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 * Deprecated all the libcrypto and libssl error string loading
27 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
28 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
29 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
30 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
31 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
32 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
33 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
34 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
35 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
36 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
38 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
39 now loads error strings automatically.
43 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
44 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
45 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
46 are to be used by servers requiring ephemeral DH keys. Instead applications
47 should consider using the built-in DH parameters that are available by
48 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
49 necessary then applications can use the alternative functions
50 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
51 replacement for the "callback" functions. The callback was originally useful
52 in order to have different parameters for export and non-export ciphersuites.
53 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
54 functions should be replaced by one of the other methods described above.
58 * The -crypt option to the passwd command line tool has been removed.
62 * The -C option to the x509, dhparam, dsaparam, and ecparam commands
67 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
69 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
70 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
71 The inverse ciphers use AES decryption for wrapping, and
72 AES encryption for unwrapping.
76 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
77 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
78 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
79 OpenSSL 3.0 these are replaced by the more generic functions
80 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
81 The old versions have been converted to deprecated macros that just call the
86 * The security callback, which can be customised by application code, supports
87 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
88 in the "other" parameter. In most places this is what is passed. All these
89 places occur server side. However there was one client side call of this
90 security operation and it passed a DH object instead. This is incorrect
91 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
92 of the other locations. Therefore this client side call has been changed to
93 pass an EVP_PKEY instead.
97 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
98 interface. Their functionality remains unchanged.
102 * Added new option for 'openssl list', '-providers', which will display the
103 list of loaded providers, their names, version and status. It optionally
104 displays their gettable parameters.
108 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
109 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
110 type is internally recognised so the workaround is no longer needed.
112 Functionality is still retained as it is, but will only work with
113 EVP_PKEYs with a legacy internal key.
117 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
118 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
122 * Changed all "STACK" functions to be macros instead of inline functions. Macro
123 parameters are still checked for type safety at compile time via helper
128 * Remove the RAND_DRBG API
130 The RAND_DRBG API did not fit well into the new provider concept as
131 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
132 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
133 and some of its API calls are rather low-level. This holds in particular
134 for the callback mechanism (RAND_DRBG_set_callbacks()).
136 Adding a compatibility layer to continue supporting the RAND_DRBG API as
137 a legacy API for a regular deprecation period turned out to come at the
138 price of complicating the new provider API unnecessarily. Since the
139 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
142 *Paul Dale and Matthias St. Pierre*
144 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
145 as well as actual hostnames.
149 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
150 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
151 conversely, silently ignore DTLS protocol version bounds when configuring
152 TLS-based contexts. The commands can be repeated to set bounds of both
153 types. The same applies with the corresponding "min_protocol" and
154 "max_protocol" command-line switches, in case some application uses both TLS
157 SSL_CTX instances that are created for a fixed protocol version (e.g.
158 TLSv1_server_method()) also silently ignore version bounds. Previously
159 attempts to apply bounds to these protocol versions would result in an
160 error. Now only the "version-flexible" SSL_CTX instances are subject to
161 limits in configuration files in command-line options.
165 * Deprecated the `ENGINE` API. Engines should be replaced with providers
170 * Reworked the recorded ERR codes to make better space for system errors.
171 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
172 given code is a system error (true) or an OpenSSL error (false).
176 * Reworked the test perl framework to better allow parallel testing.
178 *Nicola Tuveri and David von Oheimb*
180 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
181 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
185 * 'Configure' has been changed to figure out the configuration target if
186 none is given on the command line. Consequently, the 'config' script is
187 now only a mere wrapper. All documentation is changed to only mention
190 *Rich Salz and Richard Levitte*
192 * Added a library context `OSSL_LIB_CTX` that applications as well as
193 other libraries can use to form a separate context within which
194 libcrypto operations are performed.
196 There are two ways this can be used:
198 - Directly, by passing a library context to functions that take
199 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
201 - Indirectly, by creating a new library context and then assigning
202 it as the new default, with `OSSL_LIB_CTX_set0_default`.
204 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
205 apart from the functions directly related to `OSSL_LIB_CTX`, accept
206 NULL to indicate that the default library context should be used.
208 Library code that changes the default library context using
209 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
210 second call before returning to the caller.
212 _(Note: the library context was initially called `OPENSSL_CTX` and
213 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
217 * Handshake now fails if Extended Master Secret extension is dropped
222 * Dropped interactive mode from the 'openssl' program. From now on,
223 the `openssl` command without arguments is equivalent to `openssl
228 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
229 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
230 While the old function names have been retained for backward compatibility
231 they should not be used in new developments
232 because their return values are confusing: Unlike other `_cmp()` functions
233 they do not return 0 in case their arguments are equal.
237 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
238 EC_GROUP_get_field_type().
242 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
243 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
244 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
245 Applications should rely on the library automatically assigning a suitable
246 EC_METHOD internally upon EC_GROUP construction.
250 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
251 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
252 assigned internally without application intervention.
253 Users of EC_GROUP_new() should switch to a different suitable constructor.
257 * Add CAdES-BES signature verification support, mostly derived
258 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
260 *Filipe Raimundo da Silva*
262 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
266 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
267 functions are not widely used and now OpenSSL automatically perform this
268 conversion when needed.
272 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
273 EC_KEY_precompute_mult(). These functions are not widely used and
274 applications should instead switch to named curves which OpenSSL has
275 hardcoded lookup tables for.
279 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
280 should instead use the L<EC_POINT_mul(3)> function.
284 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
285 that are not applicable to the new provider model. Applications should
286 instead use EVP_default_properties_is_fips_enabled() and
287 EVP_default_properties_enable_fips().
291 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
292 is set, an unexpected EOF is ignored, it pretends a close notify was received
293 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
297 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
298 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
299 used and applications should instead use the
300 L<EC_POINT_set_affine_coordinates(3)> and
301 L<EC_POINT_get_affine_coordinates(3)> functions.
305 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
306 arrays to be more easily constructed via a series of utility functions.
307 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
308 the various push functions and finally convert to a passable OSSL_PARAM
309 array using OSSL_PARAM_BLD_to_param().
313 * The security strength of SHA1 and MD5 based signatures in TLS has been
314 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
315 working at the default security level of 1 and instead requires security
316 level 0. The security level can be changed either using the cipher string
317 with @SECLEVEL, or calling SSL_CTX_set_security_level().
321 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
322 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
323 internal keys, if they correspond to one of those built in types.
327 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
328 contain a provider side internal key.
332 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
333 They are old functions that we don't use, and that you could disable with
334 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
338 * Project text documents not yet having a proper file name extension
339 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
340 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
344 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
345 have been converted to Markdown with the goal to produce documents
346 which not only look pretty when viewed online in the browser, but
347 remain well readable inside a plain text editor.
