
12917 well as the original possibility to just replace malloc(),
12918 realloc() and free() by functions that do not know about these
12919 additional arguments. To register and find out the current
12920 settings for extended allocation functions, the following
12921 functions are provided:
12923 CRYPTO_set_mem_ex_functions
12924 CRYPTO_set_locked_mem_ex_functions
12925 CRYPTO_get_mem_ex_functions
12926 CRYPTO_get_locked_mem_ex_functions
12928 These work the same way as CRYPTO_set_mem_functions and friends.
12929 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12930 extended allocation function is enabled.
12931 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12932 a conventional allocation function is enabled.
12934 *Richard Levitte, Bodo Moeller*
12936 * Finish off removing the remaining LHASH function pointer casts.
12937 There should no longer be any prototype-casting required when using
12938 the LHASH abstraction, and any casts that remain are "bugs". See
12939 the callback types and macros at the head of lhash.h for details
12940 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12944 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12945 If /dev/[u]random devices are not available or do not return enough
12946 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12948 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12949 /etc/entropy will be queried once each in this sequence, querying stops
12950 when enough entropy was collected without querying more sockets.
12954 * Change the Unix RAND_poll() variant to be able to poll several
12955 random devices, as specified by DEVRANDOM, until a sufficient amount
12956 of data has been collected. We spend at most 10 ms on each file
12957 (select timeout) and read in non-blocking mode. DEVRANDOM now
12958 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12959 (previously it was just the string "/dev/urandom"), so on typical
12960 platforms the 10 ms delay will never occur.
12961 Also separate out the Unix variant to its own file, rand_unix.c.
12962 For VMS, there's a currently-empty rand_vms.c.
12966 * Move OCSP client related routines to ocsp_cl.c. These
12967 provide utility functions which an application needing
12968 to issue a request to an OCSP responder and analyse the
12969 response will typically need: as opposed to those which an
12970 OCSP responder itself would need which will be added later.
12972 OCSP_request_sign() signs an OCSP request with an API similar
12973 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12974 response. OCSP_response_get1_basic() extracts basic response
12975 from response. OCSP_resp_find_status(): finds and extracts status
12976 information from an OCSP_CERTID structure (which will be created
12977 when the request structure is built). These are built from lower
12978 level functions which work on OCSP_SINGLERESP structures but
12979 won't normally be used unless the application wishes to examine
12980 extensions in the OCSP response for example.
12982 Replace nonce routines with a pair of functions.
12983 OCSP_request_add1_nonce() adds a nonce value and optionally
12984 generates a random value. OCSP_check_nonce() checks the
12985 validity of the nonce in an OCSP response.
12989 * Change function OCSP_request_add() to OCSP_request_add0_id().
12990 This doesn't copy the supplied OCSP_CERTID and avoids the
12991 need to free up the newly created id. Change return type
12992 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12993 This can then be used to add extensions to the request.
12994 Deleted OCSP_request_new(), since most of its functionality
12995 is now in OCSP_REQUEST_new() (and the case insensitive name
12996 clash) apart from the ability to set the request name which
12997 will be added elsewhere.
13001 * Update OCSP API. Remove obsolete extensions argument from
13002 various functions. Extensions are now handled using the new
13003 OCSP extension code. New simple OCSP HTTP function which
13004 can be used to send requests and parse the response.
13008 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13009 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13010 uses the special reorder version of SET OF to sort the attributes
13011 and reorder them to match the encoded order. This resolves a long
13012 standing problem: a verify on a PKCS7 structure just after signing
13013 it used to fail because the attribute order did not match the
13014 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13015 it uses the received order. This is necessary to tolerate some broken
13016 software that does not order SET OF. This is handled by encoding
13017 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13018 to produce the required SET OF.
13022 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13023 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13024 files to get correct declarations of the ASN.1 item variables.
13028 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13029 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13030 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13031 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13032 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13033 ASN1_ITEM and no wrapper functions.
13037 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13038 replace the old function pointer based I/O routines. Change most of
13039 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13043 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13044 lines, recognize more "algorithms" that can be deselected, and make
13045 it complain about algorithm deselection that isn't recognised.
13049 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13050 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13051 to use new functions. Add NO_ASN1_OLD which can be set to remove
13052 some old style ASN1 functions: this can be used to determine if old
13053 code will still work when these eventually go away.
13057 * New extension functions for OCSP structures, these follow the
13058 same conventions as certificates and CRLs.
13062 * New function X509V3_add1_i2d(). This automatically encodes and
13063 adds an extension. Its behaviour can be customised with various
13064 flags to append, replace or delete. Various wrappers added for
13065 certificates and CRLs.
13069 * Fix to avoid calling the underlying ASN1 print routine when
13070 an extension cannot be parsed. Correct a typo in the
13071 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13075 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13076 entries for variables.
13080 * Add functionality to apps/openssl.c for detecting locking
13081 problems: As the program is single-threaded, all we have
13082 to do is register a locking callback using an array for
13083 storing which locks are currently held by the program.
13087 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13088 SSL_get_ex_data_X509_STORE_idx(), which is used in
13089 ssl_verify_cert_chain() and thus can be called at any time
13090 during TLS/SSL handshakes so that thread-safety is essential.
13091 Unfortunately, the ex_data design is not at all suited
13092 for multi-threaded use, so it probably should be abolished.
13096 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13098 *Broadcom, tweaked and integrated by Geoff Thorpe*
13100 * Move common extension printing code to new function
13101 X509V3_print_extensions(). Reorganise OCSP print routines and
13102 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13106 * New function X509_signature_print() to remove duplication in some
13111 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13112 set (this was treated exactly the same as SET OF previously). This
13113 is used to reorder the STACK representing the structure to match the
13114 encoding. This will be used to get round a problem where a PKCS7
13115 structure which was signed could not be verified because the STACK
13116 order did not reflect the encoded order.
13120 * Reimplement the OCSP ASN1 module using the new code.
13124 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13125 for its ASN1 operations. The old style function pointers still exist
13126 for now but they will eventually go away.
13130 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13131 completely replaces the old ASN1 functionality with a table driven
13132 encoder and decoder which interprets an ASN1_ITEM structure describing
13133 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13134 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13135 has also been converted to the new form.
13139 * Change BN_mod_exp_recp so that negative moduli are tolerated
13140 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13141 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13142 for negative moduli.
13146 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13147 of not touching the result's sign bit.
13151 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13156 * Changed the LHASH code to use prototypes for callbacks, and created
13157 macros to declare and implement thin (optionally static) functions
13158 that provide type-safety and avoid function pointer casting for the
13159 type-specific callbacks.
13163 * Added Kerberos Cipher Suites to be used with TLS, as written in
13165 *Veers Staats <staatsvr@asc.hpc.mil>,
13166 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13168 * Reformat the FAQ so the different questions and answers can be divided
13169 in sections depending on the subject.
13173 * Have the zlib compression code load ZLIB.DLL dynamically under
13178 * New function BN_mod_sqrt for computing square roots modulo a prime
13179 (using the probabilistic Tonelli-Shanks algorithm unless
13180 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13181 be handled deterministically).
13183 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13185 * Make BN_mod_inverse faster by explicitly handling small quotients
13186 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13187 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13191 * New function BN_kronecker.
13195 * Fix BN_gcd so that it works on negative inputs; the result is
13196 positive unless both parameters are zero.
13197 Previously something reasonably close to an infinite loop was
13198 possible because numbers could be growing instead of shrinking
13199 in the implementation of Euclid's algorithm.
13203 * Fix BN_is_word() and BN_is_one() macros to take into account the
13204 sign of the number in question.
13206 Fix BN_is_word(a,w) to work correctly for w == 0.
13208 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13209 because its test if the absolute value of 'a' equals 'w'.
13210 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13211 it exists mostly for use in the implementations of BN_is_zero(),
13212 BN_is_one(), and BN_is_word().
13216 * New function BN_swap.
13220 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13221 the exponentiation functions are more likely to produce reasonable
13222 results on negative inputs.
13226 * Change BN_mod_mul so that the result is always non-negative.
13227 Previously, it could be negative if one of the factors was negative;
13228 I don't think anyone really wanted that behaviour.
13232 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13233 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13234 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13235 and add new functions:
13244 BN_mod_lshift1_quick
13246 BN_mod_lshift_quick
13248 These functions always generate non-negative results.
13250 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13251 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13253 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13254 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13255 be reduced modulo `m`.
13257 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13260 The following entry accidentally appeared in the CHANGES file
13261 distributed with OpenSSL 0.9.7. The modifications described in
13262 it do *not* apply to OpenSSL 0.9.7.
13264 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13265 was actually never needed) and in BN_mul(). The removal in BN_mul()
13266 required a small change in bn_mul_part_recursive() and the addition
13267 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13268 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13269 bn_sub_words() and bn_add_words() except they take arrays with
13275 * In 'openssl passwd', verify passwords read from the terminal
13276 unless the '-salt' option is used (which usually means that
13277 verification would just waste user's time since the resulting
13278 hash is going to be compared with some given password hash)
13279 or the new '-noverify' option is used.
13281 This is an incompatible change, but it does not affect
13282 non-interactive use of 'openssl passwd' (passwords on the command
13283 line, '-stdin' option, '-in ...' option) and thus should not
13284 cause any problems.
13288 * Remove all references to RSAref, since there's no more need for it.
13292 * Make DSO load along a path given through an environment variable
13293 (SHLIB_PATH) with shl_load().
13297 * Constify the ENGINE code as a result of BIGNUM constification.
13298 Also constify the RSA code and most things related to it. In a
13299 few places, most notable in the depth of the ASN.1 code, ugly
13300 casts back to non-const were required (to be solved at a later
13305 * Make it so the openssl application has all engines loaded by default.
13309 * Constify the BIGNUM routines a little more.
13313 * Add the following functions:
13315 ENGINE_load_cswift()
13317 ENGINE_load_atalla()
13318 ENGINE_load_nuron()
13319 ENGINE_load_builtin_engines()
13321 That way, an application can itself choose if external engines that
13322 are built-in in OpenSSL shall ever be used or not. The benefit is
13323 that applications won't have to be linked with libdl or other dso
13324 libraries unless it's really needed.
13326 Changed 'openssl engine' to load all engines on demand.
13327 Changed the engine header files to avoid the duplication of some
13328 declarations (they differed!).
13332 * 'openssl engine' can now list capabilities.
13336 * Better error reporting in 'openssl engine'.
13340 * Never call load_dh_param(NULL) in s_server.
13344 * Add engine application. It can currently list engines by name and
13345 identity, and test if they are actually available.
13349 * Improve RPM specification file by forcing symbolic linking and making
13350 sure the installed documentation is also owned by root.root.
13352 *Damien Miller <djm@mindrot.org>*
13354 * Give the OpenSSL applications more possibilities to make use of
13355 keys (public as well as private) handled by engines.
13359 * Add OCSP code that comes from CertCo.
13363 * Add VMS support for the Rijndael code.
13367 * Added untested support for Nuron crypto accelerator.
13371 * Add support for external cryptographic devices. This code was
13372 previously distributed separately as the "engine" branch.
13374 *Geoff Thorpe, Richard Levitte*
13376 * Rework the filename-translation in the DSO code. It is now possible to
13377 have far greater control over how a "name" is turned into a filename
13378 depending on the operating environment and any oddities about the
13379 different shared library filenames on each system.
13383 * Support threads on FreeBSD-elf in Configure.
13387 * Fix for SHA1 assembly problem with MASM: it produces
13388 warnings about corrupt line number information when assembling
13389 with debugging information. This is caused by the overlapping
13392 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13395 NCONF_get_number() has no error checking at all. As a replacement,
13396 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13397 promoted strongly. The old NCONF_get_number is kept around for
13398 binary backward compatibility.
13399 Make it possible for methods to load from something other than a BIO,
13400 by providing a function pointer that is given a name instead of a BIO.
13401 For example, this could be used to load configuration data from an
13406 * Fix for non blocking accept BIOs. Added new I/O special reason
13407 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13408 with non blocking I/O was not possible because no retry code was
13409 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13414 * Added the beginnings of Rijndael support.
13418 * Fix for bug in DirectoryString mask setting. Add support for
13419 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13420 to allow certificate printing to more controllable, additional
13421 'certopt' option to 'x509' to allow new printing options to be
13426 * Clean old EAY MD5 hack from e_os.h.
13430 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13432 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13433 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13435 *Joe Orton, Steve Henson*
13437 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13439 * Fix additional bug revealed by the NISCC test suite:
13441 Stop bug triggering large recursion when presented with
13442 certain ASN.1 tags ([CVE-2003-0851])
13446 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13448 * Fix various bugs revealed by running the NISCC test suite:
13450 Stop out of bounds reads in the ASN1 code when presented with
13451 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13453 If verify callback ignores invalid public key errors don't try to check
13454 certificate signature with the NULL public key.
13458 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13459 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13464 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13465 extra data after the compression methods not only for TLS 1.0
13466 but also for SSL 3.0 (as required by the specification).
13468 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13470 * Change X509_certificate_type() to mark the key as exported/exportable
13471 when it's 512 *bits* long, not 512 bytes.
13475 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13477 * Countermeasure against the Klima-Pokorny-Rosa extension of
13478 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13479 a protocol version number mismatch like a decryption error
13480 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13484 * Turn on RSA blinding by default in the default implementation
13485 to avoid a timing attack. Applications that don't want it can call
13486 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13487 They would be ill-advised to do so in most cases.
13489 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13491 * Change RSA blinding code so that it works when the PRNG is not
13492 seeded (in this case, the secret RSA exponent is abused as
13493 an unpredictable seed -- if it is not unpredictable, there
13494 is no point in blinding anyway). Make RSA blinding thread-safe
13495 by remembering the creator's thread ID in rsa->blinding and
13496 having all other threads use local one-time blinding factors
13497 (this requires more computation than sharing rsa->blinding, but
13498 avoids excessive locking; and if an RSA object is not shared
13499 between threads, blinding will still be very fast).
13503 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13505 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13506 via timing by performing a MAC computation even if incorrect
13507 block cipher padding has been found. This is a countermeasure
13508 against active attacks where the attacker has to distinguish
13509 between bad padding and a MAC verification error. ([CVE-2003-0078])
13511 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13512 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13513 Martin Vuagnoux (EPFL, Ilion)*
13515 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13517 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13518 memory from its contents. This is done with a counter that will
13519 place alternating values in each byte. This can be used to solve
13520 two issues: 1) the removal of calls to memset() by highly optimizing
13521 compilers, and 2) cleansing with other values than 0, since those can
13522 be read through on certain media, for example a swap space on disk.
13526 * Bugfix: client side session caching did not work with external caching,
13527 because the session->cipher setting was not restored when reloading
13528 from the external cache. This problem was masked, when
13529 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13530 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13534 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13535 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13537 *Zeev Lieber <zeev-l@yahoo.com>*
13539 * Undo an undocumented change introduced in 0.9.6e which caused
13540 repeated calls to OpenSSL_add_all_ciphers() and
13541 OpenSSL_add_all_digests() to be ignored, even after calling
13546 * Change the default configuration reader to deal with last line not
13547 being properly terminated.
13551 * Change X509_NAME_cmp() so it applies the special rules on handling
13552 DN values that are of type PrintableString, as well as RDNs of type
13553 emailAddress where the value has the type ia5String.
13555 *stefank@valicert.com via Richard Levitte*
13557 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13558 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13559 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13560 the bitwise-OR of the two for use by the majority of applications
13561 wanting this behaviour, and update the docs. The documented
13562 behaviour and actual behaviour were inconsistent and had been
13563 changing anyway, so this is more a bug-fix than a behavioural
13566 *Geoff Thorpe, diagnosed by Nadav Har'El*
13568 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13569 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13573 * Fix initialization code race conditions in
13574 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13575 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13576 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13577 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13578 ssl2_get_cipher_by_char(),
13579 ssl3_get_cipher_by_char().
13581 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13583 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13584 the cached sessions are flushed, as the remove_cb() might use ex_data
13585 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13586 (see [openssl.org #212]).
13588 *Geoff Thorpe, Lutz Jaenicke*
13590 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13591 length, instead of the encoding length to d2i_ASN1_OBJECT.
13595 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13597 * [In 0.9.6g-engine release:]
13598 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13600 *Lynn Gazis <lgazis@rainbow.com>*
13602 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13604 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13605 and get fix the header length calculation.
13606 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13607 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13609 * Use proper error handling instead of 'assertions' in buffer
13610 overflow checks added in 0.9.6e. This prevents DoS (the
13611 assertions could call abort()).
13613 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13615 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13617 * Add various sanity checks to asn1_get_length() to reject
13618 the ASN1 length bytes if they exceed sizeof(long), will appear
13619 negative or the content length exceeds the length of the
13622 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13624 * Fix cipher selection routines: ciphers without encryption had no flags
13625 for the cipher strength set and where therefore not handled correctly
13626 by the selection routines (PR #130).
13630 * Fix EVP_dsa_sha macro.
13635 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13636 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13637 that was added in OpenSSL 0.9.6d.
13639 As the countermeasure turned out to be incompatible with some
13640 broken SSL implementations, the new option is part of SSL_OP_ALL.
13641 SSL_OP_ALL is usually employed when compatibility with weird SSL
13642 implementations is desired (e.g. '-bugs' option to 's_client' and
13643 's_server'), so the new option is automatically set in many
13648 * Changes in security patch:
13650 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13651 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13652 Air Force Materiel Command, USAF, under agreement number
13655 * Add various sanity checks to asn1_get_length() to reject
13656 the ASN1 length bytes if they exceed sizeof(long), will appear
13657 negative or the content length exceeds the length of the
13658 supplied buffer. ([CVE-2002-0659])
13660 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13662 * Assertions for various potential buffer overflows, not known to
13663 happen in practice.