349 To achieve this goal, a 'minimalistic' Markdown style has been applied
350 which avoids formatting elements that interfere too much with the
351 reading flow in the text file. For example, it
353 * avoids [ATX headings][] and uses [setext headings][] instead
354 (which works for `<h1>` and `<h2>` headings only).
355 * avoids [inline links][] and uses [reference links][] instead.
356 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
358 [ATX headings]: https://github.github.com/gfm/#atx-headings
359 [setext headings]: https://github.github.com/gfm/#setext-headings
360 [inline links]: https://github.github.com/gfm/#inline-link
361 [reference links]: https://github.github.com/gfm/#reference-link
362 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
363 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
365 *Matthias St. Pierre*
367 * The test suite is changed to preserve results of each test recipe.
368 A new directory test-runs/ with subdirectories named like the
369 test recipes are created in the build tree for this purpose.
373 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
374 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
375 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
377 *David von Oheimb, Martin Peylo*
379 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
380 The legacy OCSP-focused and only partly documented API is retained.
381 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
385 * Added `util/check-format.pl`, a tool for checking adherence to the
386 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
387 The checks performed are incomplete and yield some false positives.
388 Still the tool should be useful for detecting most typical glitches.
392 * BIO_do_connect and BIO_do_handshake have been extended:
393 If domain name resolution yields multiple IP addresses all of them are tried
394 after connect() failures.
398 * All of the low level RSA functions have been deprecated including:
400 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
401 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
402 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
403 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
404 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
405 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
406 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
407 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
408 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
409 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
410 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
411 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
412 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
413 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
414 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
415 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
416 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
417 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
418 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
419 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
420 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
421 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
422 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
423 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
424 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
425 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
426 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
427 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
429 Use of these low level functions has been informally discouraged for a long
430 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
431 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
432 L<EVP_PKEY_decrypt(3)>.
436 * X509 certificates signed using SHA1 are no longer allowed at security
438 In TLS/SSL the default security level is 1. It can be set either
439 using the cipher string with @SECLEVEL, or calling
440 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
441 a call to SSL_CTX_use_certificate() will fail if the security level is not
443 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
444 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
449 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
450 modified to use PKEY APIs. These commands are now in maintenance mode
451 and no new features will be added to them.
455 * The command line utility rsautl has been deprecated.
456 Instead use the pkeyutl program.
460 * The command line utilities genrsa and rsa have been modified to use PKEY
461 APIs. They now write PKCS#8 keys by default. These commands are now in
462 maintenance mode and no new features will be added to them.
466 * All of the low level DH functions have been deprecated including:
468 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
469 DH_new_method, DH_size, DH_security_bits, DH_get_ex_new_index,
470 DH_set_ex_data, DH_get_ex_data, DH_generate_parameters_ex,
471 DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
472 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
473 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
474 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
475 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
476 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
477 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
478 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
479 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
480 DH_meth_get_generate_params and DH_meth_set_generate_params.
482 Use of these low level functions has been informally discouraged for a long
483 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
484 and L<EVP_PKEY_derive(3)>.
488 * All of the low level DSA functions have been deprecated including:
490 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
491 DSA_get_default_method, DSA_set_method, DSA_get_method,
492 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
493 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
494 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
495 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
496 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
497 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
498 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
499 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
500 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
501 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
502 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
505 Use of these low level functions has been informally discouraged for a long
506 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
507 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
511 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
512 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
513 This means that applications don't have to look at the curve NID and
514 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
515 However, they still can, that EVP_PKEY_set_alias_type() call acts as
516 a no-op when the EVP_PKEY is already of the given type.
518 Parameter and key generation is also reworked to make it possible
519 to generate EVP_PKEY_SM2 parameters and keys without having to go
520 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
521 However, code that does the latter will still work as before.
525 * Deprecated low level ECDH and ECDSA functions. These include:
527 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
528 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
531 Use of these low level functions has been informally discouraged for a long
532 time. Instead applications should use the EVP_PKEY_derive(3),
533 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
537 * Deprecated the EC_KEY_METHOD functions. These include:
539 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
540 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
541 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
542 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
543 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
544 EC_KEY_METHOD_get_verify.
546 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
550 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
551 and EVP_PKEY_decrypt() instead.
552 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
553 and EVP_PKEY_encrypt() instead.
557 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
558 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
559 a new formulation to include all the things it can be used for,
560 as well as words of caution.
564 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
565 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
569 * All of the low level HMAC functions have been deprecated including:
571 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
572 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
575 Use of these low level functions has been informally discouraged for a long
576 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
577 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
578 and L<EVP_MAC_final(3)>.
582 * Over two thousand fixes were made to the documentation, including:
583 - Common options (such as -rand/-writerand, TLS version control, etc)
584 were refactored and point to newly-enhanced descriptions in openssl.pod.
585 - Added style conformance for all options (with help from Richard Levitte),
586 documented all reported missing options, added a CI build to check
587 that all options are documented and that no unimplemented options
589 - Documented some internals, such as all use of environment variables.
590 - Addressed all internal broken L<> references.
594 * All of the low level CMAC functions have been deprecated including:
596 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
597 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
599 Use of these low level functions has been informally discouraged for a long
600 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
601 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
602 and L<EVP_MAC_final(3)>.
606 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
607 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
610 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
611 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
612 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
613 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
614 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
615 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
616 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
617 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
618 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
619 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
621 Use of these low level functions has been informally discouraged
622 for a long time. Applications should use the EVP_DigestInit_ex(3),
623 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
627 * Corrected the documentation of the return values from the `EVP_DigestSign*`
628 set of functions. The documentation mentioned negative values for some
629 errors, but this was never the case, so the mention of negative values
632 Code that followed the documentation and thereby check with something
633 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
637 * All of the low level cipher functions have been deprecated including:
639 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
640 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
641 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
642 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
643 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
644 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
645 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
646 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
647 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
648 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
649 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
650 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
651 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
652 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
653 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
654 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
655 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
656 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
657 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
658 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
659 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
660 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
661 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
662 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
663 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
664 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
665 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
666 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
667 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
669 Use of these low level functions has been informally discouraged for
670 a long time. Applications should use the high level EVP APIs, e.g.
671 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
672 equivalently named decrypt functions instead.
674 *Matt Caswell and Paul Dale*
676 * Removed include/openssl/opensslconf.h.in and replaced it with
677 include/openssl/configuration.h.in, which differs in not including
678 <openssl/macros.h>. A short header include/openssl/opensslconf.h
679 was added to include both.
681 This allows internal hacks where one might need to modify the set
682 of configured macros, for example this if deprecated symbols are
683 still supposed to be available internally:
685 #include <openssl/configuration.h>
687 #undef OPENSSL_NO_DEPRECATED
688 #define OPENSSL_SUPPRESS_DEPRECATED
690 #include <openssl/macros.h>
692 This should not be used by applications that use the exported
693 symbols, as that will lead to linking errors.