13665 *Ben Laurie (CHATS)*
13667 * Various temporary buffers to hold ASCII versions of integers were
13668 too small for 64 bit platforms. ([CVE-2002-0655])
13669 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13671 * Remote buffer overflow in SSL3 protocol - an attacker could
13672 supply an oversized session ID to a client. ([CVE-2002-0656])
13674 *Ben Laurie (CHATS)*
13676 * Remote buffer overflow in SSL2 protocol - an attacker could
13677 supply an oversized client master key. ([CVE-2002-0656])
13679 *Ben Laurie (CHATS)*
13681 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13683 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13684 encoded as NULL) with id-dsa-with-sha1.
13686 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13688 * Check various `X509_...()` return values in apps/req.c.
13690 *Nils Larsch <nla@trustcenter.de>*
13692 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13693 an end-of-file condition would erroneously be flagged, when the CRLF
13694 was just at the end of a processed block. The bug was discovered when
13695 processing data through a buffering memory BIO handing the data to a
13696 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13697 <ptsekov@syntrex.com> and Nedelcho Stanev.
13701 * Implement a countermeasure against a vulnerability recently found
13702 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13703 before application data chunks to avoid the use of known IVs
13704 with data potentially chosen by the attacker.
13708 * Fix length checks in ssl3_get_client_hello().
13712 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13713 to prevent ssl3_read_internal() from incorrectly assuming that
13714 ssl3_read_bytes() found application data while handshake
13715 processing was enabled when in fact s->s3->in_read_app_data was
13716 merely automatically cleared during the initial handshake.
13718 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13720 * Fix object definitions for Private and Enterprise: they were not
13721 recognized in their shortname (=lowercase) representation. Extend
13722 obj_dat.pl to issue an error when using undefined keywords instead
13723 of silently ignoring the problem (Svenning Sorensen
13724 <sss@sss.dnsalias.net>).
13728 * Fix DH_generate_parameters() so that it works for 'non-standard'
13729 generators, i.e. generators other than 2 and 5. (Previously, the
13730 code did not properly initialise the 'add' and 'rem' values to
13731 BN_generate_prime().)
13733 In the new general case, we do not insist that 'generator' is
13734 actually a primitive root: This requirement is rather pointless;
13735 a generator of the order-q subgroup is just as good, if not
13740 * Map new X509 verification errors to alerts. Discovered and submitted by
13741 Tom Wu <tom@arcot.com>.
13745 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13746 returning non-zero before the data has been completely received
13747 when using non-blocking I/O.
13749 *Bodo Moeller; problem pointed out by John Hughes*
13751 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13753 *Ben Laurie, Lutz Jaenicke*
13755 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13756 Yoram Zahavi <YoramZ@gilian.com>).
13760 * Add information about CygWin 1.3 and on, and preserve proper
13761 configuration for the versions before that.
13763 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13765 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13766 check whether we deal with a copy of a session and do not delete from
13767 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13768 <izhar@checkpoint.com>.
13772 * Do not store session data into the internal session cache, if it
13773 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13774 flag is set). Proposed by Aslam <aslam@funk.com>.
13778 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13783 * [In 0.9.6d-engine release:]
13784 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13786 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13788 * Add the configuration target linux-s390x.
13790 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13792 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13793 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13794 variable as an indication that a ClientHello message has been
13795 received. As the flag value will be lost between multiple
13796 invocations of ssl3_accept when using non-blocking I/O, the
13797 function may not be aware that a handshake has actually taken
13798 place, thus preventing a new session from being added to the
13801 To avoid this problem, we now set s->new_session to 2 instead of
13802 using a local variable.
13804 *Lutz Jaenicke, Bodo Moeller*
13806 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13807 if the SSL_R_LENGTH_MISMATCH error is detected.
13809 *Geoff Thorpe, Bodo Moeller*
13811 * New 'shared_ldflag' column in Configure platform table.
13815 * Fix EVP_CIPHER_mode macro.
13817 *"Dan S. Camper" <dan@bti.net>*
13819 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13820 type, we must throw them away by setting rr->length to 0.
13822 *D P Chang <dpc@qualys.com>*
13824 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13826 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13827 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13828 worked incorrectly for those cases where range = `10..._2` and
13829 `3*range` is two bits longer than range.)
13833 * Only add signing time to PKCS7 structures if it is not already
13838 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13839 OBJ_ld_ce should be OBJ_id_ce.
13840 Also some ip-pda OIDs in crypto/objects/objects.txt were
13841 incorrect (cf. RFC 3039).
13843 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13845 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13846 returns early because it has nothing to do.
13848 *Andy Schneider <andy.schneider@bjss.co.uk>*
13850 * [In 0.9.6c-engine release:]
13851 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13853 *Andy Schneider <andy.schneider@bjss.co.uk>*
13855 * [In 0.9.6c-engine release:]
13856 Add support for Cryptographic Appliance's keyserver technology.
13857 (Use engine 'keyclient')
13859 *Cryptographic Appliances and Geoff Thorpe*
13861 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13862 is called via tools/c89.sh because arguments have to be
13863 rearranged (all '-L' options must appear before the first object
13866 *Richard Shapiro <rshapiro@abinitio.com>*
13868 * [In 0.9.6c-engine release:]
13869 Add support for Broadcom crypto accelerator cards, backported
13872 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13874 * [In 0.9.6c-engine release:]
13875 Add support for SureWare crypto accelerator cards from
13876 Baltimore Technologies. (Use engine 'sureware')
13878 *Baltimore Technologies and Mark Cox*
13880 * [In 0.9.6c-engine release:]
13881 Add support for crypto accelerator cards from Accelerated
13882 Encryption Processing, www.aep.ie. (Use engine 'aep')
13884 *AEP Inc. and Mark Cox*
13886 * Add a configuration entry for gcc on UnixWare.
13888 *Gary Benson <gbenson@redhat.com>*
13890 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13891 messages are stored in a single piece (fixed-length part and
13892 variable-length part combined) and fix various bugs found on the way.
13896 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13897 instead. BIO_gethostbyname() does not know what timeouts are
13898 appropriate, so entries would stay in cache even when they have
13900 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13902 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13903 faced with a pathologically small ClientHello fragment that does
13904 not contain client_version: Instead of aborting with an error,
13905 simply choose the highest available protocol version (i.e.,
13906 TLS 1.0 unless it is disabled). In practice, ClientHello
13907 messages are never sent like this, but this change gives us
13908 strictly correct behaviour at least for TLS.
13912 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13913 never resets s->method to s->ctx->method when called from within
13914 one of the SSL handshake functions.
13916 *Bodo Moeller; problem pointed out by Niko Baric*
13918 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13919 (sent using the client's version number) if client_version is
13920 smaller than the protocol version in use. Also change
13921 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13922 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13923 the client will at least see that alert.
13927 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13932 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13933 client receives HelloRequest while in a handshake.
13935 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13937 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13938 should end in 'break', not 'goto end' which circumvents various
13939 cleanups done in state SSL_ST_OK. But session related stuff
13940 must be disabled for SSL_ST_OK in the case that we just sent a
13943 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13944 before just sending a HelloRequest.
13946 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13948 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13949 reveal whether illegal block cipher padding was found or a MAC
13950 verification error occurred. (Neither SSLerr() codes nor alerts
13951 are directly visible to potential attackers, but the information
13952 may leak via logfiles.)
13954 Similar changes are not required for the SSL 2.0 implementation
13955 because the number of padding bytes is sent in clear for SSL 2.0,
13956 and the extra bytes are just ignored. However ssl/s2_pkt.c
13957 failed to verify that the purported number of padding bytes is in
13962 * Add OpenUNIX-8 support including shared libraries
13963 (Boyd Lynn Gerber <gerberb@zenez.com>).
13967 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13968 'wristwatch attack' using huge encoding parameters (cf.
13969 James H. Manger's CRYPTO 2001 paper). Note that the
13970 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13971 encoding parameters and hence was not vulnerable.
13975 * BN_sqr() bug fix.
13977 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13979 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13980 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13981 followed by modular reduction.
13983 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13985 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13986 equivalent based on BN_pseudo_rand() instead of BN_rand().
13990 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13991 This function was broken, as the check for a new client hello message
13992 to handle SGC did not allow these large messages.
13993 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13997 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14001 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14002 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14006 * Rework the configuration and shared library support for Tru64 Unix.
14007 The configuration part makes use of modern compiler features and
14008 still retains old compiler behavior for those that run older versions
14009 of the OS. The shared library support part includes a variant that
14010 uses the RPATH feature, and is available through the special
14011 configuration target "alpha-cc-rpath", which will never be selected
14014 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14016 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14017 with the same message size as in ssl3_get_certificate_request().
14018 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14019 messages might inadvertently be reject as too long.
14021 *Petr Lampa <lampa@fee.vutbr.cz>*
14023 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14027 * Modified SSL library such that the verify_callback that has been set
14028 specifically for an SSL object with SSL_set_verify() is actually being
14029 used. Before the change, a verify_callback set with this function was
14030 ignored and the verify_callback() set in the SSL_CTX at the time of
14031 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14032 to allow the necessary settings.
14036 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14037 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14038 done automatically (in contradiction to the requirements of the C
14039 standard). This made problems when used from OpenSSH.
14043 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14044 dh->length and always used
14046 BN_rand_range(priv_key, dh->p).
14048 BN_rand_range() is not necessary for Diffie-Hellman, and this
14049 specific range makes Diffie-Hellman unnecessarily inefficient if
14050 dh->length (recommended exponent length) is much smaller than the
14051 length of dh->p. We could use BN_rand_range() if the order of
14052 the subgroup was stored in the DH structure, but we only have
14057 BN_rand(priv_key, l, ...)
14059 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14066 RSA_eay_public_encrypt
14067 RSA_eay_private_decrypt
14068 RSA_eay_private_encrypt (signing)
14069 RSA_eay_public_decrypt (signature verification)
14071 (default implementations for RSA_public_encrypt,
14072 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14073 always reject numbers >= n.
14077 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14078 to synchronize access to 'locking_thread'. This is necessary on
14079 systems where access to 'locking_thread' (an 'unsigned long'
14080 variable) is not atomic.
14084 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14085 *before* setting the 'crypto_lock_rand' flag. The previous code had
14086 a race condition if 0 is a valid thread ID.
14088 *Travis Vitek <vitek@roguewave.com>*
14090 * Add support for shared libraries under Irix.
14092 *Albert Chin-A-Young <china@thewrittenword.com>*
14094 * Add configuration option to build on Linux on both big-endian and
14095 little-endian MIPS.
14097 *Ralf Baechle <ralf@uni-koblenz.de>*
14099 * Add the possibility to create shared libraries on HP-UX.
14103 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14105 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14106 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14107 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14108 PRNG state recovery was possible based on the output of
14109 one PRNG request appropriately sized to gain knowledge on
14110 'md' followed by enough consecutive 1-byte PRNG requests
14111 to traverse all of 'state'.
14113 1. When updating 'md_local' (the current thread's copy of 'md')
14114 during PRNG output generation, hash all of the previous
14115 'md_local' value, not just the half used for PRNG output.
14117 2. Make the number of bytes from 'state' included into the hash
14118 independent from the number of PRNG bytes requested.
14120 The first measure alone would be sufficient to avoid
14121 Markku-Juhani's attack. (Actually it had never occurred
14122 to me that the half of 'md_local' used for chaining was the
14123 half from which PRNG output bytes were taken -- I had always
14124 assumed that the secret half would be used.) The second
14125 measure makes sure that additional data from 'state' is never
14126 mixed into 'md_local' in small portions; this heuristically
14127 further strengthens the PRNG.
14131 * Fix crypto/bn/asm/mips3.s.
14135 * When only the key is given to "enc", the IV is undefined. Print out
14136 an error message in this case.
14140 * Handle special case when X509_NAME is empty in X509 printing routines.
14144 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14145 positive and less than q.
14149 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14150 used: it isn't thread safe and the add_lock_callback should handle
14153 *Paul Rose <Paul.Rose@bridge.com>*
14155 * Verify that incoming data obeys the block size in
14156 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14162 *Ulf Möller, Bodo Möller*
14164 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14165 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14166 when fixing the server behaviour for backwards-compatible 'client
14167 hello' messages. (Note that the attack is impractical against
14168 SSL 3.0 and TLS 1.0 anyway because length and version checking
14169 means that the probability of guessing a valid ciphertext is
14170 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14173 Before 0.9.5, the countermeasure (hide the error by generating a
14174 random 'decryption result') did not work properly because
14175 ERR_clear_error() was missing, meaning that SSL_get_error() would
14176 detect the supposedly ignored error.
14178 Both problems are now fixed.
14182 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14183 (previously it was 1024).
14187 * Fix for compatibility mode trust settings: ignore trust settings
14188 unless some valid trust or reject settings are present.
14192 * Fix for blowfish EVP: its a variable length cipher.
14196 * Fix various bugs related to DSA S/MIME verification. Handle missing
14197 parameters in DSA public key structures and return an error in the
14198 DSA routines if parameters are absent.
14202 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14203 in the current directory if neither $RANDFILE nor $HOME was set.
14204 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14205 caused some confusion to Windows users who haven't defined $HOME.
14206 Thus RAND_file_name() is changed again: e_os.h can define a
14207 DEFAULT_HOME, which will be used if $HOME is not set.
14208 For Windows, we use "C:"; on other platforms, we still require
14209 environment variables.
14211 * Move 'if (!initialized) RAND_poll()' into regions protected by
14212 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14213 having multiple threads call RAND_poll() concurrently.
14217 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14218 combination of a flag and a thread ID variable.
14219 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14220 flag), *other* threads can enter ssleay_add_bytes without obeying
14221 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14222 that they do not hold after the first thread unsets add_do_not_lock).
14226 * Change bctest again: '-x' expressions are not available in all
14227 versions of 'test'.
14231 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14233 * Fix a couple of memory leaks in PKCS7_dataDecode()
14235 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14237 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14238 the default extension for executables, if any. Also, make the perl
14239 scripts that use symlink() to test if it really exists and use "cp"
14240 if it doesn't. All this made OpenSSL compilable and installable in
14245 * Fix for asn1_GetSequence() for indefinite length constructed data.
14246 If SEQUENCE is length is indefinite just set c->slen to the total
14247 amount of data available.
14249 *Steve Henson, reported by shige@FreeBSD.org*
14251 *This change does not apply to 0.9.7.*
14253 * Change bctest to avoid here-documents inside command substitution
14254 (workaround for FreeBSD /bin/sh bug).
14255 For compatibility with Ultrix, avoid shell functions (introduced
14256 in the bctest version that searches along $PATH).
14260 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14261 with des_encrypt() defined on some operating systems, like Solaris
14266 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14267 On the Importance of Eliminating Errors in Cryptographic
14268 Computations, J. Cryptology 14 (2001) 2, 101-119,
14269 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14273 * MIPS assembler BIGNUM division bug fix.
14277 * Disabled incorrect Alpha assembler code.
14281 * Fix PKCS#7 decode routines so they correctly update the length
14282 after reading an EOC for the EXPLICIT tag.
14286 *This change does not apply to 0.9.7.*
14288 * Fix bug in PKCS#12 key generation routines. This was triggered
14289 if a 3DES key was generated with a 0 initial byte. Include
14290 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14291 (but broken) behaviour.
14295 * Enhance bctest to search for a working bc along $PATH and print
14298 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14300 * Fix memory leaks in err.c: free err_data string if necessary;
14301 don't write to the wrong index in ERR_set_error_data.
14305 * Implement ssl23_peek (analogous to ssl23_read), which previously
14310 * Replace rdtsc with `_emit` statements for VC++ version 5.
14312 *Jeremy Cooper <jeremy@baymoo.org>*
14314 * Make it possible to reuse SSLv2 sessions.
14318 * In copy_email() check for >= 0 as a return value for
14319 X509_NAME_get_index_by_NID() since 0 is a valid index.
14321 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14323 * Avoid coredump with unsupported or invalid public keys by checking if
14324 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14325 PKCS7_verify() fails with non detached data.
14329 * Don't use getenv in library functions when run as setuid/setgid.
14330 New function OPENSSL_issetugid().
14334 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14335 due to incorrect handling of multi-threading:
14337 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14339 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14341 3. Count how many times MemCheck_off() has been called so that
14342 nested use can be treated correctly. This also avoids
14343 inband-signalling in the previous code (which relied on the
14344 assumption that thread ID 0 is impossible).
14348 * Add "-rand" option also to s_client and s_server.
14352 * Fix CPU detection on Irix 6.x.
14353 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14354 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14356 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14361 *This change does not apply to 0.9.7.*
14363 * Use the cached encoding of an X509_NAME structure rather than
14364 copying it. This is apparently the reason for the libsafe "errors"
14365 but the code is actually correct.
14369 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14370 Bleichenbacher's DSA attack.
14371 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14372 to be set and top=0 forces the highest bit to be set; top=-1 is new
14373 and leaves the highest bit random.
14375 *Ulf Moeller, Bodo Moeller*
14377 * In the `NCONF_...`-based implementations for `CONF_...` queries
14378 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14379 a temporary CONF structure with the data component set to NULL
14380 (which gives segmentation faults in lh_retrieve).
14381 Instead, use NULL for the CONF pointer in CONF_get_string and
14382 CONF_get_number (which may use environment variables) and directly
14383 return NULL from CONF_get_section.
14387 * Fix potential buffer overrun for EBCDIC.
14391 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14392 keyUsage if basicConstraints absent for a CA.
14396 * Make SMIME_write_PKCS7() write mail header values with a format that
14397 is more generally accepted (no spaces before the semicolon), since
14398 some programs can't parse those values properly otherwise. Also make
14399 sure BIO's that break lines after each write do not create invalid
14404 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14405 macros previously used would not encode an empty SEQUENCE OF
14406 and break the signature.
14410 *This change does not apply to 0.9.7.*
14412 * Zero the premaster secret after deriving the master secret in
14417 * Add some EVP_add_digest_alias registrations (as found in
14418 OpenSSL_add_all_digests()) to SSL_library_init()
14419 aka OpenSSL_add_ssl_algorithms(). This provides improved
14420 compatibility with peers using X.509 certificates
14421 with unconventional AlgorithmIdentifier OIDs.
14425 * Fix for Irix with NO_ASM.
14427 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14429 * ./config script fixes.
14431 *Ulf Moeller, Richard Levitte*
14433 * Fix 'openssl passwd -1'.
14437 * Change PKCS12_key_gen_asc() so it can cope with non null
14438 terminated strings whose length is passed in the passlen
14439 parameter, for example from PEM callbacks. This was done
14440 by adding an extra length parameter to asc2uni().
14442 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14444 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14445 call failed, free the DSA structure.
14449 * Fix to uni2asc() to cope with zero length Unicode strings.
14450 These are present in some PKCS#12 files.
14454 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14455 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14456 when writing a 32767 byte record.
14458 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14460 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14461 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14463 (RSA objects have a reference count access to which is protected
14464 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14465 so they are meant to be shared between threads.)
14466 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14467 "Reddie, Steven" <Steven.Reddie@ca.com>*
14469 * Fix a deadlock in CRYPTO_mem_leaks().
14473 * Use better test patterns in bntest.
14477 * rand_win.c fix for Borland C.
14481 * BN_rshift bugfix for n == 0.
14485 * Add a 'bctest' script that checks for some known 'bc' bugs
14486 so that 'make test' does not abort just because 'bc' is broken.
14490 * Store verify_result within SSL_SESSION also for client side to
14491 avoid potential security hole. (Re-used sessions on the client side
14492 always resulted in verify_result==X509_V_OK, not using the original
14493 result of the server certificate verification.)
14497 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14498 SSL3_RT_APPLICATION_DATA, return 0.
14499 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14504 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14505 releases, have been re-implemented by renaming the previous
14506 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14507 and ssl3_read_internal, respectively, and adding 'peek' parameters
14508 to them. The new ssl[23]_{read,peek} functions are calls to
14509 ssl[23]_read_internal with the 'peek' flag set appropriately.
14510 A 'peek' parameter has also been added to ssl3_read_bytes, which
14511 does the actual work for ssl3_read_internal.
14515 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14516 the method-specific "init()" handler. Also clean up ex_data after
14517 calling the method-specific "finish()" handler. Previously, this was
14518 happening the other way round.
14522 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14523 The previous value, 12, was not always sufficient for BN_mod_exp().
14527 * Make sure that shared libraries get the internal name engine with
14528 the full version number and not just 0. This should mark the
14529 shared libraries as not backward compatible. Of course, this should
14530 be changed again when we can guarantee backward binary compatibility.
14534 * Fix typo in get_cert_by_subject() in by_dir.c
14536 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14538 * Rework the system to generate shared libraries:
14540 - Make note of the expected extension for the shared libraries and
14541 if there is a need for symbolic links from for example libcrypto.so.0
14542 to libcrypto.so.0.9.7. There is extended info in Configure for
14545 - Make as few rebuilds of the shared libraries as possible.
14547 - Still avoid linking the OpenSSL programs with the shared libraries.
14549 - When installing, install the shared libraries separately from the
14554 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14556 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14557 and not in SSL_clear because the latter is also used by the
14558 accept/connect functions; previously, the settings made by
14559 SSL_set_read_ahead would be lost during the handshake.
14561 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14563 * Correct util/mkdef.pl to be selective about disabled algorithms.
14564 Previously, it would create entries for disabled algorithms no
14569 * Added several new manual pages for SSL_* function.
14573 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14575 * In ssl23_get_client_hello, generate an error message when faced
14576 with an initial SSL 3.0/TLS record that is too small to contain the
14577 first two bytes of the ClientHello message, i.e. client_version.
14578 (Note that this is a pathologic case that probably has never happened
14579 in real life.) The previous approach was to use the version number
14580 from the record header as a substitute; but our protocol choice
14581 should not depend on that one because it is not authenticated
14582 by the Finished messages.
14586 * More robust randomness gathering functions for Windows.
14588 *Jeffrey Altman <jaltman@columbia.edu>*
14590 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14591 not set then we don't setup the error code for issuer check errors
14592 to avoid possibly overwriting other errors which the callback does
14593 handle. If an application does set the flag then we assume it knows
14594 what it is doing and can handle the new informational codes
14599 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14600 a general "ANY" type, as such it should be able to decode anything
14601 including tagged types. However it didn't check the class so it would
14602 wrongly interpret tagged types in the same way as their universal
14603 counterpart and unknown types were just rejected. Changed so that the
14604 tagged and unknown types are handled in the same way as a SEQUENCE:
14605 that is the encoding is stored intact. There is also a new type
14606 "V_ASN1_OTHER" which is used when the class is not universal, in this
14607 case we have no idea what the actual type is so we just lump them all
14612 * On VMS, stdout may very well lead to a file that is written to
14613 in a record-oriented fashion. That means that every write() will
14614 write a separate record, which will be read separately by the
14615 programs trying to read from it. This can be very confusing.
14617 The solution is to put a BIO filter in the way that will buffer
14618 text until a linefeed is reached, and then write everything a
14619 line at a time, so every record written will be an actual line,
14620 not chunks of lines and not (usually doesn't happen, but I've
14621 seen it once) several lines in one record. BIO_f_linebuffer() is
14624 Currently, it's a VMS-only method, because that's where it has
14625 been tested well enough.
14629 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14630 it can return incorrect results.
14631 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14632 but it was in 0.9.6-beta[12].)
14636 * Disable the check for content being present when verifying detached
14637 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14638 include zero length content when signing messages.
14642 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14643 BIO_ctrl (for BIO pairs).
14647 * Add DSO method for VMS.
14651 * Bug fix: Montgomery multiplication could produce results with the
14656 * Add RPM specification openssl.spec and modify it to build three
14657 packages. The default package contains applications, application
14658 documentation and run-time libraries. The devel package contains
14659 include files, static libraries and function documentation. The
14660 doc package contains the contents of the doc directory. The original
14661 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14665 * Add a large number of documentation files for many SSL routines.
14667 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14669 * Add a configuration entry for Sony News 4.
14671 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14673 * Don't set the two most significant bits to one when generating a
14674 random number < q in the DSA library.
14678 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14679 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14680 the underlying transport is blocking) if a handshake took place.
14681 (The default behaviour is needed by applications such as s_client
14682 and s_server that use select() to determine when to use SSL_read;
14683 but for applications that know in advance when to expect data, it
14684 just makes things more complicated.)
14688 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14693 * Add a few more EBCDIC conditionals that make `req` and `x509`
14694 work better on such systems.
14696 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14698 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14699 Update PKCS12_parse() so it copies the friendlyName and the
14700 keyid to the certificates aux info.
14704 * Fix bug in PKCS7_verify() which caused an infinite loop
14705 if there was more than one signature.
14707 *Sven Uszpelkat <su@celocom.de>*
14709 * Major change in util/mkdef.pl to include extra information
14710 about each symbol, as well as presenting variables as well
14711 as functions. This change means that there's n more need
14712 to rebuild the .num files when some algorithms are excluded.
14716 * Allow the verify time to be set by an application,
14717 rather than always using the current time.
14721 * Phase 2 verify code reorganisation. The certificate
14722 verify code now looks up an issuer certificate by a
14723 number of criteria: subject name, authority key id
14724 and key usage. It also verifies self signed certificates
14725 by the same criteria. The main comparison function is
14726 X509_check_issued() which performs these checks.
14728 Lot of changes were necessary in order to support this
14729 without completely rewriting the lookup code.
14731 Authority and subject key identifier are now cached.
14733 The LHASH 'certs' is X509_STORE has now been replaced
14734 by a STACK_OF(X509_OBJECT). This is mainly because an
14735 LHASH can't store or retrieve multiple objects with
14736 the same hash value.
14738 As a result various functions (which were all internal
14739 use only) have changed to handle the new X509_STORE
14740 structure. This will break anything that messed round
14741 with X509_STORE internally.
14743 The functions X509_STORE_add_cert() now checks for an
14744 exact match, rather than just subject name.
14746 The X509_STORE API doesn't directly support the retrieval
14747 of multiple certificates matching a given criteria, however
14748 this can be worked round by performing a lookup first
14749 (which will fill the cache with candidate certificates)
14750 and then examining the cache for matches. This is probably
14751 the best we can do without throwing out X509_LOOKUP
14752 entirely (maybe later...).
14754 The X509_VERIFY_CTX structure has been enhanced considerably.
14756 All certificate lookup operations now go via a get_issuer()
14757 callback. Although this currently uses an X509_STORE it
14758 can be replaced by custom lookups. This is a simple way
14759 to bypass the X509_STORE hackery necessary to make this
14760 work and makes it possible to use more efficient techniques
14761 in future. A very simple version which uses a simple
14762 STACK for its trusted certificate store is also provided
14763 using X509_STORE_CTX_trusted_stack().
14765 The verify_cb() and verify() callbacks now have equivalents
14766 in the X509_STORE_CTX structure.
14768 X509_STORE_CTX also has a 'flags' field which can be used
14769 to customise the verify behaviour.
14773 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14774 excludes S/MIME capabilities.
14778 * When a certificate request is read in keep a copy of the
14779 original encoding of the signed data and use it when outputting
14780 again. Signatures then use the original encoding rather than
14781 a decoded, encoded version which may cause problems if the
14782 request is improperly encoded.
14786 * For consistency with other BIO_puts implementations, call
14787 buffer_write(b, ...) directly in buffer_puts instead of calling
14790 In BIO_puts, increment b->num_write as in BIO_write.
14792 *Peter.Sylvester@EdelWeb.fr*
14794 * Fix BN_mul_word for the case where the word is 0. (We have to use
14795 BN_zero, we may not return a BIGNUM with an array consisting of
14796 words set to zero.)
14800 * Avoid calling abort() from within the library when problems are
14801 detected, except if preprocessor symbols have been defined
14802 (such as REF_CHECK, BN_DEBUG etc.).
14806 * New openssl application 'rsautl'. This utility can be
14807 used for low level RSA operations. DER public key
14808 BIO/fp routines also added.
14812 * New Configure entry and patches for compiling on QNX 4.
14814 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14816 * A demo state-machine implementation was sponsored by
14817 Nuron (<http://www.nuron.com/>) and is now available in
14818 demos/state_machine.
14822 * New options added to the 'dgst' utility for signature
14823 generation and verification.
14827 * Unrecognized PKCS#7 content types are now handled via a
14828 catch all ASN1_TYPE structure. This allows unsupported
14829 types to be stored as a "blob" and an application can
14830 encode and decode it manually.
14834 * Fix various signed/unsigned issues to make a_strex.c
14835 compile under VC++.
14837 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14839 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14840 length if passed a buffer. ASN1_INTEGER_to_BN failed
14841 if passed a NULL BN and its argument was negative.
14843 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14845 * Modification to PKCS#7 encoding routines to output definite
14846 length encoding. Since currently the whole structures are in
14847 memory there's not real point in using indefinite length
14848 constructed encoding. However if OpenSSL is compiled with
14849 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14853 * Added BIO_vprintf() and BIO_vsnprintf().
14857 * Added more prefixes to parse for in the strings written
14858 through a logging bio, to cover all the levels that are available
14859 through syslog. The prefixes are now:
14861 PANIC, EMERG, EMR => LOG_EMERG
14862 ALERT, ALR => LOG_ALERT
14863 CRIT, CRI => LOG_CRIT
14864 ERROR, ERR => LOG_ERR
14865 WARNING, WARN, WAR => LOG_WARNING
14866 NOTICE, NOTE, NOT => LOG_NOTICE
14867 INFO, INF => LOG_INFO
14868 DEBUG, DBG => LOG_DEBUG
14870 and as before, if none of those prefixes are present at the
14871 beginning of the string, LOG_ERR is chosen.
14873 On Win32, the `LOG_*` levels are mapped according to this:
14875 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14876 LOG_WARNING => EVENTLOG_WARNING_TYPE
14877 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14881 * Made it possible to reconfigure with just the configuration
14882 argument "reconf" or "reconfigure". The command line arguments
14883 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14884 and are retrieved from there when reconfiguring.
14890 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14892 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14896 * The obj_dat.pl script was messing up the sorting of object
14897 names. The reason was that it compared the quoted version
14898 of strings as a result "OCSP" > "OCSP Signing" because
14899 " > SPACE. Changed script to store unquoted versions of
14900 names and add quotes on output. It was also omitting some
14901 names from the lookup table if they were given a default
14902 value (that is if SN is missing it is given the same
14903 value as LN and vice versa), these are now added on the
14904 grounds that if an object has a name we should be able to
14905 look it up. Finally added warning output when duplicate
14906 short or long names are found.
14910 * Changes needed for Tandem NSK.
14912 *Scott Uroff <scott@xypro.com>*
14914 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14915 RSA_padding_check_SSLv23(), special padding was never detected
14916 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14917 version rollback attacks was not effective.
14919 In s23_clnt.c, don't use special rollback-attack detection padding
14920 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14921 client; similarly, in s23_srvr.c, don't do the rollback check if
14922 SSL 2.0 is the only protocol enabled in the server.
14926 * Make it possible to get hexdumps of unprintable data with 'openssl
14927 asn1parse'. By implication, the functions ASN1_parse_dump() and
14928 BIO_dump_indent() are added.
14932 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14933 these print out strings and name structures based on various
14934 flags including RFC2253 support and proper handling of
14935 multibyte characters. Added options to the 'x509' utility
14936 to allow the various flags to be set.
14940 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14941 Also change the functions X509_cmp_current_time() and
14942 X509_gmtime_adj() work with an ASN1_TIME structure,
14943 this will enable certificates using GeneralizedTime in validity
14944 dates to be checked.
14948 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14949 negative public key encodings) on by default,
14950 NO_NEG_PUBKEY_BUG can be set to disable it.
14954 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14955 content octets. An i2c_ASN1_OBJECT is unnecessary because
14956 the encoding can be trivially obtained from the structure.
14960 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14961 not read locks (`CRYPTO_r_[un]lock`).
14965 * A first attempt at creating official support for shared
14966 libraries through configuration. I've kept it so the
14967 default is static libraries only, and the OpenSSL programs
14968 are always statically linked for now, but there are
14969 preparations for dynamic linking in place.
14970 This has been tested on Linux and Tru64.
14974 * Randomness polling function for Win9x, as described in:
14975 Peter Gutmann, Software Generation of Practically Strong
14980 * Fix so PRNG is seeded in req if using an already existing
14985 * New options to smime application. -inform and -outform
14986 allow alternative formats for the S/MIME message including
14987 PEM and DER. The -content option allows the content to be
14988 specified separately. This should allow things like Netscape
14989 form signing output easier to verify.
14993 * Fix the ASN1 encoding of tags using the 'long form'.
14997 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14998 STRING types. These convert content octets to and from the
14999 underlying type. The actual tag and length octets are
15000 already assumed to have been read in and checked. These
15001 are needed because all other string types have virtually
15002 identical handling apart from the tag. By having versions
15003 of the ASN1 functions that just operate on content octets
15004 IMPLICIT tagging can be handled properly. It also allows
15005 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15006 and ASN1_INTEGER are identical apart from the tag.
15010 * Change the handling of OID objects as follows:
15012 - New object identifiers are inserted in objects.txt, following
15013 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15014 - objects.pl is used to process obj_mac.num and create a new
15016 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15019 This is currently kind of a hack, and the perl code in objects.pl
15020 isn't very elegant, but it works as I intended. The simplest way
15021 to check that it worked correctly is to look in obj_dat.h and
15022 check the array nid_objs and make sure the objects haven't moved
15023 around (this is important!). Additions are OK, as well as
15024 consistent name changes.
15028 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15032 * Addition of the command line parameter '-rand file' to 'openssl req'.
15033 The given file adds to whatever has already been seeded into the
15034 random pool through the RANDFILE configuration file option or
15035 environment variable, or the default random state file.
15039 * mkstack.pl now sorts each macro group into lexical order.
15040 Previously the output order depended on the order the files
15041 appeared in the directory, resulting in needless rewriting
15046 * Patches to make OpenSSL compile under Win32 again. Mostly
15047 work arounds for the VC++ problem that it treats func() as
15048 func(void). Also stripped out the parts of mkdef.pl that
15049 added extra typesafe functions: these no longer exist.
15053 * Reorganisation of the stack code. The macros are now all
15054 collected in safestack.h . Each macro is defined in terms of
15055 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15056 DEBUG_SAFESTACK is now handled in terms of function casts,
15057 this has the advantage of retaining type safety without the
15058 use of additional functions. If DEBUG_SAFESTACK is not defined
15059 then the non typesafe macros are used instead. Also modified the
15060 mkstack.pl script to handle the new form. Needs testing to see
15061 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15062 the default if no major problems. Similar behaviour for ASN1_SET_OF
15063 and PKCS12_STACK_OF.
15067 * When some versions of IIS use the 'NET' form of private key the
15068 key derivation algorithm is different. Normally MD5(password) is
15069 used as a 128 bit RC4 key. In the modified case
15070 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15071 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15072 as the old Netscape_RSA functions except they have an additional
15073 'sgckey' parameter which uses the modified algorithm. Also added
15074 an -sgckey command line option to the rsa utility. Thanks to
15075 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15076 algorithm to openssl-dev.