697 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
698 used in exponentiation with 512-bit moduli. No EC algorithms are
699 affected. Analysis suggests that attacks against 2-prime RSA1024,
700 3-prime RSA1536, and DSA1024 as a result of this defect would be very
701 difficult to perform and are not believed likely. Attacks against DH512
702 are considered just feasible. However, for an attack the target would
703 have to re-use the DH512 private key, which is not recommended anyway.
704 Also applications directly using the low level API BN_mod_exp may be
705 affected if they use BN_FLG_CONSTTIME.
710 * Most memory-debug features have been deprecated, and the functionality
711 replaced with no-ops.
715 * Added documentation for the STACK API.
719 * Introduced a new method type and API, OSSL_ENCODER, to
720 represent generic encoders. An implementation is expected to
721 be able to encode an object associated with a given name (such
722 as an algorithm name for an asymmetric key) into forms given by
723 implementation properties.
725 Encoders are primarily used from inside libcrypto, through
726 calls to functions like EVP_PKEY_print_private(),
727 PEM_write_bio_PrivateKey() and similar.
729 Encoders are specified in such a way that they can be made to
730 directly handle the provider side portion of an object, if this
731 provider side part comes from the same provider as the encoder
732 itself, but can also be made to handle objects in parametrized
733 form (as an OSSL_PARAM array of data). This allows a provider to
734 offer generic encoders as a service for any other provider.
738 * Added a .pragma directive to the syntax of configuration files, to
739 allow varying behavior in a supported and predictable manner.
740 Currently added pragma:
744 This allows dollar signs to be a keyword character unless it's
745 followed by a opening brace or parenthesis. This is useful for
746 platforms where dollar signs are commonly used in names, such as
747 volume names and system directory names on VMS.
751 * Added functionality to create an EVP_PKEY from user data. This
752 is effectively the same as creating a RSA, DH or DSA object and
753 then assigning them to an EVP_PKEY, but directly using algorithm
754 agnostic EVP functions. A benefit is that this should be future
755 proof for public key algorithms to come.
759 * Change the interpretation of the '--api' configuration option to
760 mean that this is a desired API compatibility level with no
761 further meaning. The previous interpretation, that this would
762 also mean to remove all deprecated symbols up to and including
763 the given version, no requires that 'no-deprecated' is also used
764 in the configuration.
766 When building applications, the desired API compatibility level
767 can be set with the OPENSSL_API_COMPAT macro like before. For
768 API compatibility version below 3.0, the old style numerical
769 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
770 For version 3.0 and on, the value is expected to be the decimal
771 value calculated from the major and minor version like this:
773 MAJOR * 10000 + MINOR * 100
777 -DOPENSSL_API_COMPAT=30000 For 3.0
778 -DOPENSSL_API_COMPAT=30200 For 3.2
780 To hide declarations that are deprecated up to and including the
781 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
782 given when building the application as well.
786 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
787 access to certificate and CRL stores via URIs and OSSL_STORE
790 This adds the following functions:
792 - X509_LOOKUP_store()
793 - X509_STORE_load_file()
794 - X509_STORE_load_path()
795 - X509_STORE_load_store()
796 - SSL_add_store_cert_subjects_to_stack()
797 - SSL_CTX_set_default_verify_store()
798 - SSL_CTX_load_verify_file()
799 - SSL_CTX_load_verify_dir()
800 - SSL_CTX_load_verify_store()
804 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
805 The presence of this system service is determined at run-time.
809 * Added functionality to create an EVP_PKEY context based on data
810 for methods from providers. This takes an algorithm name and a
811 property query string and simply stores them, with the intent
812 that any operation that uses this context will use those strings
813 to fetch the needed methods implicitly, thereby making the port
814 of application written for pre-3.0 OpenSSL easier.
818 * The undocumented function NCONF_WIN32() has been deprecated; for
819 conversion details see the HISTORY section of doc/man5/config.pod
823 * Introduced the new functions EVP_DigestSignInit_ex() and
824 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
825 EVP_DigestVerifyUpdate() have been converted to functions. See the man
826 pages for further details.
830 * Over two thousand fixes were made to the documentation, including:
831 adding missing command flags, better style conformance, documentation
834 *Rich Salz, Richard Levitte*
836 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
837 X25519, X448, Ed25519 and Ed448.
841 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
846 * Deprecated the public definition of ERR_STATE as well as the function
847 ERR_get_state(). This is done in preparation of making ERR_STATE an
852 * Added ERR functionality to give callers access to the stored function
853 names that have replaced the older function code based functions.
855 New functions are ERR_get_error_func(), ERR_peek_error_func(),
856 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
857 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
858 and ERR_peek_last_error_all().
860 These functions have become deprecated: ERR_get_error_line_data(),
861 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
862 ERR_func_error_string().
866 * Extended testing to be verbose for failing tests only. The make variables
867 VERBOSE_FAILURE or VF can be used to enable this:
869 $ make VF=1 test # Unix
870 $ mms /macro=(VF=1) test ! OpenVMS
871 $ nmake VF=1 test # Windows
875 * Added several checks to X509_verify_cert() according to requirements in
876 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
877 (which may be done by using the CLI option `-x509_strict`):
878 * The basicConstraints of CA certificates must be marked critical.
879 * CA certificates must explicitly include the keyUsage extension.
880 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
881 * The issuer name of any certificate must not be empty.
882 * The subject name of CA certs, certs with keyUsage crlSign,
883 and certs without subjectAlternativeName must not be empty.
884 * If a subjectAlternativeName extension is given it must not be empty.
885 * The signatureAlgorithm field and the cert signature must be consistent.
886 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
887 must not be marked critical.
888 * The authorityKeyIdentifier must be given for X.509v3 certs
889 unless they are self-signed.
890 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
894 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
895 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
899 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
900 used even when parsing explicit parameters, when loading a encoded key
901 or calling `EC_GROUP_new_from_ecpkparameters()`/
902 `EC_GROUP_new_from_ecparameters()`.
903 This prevents bypass of security hardening and performance gains,
904 especially for curves with specialized EC_METHODs.
905 By default, if a key encoded with explicit parameters is loaded and later
906 encoded, the output is still encoded with explicit parameters, even if
907 internally a "named" EC_GROUP is used for computation.
911 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
912 this change, EC_GROUP_set_generator would accept order and/or cofactor as
913 NULL. After this change, only the cofactor parameter can be NULL. It also
914 does some minimal sanity checks on the passed order.
919 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
920 An attack is simple, if the first CMS_recipientInfo is valid but the
921 second CMS_recipientInfo is chosen ciphertext. If the second
922 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
923 encryption key will be replaced by garbage, and the message cannot be
924 decoded, but if the RSA decryption fails, the correct encryption key is
925 used and the recipient will not notice the attack.