15080 * The evp_local.h macros were using 'c.##kname' which resulted in
15081 invalid expansion on some systems (SCO 5.0.5 for example).
15082 Corrected to 'c.kname'.
15084 *Phillip Porch <root@theporch.com>*
15086 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15087 a STACK of email addresses from a certificate or request, these look
15088 in the subject name and the subject alternative name extensions and
15089 omit any duplicate addresses.
15093 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15094 This makes DSA verification about 2 % faster.
15098 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15099 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15100 plus overhead for 1024 bit moduli).
15101 This makes exponentiations about 0.5 % faster for 1024 bit
15102 exponents (as measured by "openssl speed rsa2048").
15106 * Rename memory handling macros to avoid conflicts with other
15108 Malloc => OPENSSL_malloc
15109 Malloc_locked => OPENSSL_malloc_locked
15110 Realloc => OPENSSL_realloc
15111 Free => OPENSSL_free
15115 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15116 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15120 * CygWin32 support.
15122 *John Jarvie <jjarvie@newsguy.com>*
15124 * The type-safe stack code has been rejigged. It is now only compiled
15125 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15126 by default all type-specific stack functions are "#define"d back to
15127 standard stack functions. This results in more streamlined output
15128 but retains the type-safety checking possibilities of the original
15133 * The STACK code has been cleaned up, and certain type declarations
15134 that didn't make a lot of sense have been brought in line. This has
15135 also involved a cleanup of sorts in safestack.h to more correctly
15136 map type-safe stack functions onto their plain stack counterparts.
15137 This work has also resulted in a variety of "const"ifications of
15138 lots of the code, especially `_cmp` operations which should normally
15139 be prototyped with "const" parameters anyway.
15143 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15144 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15145 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15146 where all of 'md' is used each time the PRNG is used, but 'state'
15147 is used only indexed by a cyclic counter. As entropy may not be
15148 well distributed from the beginning, 'md' is important as a
15149 chaining variable. However, the output function chains only half
15150 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15151 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15152 in all of 'state' being rewritten, with the new values depending
15153 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15157 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15158 the handshake is continued after ssl_verify_cert_chain();
15159 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15160 can lead to 'unexplainable' connection aborts later.
15162 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15164 * Major EVP API cipher revision.
15165 Add hooks for extra EVP features. This allows various cipher
15166 parameters to be set in the EVP interface. Support added for variable
15167 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15168 setting of RC2 and RC5 parameters.
15170 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15173 Remove lots of duplicated code from the EVP library. For example *every*
15174 cipher init() function handles the 'iv' in the same way according to the
15175 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15176 for CFB and OFB modes they zero ctx->num.
15178 New functionality allows removal of S/MIME code RC2 hack.
15180 Most of the routines have the same form and so can be declared in terms
15183 By shifting this to the top level EVP_CipherInit() it can be removed from
15184 all individual ciphers. If the cipher wants to handle IVs or keys
15185 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15188 Change lots of functions like EVP_EncryptUpdate() to now return a
15189 value: although software versions of the algorithms cannot fail
15190 any installed hardware versions can.
15194 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15195 this option is set, tolerate broken clients that send the negotiated
15196 protocol version number instead of the requested protocol version
15201 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15202 i.e. non-zero for export ciphersuites, zero otherwise.
15203 Previous versions had this flag inverted, inconsistent with
15204 rsa_tmp_cb (..._TMP_RSA_CB).
15206 *Bodo Moeller; problem reported by Amit Chopra*
15208 * Add missing DSA library text string. Work around for some IIS
15209 key files with invalid SEQUENCE encoding.
15213 * Add a document (doc/standards.txt) that list all kinds of standards
15214 and so on that are implemented in OpenSSL.
15218 * Enhance c_rehash script. Old version would mishandle certificates
15219 with the same subject name hash and wouldn't handle CRLs at all.
15220 Added -fingerprint option to crl utility, to support new c_rehash
15225 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15229 * Fix for SSL server purpose checking. Server checking was
15230 rejecting certificates which had extended key usage present
15231 but no ssl client purpose.
15233 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15235 * Make PKCS#12 code work with no password. The PKCS#12 spec
15236 is a little unclear about how a blank password is handled.
15237 Since the password in encoded as a BMPString with terminating
15238 double NULL a zero length password would end up as just the
15239 double NULL. However no password at all is different and is
15240 handled differently in the PKCS#12 key generation code. NS
15241 treats a blank password as zero length. MSIE treats it as no
15242 password on export: but it will try both on import. We now do
15243 the same: PKCS12_parse() tries zero length and no password if
15244 the password is set to "" or NULL (NULL is now a valid password:
15245 it wasn't before) as does the pkcs12 application.
15249 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15250 perror when PEM_read_bio_X509_REQ fails, the error message must
15251 be obtained from the error queue.
15255 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15256 it in ERR_remove_state if appropriate, and change ERR_get_state
15257 accordingly to avoid race conditions (this is necessary because
15258 thread_hash is no longer constant once set).
15262 * Bugfix for linux-elf makefile.one.
15266 * RSA_get_default_method() will now cause a default
15267 RSA_METHOD to be chosen if one doesn't exist already.
15268 Previously this was only set during a call to RSA_new()
15269 or RSA_new_method(NULL) meaning it was possible for
15270 RSA_get_default_method() to return NULL.
15274 * Added native name translation to the existing DSO code
15275 that will convert (if the flag to do so is set) filenames
15276 that are sufficiently small and have no path information
15277 into a canonical native form. Eg. "blah" converted to
15278 "libblah.so" or "blah.dll" etc.
15282 * New function ERR_error_string_n(e, buf, len) which is like
15283 ERR_error_string(e, buf), but writes at most 'len' bytes
15284 including the 0 terminator. For ERR_error_string_n, 'buf'
15287 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15289 * CONF library reworked to become more general. A new CONF
15290 configuration file reader "class" is implemented as well as a
15291 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15292 old `CONF_*` functions are still there, but are reimplemented to
15293 work in terms of the new functions. Also, a set of functions
15294 to handle the internal storage of the configuration data is
15295 provided to make it easier to write new configuration file
15296 reader "classes" (I can definitely see something reading a
15297 configuration file in XML format, for example), called `_CONF_*`,
15298 or "the configuration storage API"...
15300 The new configuration file reading functions are:
15302 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15303 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15305 NCONF_default, NCONF_WIN32
15307 NCONF_dump_fp, NCONF_dump_bio
15309 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15310 NCONF_new creates a new CONF object. This works in the same way
15311 as other interfaces in OpenSSL, like the BIO interface.
15312 `NCONF_dump_*` dump the internal storage of the configuration file,
15313 which is useful for debugging. All other functions take the same
15314 arguments as the old `CONF_*` functions with the exception of the
15315 first that must be a `CONF *` instead of a `LHASH *`.
15317 To make it easier to use the new classes with the old `CONF_*` functions,
15318 the function CONF_set_default_method is provided.
15322 * Add '-tls1' option to 'openssl ciphers', which was already
15323 mentioned in the documentation but had not been implemented.
15324 (This option is not yet really useful because even the additional
15325 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15329 * Initial DSO code added into libcrypto for letting OpenSSL (and
15330 OpenSSL-based applications) load shared libraries and bind to
15331 them in a portable way.
15333 *Geoff Thorpe, with contributions from Richard Levitte*
15335 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15337 * Make sure _lrotl and _lrotr are only used with MSVC.
15339 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15340 (the default implementation of RAND_status).
15342 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15343 to '-clrext' (= clear extensions), as intended and documented.
15344 *Bodo Moeller; inconsistency pointed out by Michael Attili
15345 <attili@amaxo.com>*
15347 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15348 was larger than the MD block size.
15350 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15352 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15353 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15354 using the passed key: if the passed key was a private key the result
15355 of X509_print(), for example, would be to print out all the private key
15360 * des_quad_cksum() byte order bug fix.
15361 *Ulf Möller, using the problem description in krb4-0.9.7, where
15362 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15364 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15367 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15369 * For easily testing in shell scripts whether some command
15370 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15371 returns with exit code 0 iff no command of the given name is available.
15372 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15373 the output goes to stdout and nothing is printed to stderr.
15374 Additional arguments are always ignored.
15376 Since for each cipher there is a command of the same name,
15377 the 'no-cipher' compilation switches can be tested this way.
15379 ('openssl no-XXX' is not able to detect pseudo-commands such
15380 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15384 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15388 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15389 is set; it will be thrown away anyway because each handshake creates
15391 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15392 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15393 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15394 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15398 * New s_client option -ign_eof: EOF at stdin is ignored, and
15399 'Q' and 'R' lose their special meanings (quit/renegotiate).
15400 This is part of what -quiet does; unlike -quiet, -ign_eof
15401 does not suppress any output.
15405 * Add compatibility options to the purpose and trust code. The
15406 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15407 accepts a certificate or CA, this was the previous behaviour,
15408 with all the associated security issues.
15410 X509_TRUST_COMPAT is the old trust behaviour: only and
15411 automatically trust self signed roots in certificate store. A
15412 new trust setting X509_TRUST_DEFAULT is used to specify that
15413 a purpose has no associated trust setting and it should instead
15414 use the value in the default purpose.
15418 * Fix the PKCS#8 DSA private key code so it decodes keys again
15419 and fix a memory leak.
15423 * In util/mkerr.pl (which implements 'make errors'), preserve
15424 reason strings from the previous version of the .c file, as
15425 the default to have only downcase letters (and digits) in
15426 automatically generated reasons codes is not always appropriate.
15430 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15431 using strerror. Previously, ERR_reason_error_string() returned
15432 library names as reason strings for SYSerr; but SYSerr is a special
15433 case where small numbers are errno values, not library numbers.
15437 * Add '-dsaparam' option to 'openssl dhparam' application. This
15438 converts DSA parameters into DH parameters. (When creating parameters,
15439 DSA_generate_parameters is used.)
15443 * Include 'length' (recommended exponent length) in C code generated
15444 by 'openssl dhparam -C'.
15448 * The second argument to set_label in perlasm was already being used
15449 so couldn't be used as a "file scope" flag. Moved to third argument
15454 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15455 instead of RAND_bytes for encryption IVs and salts.
15459 * Include RAND_status() into RAND_METHOD instead of implementing
15460 it only for md_rand.c Otherwise replacing the PRNG by calling
15461 RAND_set_rand_method would be impossible.
15465 * Don't let DSA_generate_key() enter an infinite loop if the random
15466 number generation fails.
15470 * New 'rand' application for creating pseudo-random output.
15474 * Added configuration support for Linux/IA64
15476 *Rolf Haberrecker <rolf@suse.de>*
15478 * Assembler module support for Mingw32.
15482 * Shared library support for HPUX (in shlib/).
15484 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15486 * Shared library support for Solaris gcc.
15488 *Lutz Behnke <behnke@trustcenter.de>*
15490 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15492 * PKCS7_encrypt() was adding text MIME headers twice because they
15493 were added manually and by SMIME_crlf_copy().
15497 * In bntest.c don't call BN_rand with zero bits argument.
15499 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15501 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15502 case was implemented. This caused BN_div_recp() to fail occasionally.
15506 * Add an optional second argument to the set_label() in the perl
15507 assembly language builder. If this argument exists and is set
15508 to 1 it signals that the assembler should use a symbol whose
15509 scope is the entire file, not just the current function. This
15510 is needed with MASM which uses the format label:: for this scope.
15512 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15514 * Change the ASN1 types so they are typedefs by default. Before
15515 almost all types were #define'd to ASN1_STRING which was causing
15516 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15521 * Change names of new functions to the new get1/get0 naming
15522 convention: After 'get1', the caller owns a reference count
15523 and has to call `..._free`; 'get0' returns a pointer to some
15524 data structure without incrementing reference counters.
15525 (Some of the existing 'get' functions increment a reference
15526 counter, some don't.)
15527 Similarly, 'set1' and 'add1' functions increase reference
15528 counters or duplicate objects.
15532 * Allow for the possibility of temp RSA key generation failure:
15533 the code used to assume it always worked and crashed on failure.
15537 * Fix potential buffer overrun problem in BIO_printf().
15538 *Ulf Möller, using public domain code by Patrick Powell; problem
15539 pointed out by David Sacerdote <das33@cornell.edu>*
15541 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15542 RAND_egd() and RAND_status(). In the command line application,
15543 the EGD socket can be specified like a seed file using RANDFILE
15548 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15549 Some CAs (e.g. Verisign) distribute certificates in this form.
15553 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15554 list to exclude them. This means that no special compilation option
15555 is needed to use anonymous DH: it just needs to be included in the
15560 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15561 EVP_MD_type. The old functionality is available in a new macro called
15562 EVP_MD_md(). Change code that uses it and update docs.
15566 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15567 where the `void *` argument is replaced by a function pointer argument.
15568 Previously `void *` was abused to point to functions, which works on
15569 many platforms, but is not correct. As these functions are usually
15570 called by macros defined in OpenSSL header files, most source code
15571 should work without changes.
15575 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15576 sections with information on -D... compiler switches used for
15577 compiling the library so that applications can see them. To enable
15578 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15579 must be defined. E.g.,
15580 #define OPENSSL_ALGORITHM_DEFINES
15581 #include <openssl/opensslconf.h>
15582 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15584 *Richard Levitte, Ulf and Bodo Möller*
15586 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15591 * Change the 'other' type in certificate aux info to a STACK_OF
15592 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15593 the required ASN1 format: arbitrary types determined by an OID.
15597 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15598 argument to 'req'. This is not because the function is newer or
15599 better than others it just uses the work 'NEW' in the certificate
15600 request header lines. Some software needs this.
15604 * Reorganise password command line arguments: now passwords can be
15605 obtained from various sources. Delete the PEM_cb function and make
15606 it the default behaviour: i.e. if the callback is NULL and the
15607 usrdata argument is not NULL interpret it as a null terminated pass
15608 phrase. If usrdata and the callback are NULL then the pass phrase
15609 is prompted for as usual.
15613 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15614 the support is automatically enabled. The resulting binaries will
15615 autodetect the card and use it if present.
15617 *Ben Laurie and Compaq Inc.*
15619 * Work around for Netscape hang bug. This sends certificate request
15620 and server done in one record. Since this is perfectly legal in the
15621 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15622 the bugs/SSLv3 entry for more info.
15626 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15630 * Add -rand argument to smime and pkcs12 applications and read/write
15635 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15639 * Add command line password options to the remaining applications.
15643 * Bug fix for BN_div_recp() for numerators with an even number of
15648 * More tests in bntest.c, and changed test_bn output.
15652 * ./config recognizes MacOS X now.
15656 * Bug fix for BN_div() when the first words of num and divisor are
15657 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15661 * Add support for various broken PKCS#8 formats, and command line
15662 options to produce them.
15666 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15667 get temporary BIGNUMs from a BN_CTX.
15671 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15676 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15677 include a #define from the old name to the new. The original intent
15678 was that statically linked binaries could for example just call
15679 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15680 link with digests. This never worked because SSLeay_add_all_digests()
15681 and SSLeay_add_all_ciphers() were in the same source file so calling
15682 one would link with the other. They are now in separate source files.
15686 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15690 * Use a less unusual form of the Miller-Rabin primality test (it used
15691 a binary algorithm for exponentiation integrated into the Miller-Rabin
15692 loop, our standard modexp algorithms are faster).
15696 * Support for the EBCDIC character set completed.
15698 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15700 * Source code cleanups: use const where appropriate, eliminate casts,
15701 use `void *` instead of `char *` in lhash.
15705 * Bugfix: ssl3_send_server_key_exchange was not restartable
15706 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15707 this the server could overwrite ephemeral keys that the client
15712 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15713 using 50 iterations of the Rabin-Miller test.
15715 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15716 iterations of the Rabin-Miller test as required by the appendix
15717 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15718 As BN_is_prime_fasttest includes trial division, DSA parameter
15719 generation becomes much faster.
15721 This implies a change for the callback functions in DSA_is_prime
15722 and DSA_generate_parameters: The callback function is called once
15723 for each positive witness in the Rabin-Miller test, not just
15724 occasionally in the inner loop; and the parameters to the
15725 callback function now provide an iteration count for the outer
15726 loop rather than for the current invocation of the inner loop.
15727 DSA_generate_parameters additionally can call the callback
15728 function with an 'iteration count' of -1, meaning that a
15729 candidate has passed the trial division test (when q is generated
15730 from an application-provided seed, trial division is skipped).
15734 * New function BN_is_prime_fasttest that optionally does trial
15735 division before starting the Rabin-Miller test and has
15736 an additional BN_CTX * argument (whereas BN_is_prime always
15737 has to allocate at least one BN_CTX).
15738 'callback(1, -1, cb_arg)' is called when a number has passed the
15739 trial division stage.
15743 * Fix for bug in CRL encoding. The validity dates weren't being handled
15748 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15752 * New function BN_pseudo_rand().
15756 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15757 bignum version of BN_from_montgomery() with the working code from
15758 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15763 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15764 made it impossible to use the same SSL_SESSION data structure in
15765 SSL2 clients in multiple threads.
15769 * The return value of RAND_load_file() no longer counts bytes obtained
15770 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15771 to seed the PRNG (previously an explicit byte count was required).
15773 *Ulf Möller, Bodo Möller*
15775 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15776 used `char *` instead of `void *` and had casts all over the place.
15780 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15784 * Retain source code compatibility for BN_prime_checks macro:
15785 BN_is_prime(..., BN_prime_checks, ...) now uses
15786 BN_prime_checks_for_size to determine the appropriate number of
15787 Rabin-Miller iterations.
15791 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15792 DH_CHECK_P_NOT_SAFE_PRIME.
15793 (Check if this is true? OpenPGP calls them "strong".)
15797 * Merge the functionality of "dh" and "gendh" programs into a new program
15798 "dhparam". The old programs are retained for now but will handle DH keys
15799 (instead of parameters) in future.
15803 * Make the ciphers, s_server and s_client programs check the return values
15804 when a new cipher list is set.
15808 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15809 ciphers. Before when the 56bit ciphers were enabled the sorting was
15812 The syntax for the cipher sorting has been extended to support sorting by
15813 cipher-strength (using the strength_bits hard coded in the tables).
15814 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15816 Fix a bug in the cipher-command parser: when supplying a cipher command
15817 string with an "undefined" symbol (neither command nor alphanumeric
15818 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15819 an error is flagged.
15821 Due to the strength-sorting extension, the code of the
15822 ssl_create_cipher_list() function was completely rearranged. I hope that
15823 the readability was also increased :-)
15825 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15827 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15828 for the first serial number and places 2 in the serial number file. This
15829 avoids problems when the root CA is created with serial number zero and
15830 the first user certificate has the same issuer name and serial number
15835 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15836 the new code. Add documentation for this stuff.
15840 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15841 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15842 structures and behave in an analogous way to the X509v3 functions:
15843 they shouldn't be called directly but wrapper functions should be used
15846 So we also now have some wrapper functions that call the X509at functions
15847 when passed certificate requests. (TO DO: similar things can be done with
15848 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15849 things. Some of these need some d2i or i2d and print functionality
15850 because they handle more complex structures.)
15854 * Add missing #ifndefs that caused missing symbols when building libssl
15855 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15856 NO_RSA in `ssl/s2*.c`.
15858 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15860 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15861 has a return value which indicates the quality of the random data
15862 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15863 error queue. New function RAND_pseudo_bytes() generates output that is
15864 guaranteed to be unique but not unpredictable. RAND_add is like
15865 RAND_seed, but takes an extra argument for an entropy estimate
15866 (RAND_seed always assumes full entropy).
15870 * Do more iterations of Rabin-Miller probable prime test (specifically,
15871 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15872 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15873 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15874 false-positive rate of at most 2^-80 for random input.
15878 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15882 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15883 in the 0.9.5 release), this returns the chain
15884 from an X509_CTX structure with a dup of the stack and all
15885 the X509 reference counts upped: so the stack will exist
15886 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15889 Also make SSL_SESSION_print() print out the verify return
15894 * Add manpage for the pkcs12 command. Also change the default
15895 behaviour so MAC iteration counts are used unless the new
15896 -nomaciter option is used. This improves file security and
15897 only older versions of MSIE (4.0 for example) need it.
15901 * Honor the no-xxx Configure options when creating .DEF files.
15905 * Add PKCS#10 attributes to field table: challengePassword,
15906 unstructuredName and unstructuredAddress. These are taken from
15907 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15908 international characters are used.
15910 More changes to X509_ATTRIBUTE code: allow the setting of types
15911 based on strings. Remove the 'loc' parameter when adding
15912 attributes because these will be a SET OF encoding which is sorted
15917 * Initial changes to the 'req' utility to allow request generation
15918 automation. This will allow an application to just generate a template
15919 file containing all the field values and have req construct the
15922 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15923 used all over the place including certificate requests and PKCS#7
15924 structures. They are currently handled manually where necessary with
15925 some primitive wrappers for PKCS#7. The new functions behave in a
15926 manner analogous to the X509 extension functions: they allow
15927 attributes to be looked up by NID and added.
15929 Later something similar to the X509V3 code would be desirable to
15930 automatically handle the encoding, decoding and printing of the
15931 more complex types. The string types like challengePassword can
15932 be handled by the string table functions.
15934 Also modified the multi byte string table handling. Now there is
15935 a 'global mask' which masks out certain types. The table itself
15936 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15937 is useful when for example there is only one permissible type
15938 (as in countryName) and using the mask might result in no valid
15943 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15944 SSL_get_peer_finished to allow applications to obtain the latest
15945 Finished messages sent to the peer or expected from the peer,
15946 respectively. (SSL_get_peer_finished is usually the Finished message
15947 actually received from the peer, otherwise the protocol will be aborted.)
15949 As the Finished message are message digests of the complete handshake
15950 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15951 be used for external authentication procedures when the authentication
15952 provided by SSL/TLS is not desired or is not enough.
15956 * Enhanced support for Alpha Linux is added. Now ./config checks if
15957 the host supports BWX extension and if Compaq C is present on the
15958 $PATH. Just exploiting of the BWX extension results in 20-30%
15959 performance kick for some algorithms, e.g. DES and RC4 to mention
15960 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15965 * Add support for MS "fast SGC". This is arguably a violation of the
15966 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15967 weak crypto and after checking the certificate is SGC a second one
15968 with strong crypto. MS SGC stops the first handshake after receiving
15969 the server certificate message and sends a second client hello. Since
15970 a server will typically do all the time consuming operations before
15971 expecting any further messages from the client (server key exchange
15972 is the most expensive) there is little difference between the two.
15974 To get OpenSSL to support MS SGC we have to permit a second client
15975 hello message after we have sent server done. In addition we have to
15976 reset the MAC if we do get this second client hello.
15980 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15981 if a DER encoded private key is RSA or DSA traditional format. Changed
15982 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15983 format DER encoded private key. Newer code should use PKCS#8 format which
15984 has the key type encoded in the ASN1 structure. Added DER private key
15985 support to pkcs8 application.
15989 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15990 ciphersuites has been selected (as required by the SSL 3/TLS 1
15991 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15992 is set, we interpret this as a request to violate the specification
15993 (the worst that can happen is a handshake failure, and 'correct'
15994 behaviour would result in a handshake failure anyway).
15998 * In SSL_CTX_add_session, take into account that there might be multiple
15999 SSL_SESSION structures with the same session ID (e.g. when two threads
16000 concurrently obtain them from an external cache).
16001 The internal cache can handle only one SSL_SESSION with a given ID,
16002 so if there's a conflict, we now throw out the old one to achieve
16007 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16008 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16009 some routines that use cipher OIDs: some ciphers do not have OIDs
16010 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16015 * Simplify the trust setting structure and code. Now we just have
16016 two sequences of OIDs for trusted and rejected settings. These will
16017 typically have values the same as the extended key usage extension
16018 and any application specific purposes.
16020 The trust checking code now has a default behaviour: it will just
16021 check for an object with the same NID as the passed id. Functions can
16022 be provided to override either the default behaviour or the behaviour
16023 for a given id. SSL client, server and email already have functions
16024 in place for compatibility: they check the NID and also return "trusted"
16025 if the certificate is self signed.
16029 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16030 traditional format into an EVP_PKEY structure.
16034 * Add a password callback function PEM_cb() which either prompts for
16035 a password if usr_data is NULL or otherwise assumes it is a null
16036 terminated password. Allow passwords to be passed on command line
16037 environment or config files in a few more utilities.
16041 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16042 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16043 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16044 Update documentation.
16048 * Support for ASN1 "NULL" type. This could be handled before by using
16049 ASN1_TYPE but there wasn't any function that would try to read a NULL
16050 and produce an error if it couldn't. For compatibility we also have
16051 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16052 don't allocate anything because they don't need to.
16056 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16059 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16061 * Rebuild of the memory allocation routines used by OpenSSL code and
16062 possibly others as well. The purpose is to make an interface that
16063 provide hooks so anyone can build a separate set of allocation and
16064 deallocation routines to be used by OpenSSL, for example memory
16065 pool implementations, or something else, which was previously hard
16066 since Malloc(), Realloc() and Free() were defined as macros having
16067 the values malloc, realloc and free, respectively (except for Win32
16068 compilations). The same is provided for memory debugging code.
16069 OpenSSL already comes with functionality to find memory leaks, but
16070 this gives people a chance to debug other memory problems.
16072 With these changes, a new set of functions and macros have appeared:
16074 CRYPTO_set_mem_debug_functions() [F]
16075 CRYPTO_get_mem_debug_functions() [F]
16076 CRYPTO_dbg_set_options() [F]
16077 CRYPTO_dbg_get_options() [F]
16078 CRYPTO_malloc_debug_init() [M]
16080 The memory debug functions are NULL by default, unless the library
16081 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16082 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16083 gives the standard debugging functions that come with OpenSSL) or
16084 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16085 provided by the library user) must be used. When the standard
16086 debugging functions are used, CRYPTO_dbg_set_options can be used to
16087 request additional information:
16088 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16089 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16091 Also, things like CRYPTO_set_mem_functions will always give the
16092 expected result (the new set of functions is used for allocation
16093 and deallocation) at all times, regardless of platform and compiler
16096 To finish it up, some functions that were never use in any other
16097 way than through macros have a new API and new semantic:
16099 CRYPTO_dbg_malloc()
16100 CRYPTO_dbg_realloc()
16103 All macros of value have retained their old syntax.
16105 *Richard Levitte and Bodo Moeller*
16107 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16108 ordering of SMIMECapabilities wasn't in "strength order" and there
16109 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16114 * Some ASN1 types with illegal zero length encoding (INTEGER,
16115 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16117 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16119 * Merge in my S/MIME library for OpenSSL. This provides a simple
16120 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16121 functionality to handle multipart/signed properly) and a utility
16122 called 'smime' to call all this stuff. This is based on code I
16123 originally wrote for Celo who have kindly allowed it to be
16124 included in OpenSSL.
16128 * Add variants des_set_key_checked and des_set_key_unchecked of
16129 des_set_key (aka des_key_sched). Global variable des_check_key
16130 decides which of these is called by des_set_key; this way
16131 des_check_key behaves as it always did, but applications and
16132 the library itself, which was buggy for des_check_key == 1,
16133 have a cleaner way to pick the version they need.
16137 * New function PKCS12_newpass() which changes the password of a
16142 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16143 dynamic mix. In both cases the ids can be used as an index into the
16144 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16145 functions so they accept a list of the field values and the
16146 application doesn't need to directly manipulate the X509_TRUST
16151 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16156 * Modify the way the V3 extension code looks up extensions. This now
16157 works in a similar way to the object code: we have some "standard"
16158 extensions in a static table which is searched with OBJ_bsearch()
16159 and the application can add dynamic ones if needed. The file
16160 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16161 updated whenever a new extension is added to the core code and kept
16162 in ext_nid order. There is a simple program 'tabtest.c' which checks
16163 this. New extensions are not added too often so this file can readily
16164 be maintained manually.
16166 There are two big advantages in doing things this way. The extensions
16167 can be looked up immediately and no longer need to be "added" using
16168 X509V3_add_standard_extensions(): this function now does nothing.
16169 Side note: I get *lots* of email saying the extension code doesn't
16170 work because people forget to call this function.
16171 Also no dynamic allocation is done unless new extensions are added:
16172 so if we don't add custom extensions there is no need to call
16173 X509V3_EXT_cleanup().
16177 * Modify enc utility's salting as follows: make salting the default. Add a
16178 magic header, so unsalted files fail gracefully instead of just decrypting
16179 to garbage. This is because not salting is a big security hole, so people
16180 should be discouraged from doing it.
16184 * Fixes and enhancements to the 'x509' utility. It allowed a message
16185 digest to be passed on the command line but it only used this
16186 parameter when signing a certificate. Modified so all relevant
16187 operations are affected by the digest parameter including the
16188 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16189 DSA key was used because it didn't fix the digest.
16193 * Initial certificate chain verify code. Currently tests the untrusted
16194 certificates for consistency with the verify purpose (which is set
16195 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16197 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16198 this is because it will reject chains with invalid extensions whereas
16199 every previous version of OpenSSL and SSLeay made no checks at all.
16201 Trust code: checks the root CA for the relevant trust settings. Trust
16202 settings have an initial value consistent with the verify purpose: e.g.
16203 if the verify purpose is for SSL client use it expects the CA to be
16204 trusted for SSL client use. However the default value can be changed to
16205 permit custom trust settings: one example of this would be to only trust
16206 certificates from a specific "secure" set of CAs.
16208 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16209 which should be used for version portability: especially since the
16210 verify structure is likely to change more often now.
16212 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16213 to set them. If not set then assume SSL clients will verify SSL servers
16216 Two new options to the verify program: -untrusted allows a set of
16217 untrusted certificates to be passed in and -purpose which sets the
16218 intended purpose of the certificate. If a purpose is set then the
16219 new chain verify code is used to check extension consistency.
16223 * Support for the authority information access extension.
16227 * Modify RSA and DSA PEM read routines to transparently handle
16228 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16229 public keys in a format compatible with certificate
16230 SubjectPublicKeyInfo structures. Unfortunately there were already
16231 functions called *_PublicKey_* which used various odd formats so
16232 these are retained for compatibility: however the DSA variants were
16233 never in a public release so they have been deleted. Changed dsa/rsa
16234 utilities to handle the new format: note no releases ever handled public
16235 keys so we should be OK.
16237 The primary motivation for this change is to avoid the same fiasco
16238 that dogs private keys: there are several incompatible private key
16239 formats some of which are standard and some OpenSSL specific and
16240 require various evil hacks to allow partial transparent handling and
16241 even then it doesn't work with DER formats. Given the option anything
16242 other than PKCS#8 should be dumped: but the other formats have to
16243 stay in the name of compatibility.
16245 With public keys and the benefit of hindsight one standard format
16246 is used which works with EVP_PKEY, RSA or DSA structures: though
16247 it clearly returns an error if you try to read the wrong kind of key.
16249 Added a -pubkey option to the 'x509' utility to output the public key.
16250 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16251 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16252 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16253 that do the same as the `EVP_PKEY_assign_*()` except they up the
16254 reference count of the added key (they don't "swallow" the
16259 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16260 CRLs would fail if the file contained no certificates or no CRLs:
16261 added a new function to read in both types and return the number
16262 read: this means that if none are read it will be an error. The
16263 DER versions of the certificate and CRL reader would always fail
16264 because it isn't possible to mix certificates and CRLs in DER format
16265 without choking one or the other routine. Changed this to just read
16266 a certificate: this is the best we can do. Also modified the code
16267 in apps/verify.c to take notice of return codes: it was previously
16268 attempting to read in certificates from NULL pointers and ignoring
16269 any errors: this is one reason why the cert and CRL reader seemed
16270 to work. It doesn't check return codes from the default certificate
16271 routines: these may well fail if the certificates aren't installed.
16275 * Code to support otherName option in GeneralName.
16279 * First update to verify code. Change the verify utility
16280 so it warns if it is passed a self signed certificate:
16281 for consistency with the normal behaviour. X509_verify
16282 has been modified to it will now verify a self signed
16283 certificate if *exactly* the same certificate appears
16284 in the store: it was previously impossible to trust a
16285 single self signed certificate. This means that:
16286 openssl verify ss.pem
16287 now gives a warning about a self signed certificate but
16288 openssl verify -CAfile ss.pem ss.pem
16293 * For servers, store verify_result in SSL_SESSION data structure
16294 (and add it to external session representation).
16295 This is needed when client certificate verifications fails,
16296 but an application-provided verification callback (set by
16297 SSL_CTX_set_cert_verify_callback) allows accepting the session
16298 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16299 but returns 1): When the session is reused, we have to set
16300 ssl->verify_result to the appropriate error code to avoid
16303 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16305 * Fix a bug in the new PKCS#7 code: it didn't consider the
16306 case in PKCS7_dataInit() where the signed PKCS7 structure
16307 didn't contain any existing data because it was being created.
16309 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16311 * Add a salt to the key derivation routines in enc.c. This
16312 forms the first 8 bytes of the encrypted file. Also add a
16313 -S option to allow a salt to be input on the command line.
16317 * New function X509_cmp(). Oddly enough there wasn't a function
16318 to compare two certificates. We do this by working out the SHA1
16319 hash and comparing that. X509_cmp() will be needed by the trust
16324 * SSL_get1_session() is like SSL_get_session(), but increments
16325 the reference count in the SSL_SESSION returned.
16327 *Geoff Thorpe <geoff@eu.c2.net>*
16329 * Fix for 'req': it was adding a null to request attributes.
16330 Also change the X509_LOOKUP and X509_INFO code to handle
16331 certificate auxiliary information.
16335 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16340 * Add the possibility to add extra information to the memory leak
16341 detecting output, to form tracebacks, showing from where each
16342 allocation was originated: CRYPTO_push_info("constant string") adds
16343 the string plus current file name and line number to a per-thread
16344 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16345 is like calling CYRPTO_pop_info() until the stack is empty.
16346 Also updated memory leak detection code to be multi-thread-safe.
16350 * Add options -text and -noout to pkcs7 utility and delete the
16351 encryption options which never did anything. Update docs.
16355 * Add options to some of the utilities to allow the pass phrase
16356 to be included on either the command line (not recommended on
16357 OSes like Unix) or read from the environment. Update the
16358 manpages and fix a few bugs.
16362 * Add a few manpages for some of the openssl commands.
16366 * Fix the -revoke option in ca. It was freeing up memory twice,
16367 leaking and not finding already revoked certificates.
16371 * Extensive changes to support certificate auxiliary information.
16372 This involves the use of X509_CERT_AUX structure and X509_AUX
16373 functions. An X509_AUX function such as PEM_read_X509_AUX()
16374 can still read in a certificate file in the usual way but it
16375 will also read in any additional "auxiliary information". By
16376 doing things this way a fair degree of compatibility can be
16377 retained: existing certificates can have this information added
16378 using the new 'x509' options.