926 As a work around for this potential attack the length of the decrypted
927 key must be equal to the cipher default key length, in case the
928 certifiate is not given and all recipientInfo are tried out.
929 The old behaviour can be re-enabled in the CMS code by setting the
930 CMS_DEBUG_DECRYPT flag.
934 * Early start up entropy quality from the DEVRANDOM seed source has been
935 improved for older Linux systems. The RAND subsystem will wait for
936 /dev/random to be producing output before seeding from /dev/urandom.
937 The seeded state is stored for future library initialisations using
938 a system global shared memory segment. The shared memory identifier
939 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
940 the desired value. The default identifier is 114.
944 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
945 when primes for RSA keys are computed.
946 Since we previously always generated primes == 2 (mod 3) for RSA keys,
947 the 2-prime and 3-prime RSA modules were easy to distinguish, since
948 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
949 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
950 This avoids possible fingerprinting of newly generated RSA modules.
954 * Correct the extended master secret constant on EBCDIC systems. Without this
955 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
956 negotiate EMS will fail. Unfortunately this also means that TLS connections
957 between EBCDIC systems with this fix, and EBCDIC systems without this
958 fix will fail if they negotiate EMS.
962 * Changed the library initialisation so that the config file is now loaded
963 by default. This was already the case for libssl. It now occurs for both
964 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
965 OPENSSL_init_crypto() to suppress automatic loading of a config file.
969 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
970 where the former acts as a replacement for ERR_put_error(), and the
971 latter replaces the combination ERR_put_error()+ERR_add_error_data().
972 ERR_raise_data() adds more flexibility by taking a format string and
973 an arbitrary number of arguments following it, to be processed with
978 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
979 to check if a named provider is loaded and available. When called, it
980 will also activate all fallback providers if such are still present.
984 * Enforce a minimum DH modulus size of 512 bits.
988 * Changed DH parameters to generate the order q subgroup instead of 2q.
989 Previously generated DH parameters are still accepted by DH_check
990 but DH_generate_key works around that by clearing bit 0 of the
991 private key for those. This avoids leaking bit 0 of the private key.
995 * Significantly reduce secure memory usage by the randomness pools.
999 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1004 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1005 algorithms. An implementation of a key exchange algorithm can be obtained
1006 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1007 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1008 the older EVP_PKEY_derive_init() function. See the man pages for the new
1009 functions for further details.
1013 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1017 * Removed the function names from error messages and deprecated the
1020 * Removed NextStep support and the macro OPENSSL_UNISTD
1024 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1025 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1026 Also removed "export var as function" capability; we do not export
1027 variables, only functions.
1031 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1032 an error and 1 indicating success. In previous versions of OpenSSL this
1033 was a void type. If a key was set longer than the maximum possible this
1038 * Support SM2 signing and verification schemes with X509 certificate.
1042 * Use SHA256 as the default digest for TS query in the ts app.
1046 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1047 This checks that the salt length is at least 128 bits, the derived key
1048 length is at least 112 bits, and that the iteration count is at least 1000.
1049 For backwards compatibility these checks are disabled by default in the
1050 default provider, but are enabled by default in the fips provider.
1051 To enable or disable these checks use the control
1052 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1056 * Default cipher lists/suites are now available via a function, the
1057 #defines are deprecated.
1061 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1062 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1063 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1067 * Join the directories crypto/x509 and crypto/x509v3
1071 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1072 This changes the size when using the genpkey app when no size is given. It
1073 fixes an omission in earlier changes that changed all RSA, DSA and DH
1074 generation apps to use 2048 bits by default.
1078 * Added command 'openssl kdf' that uses the EVP_KDF API.
1082 * Added command 'openssl mac' that uses the EVP_MAC API.
1086 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1087 as default directories. Also added the command 'openssl info'
1088 for scripting purposes.
1092 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1093 deprecated. These undocumented functions were never integrated into the EVP
1094 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1095 Bi-directional IGE mode. These modes were never formally standardised and
1096 usage of these functions is believed to be very small. In particular
1097 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1098 is ever used. The security implications are believed to be minimal, but
1099 this issue was never fixed for backwards compatibility reasons. New code
1100 should not use these modes.
1104 * Add prediction resistance to the DRBG reseeding process.
1108 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1109 mandated by IEEE Std 1619-2018.
1113 * Added newline escaping functionality to a filename when using openssl dgst.
1114 This output format is to replicate the output format found in the `*sum`
1115 checksum programs. This aims to preserve backward compatibility.
1117 *Matt Eaton, Richard Levitte, and Paul Dale*
1119 * Removed the heartbeat message in DTLS feature, as it has very
1120 little usage and doesn't seem to fulfill a valuable purpose.
1121 The configuration option is now deprecated.
1125 * Changed the output of 'openssl {digestname} < file' to display the
1126 digest name in its output.
1130 * Added a new generic trace API which provides support for enabling
1131 instrumentation through trace output. This feature is mainly intended
1132 as an aid for developers and is disabled by default. To utilize it,
1133 OpenSSL needs to be configured with the `enable-trace` option.
1135 If the tracing API is enabled, the application can activate trace output
1136 by registering BIOs as trace channels for a number of tracing and debugging
1139 The 'openssl' application has been expanded to enable any of the types
1140 available via environment variables defined by the user, and serves as
1141 one possible example on how to use this functionality.
1143 *Richard Levitte & Matthias St. Pierre*
1145 * Added build tests for C++. These are generated files that only do one
1146 thing, to include one public OpenSSL head file each. This tests that
1147 the public header files can be usefully included in a C++ application.
1149 This test isn't enabled by default. It can be enabled with the option
1150 'enable-buildtest-c++'.
1154 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1158 * Add KMAC to EVP_MAC.
1162 * Added property based algorithm implementation selection framework to
1167 * Added SCA hardening for modular field inversion in EC_GROUP through
1168 a new dedicated field_inv() pointer in EC_METHOD.
1169 This also addresses a leakage affecting conversions from projective
1170 to affine coordinates.
1172 *Billy Bob Brumley, Nicola Tuveri*
1174 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1175 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1176 those algorithms that were already supported through the EVP_PKEY API
1177 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1178 and scrypt are now wrappers that call EVP_KDF.
1182 * Build devcrypto engine as a dynamic engine.
1184 *Eneas U de Queiroz*
1186 * Add keyed BLAKE2 to EVP_MAC.
1190 * Fix a bug in the computation of the endpoint-pair shared secret used
1191 by DTLS over SCTP. This breaks interoperability with older versions
1192 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1193 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1194 interoperability with such broken implementations. However, enabling
1195 this switch breaks interoperability with correct implementations.