16380 Current auxiliary information includes an "alias" and some trust
16381 settings. The trust settings will ultimately be used in enhanced
16382 certificate chain verification routines: currently a certificate
16383 can only be trusted if it is self signed and then it is trusted
16388 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16389 The problem was that one of the replacement routines had not been working
16390 since SSLeay releases. For now the offending routine has been replaced
16391 with non-optimised assembler. Even so, this now gives around 95%
16392 performance improvement for 1024 bit RSA signs.
16396 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16397 handling. Most clients have the effective key size in bits equal to
16398 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16399 A few however don't do this and instead use the size of the decrypted key
16400 to determine the RC2 key length and the AlgorithmIdentifier to determine
16401 the effective key length. In this case the effective key length can still
16402 be 40 bits but the key length can be 168 bits for example. This is fixed
16403 by manually forcing an RC2 key into the EVP_PKEY structure because the
16404 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16405 the key length and effective key length are equal.
16409 * Add a bunch of functions that should simplify the creation of
16410 X509_NAME structures. Now you should be able to do:
16411 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16412 and have it automatically work out the correct field type and fill in
16413 the structures. The more adventurous can try:
16414 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16415 and it will (hopefully) work out the correct multibyte encoding.
16419 * Change the 'req' utility to use the new field handling and multibyte
16420 copy routines. Before the DN field creation was handled in an ad hoc
16421 way in req, ca, and x509 which was rather broken and didn't support
16422 BMPStrings or UTF8Strings. Since some software doesn't implement
16423 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16424 using the dirstring_type option. See the new comment in the default
16425 openssl.cnf for more info.
16429 * Make crypto/rand/md_rand.c more robust:
16430 - Assure unique random numbers after fork().
16431 - Make sure that concurrent threads access the global counter and
16432 md serializably so that we never lose entropy in them
16433 or use exactly the same state in multiple threads.
16434 Access to the large state is not always serializable because
16435 the additional locking could be a performance killer, and
16436 md should be large enough anyway.
16440 * New file apps/app_rand.c with commonly needed functionality
16441 for handling the random seed file.
16443 Use the random seed file in some applications that previously did not:
16445 dsaparam -genkey (which also ignored its '-rand' option),
16448 x509 (when signing).
16449 Except on systems with /dev/urandom, it is crucial to have a random
16450 seed file at least for key creation, DSA signing, and for DH exchanges;
16451 for RSA signatures we could do without one.
16453 gendh and gendsa (unlike genrsa) used to read only the first byte
16454 of each file listed in the '-rand' option. The function as previously
16455 found in genrsa is now in app_rand.c and is used by all programs
16456 that support '-rand'.
16460 * In RAND_write_file, use mode 0600 for creating files;
16461 don't just chmod when it may be too late.
16465 * Report an error from X509_STORE_load_locations
16466 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16470 * New function ASN1_mbstring_copy() this copies a string in either
16471 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16472 into an ASN1_STRING type. A mask of permissible types is passed
16473 and it chooses the "minimal" type to use or an error if not type
16478 * Add function equivalents to the various macros in asn1.h. The old
16479 macros are retained with an `M_` prefix. Code inside the library can
16480 use the `M_` macros. External code (including the openssl utility)
16481 should *NOT* in order to be "shared library friendly".
16485 * Add various functions that can check a certificate's extensions
16486 to see if it usable for various purposes such as SSL client,
16487 server or S/MIME and CAs of these types. This is currently
16488 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16489 verification. Also added a -purpose flag to x509 utility to
16490 print out all the purposes.
16494 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16499 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16500 for, obtain and decode and extension and obtain its critical flag.
16501 This allows all the necessary extension code to be handled in a
16502 single function call.
16506 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16507 platforms. See crypto/rc4/rc4_enc.c for further details.
16511 * New -noout option to asn1parse. This causes no output to be produced
16512 its main use is when combined with -strparse and -out to extract data
16513 from a file (which may not be in ASN.1 format).
16517 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16518 when producing the local key id.
16520 *Richard Levitte <levitte@stacken.kth.se>*
16522 * New option -dhparam in s_server. This allows a DH parameter file to be
16523 stated explicitly. If it is not stated then it tries the first server
16524 certificate file. The previous behaviour hard coded the filename
16529 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16530 a public key to be input or output. For example:
16531 openssl rsa -in key.pem -pubout -out pubkey.pem
16532 Also added necessary DSA public key functions to handle this.
16536 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16537 in the message. This was handled by allowing
16538 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16540 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16542 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16543 to the end of the strings whereas this didn't. This would cause problems
16544 if strings read with d2i_ASN1_bytes() were later modified.
16546 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16548 * Fix for base64 decode bug. When a base64 bio reads only one line of
16549 data and it contains EOF it will end up returning an error. This is
16550 caused by input 46 bytes long. The cause is due to the way base64
16551 BIOs find the start of base64 encoded data. They do this by trying a
16552 trial decode on each line until they find one that works. When they
16553 do a flag is set and it starts again knowing it can pass all the
16554 data directly through the decoder. Unfortunately it doesn't reset
16555 the context it uses. This means that if EOF is reached an attempt
16556 is made to pass two EOFs through the context and this causes the
16557 resulting error. This can also cause other problems as well. As is
16558 usual with these problems it takes *ages* to find and the fix is
16559 trivial: move one line.
16561 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16563 * Ugly workaround to get s_client and s_server working under Windows. The
16564 old code wouldn't work because it needed to select() on sockets and the
16565 tty (for keypresses and to see if data could be written). Win32 only
16566 supports select() on sockets so we select() with a 1s timeout on the
16567 sockets and then see if any characters are waiting to be read, if none
16568 are present then we retry, we also assume we can always write data to
16569 the tty. This isn't nice because the code then blocks until we've
16570 received a complete line of data and it is effectively polling the
16571 keyboard at 1s intervals: however it's quite a bit better than not
16572 working at all :-) A dedicated Windows application might handle this
16573 with an event loop for example.
16577 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16578 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16579 will be called when RSA_sign() and RSA_verify() are used. This is useful
16580 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16581 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16582 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16583 This necessitated the support of an extra signature type NID_md5_sha1
16584 for SSL signatures and modifications to the SSL library to use it instead
16585 of calling RSA_public_decrypt() and RSA_private_encrypt().
16589 * Add new -verify -CAfile and -CApath options to the crl program, these
16590 will lookup a CRL issuers certificate and verify the signature in a
16591 similar way to the verify program. Tidy up the crl program so it
16592 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16593 less strict. It will now permit CRL extensions even if it is not
16594 a V2 CRL: this will allow it to tolerate some broken CRLs.
16598 * Initialize all non-automatic variables each time one of the openssl
16599 sub-programs is started (this is necessary as they may be started
16600 multiple times from the "OpenSSL>" prompt).
16602 *Lennart Bang, Bodo Moeller*
16604 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16605 removing all other RSA functionality (this is what NO_RSA does). This
16606 is so (for example) those in the US can disable those operations covered
16607 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16612 * Non-copying interface to BIO pairs.
16613 (still largely untested)
16617 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16618 ASCII string. This was handled independently in various places before.
16622 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16623 UTF8 strings a character at a time.
16627 * Use client_version from client hello to select the protocol
16628 (s23_srvr.c) and for RSA client key exchange verification
16629 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16633 * Add various utility functions to handle SPKACs, these were previously
16634 handled by poking round in the structure internals. Added new function
16635 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16636 print, verify and generate SPKACs. Based on an original idea from
16637 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16641 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16645 * Allow the config file extension section to be overwritten on the
16646 command line. Based on an original idea from Massimiliano Pala
16647 <madwolf@comune.modena.it>. The new option is called -extensions
16648 and can be applied to ca, req and x509. Also -reqexts to override
16649 the request extensions in req and -crlexts to override the crl extensions
16654 * Add new feature to the SPKAC handling in ca. Now you can include
16655 the same field multiple times by preceding it by "XXXX." for example:
16658 this is the same syntax as used in the req config file.
16662 * Allow certificate extensions to be added to certificate requests. These
16663 are specified in a 'req_extensions' option of the req section of the
16664 config file. They can be printed out with the -text option to req but
16665 are otherwise ignored at present.
16669 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16670 data read consists of only the final block it would not decrypted because
16671 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16672 A misplaced 'break' also meant the decrypted final block might not be
16673 copied until the next read.
16677 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16678 a few extra parameters to the DH structure: these will be useful if
16679 for example we want the value of 'q' or implement X9.42 DH.
16683 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16684 provides hooks that allow the default DSA functions or functions on a
16685 "per key" basis to be replaced. This allows hardware acceleration and
16686 hardware key storage to be handled without major modification to the
16687 library. Also added low level modexp hooks and CRYPTO_EX structure and
16688 associated functions.
16692 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16693 as "read only": it can't be written to and the buffer it points to will
16694 not be freed. Reading from a read only BIO is much more efficient than
16695 a normal memory BIO. This was added because there are several times when
16696 an area of memory needs to be read from a BIO. The previous method was
16697 to create a memory BIO and write the data to it, this results in two
16698 copies of the data and an O(n^2) reading algorithm. There is a new
16699 function BIO_new_mem_buf() which creates a read only memory BIO from
16700 an area of memory. Also modified the PKCS#7 routines to use read only
16705 * Bugfix: ssl23_get_client_hello did not work properly when called in
16706 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16707 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16708 but a retry condition occurred while trying to read the rest.
16712 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16713 NID_pkcs7_encrypted by default: this was wrong since this should almost
16714 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16715 the encrypted data type: this is a more sensible place to put it and it
16716 allows the PKCS#12 code to be tidied up that duplicated this
16721 * Changed obj_dat.pl script so it takes its input and output files on
16722 the command line. This should avoid shell escape redirection problems
16727 * Initial support for certificate extension requests, these are included
16728 in things like Xenroll certificate requests. Included functions to allow
16729 extensions to be obtained and added.
16733 * -crlf option to s_client and s_server for sending newlines as
16734 CRLF (as required by many protocols).
16738 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16740 * Install libRSAglue.a when OpenSSL is built with RSAref.
16742 *Ralf S. Engelschall*
16744 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16746 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16748 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16753 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16754 DH parameters/keys (q is lost during that conversion, but the resulting
16755 DH parameters contain its length).
16757 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16758 much faster than DH_generate_parameters (which creates parameters
16759 where `p = 2*q + 1`), and also the smaller q makes DH computations
16760 much more efficient (160-bit exponentiation instead of 1024-bit
16761 exponentiation); so this provides a convenient way to support DHE
16762 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16763 utter importance to use
16764 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16766 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16767 when such DH parameters are used, because otherwise small subgroup
16768 attacks may become possible!
16772 * Avoid memory leak in i2d_DHparams.
16776 * Allow the -k option to be used more than once in the enc program:
16777 this allows the same encrypted message to be read by multiple recipients.
16781 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16782 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16783 it will always use the numerical form of the OID, even if it has a short
16788 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16789 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16790 otherwise bn_mod_exp was called. In the case of hardware keys for example
16791 no private key components need be present and it might store extra data
16792 in the RSA structure, which cannot be accessed from bn_mod_exp.
16793 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16794 private key operations.
16798 * Added support for SPARC Linux.
16802 * pem_password_cb function type incompatibly changed from
16803 typedef int pem_password_cb(char *buf, int size, int rwflag);
16805 ....(char *buf, int size, int rwflag, void *userdata);
16806 so that applications can pass data to their callbacks:
16807 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16808 additional void * argument, which is just handed through whenever
16809 the password callback is called.
16811 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16813 New function SSL_CTX_set_default_passwd_cb_userdata.
16815 Compatibility note: As many C implementations push function arguments
16816 onto the stack in reverse order, the new library version is likely to
16817 interoperate with programs that have been compiled with the old
16818 pem_password_cb definition (PEM_whatever takes some data that
16819 happens to be on the stack as its last argument, and the callback
16820 just ignores this garbage); but there is no guarantee whatsoever that
16823 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16824 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16825 problems not only on Windows, but also on some Unix platforms.
16826 To avoid problematic command lines, these definitions are now in an
16827 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16828 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16832 * MIPS III/IV assembler module is reimplemented.
16836 * More DES library cleanups: remove references to srand/rand and
16837 delete an unused file.
16841 * Add support for the free Netwide assembler (NASM) under Win32,
16842 since not many people have MASM (ml) and it can be hard to obtain.
16843 This is currently experimental but it seems to work OK and pass all
16844 the tests. Check out INSTALL.W32 for info.
16848 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16849 without temporary keys kept an extra copy of the server key,
16850 and connections with temporary keys did not free everything in case
16855 * New function RSA_check_key and new openssl rsa option -check
16856 for verifying the consistency of RSA keys.
16858 *Ulf Moeller, Bodo Moeller*
16860 * Various changes to make Win32 compile work:
16861 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16862 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16863 comparison" warnings.
16864 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16868 * Add a debugging option to PKCS#5 v2 key generation function: when
16869 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16870 derived keys are printed to stderr.
16874 * Copy the flags in ASN1_STRING_dup().
16876 *Roman E. Pavlov <pre@mo.msk.ru>*
16878 * The x509 application mishandled signing requests containing DSA
16879 keys when the signing key was also DSA and the parameters didn't match.
16881 It was supposed to omit the parameters when they matched the signing key:
16882 the verifying software was then supposed to automatically use the CA's
16883 parameters if they were absent from the end user certificate.
16885 Omitting parameters is no longer recommended. The test was also
16886 the wrong way round! This was probably due to unusual behaviour in
16887 EVP_cmp_parameters() which returns 1 if the parameters match.
16888 This meant that parameters were omitted when they *didn't* match and
16889 the certificate was useless. Certificates signed with 'ca' didn't have
16892 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16894 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16895 The interface is as follows:
16896 Applications can use
16897 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16898 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16899 "off" is now the default.
16900 The library internally uses
16901 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16902 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16903 to disable memory-checking temporarily.
16905 Some inconsistent states that previously were possible (and were
16906 even the default) are now avoided.
16908 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16909 with each memory chunk allocated; this is occasionally more helpful
16910 than just having a counter.
16912 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16914 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16919 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16920 which largely parallels "options", but is for changing API behaviour,
16921 whereas "options" are about protocol behaviour.
16922 Initial "mode" flags are:
16924 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16925 a single record has been written.
16926 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16927 retries use the same buffer location.
16928 (But all of the contents must be
16933 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16936 * Fix problems with no-hmac etc.
16938 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16940 * New functions RSA_get_default_method(), RSA_set_method() and
16941 RSA_get_method(). These allows replacement of RSA_METHODs without having
16942 to mess around with the internals of an RSA structure.
16946 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16947 Also really enable memory leak checks in openssl.c and in some
16950 *Chad C. Mulligan, Bodo Moeller*
16952 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16953 up the length of negative integers. This has now been simplified to just
16954 store the length when it is first determined and use it later, rather
16955 than trying to keep track of where data is copied and updating it to
16957 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16959 * Add a new function PKCS7_signatureVerify. This allows the verification
16960 of a PKCS#7 signature but with the signing certificate passed to the
16961 function itself. This contrasts with PKCS7_dataVerify which assumes the
16962 certificate is present in the PKCS#7 structure. This isn't always the
16963 case: certificates can be omitted from a PKCS#7 structure and be
16964 distributed by "out of band" means (such as a certificate database).
16968 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16969 function prototypes in pem.h, also change util/mkdef.pl to add the
16970 necessary function names.
16974 * mk1mf.pl (used by Windows builds) did not properly read the
16975 options set by Configure in the top level Makefile, and Configure
16976 was not even able to write more than one option correctly.
16977 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16981 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16982 file to be loaded from a BIO or FILE pointer. The BIO version will
16983 for example allow memory BIOs to contain config info.
16987 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16988 Whoever hopes to achieve shared-library compatibility across versions
16989 must use this, not the compile-time macro.
16990 (Exercise 0.9.4: Which is the minimum library version required by
16992 Note: All this applies only to multi-threaded programs, others don't
16997 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16998 through a BIO pair triggered the default case, i.e.
16999 SSLerr(...,SSL_R_UNKNOWN_STATE).
17003 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17004 can use the SSL library even if none of the specific BIOs is
17009 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17010 for the encoded length.
17012 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17014 * Add initial documentation of the X509V3 functions.
17018 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17019 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17020 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17021 secure PKCS#8 private key format with a high iteration count.
17025 * Fix determination of Perl interpreter: A perl or perl5
17026 *directory* in $PATH was also accepted as the interpreter.
17028 *Ralf S. Engelschall*
17030 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17031 wrong with it but it was very old and did things like calling
17032 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17033 unusual formatting.
17037 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17038 to use the new extension code.
17042 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17043 with macros. This should make it easier to change their form, add extra
17044 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17049 * Add to configuration table a new entry that can specify an alternative
17050 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17051 according to Mark Crispin <MRC@Panda.COM>.
17055 * DES CBC did not update the IV. Weird.
17059 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17060 Changing the behaviour of the former might break existing programs --
17061 where IV updating is needed, des_ncbc_encrypt can be used.
17064 * When bntest is run from "make test" it drives bc to check its
17065 calculations, as well as internally checking them. If an internal check
17066 fails, it needs to cause bc to give a non-zero result or make test carries
17067 on without noticing the failure. Fixed.
17071 * DES library cleanups.
17075 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17076 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17077 ciphers. NOTE: although the key derivation function has been verified
17078 against some published test vectors it has not been extensively tested
17079 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17084 * Instead of "mkdir -p", which is not fully portable, use new
17085 Perl script "util/mkdir-p.pl".
17089 * Rewrite the way password based encryption (PBE) is handled. It used to
17090 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17091 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17092 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17093 the 'parameter' field of the AlgorithmIdentifier is passed to the
17094 underlying key generation function so it must do its own ASN1 parsing.
17095 This has also changed the EVP_PBE_CipherInit() function which now has a
17096 'parameter' argument instead of literal salt and iteration count values
17097 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17101 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17102 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17103 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17104 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17105 value was just used as a "magic string" and not used directly its
17106 value doesn't matter.
17110 * Introduce some semblance of const correctness to BN. Shame C doesn't
17115 * "linux-sparc64" configuration (ultrapenguin).
17117 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17118 "linux-sparc" configuration.
17120 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17122 * config now generates no-xxx options for missing ciphers.
17126 * Support the EBCDIC character set (work in progress).
17127 File ebcdic.c not yet included because it has a different license.
17129 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17131 * Support BS2000/OSD-POSIX.
17133 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17135 * Make callbacks for key generation use `void *` instead of `char *`.
17139 * Make S/MIME samples compile (not yet tested).
17143 * Additional typesafe stacks.
17147 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17151 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17153 * New configuration variant "sco5-gcc".
17155 * Updated some demos.
17157 *Sean O Riordain, Wade Scholine*
17159 * Add missing BIO_free at exit of pkcs12 application.
17163 * Fix memory leak in conf.c.
17167 * Updates for Win32 to assembler version of MD5.
17171 * Set #! path to perl in apps/der_chop to where we found it
17172 instead of using a fixed path.
17176 * SHA library changes for irix64-mips4-cc.
17180 * Improvements for VMS support.
17184 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17186 * Bignum library bug fix. IRIX 6 passes "make test" now!
17187 This also avoids the problems with SC4.2 and unpatched SC5.
17189 *Andy Polyakov <appro@fy.chalmers.se>*
17191 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17192 These are required because of the typesafe stack would otherwise break
17193 existing code. If old code used a structure member which used to be STACK
17194 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17195 sk_num or sk_value it would produce an error because the num, data members
17196 are not present in STACK_OF. Now it just produces a warning. sk_set
17197 replaces the old method of assigning a value to sk_value
17198 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17199 that does this will no longer work (and should use sk_set instead) but
17200 this could be regarded as a "questionable" behaviour anyway.
17204 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17205 correctly handle encrypted S/MIME data.
17209 * Change type of various DES function arguments from des_cblock
17210 (which means, in function argument declarations, pointer to char)
17211 to des_cblock * (meaning pointer to array with 8 char elements),
17212 which allows the compiler to do more typechecking; it was like
17213 that back in SSLeay, but with lots of ugly casts.
17215 Introduce new type const_des_cblock.
17219 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17220 problems: find RecipientInfo structure that matches recipient certificate
17221 and initialise the ASN1 structures properly based on passed cipher.
17225 * Belatedly make the BN tests actually check the results.
17229 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17230 to and from BNs: it was completely broken. New compilation option
17231 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17232 key elements as negative integers.
17236 * Reorganize and speed up MD5.
17238 *Andy Polyakov <appro@fy.chalmers.se>*
17242 *Richard Levitte <richard@levitte.org>*
17244 * New option -out to asn1parse to allow the parsed structure to be
17245 output to a file. This is most useful when combined with the -strparse
17246 option to examine the output of things like OCTET STRINGS.
17250 * Make SSL library a little more fool-proof by not requiring any longer
17251 that `SSL_set_{accept,connect}_state` be called before
17252 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17253 in many applications because usually everything *appeared* to work as
17254 intended anyway -- now it really works as intended).
17258 * Move openssl.cnf out of lib/.
17262 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17263 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17264 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17266 *Ralf S. Engelschall*
17268 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17269 handle PKCS#7 enveloped data properly.
17271 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17273 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17274 copying pointers. The cert_st handling is changed by this in
17275 various ways (and thus what used to be known as ctx->default_cert
17276 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17277 any longer when s->cert does not give us what we need).
17278 ssl_cert_instantiate becomes obsolete by this change.
17279 As soon as we've got the new code right (possibly it already is?),
17280 we have solved a couple of bugs of the earlier code where s->cert
17281 was used as if it could not have been shared with other SSL structures.
17283 Note that using the SSL API in certain dirty ways now will result
17284 in different behaviour than observed with earlier library versions:
17285 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17286 does not influence s as it used to.
17288 In order to clean up things more thoroughly, inside SSL_SESSION
17289 we don't use CERT any longer, but a new structure SESS_CERT
17290 that holds per-session data (if available); currently, this is
17291 the peer's certificate chain and, for clients, the server's certificate
17292 and temporary key. CERT holds only those values that can have
17293 meaningful defaults in an SSL_CTX.
17297 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17298 from the internal representation. Various PKCS#7 fixes: remove some
17299 evil casts and set the enc_dig_alg field properly based on the signing
17304 * Allow PKCS#12 password to be set from the command line or the
17305 environment. Let 'ca' get its config file name from the environment
17306 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17311 * Allow certificate policies extension to use an IA5STRING for the
17312 organization field. This is contrary to the PKIX definition but
17313 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17318 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17319 without disallowing inline assembler and the like for non-pedantic builds.
17323 * Support Borland C++ builder.
17325 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17331 * SHA-1 cleanups and performance enhancements.
17333 *Andy Polyakov <appro@fy.chalmers.se>*
17335 * Sparc v8plus assembler for the bignum library.
17337 *Andy Polyakov <appro@fy.chalmers.se>*
17339 * Accept any -xxx and +xxx compiler options in Configure.
17343 * Update HPUX configuration.
17347 * Add missing `sk_<type>_unshift()` function to safestack.h
17349 *Ralf S. Engelschall*
17351 * New function SSL_CTX_use_certificate_chain_file that sets the
17352 "extra_cert"s in addition to the certificate. (This makes sense
17353 only for "PEM" format files, as chains as a whole are not
17358 * Support verify_depth from the SSL API.
17359 x509_vfy.c had what can be considered an off-by-one-error:
17360 Its depth (which was not part of the external interface)
17361 was actually counting the number of certificates in a chain;
17362 now it really counts the depth.
17366 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17367 instead of X509err, which often resulted in confusing error
17368 messages since the error codes are not globally unique
17369 (e.g. an alleged error in ssl3_accept when a certificate
17370 didn't match the private key).
17372 * New function SSL_CTX_set_session_id_context that allows to set a default
17373 value (so that you don't need SSL_set_session_id_context for each
17374 connection using the SSL_CTX).
17378 * OAEP decoding bug fix.
17382 * Support INSTALL_PREFIX for package builders, as proposed by
17387 * New Configure options "threads" and "no-threads". For systems
17388 where the proper compiler options are known (currently Solaris
17389 and Linux), "threads" is the default.
17393 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17397 * Install various scripts to $(OPENSSLDIR)/misc, not to
17398 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17399 such as /usr/local/bin.
17403 * "make linux-shared" to build shared libraries.
17405 *Niels Poppe <niels@netbox.org>*
17407 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17411 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17412 extension adding in x509 utility.
17416 * Remove NOPROTO sections and error code comments.
17420 * Partial rewrite of the DEF file generator to now parse the ANSI
17425 * New Configure options --prefix=DIR and --openssldir=DIR.
17429 * Complete rewrite of the error code script(s). It is all now handled
17430 by one script at the top level which handles error code gathering,
17431 header rewriting and C source file generation. It should be much better
17432 than the old method: it now uses a modified version of Ulf's parser to
17433 read the ANSI prototypes in all header files (thus the old K&R definitions
17434 aren't needed for error creation any more) and do a better job of
17435 translating function codes into names. The old 'ASN1 error code embedded
17436 in a comment' is no longer necessary and it doesn't use .err files which
17437 have now been deleted. Also the error code call doesn't have to appear all
17438 on one line (which resulted in some large lines...).
17442 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17446 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17447 0 (which usually indicates a closed connection), but continue reading.
17451 * Fix some race conditions.
17455 * Add support for CRL distribution points extension. Add Certificate
17456 Policies and CRL distribution points documentation.
17460 * Move the autogenerated header file parts to crypto/opensslconf.h.
17464 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17465 8 of keying material. Merlin has also confirmed interop with this fix
17466 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17468 *Merlin Hughes <merlin@baltimore.ie>*
17470 * Fix lots of warnings.
17472 *Richard Levitte <levitte@stacken.kth.se>*
17474 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17475 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17477 *Richard Levitte <levitte@stacken.kth.se>*
17479 * Fix problems with sizeof(long) == 8.
17481 *Andy Polyakov <appro@fy.chalmers.se>*
17483 * Change functions to ANSI C.
17487 * Fix typos in error codes.
17489 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17491 * Remove defunct assembler files from Configure.
17495 * SPARC v8 assembler BIGNUM implementation.
17497 *Andy Polyakov <appro@fy.chalmers.se>*
17499 * Support for Certificate Policies extension: both print and set.
17500 Various additions to support the r2i method this uses.
17504 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17505 return a const string when you are expecting an allocated buffer.
17509 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17510 types DirectoryString and DisplayText.
17514 * Add code to allow r2i extensions to access the configuration database,
17515 add an LHASH database driver and add several ctx helper functions.
17519 * Fix an evil bug in bn_expand2() which caused various BN functions to
17520 fail when they extended the size of a BIGNUM.
17524 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17525 support typesafe stack.
17529 * Fix typo in SSL_[gs]et_options().
17531 *Nils Frostberg <nils@medcom.se>*
17533 * Delete various functions and files that belonged to the (now obsolete)
17534 old X509V3 handling code.
17538 * New Configure option "rsaref".
17542 * Don't auto-generate pem.h.
17546 * Introduce type-safe ASN.1 SETs.
17550 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17552 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17554 * Introduce type-safe STACKs. This will almost certainly break lots of code
17555 that links with OpenSSL (well at least cause lots of warnings), but fear
17556 not: the conversion is trivial, and it eliminates loads of evil casts. A
17557 few STACKed things have been converted already. Feel free to convert more.
17558 In the fullness of time, I'll do away with the STACK type altogether.
17562 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17563 specified in `<certfile>` by updating the entry in the index.txt file.
17564 This way one no longer has to edit the index.txt file manually for
17565 revoking a certificate. The -revoke option does the gory details now.
17567 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17569 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17570 `-text` option at all and this way the `-noout -text` combination was
17571 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17573 *Ralf S. Engelschall*
17575 * Make sure a corresponding plain text error message exists for the
17576 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17577 verify callback function determined that a certificate was revoked.
17579 *Ralf S. Engelschall*
17581 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17582 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17583 all available ciphers including rc5, which was forgotten until now.
17584 In order to let the testing shell script know which algorithms
17585 are available, a new (up to now undocumented) command
17586 `openssl list-cipher-commands` is used.
17590 * Bugfix: s_client occasionally would sleep in select() when
17591 it should have checked SSL_pending() first.
17595 * New functions DSA_do_sign and DSA_do_verify to provide access to
17596 the raw DSA values prior to ASN.1 encoding.
17600 * Tweaks to Configure
17602 *Niels Poppe <niels@netbox.org>*
17604 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17609 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17613 * New config option to avoid instructions that are illegal on the 80386.
17614 The default code is faster, but requires at least a 486.
17618 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17619 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17620 same as SSL2_VERSION anyway.
17624 * New "-showcerts" option for s_client.
17628 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17629 application. Various cleanups and fixes.
17633 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17634 modify error routines to work internally. Add error codes and PBE init
17635 to library startup routines.
17639 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17640 packing functions to asn1 and evp. Changed function names and error
17641 codes along the way.
17645 * PKCS12 integration: and so it begins... First of several patches to
17646 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17647 objects to objects.h
17651 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17652 and display support for Thawte strong extranet extension.
17656 * Add LinuxPPC support.
17658 *Jeff Dubrule <igor@pobox.org>*
17660 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17661 bn_div_words in alpha.s.
17663 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17665 * Make sure the RSA OAEP test is skipped under -DRSAref because
17666 OAEP isn't supported when OpenSSL is built with RSAref.
17668 *Ulf Moeller <ulf@fitug.de>*
17670 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17671 so they no longer are missing under -DNOPROTO.
17673 *Soren S. Jorvang <soren@t.dk>*
17675 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17677 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17678 doesn't work when the session is reused. Coming soon!
17682 * Fix a security hole, that allows sessions to be reused in the wrong
17683 context thus bypassing client cert protection! All software that uses
17684 client certs and session caches in multiple contexts NEEDS PATCHING to
17685 allow session reuse! A fuller solution is in the works.
17687 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17689 * Some more source tree cleanups (removed obsolete files
17690 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17691 permission on "config" script to be executable) and a fix for the INSTALL
17694 *Ulf Moeller <ulf@fitug.de>*
17696 * Remove some legacy and erroneous uses of malloc, free instead of
17699 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17701 * Make rsa_oaep_test return non-zero on error.
17703 *Ulf Moeller <ulf@fitug.de>*
17705 * Add support for native Solaris shared libraries. Configure
17706 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17707 if someone would make that last step automatic.
17709 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17711 * ctx_size was not built with the right compiler during "make links". Fixed.
17715 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17716 except NULL ciphers". This means the default cipher list will no longer
17717 enable NULL ciphers. They need to be specifically enabled e.g. with
17718 the string "DEFAULT:eNULL".
17722 * Fix to RSA private encryption routines: if p < q then it would
17723 occasionally produce an invalid result. This will only happen with
17724 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17728 * Be less restrictive and allow also `perl util/perlpath.pl
17729 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17730 because this way one can also use an interpreter named `perl5` (which is
17731 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17732 installed as `perl`).
17734 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17736 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17738 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17740 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17741 advapi32.lib to Win32 build and change the pem test comparison
17742 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17743 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17744 and crypto/des/ede_cbcm_enc.c.
17748 * DES quad checksum was broken on big-endian architectures. Fixed.
17752 * Comment out two functions in bio.h that aren't implemented. Fix up the
17753 Win32 test batch file so it (might) work again. The Win32 test batch file
17754 is horrible: I feel ill....
17758 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17759 in e_os.h. Audit of header files to check ANSI and non ANSI
17760 sections: 10 functions were absent from non ANSI section and not exported
17761 from Windows DLLs. Fixed up libeay.num for new functions.
17765 * Make `openssl version` output lines consistent.
17767 *Ralf S. Engelschall*
17769 * Fix Win32 symbol export lists for BIO functions: Added
17770 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17771 to ms/libeay{16,32}.def.
17773 *Ralf S. Engelschall*
17775 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17776 fine under Unix and passes some trivial tests I've now added. But the
17777 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17778 added to make sure no one expects that this stuff really works in the
17779 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17780 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17783 *Ralf S. Engelschall*
17785 * Fix the generation of two part addresses in perl.
17787 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17789 * Add config entry for Linux on MIPS.
17791 *John Tobey <jtobey@channel1.com>*
17793 * Make links whenever Configure is run, unless we are on Windoze.
17797 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17798 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17803 * Add a useful kludge to allow package maintainers to specify compiler and
17804 other platforms details on the command line without having to patch the
17805 Configure script every time: One now can use
17806 `perl Configure <id>:<details>`,
17807 i.e. platform ids are allowed to have details appended
17808 to them (separated by colons). This is treated as there would be a static
17809 pre-configured entry in Configure's %table under key `<id>` with value
17810 `<details>` and `perl Configure <id>` is called. So, when you want to
17811 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17812 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17813 now, which overrides the FreeBSD-elf entry on-the-fly.
17815 *Ralf S. Engelschall*
17817 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17821 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17822 on the `perl Configure ...` command line. This way one can compile
17823 OpenSSL libraries with Position Independent Code (PIC) which is needed
17824 for linking it into DSOs.
17826 *Ralf S. Engelschall*
17828 * Remarkably, export ciphers were totally broken and no-one had noticed!
17833 * Cleaned up the LICENSE document: The official contact for any license
17834 questions now is the OpenSSL core team under openssl-core@openssl.org.
17835 And add a paragraph about the dual-license situation to make sure people
17836 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17837 to the OpenSSL toolkit.
17839 *Ralf S. Engelschall*
17841 * General source tree makefile cleanups: Made `making xxx in yyy...`
17842 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17843 Additionally cleaned up the `make links` target: Remove unnecessary
17844 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17845 to speed processing and no longer clutter the display with confusing
17846 stuff. Instead only the actually done links are displayed.
17848 *Ralf S. Engelschall*
17850 * Permit null encryption ciphersuites, used for authentication only. It used
17851 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17852 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17857 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17858 signed attributes when verifying signatures (this would break them),
17859 the detached data encoding was wrong and public keys obtained using
17860 X509_get_pubkey() weren't freed.
17864 * Add text documentation for the BUFFER functions. Also added a work around
17865 to a Win95 console bug. This was triggered by the password read stuff: the
17866 last character typed gets carried over to the next fread(). If you were
17867 generating a new cert request using 'req' for example then the last
17868 character of the passphrase would be CR which would then enter the first
17873 * Added the new 'Includes OpenSSL Cryptography Software' button as
17874 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17875 button and can be used by applications based on OpenSSL to show the
17876 relationship to the OpenSSL project.
17878 *Ralf S. Engelschall*
17880 * Remove confusing variables in function signatures in files
17881 ssl/ssl_lib.c and ssl/ssl.h.
17883 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17885 * Don't install bss_file.c under PREFIX/include/
17887 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17889 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17890 functions that return function pointers and has support for NT specific
17891 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17892 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17893 unsigned to signed types: this was killing the Win32 compile.
17897 * Add new certificate file to stack functions,
17898 SSL_add_dir_cert_subjects_to_stack() and
17899 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17900 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17901 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17902 This means that Apache-SSL and similar packages don't have to mess around
17903 to add as many CAs as they want to the preferred list.
17907 * Experiment with doxygen documentation. Currently only partially applied to
17909 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17910 openssl.doxy as the configuration file.