1197 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1198 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1202 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1206 * Change the license to the Apache License v2.0.
1210 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1212 - Major releases (indicated by incrementing the MAJOR release number)
1213 may introduce incompatible API/ABI changes.
1214 - Minor releases (indicated by incrementing the MINOR release number)
1215 may introduce new features but retain API/ABI compatibility.
1216 - Patch releases (indicated by incrementing the PATCH number)
1217 are intended for bug fixes and other improvements of existing
1218 features only (like improving performance or adding documentation)
1219 and retain API/ABI compatibility.
1223 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1227 * Remove the 'dist' target and add a tarball building script. The
1228 'dist' target has fallen out of use, and it shouldn't be
1229 necessary to configure just to create a source distribution.
1233 * Recreate the OS390-Unix config target. It no longer relies on a
1234 special script like it did for OpenSSL pre-1.1.0.
1238 * Instead of having the source directories listed in Configure, add
1239 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1244 * Add GMAC to EVP_MAC.
1248 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1252 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1253 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1254 to facilitate the continued use of MACs through raw private keys in
1255 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1259 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1260 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1264 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1265 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1266 are retained for backwards compatibility.
1270 * AES-XTS mode now enforces that its two keys are different to mitigate
1271 the attacked described in "Efficient Instantiations of Tweakable
1272 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1273 Details of this attack can be obtained from:
1274 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1278 * Rename the object files, i.e. give them other names than in previous
1279 versions. Their names now include the name of the final product, as
1280 well as its type mnemonic (bin, lib, shlib).
1284 * Added new option for 'openssl list', '-objects', which will display the
1285 list of built in objects, i.e. OIDs with names.
1289 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1290 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1295 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1296 improves application performance by removing data copies and providing
1297 applications with zero-copy system calls such as sendfile and splice.
1301 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1302 option is set, openssl cleanses (zeroize) plaintext bytes from
1303 internal buffers after delivering them to the application. Note,
1304 the application is still responsible for cleansing other copies
1305 (e.g.: data received by SSL_read(3)).
1309 * `PKCS12_parse` now maintains the order of the parsed certificates
1310 when outputting them via `*ca` (rather than reversing it).
1314 * Deprecated pthread fork support methods. These were unused so no
1315 replacement is required.
1317 - OPENSSL_fork_prepare()
1318 - OPENSSL_fork_parent()
1319 - OPENSSL_fork_child()
1326 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1330 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1332 * Certificates with explicit curve parameters are now disallowed in
1333 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1337 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1338 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1339 conversely, silently ignore DTLS protocol version bounds when configuring
1340 TLS-based contexts. The commands can be repeated to set bounds of both
1341 types. The same applies with the corresponding "min_protocol" and
1342 "max_protocol" command-line switches, in case some application uses both TLS
1345 SSL_CTX instances that are created for a fixed protocol version (e.g.
1346 TLSv1_server_method()) also silently ignore version bounds. Previously
1347 attempts to apply bounds to these protocol versions would result in an
1348 error. Now only the "version-flexible" SSL_CTX instances are subject to
1349 limits in configuration files in command-line options.
1353 * Handshake now fails if Extended Master Secret extension is dropped
1358 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1360 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1362 * Fixed segmentation fault in SSL_check_chain()
1363 Server or client applications that call the SSL_check_chain() function
1364 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1365 dereference as a result of incorrect handling of the
1366 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1367 or unrecognised signature algorithm is received from the peer. This could
1368 be exploited by a malicious peer in a Denial of Service attack.
1373 * Added AES consttime code for no-asm configurations
1374 an optional constant time support for AES was added
1375 when building openssl for no-asm.
1376 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1377 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1378 At this time this feature is by default disabled.
1379 It will be enabled by default in 3.0.
1383 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1385 * Revert the change of EOF detection while reading in libssl to avoid
1386 regressions in applications depending on the current way of reporting
1387 the EOF. As the existing method is not fully accurate the change to
1388 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1389 branch and will be present in the 3.0 release.
1393 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1394 when primes for RSA keys are computed.
1395 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1396 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1397 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1398 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1399 This avoids possible fingerprinting of newly generated RSA modules.
1403 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1405 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1406 while reading in libssl then we would report an error back to the
1407 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1408 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1409 therefore give a hint as to what went wrong.
1413 * Check that ed25519 and ed448 are allowed by the security level. Previously
1414 signature algorithms not using an MD were not being checked that they were
1415 allowed by the security level.
1419 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1420 was not quite right. The behaviour was not consistent between resumption
1421 and normal handshakes, and also not quite consistent with historical
1422 behaviour. The behaviour in various scenarios has been clarified and
1423 it has been updated to make it match historical behaviour as closely as
1428 * *[VMS only]* The header files that the VMS compilers include automatically,
1429 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1430 that the C++ compiler doesn't understand. This is a shortcoming in the
1431 compiler, but can be worked around with `__cplusplus` guards.
1433 C++ applications that use OpenSSL libraries must be compiled using the
1434 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1435 functions. Otherwise, only functions with symbols of less than 31
1436 characters can be used, as the linker will not be able to successfully
1437 resolve symbols with longer names.
1441 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1442 The presence of this system service is determined at run-time.
1446 * Added newline escaping functionality to a filename when using openssl dgst.
1447 This output format is to replicate the output format found in the `*sum`
1448 checksum programs. This aims to preserve backward compatibility.
1450 *Matt Eaton, Richard Levitte, and Paul Dale*
1452 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1457 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1459 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1460 number generator (RNG). This was intended to include protection in the
1461 event of a fork() system call in order to ensure that the parent and child
1462 processes did not share the same RNG state. However this protection was not
1463 being used in the default case.
1465 A partial mitigation for this issue is that the output from a high
1466 precision timer is mixed into the RNG state so the likelihood of a parent
1467 and child process sharing state is significantly reduced.
1469 If an application already calls OPENSSL_init_crypto() explicitly using
1470 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1473 *Matthias St. Pierre*
1475 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1476 used even when parsing explicit parameters, when loading a encoded key
1477 or calling `EC_GROUP_new_from_ecpkparameters()`/
1478 `EC_GROUP_new_from_ecparameters()`.
1479 This prevents bypass of security hardening and performance gains,
1480 especially for curves with specialized EC_METHODs.
1481 By default, if a key encoded with explicit parameters is loaded and later
1482 encoded, the output is still encoded with explicit parameters, even if
1483 internally a "named" EC_GROUP is used for computation.
1487 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1488 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1489 NULL. After this change, only the cofactor parameter can be NULL. It also
1490 does some minimal sanity checks on the passed order.
1495 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1496 An attack is simple, if the first CMS_recipientInfo is valid but the
1497 second CMS_recipientInfo is chosen ciphertext. If the second
1498 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1499 encryption key will be replaced by garbage, and the message cannot be
1500 decoded, but if the RSA decryption fails, the correct encryption key is
1501 used and the recipient will not notice the attack.