17914 * Get rid of remaining C++-style comments which strict C compilers hate.
17916 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17918 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17919 compiled in by default: it has problems with large keys.
17923 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17924 DH private keys and/or callback functions which directly correspond to
17925 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17926 is needed for applications which have to configure certificates on a
17927 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17929 For the RSA certificate situation is makes no difference, but
17930 for the DSA certificate situation this fixes the "no shared cipher"
17931 problem where the OpenSSL cipher selection procedure failed because the
17932 temporary keys were not overtaken from the context and the API provided
17933 no way to reconfigure them.
17934 The new functions now let applications reconfigure the stuff and they
17935 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17936 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17937 non-public-API function ssl_cert_instantiate() is used as a helper
17938 function and also to reduce code redundancy inside ssl_rsa.c.
17940 *Ralf S. Engelschall*
17942 * Move s_server -dcert and -dkey options out of the undocumented feature
17943 area because they are useful for the DSA situation and should be
17944 recognized by the users.
17946 *Ralf S. Engelschall*
17948 * Fix the cipher decision scheme for export ciphers: the export bits are
17949 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17950 SSL_EXP_MASK. So, the original variable has to be used instead of the
17951 already masked variable.
17953 *Richard Levitte <levitte@stacken.kth.se>*
17955 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17957 *Richard Levitte <levitte@stacken.kth.se>*
17959 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17960 from `int` to `unsigned int` because it is a length and initialized by
17961 EVP_DigestFinal() which expects an `unsigned int *`.
17963 *Richard Levitte <levitte@stacken.kth.se>*
17965 * Don't hard-code path to Perl interpreter on shebang line of Configure
17966 script. Instead use the usual Shell->Perl transition trick.
17968 *Ralf S. Engelschall*
17970 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17971 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17972 -noout -modulus` as it's already the case for `openssl rsa -noout
17973 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17974 currently the public key is printed (a decision which was already done by
17975 `openssl dsa -modulus` in the past) which serves a similar purpose.
17976 Additionally the NO_RSA no longer completely removes the whole -modulus
17977 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17980 *Ralf S. Engelschall*
17982 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17983 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17985 *Arne Ansper <arne@ats.cyber.ee>*
17987 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17988 to be added. Now both 'req' and 'ca' can use new objects defined in the
17993 * Add cool BIO that does syslog (or event log on NT).
17995 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17997 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17998 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17999 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18000 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18004 * Add preliminary config info for new extension code.
18008 * Make RSA_NO_PADDING really use no padding.
18010 *Ulf Moeller <ulf@fitug.de>*
18012 * Generate errors when private/public key check is done.
18016 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18017 for some CRL extensions and new objects added.
18021 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18022 key usage extension and fuller support for authority key id.
18026 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18027 padding method for RSA, which is recommended for new applications in PKCS
18028 #1 v2.0 (RFC 2437, October 1998).
18029 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18030 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18031 against Bleichbacher's attack on RSA.
18032 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18035 * Updates to the new SSL compression code
18037 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18039 * Fix so that the version number in the master secret, when passed
18040 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18041 (because the server will not accept higher), that the version number
18042 is 0x03,0x01, not 0x03,0x00
18044 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18046 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
18047 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
18048 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
18052 * Support for RAW extensions where an arbitrary extension can be
18053 created by including its DER encoding. See apps/openssl.cnf for
18058 * Make sure latest Perl versions don't interpret some generated C array
18059 code as Perl array code in the crypto/err/err_genc.pl script.
18061 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18063 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18064 not many people have the assembler. Various Win32 compilation fixes and
18065 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18066 build instructions.
18070 * Modify configure script 'Configure' to automatically create crypto/date.h
18071 file under Win32 and also build pem.h from pem.org. New script
18072 util/mkfiles.pl to create the MINFO file on environments that can't do a
18073 'make files': perl util/mkfiles.pl >MINFO should work.
18077 * Major rework of DES function declarations, in the pursuit of correctness
18078 and purity. As a result, many evil casts evaporated, and some weirdness,
18079 too. You may find this causes warnings in your code. Zapping your evil
18080 casts will probably fix them. Mostly.
18084 * Fix for a typo in asn1.h. Bug fix to object creation script
18085 obj_dat.pl. It considered a zero in an object definition to mean
18086 "end of object": none of the objects in objects.h have any zeros
18087 so it wasn't spotted.
18089 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18091 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18092 Masking (CBCM). In the absence of test vectors, the best I have been able
18093 to do is check that the decrypt undoes the encrypt, so far. Send me test
18094 vectors if you have them.
18098 * Correct calculation of key length for export ciphers (too much space was
18099 allocated for null ciphers). This has not been tested!
18103 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18104 message is now correct (it understands "crypto" and "ssl" on its
18105 command line). There is also now an "update" option. This will update
18106 the util/ssleay.num and util/libeay.num files with any new functions.
18108 perl util/mkdef.pl crypto ssl update
18109 it will update them.
18113 * Overhauled the Perl interface:
18114 - ported BN stuff to OpenSSL's different BN library
18115 - made the perl/ source tree CVS-aware
18116 - renamed the package from SSLeay to OpenSSL (the files still contain
18117 their history because I've copied them in the repository)
18118 - removed obsolete files (the test scripts will be replaced
18119 by better Test::Harness variants in the future)
18121 *Ralf S. Engelschall*
18123 * First cut for a very conservative source tree cleanup:
18124 1. merge various obsolete readme texts into doc/ssleay.txt
18125 where we collect the old documents and readme texts.
18126 2. remove the first part of files where I'm already sure that we no
18127 longer need them because of three reasons: either they are just temporary
18128 files which were left by Eric or they are preserved original files where
18129 I've verified that the diff is also available in the CVS via "cvs diff
18130 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18131 the crypto/md/ stuff).
18133 *Ralf S. Engelschall*
18135 * More extension code. Incomplete support for subject and issuer alt
18136 name, issuer and authority key id. Change the i2v function parameters
18137 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18138 what that's for :-) Fix to ASN1 macro which messed up
18139 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18143 * Preliminary support for ENUMERATED type. This is largely copied from the
18148 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18150 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18152 * Make sure `make rehash` target really finds the `openssl` program.
18154 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18156 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18157 like to hear about it if this slows down other processors.
18161 * Add CygWin32 platform information to Configure script.
18163 *Alan Batie <batie@aahz.jf.intel.com>*
18165 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18167 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18169 * New program nseq to manipulate netscape certificate sequences
18173 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18178 * Fixes to BN code. Previously the default was to define BN_RECURSION
18179 but the BN code had some problems that would cause failures when
18180 doing certificate verification and some other functions.
18182 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18184 * Add ASN1 and PEM code to support netscape certificate sequences.
18188 * Add ASN1 and PEM code to support netscape certificate sequences.
18192 * Add several PKIX and private extended key usage OIDs.
18196 * Modify the 'ca' program to handle the new extension code. Modify
18197 openssl.cnf for new extension format, add comments.
18201 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18202 and add a sample to openssl.cnf so req -x509 now adds appropriate
18207 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18208 error code, add initial support to X509_print() and x509 application.
18212 * Takes a deep breath and start adding X509 V3 extension support code. Add
18213 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18214 stuff is currently isolated and isn't even compiled yet.
18218 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18219 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18220 Removed the versions check from X509 routines when loading extensions:
18221 this allows certain broken certificates that don't set the version
18222 properly to be processed.
18226 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18227 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18228 can still be regenerated with "make depend".
18232 * Spelling mistake in C version of CAST-128.
18234 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18236 * Changes to the error generation code. The perl script err-code.pl
18237 now reads in the old error codes and retains the old numbers, only
18238 adding new ones if necessary. It also only changes the .err files if new
18239 codes are added. The makefiles have been modified to only insert errors
18240 when needed (to avoid needlessly modifying header files). This is done
18241 by only inserting errors if the .err file is newer than the auto generated
18242 C file. To rebuild all the error codes from scratch (the old behaviour)
18243 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18244 or delete all the .err files.
18248 * CAST-128 was incorrectly implemented for short keys. The C version has
18249 been fixed, but is untested. The assembler versions are also fixed, but
18250 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18251 to regenerate it if needed.
18252 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18253 Hagino <itojun@kame.net>*
18255 * File was opened incorrectly in randfile.c.
18257 *Ulf Möller <ulf@fitug.de>*
18259 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18260 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18261 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18262 al: it's just almost always a UTCTime. Note this patch adds new error
18263 codes so do a "make errors" if there are problems.
18267 * Correct Linux 1 recognition in config.
18269 *Ulf Möller <ulf@fitug.de>*
18271 * Remove pointless MD5 hash when using DSA keys in ca.
18273 *Anonymous <nobody@replay.com>*
18275 * Generate an error if given an empty string as a cert directory. Also
18276 generate an error if handed NULL (previously returned 0 to indicate an
18277 error, but didn't set one).
18279 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18281 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18285 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18286 parameters. This was causing a warning which killed off the Win32 compile.
18290 * Remove C++ style comments from crypto/bn/bn_local.h.
18292 *Neil Costigan <neil.costigan@celocom.com>*
18294 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18295 based on a text string, looking up short and long names and finally
18296 "dot" format. The "dot" format stuff didn't work. Added new function
18297 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18298 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18299 OID is not part of the table.
18303 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18304 X509_LOOKUP_by_alias().
18308 * Sort openssl functions by name.
18312 * Get the gendsa program working (hopefully) and add it to app list. Remove
18313 encryption from sample DSA keys (in case anyone is interested the password
18318 * Make *all* `*_free` functions accept a NULL pointer.
18320 *Frans Heymans <fheymans@isaserver.be>*
18322 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18325 *Anonymous <nobody@replay.com>*
18327 * s_server should send the CAfile as acceptable CAs, not its own cert.
18329 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18331 * Don't blow it for numeric -newkey arguments to apps/req.
18333 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18335 * Temp key "for export" tests were wrong in s3_srvr.c.
18337 *Anonymous <nobody@replay.com>*
18339 * Add prototype for temp key callback functions
18340 SSL_CTX_set_tmp_{rsa,dh}_callback().
18344 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18345 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18349 * X509_name_add_entry() freed the wrong thing after an error.
18351 *Arne Ansper <arne@ats.cyber.ee>*
18353 * rsa_eay.c would attempt to free a NULL context.
18355 *Arne Ansper <arne@ats.cyber.ee>*
18357 * BIO_s_socket() had a broken should_retry() on Windoze.
18359 *Arne Ansper <arne@ats.cyber.ee>*
18361 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18363 *Arne Ansper <arne@ats.cyber.ee>*
18365 * Make sure the already existing X509_STORE->depth variable is initialized
18366 in X509_STORE_new(), but document the fact that this variable is still
18367 unused in the certificate verification process.
18369 *Ralf S. Engelschall*
18371 * Fix the various library and apps files to free up pkeys obtained from
18372 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18376 * Fix reference counting in X509_PUBKEY_get(). This makes
18377 demos/maurice/example2.c work, amongst others, probably.
18379 *Steve Henson and Ben Laurie*
18381 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18382 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18383 are no longer created. This way we have a single and consistent command
18384 line interface `openssl <command>`, similar to `cvs <command>`.
18386 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18388 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18389 BIT STRING wrapper always have zero unused bits.
18393 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18397 * Make the top-level INSTALL documentation easier to understand.
18401 * Makefiles updated to exit if an error occurs in a sub-directory
18402 make (including if user presses ^C) [Paul Sutton]
18404 * Make Montgomery context stuff explicit in RSA data structure.
18408 * Fix build order of pem and err to allow for generated pem.h.
18412 * Fix renumbering bug in X509_NAME_delete_entry().
18416 * Enhanced the err-ins.pl script so it makes the error library number
18417 global and can add a library name. This is needed for external ASN1 and
18418 other error libraries.
18422 * Fixed sk_insert which never worked properly.
18426 * Fix ASN1 macros so they can handle indefinite length constructed
18427 EXPLICIT tags. Some non standard certificates use these: they can now
18432 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18433 into a single doc/ssleay.txt bundle. This way the information is still
18434 preserved but no longer messes up this directory. Now it's new room for
18435 the new set of documentation files.
18437 *Ralf S. Engelschall*
18439 * SETs were incorrectly DER encoded. This was a major pain, because they
18440 shared code with SEQUENCEs, which aren't coded the same. This means that
18441 almost everything to do with SETs or SEQUENCEs has either changed name or
18442 number of arguments.
18444 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18446 * Fix test data to work with the above.
18450 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18451 was already fixed by Eric for 0.9.1 it seems.
18453 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18455 * Autodetect FreeBSD3.
18459 * Fix various bugs in Configure. This affects the following platforms:
18463 unixware-2.0-pentium
18468 * Eliminate generated files from CVS. Reorder tests to regenerate files
18469 before they are needed.
18473 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18477 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18479 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18480 changed SSLeay to OpenSSL in version strings.
18482 *Ralf S. Engelschall*
18484 * Some fixups to the top-level documents.
18488 * Fixed the nasty bug where rsaref.h was not found under compile-time
18489 because the symlink to include/ was missing.
18491 *Ralf S. Engelschall*
18493 * Incorporated the popular no-RSA/DSA-only patches
18494 which allow to compile a RSA-free SSLeay.
18496 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18498 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18499 when "ssleay" is still not found.
18501 *Ralf S. Engelschall*
18503 * Added more platforms to Configure: Cray T3E, HPUX 11,
18505 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18507 * Updated the README file.
18509 *Ralf S. Engelschall*
18511 * Added various .cvsignore files in the CVS repository subdirs
18512 to make a "cvs update" really silent.
18514 *Ralf S. Engelschall*
18516 * Recompiled the error-definition header files and added
18517 missing symbols to the Win32 linker tables.
18519 *Ralf S. Engelschall*
18521 * Cleaned up the top-level documents;
18522 o new files: CHANGES and LICENSE
18523 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18524 o merged COPYRIGHT into LICENSE
18525 o removed obsolete TODO file
18526 o renamed MICROSOFT to INSTALL.W32
18528 *Ralf S. Engelschall*
18530 * Removed dummy files from the 0.9.1b source tree:
18531 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18532 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18533 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18534 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18535 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18537 *Ralf S. Engelschall*
18539 * Added various platform portability fixes.
18543 * The Genesis of the OpenSSL rpject:
18544 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18545 Young and Tim J. Hudson created while they were working for C2Net until
18548 *The OpenSSL Project*
18550 ### Changes between 0.9.0b and 0.9.1b [not released]
18552 * Updated a few CA certificates under certs/
18556 * Changed some BIGNUM api stuff.
18560 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18561 DGUX x86, Linux Alpha, etc.
18565 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18566 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18571 * Add -strparse option to asn1pars program which parses nested
18574 *Dr Stephen Henson <shenson@bigfoot.com>*
18576 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18580 * DSA fix for "ca" program.
18584 * Added "-genkey" option to "dsaparam" program.
18588 * Added RIPE MD160 (rmd160) message digest.
18592 * Added -a (all) option to "ssleay version" command.
18596 * Added PLATFORM define which is the id given to Configure.
18600 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18604 * Extended the ASN.1 parser routines.
18608 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18612 * Added a BN_CTX to the BN library.
18616 * Fixed the weak key values in DES library
18620 * Changed API in EVP library for cipher aliases.
18624 * Added support for RC2/64bit cipher.
18628 * Converted the lhash library to the crypto/mem.c functions.
18632 * Added more recognized ASN.1 object ids.
18636 * Added more RSA padding checks for SSL/TLS.
18640 * Added BIO proxy/filter functionality.
18644 * Added extra_certs to SSL_CTX which can be used
18645 send extra CA certificates to the client in the CA cert chain sending
18646 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18650 * Now Fortezza is denied in the authentication phase because
18651 this is key exchange mechanism is not supported by SSLeay at all.
18655 * Additional PKCS1 checks.
18659 * Support the string "TLSv1" for all TLS v1 ciphers.
18663 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18664 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18668 * Fixed a few memory leaks.
18672 * Fixed various code and comment typos.
18676 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18677 bytes sent in the client random.
18679 *Edward Bishop <ebishop@spyglass.com>*
18683 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18684 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18685 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18686 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18687 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18688 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18689 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18690 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18691 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18692 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18693 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18694 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18695 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18696 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18697 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18698 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18699 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18700 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18701 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18702 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18703 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18704 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18705 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18706 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18707 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18708 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18709 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18710 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18711 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18712 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18713 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18714 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18715 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18716 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18717 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18718 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18719 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18720 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18721 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18722 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18723 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18724 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18725 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18726 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18727 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18728 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18729 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18730 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18731 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18732 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18733 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18734 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18735 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18736 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18737 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18738 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18739 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18740 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18741 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18742 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18743 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18744 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18745 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18746 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18747 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18748 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18749 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18750 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18751 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18752 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18753 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18754 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18755 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18756 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18757 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18758 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18759 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18760 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18761 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18762 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18763 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18764 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18765 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18766 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18767 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18768 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18769 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18770 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18771 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18772 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18773 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18774 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18775 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18776 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18777 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18778 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18779 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18780 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18781 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18782 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18783 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18784 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18785 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18786 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18787 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18788 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18789 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18790 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18791 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18792 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18793 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18794 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18795 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18796 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18797 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18798 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18799 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18800 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18801 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18802 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18803 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18804 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18805 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18806 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18807 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18808 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18809 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18810 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18811 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18812 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18813 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18814 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18815 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18816 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18817 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18818 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18819 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18820 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18821 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18822 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18823 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18824 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18825 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18826 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18827 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18828 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18829 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18830 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18831 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18832 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18833 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18834 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18835 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18836 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18837 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18838 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18839 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18840 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18841 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18842 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18843 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18844 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655