1502 As a work around for this potential attack the length of the decrypted
1503 key must be equal to the cipher default key length, in case the
1504 certifiate is not given and all recipientInfo are tried out.
1505 The old behaviour can be re-enabled in the CMS code by setting the
1506 CMS_DEBUG_DECRYPT flag.
1511 * Early start up entropy quality from the DEVRANDOM seed source has been
1512 improved for older Linux systems. The RAND subsystem will wait for
1513 /dev/random to be producing output before seeding from /dev/urandom.
1514 The seeded state is stored for future library initialisations using
1515 a system global shared memory segment. The shared memory identifier
1516 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1517 the desired value. The default identifier is 114.
1521 * Correct the extended master secret constant on EBCDIC systems. Without this
1522 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1523 negotiate EMS will fail. Unfortunately this also means that TLS connections
1524 between EBCDIC systems with this fix, and EBCDIC systems without this
1525 fix will fail if they negotiate EMS.
1529 * Use Windows installation paths in the mingw builds
1531 Mingw isn't a POSIX environment per se, which means that Windows
1532 paths should be used for installation.
1537 * Changed DH_check to accept parameters with order q and 2q subgroups.
1538 With order 2q subgroups the bit 0 of the private key is not secret
1539 but DH_generate_key works around that by clearing bit 0 of the
1540 private key for those. This avoids leaking bit 0 of the private key.
1544 * Significantly reduce secure memory usage by the randomness pools.
1548 * Revert the DEVRANDOM_WAIT feature for Linux systems
1550 The DEVRANDOM_WAIT feature added a select() call to wait for the
1551 /dev/random device to become readable before reading from the
1552 /dev/urandom device.
1554 It turned out that this change had negative side effects on
1555 performance which were not acceptable. After some discussion it
1556 was decided to revert this feature and leave it up to the OS
1557 resp. the platform maintainer to ensure a proper initialization
1558 during early boot time.
1560 *Matthias St. Pierre*
1562 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1564 * Add build tests for C++. These are generated files that only do one
1565 thing, to include one public OpenSSL head file each. This tests that
1566 the public header files can be usefully included in a C++ application.
1568 This test isn't enabled by default. It can be enabled with the option
1569 'enable-buildtest-c++'.
1573 * Enable SHA3 pre-hashing for ECDSA and DSA.
1577 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1578 This changes the size when using the genpkey app when no size is given. It
1579 fixes an omission in earlier changes that changed all RSA, DSA and DH
1580 generation apps to use 2048 bits by default.
1584 * Reorganize the manual pages to consistently have RETURN VALUES,
1585 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1586 util/fix-doc-nits accordingly.
1588 *Paul Yang, Joshua Lock*
1590 * Add the missing accessor EVP_PKEY_get0_engine()
1594 * Have apps like 's_client' and 's_server' output the signature scheme
1595 along with other cipher suite parameters when debugging.
1599 * Make OPENSSL_config() error agnostic again.
1603 * Do the error handling in RSA decryption constant time.
1607 * Prevent over long nonces in ChaCha20-Poly1305.
1609 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1610 for every encryption operation. RFC 7539 specifies that the nonce value
1611 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1612 and front pads the nonce with 0 bytes if it is less than 12
1613 bytes. However it also incorrectly allows a nonce to be set of up to 16
1614 bytes. In this case only the last 12 bytes are significant and any
1615 additional leading bytes are ignored.
1617 It is a requirement of using this cipher that nonce values are
1618 unique. Messages encrypted using a reused nonce value are susceptible to
1619 serious confidentiality and integrity attacks. If an application changes
1620 the default nonce length to be longer than 12 bytes and then makes a
1621 change to the leading bytes of the nonce expecting the new value to be a
1622 new unique nonce then such an application could inadvertently encrypt
1623 messages with a reused nonce.
1625 Additionally the ignored bytes in a long nonce are not covered by the
1626 integrity guarantee of this cipher. Any application that relies on the
1627 integrity of these ignored leading bytes of a long nonce may be further
1628 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1629 is safe because no such use sets such a long nonce value. However user
1630 applications that use this cipher directly and set a non-default nonce
1631 length to be longer than 12 bytes may be vulnerable.
1633 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1639 * Add DEVRANDOM_WAIT feature for Linux systems
1641 On older Linux systems where the getrandom() system call is not available,
1642 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1643 Contrary to getrandom(), the /dev/urandom device will not block during
1644 early boot when the kernel CSPRNG has not been seeded yet.
1646 To mitigate this known weakness, use select() to wait for /dev/random to
1647 become readable before reading from /dev/urandom.
1649 * Ensure that SM2 only uses SM3 as digest algorithm
1653 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1655 * Change the info callback signals for the start and end of a post-handshake
1656 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1657 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1658 confused by this and assume that a TLSv1.2 renegotiation has started. This
1659 can break KeyUpdate handling. Instead we no longer signal the start and end
1660 of a post handshake message exchange (although the messages themselves are
1661 still signalled). This could break some applications that were expecting
1662 the old signals. However without this KeyUpdate is not usable for many
1667 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1669 * Timing vulnerability in DSA signature generation
1671 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1672 timing side channel attack. An attacker could use variations in the signing
1673 algorithm to recover the private key.
1675 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1680 * Timing vulnerability in ECDSA signature generation
1682 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1683 timing side channel attack. An attacker could use variations in the signing
1684 algorithm to recover the private key.
1686 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1691 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1692 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1693 of two gigabytes and the error handling improved.
1695 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1696 categorized as a normal bug, not a security issue, because the DRBG reseeds
1697 automatically and is fully functional even without additional randomness
1698 provided by the application.
1700 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1702 * Add a new ClientHello callback. Provides a callback interface that gives
1703 the application the ability to adjust the nascent SSL object at the
1704 earliest stage of ClientHello processing, immediately after extensions have
1705 been collected but before they have been processed. In particular, this
1706 callback can adjust the supported TLS versions in response to the contents
1711 * Add SM2 base algorithm support.
1715 * s390x assembly pack: add (improved) hardware-support for the following
1716 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1717 aes-cfb/cfb8, aes-ecb.
1721 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1722 parameter is no longer accepted, as it leads to a corrupt table. NULL
1723 pem_str is reserved for alias entries only.
1727 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1728 step for prime curves. The new implementation is based on formulae from
1729 differential addition-and-doubling in homogeneous projective coordinates
1730 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1731 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1732 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1733 to work in projective coordinates.
1735 *Billy Bob Brumley, Nicola Tuveri*
1737 * Change generating and checking of primes so that the error rate of not
1738 being prime depends on the intended use based on the size of the input.
1739 For larger primes this will result in more rounds of Miller-Rabin.
1740 The maximal error rate for primes with more than 1080 bits is lowered
1743 *Kurt Roeckx, Annie Yousar*
1745 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1749 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1750 moving between systems, and to avoid confusion when a Windows build is
1751 done with mingw vs with MSVC. For POSIX installs, there's still a
1752 symlink or copy named 'tsget' to avoid that confusion as well.
1756 * Revert blinding in ECDSA sign and instead make problematic addition
1757 length-invariant. Switch even to fixed-length Montgomery multiplication.
1761 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1762 step for binary curves. The new implementation is based on formulae from
1763 differential addition-and-doubling in mixed Lopez-Dahab projective
1764 coordinates, modified to independently blind the operands.
1766 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1768 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1769 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1770 EC_METHODs to implement their own specialized "ladder step", to take
1771 advantage of more favorable coordinate systems or more efficient
1772 differential addition-and-doubling algorithms.
1774 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1776 * Modified the random device based seed sources to keep the relevant
1777 file descriptors open rather than reopening them on each access.
1778 This allows such sources to operate in a chroot() jail without
1779 the associated device nodes being available. This behaviour can be
1780 controlled using RAND_keep_random_devices_open().
1784 * Numerous side-channel attack mitigations have been applied. This may have
1785 performance impacts for some algorithms for the benefit of improved
1786 security. Specific changes are noted in this change log by their respective
1791 * AIX shared library support overhaul. Switch to AIX "natural" way of
1792 handling shared libraries, which means collecting shared objects of
1793 different versions and bitnesses in one common archive. This allows to
1794 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1795 doesn't affect the way 3rd party applications are linked, only how
1796 multi-version installation is managed.
1800 * Make ec_group_do_inverse_ord() more robust and available to other
1801 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1802 mitigations are applied to the fallback BN_mod_inverse().
1803 When using this function rather than BN_mod_inverse() directly, new
1804 EC cryptosystem implementations are then safer-by-default.
1808 * Add coordinate blinding for EC_POINT and implement projective
1809 coordinate blinding for generic prime curves as a countermeasure to
1810 chosen point SCA attacks.
1812 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1814 * Add blinding to ECDSA and DSA signatures to protect against side channel
1815 attacks discovered by Keegan Ryan (NCC Group).
1819 * Enforce checking in the pkeyutl command line app to ensure that the input
1820 length does not exceed the maximum supported digest length when performing
1821 a sign, verify or verifyrecover operation.
1825 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1826 I/O in combination with something like select() or poll() will hang. This
1827 can be turned off again using SSL_CTX_clear_mode().
1828 Many applications do not properly handle non-application data records, and
1829 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1830 around the problems in those applications, but can also break some.
1831 It's recommended to read the manpages about SSL_read(), SSL_write(),
1832 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1833 SSL_CTX_set_read_ahead() again.
1837 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1838 now allow empty (zero character) pass phrases.
1842 * Apply blinding to binary field modular inversion and remove patent
1843 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1847 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1848 binary and prime elliptic curves.
1852 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1853 constant time fixed point multiplication.
1857 * Revise elliptic curve scalar multiplication with timing attack
1858 defenses: ec_wNAF_mul redirects to a constant time implementation
1859 when computing fixed point and variable point multiplication (which
1860 in OpenSSL are mostly used with secret scalars in keygen, sign,
1861 ECDH derive operations).
1862 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1865 * Updated CONTRIBUTING
1869 * Updated DRBG / RAND to request nonce and additional low entropy
1870 randomness from the system.
1872 *Matthias St. Pierre*
1874 * Updated 'openssl rehash' to use OpenSSL consistent default.
1878 * Moved the load of the ssl_conf module to libcrypto, which helps
1879 loading engines that libssl uses before libssl is initialised.
1883 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1887 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1889 *Ingo Schwarze, Rich Salz*
1891 * Added output of accepting IP address and port for 'openssl s_server'
1895 * Added a new API for TLSv1.3 ciphersuites:
1896 SSL_CTX_set_ciphersuites()
1897 SSL_set_ciphersuites()
1901 * Memory allocation failures consistently add an error to the error
1906 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1907 in libcrypto when run as setuid/setgid.
1911 * Load any config file by default when libssl is used.
1915 * Added new public header file <openssl/rand_drbg.h> and documentation
1916 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1918 *Matthias St. Pierre*
1920 * QNX support removed (cannot find contributors to get their approval
1921 for the license change).
1925 * TLSv1.3 replay protection for early data has been implemented. See the
1926 SSL_read_early_data() man page for further details.
1930 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1931 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1932 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1933 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1934 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1935 configuration has been separated out. See the ciphers man page or the
1936 SSL_CTX_set_ciphersuites() man page for more information.
1940 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1941 in responder mode now supports the new "-multi" option, which
1942 spawns the specified number of child processes to handle OCSP
1943 requests. The "-timeout" option now also limits the OCSP
1944 responder's patience to wait to receive the full client request
1945 on a newly accepted connection. Child processes are respawned
1946 as needed, and the CA index file is automatically reloaded
1947 when changed. This makes it possible to run the "ocsp" responder
1948 as a long-running service, making the OpenSSL CA somewhat more
1949 feature-complete. In this mode, most diagnostic messages logged
1950 after entering the event loop are logged via syslog(3) rather than
1955 * Added support for X448 and Ed448. Heavily based on original work by
1960 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1961 objects loaded. This adds the functions OSSL_STORE_expect() and
1962 OSSL_STORE_find() as well as needed tools to construct searches and
1963 get the search data out of them.
1967 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1968 version of OpenSSL should review their configuration settings to ensure
1969 that they are still appropriate for TLSv1.3. For further information see:
1970 <https://wiki.openssl.org/index.php/TLS1.3>
1974 * Grand redesign of the OpenSSL random generator
1976 The default RAND method now utilizes an AES-CTR DRBG according to
1977 NIST standard SP 800-90Ar1. The new random generator is essentially
1978 a port of the default random generator from the OpenSSL FIPS 2.0
1979 object module. It is a hybrid deterministic random bit generator
1980 using an AES-CTR bit stream and which seeds and reseeds itself
1981 automatically using trusted system entropy sources.
1983 Some of its new features are:
1984 - Support for multiple DRBG instances with seed chaining.
1985 - The default RAND method makes use of a DRBG.
1986 - There is a public and private DRBG instance.
1987 - The DRBG instances are fork-safe.
1988 - Keep all global DRBG instances on the secure heap if it is enabled.
1989 - The public and private DRBG instance are per thread for lock free
1992 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
1994 * Changed Configure so it only says what it does and doesn't dump
1995 so much data. Instead, ./configdata.pm should be used as a script
1996 to display all sorts of configuration data.
2000 * Added processing of "make variables" to Configure.
2004 * Added SHA512/224 and SHA512/256 algorithm support.
2008 * The last traces of Netware support, first removed in 1.1.0, have
2013 * Get rid of Makefile.shared, and in the process, make the processing
2014 of certain files (rc.obj, or the .def/.map/.opt files produced from
2015 the ordinal files) more visible and hopefully easier to trace and
2016 debug (or make silent).
2020 * Make it possible to have environment variable assignments as
2021 arguments to config / Configure.
2025 * Add multi-prime RSA (RFC 8017) support.
2029 * Add SM3 implemented according to GB/T 32905-2016
2030 *Jack Lloyd <jack.lloyd@ribose.com>,*
2031 *Ronald Tse <ronald.tse@ribose.com>,*
2032 *Erick Borsboom <erick.borsboom@ribose.com>*
2034 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2035 as documented in RFC6066.
2036 Based on a patch from Tomasz Moń
2038 *Filipe Raimundo da Silva*
2040 * Add SM4 implemented according to GB/T 32907-2016.
2041 *Jack Lloyd <jack.lloyd@ribose.com>,*
2042 *Ronald Tse <ronald.tse@ribose.com>,*
2043 *Erick Borsboom <erick.borsboom@ribose.com>*
2045 * Reimplement -newreq-nodes and ERR_error_string_n; the
2046 original author does not agree with the license change.
2050 * Add ARIA AEAD TLS support.
2054 * Some macro definitions to support VS6 have been removed. Visual
2055 Studio 6 has not worked since 1.1.0
2059 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2060 without clearing the errors.
2064 * Add "atfork" functions. If building on a system that without
2065 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2066 requirements. The RAND facility now uses/requires this.
2074 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2075 not possible to disable entirely. However, it's still possible to
2076 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2079 To disable, configure with 'no-ui-console'. 'no-ui' is still
2080 possible to use as an alias. Check at compile time with the
2081 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2082 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2086 * Add a STORE module, which implements a uniform and URI based reader of
2087 stores that can contain keys, certificates, CRLs and numerous other
2088 objects. The main API is loosely based on a few stdio functions,
2089 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2090 OSSL_STORE_error and OSSL_STORE_close.
2091 The implementation uses backends called "loaders" to implement arbitrary
2092 URI schemes. There is one built in "loader" for the 'file' scheme.
2096 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2097 then adjusted to work on FreeBSD 8.4 as well.
2098 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2099 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2103 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2104 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2105 error code calls like this:
2107 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2109 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2110 that can be encoded in C. For the foreseeable future, this will only
2113 *Richard Levitte and Tim Hudson*
2115 * Removed BSD cryptodev engine.
2119 * Add a build target 'build_all_generated', to build all generated files
2120 and only that. This can be used to prepare everything that requires
2121 things like perl for a system that lacks perl and then move everything
2122 to that system and do the rest of the build there.
2126 * In the UI interface, make it possible to duplicate the user data. This
2127 can be used by engines that need to retain the data for a longer time
2128 than just the call where this user data is passed.
2132 * Ignore the '-named_curve auto' value for compatibility of applications
2135 *Tomas Mraz <tmraz@fedoraproject.org>*
2137 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2138 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2139 alerts across multiple records (some of which could be empty). In practice
2140 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2141 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2142 support this at all. Supporting it adds significant complexity to the
2143 record layer, and its removal is unlikely to cause interoperability
2148 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2149 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2150 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2155 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2156 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2158 *Richard Levitte, Andy Polyakov*
2160 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2165 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2166 platform rather than 'mingw'.
2170 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2171 success if they are asked to add an object which already exists
2172 in the store. This change cascades to other functions which load
2173 certificates and CRLs.
2177 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2178 facilitate stack unwinding even from assembly subroutines.
2182 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2183 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2187 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2188 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2189 which is the minimum version we support.
2193 * Certificate time validation (X509_cmp_time) enforces stricter
2194 compliance with RFC 5280. Fractional seconds and timezone offsets
2195 are no longer allowed.
2199 * Add support for ARIA
2203 * s_client will now send the Server Name Indication (SNI) extension by
2204 default unless the new "-noservername" option is used. The server name is
2205 based on the host provided to the "-connect" option unless overridden by
2206 using "-servername".
2210 * Add support for SipHash
2214 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2215 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2216 prevent issues where no progress is being made and the peer continually
2217 sends unrecognised record types, using up resources processing them.
2221 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2222 using the algorithm defined in
2223 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2227 * Heartbeat support has been removed; the ABI is changed for now.
2229 *Richard Levitte, Rich Salz*
2231 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2235 * The RSA "null" method, which was partially supported to avoid patent
2236 issues, has been replaced to always returns NULL.
2243 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2245 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2246 used even when parsing explicit parameters, when loading a encoded key
2247 or calling `EC_GROUP_new_from_ecpkparameters()`/
2248 `EC_GROUP_new_from_ecparameters()`.
2249 This prevents bypass of security hardening and performance gains,
2250 especially for curves with specialized EC_METHODs.
2251 By default, if a key encoded with explicit parameters is loaded and later
2252 encoded, the output is still encoded with explicit parameters, even if
2253 internally a "named" EC_GROUP is used for computation.
2257 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2258 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2259 NULL. After this change, only the cofactor parameter can be NULL. It also
2260 does some minimal sanity checks on the passed order.
2265 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2266 An attack is simple, if the first CMS_recipientInfo is valid but the
2267 second CMS_recipientInfo is chosen ciphertext. If the second
2268 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2269 encryption key will be replaced by garbage, and the message cannot be
2270 decoded, but if the RSA decryption fails, the correct encryption key is
2271 used and the recipient will not notice the attack.
2272 As a work around for this potential attack the length of the decrypted
2273 key must be equal to the cipher default key length, in case the
2274 certifiate is not given and all recipientInfo are tried out.
2275 The old behaviour can be re-enabled in the CMS code by setting the
2276 CMS_DEBUG_DECRYPT flag.
2281 * Use Windows installation paths in the mingw builds
2283 Mingw isn't a POSIX environment per se, which means that Windows
2284 paths should be used for installation.
2289 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2291 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2292 This changes the size when using the genpkey app when no size is given. It
2293 fixes an omission in earlier changes that changed all RSA, DSA and DH
2294 generation apps to use 2048 bits by default.
2298 * Prevent over long nonces in ChaCha20-Poly1305.
2300 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2301 for every encryption operation. RFC 7539 specifies that the nonce value
2302 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2303 and front pads the nonce with 0 bytes if it is less than 12
2304 bytes. However it also incorrectly allows a nonce to be set of up to 16
2305 bytes. In this case only the last 12 bytes are significant and any
2306 additional leading bytes are ignored.
2308 It is a requirement of using this cipher that nonce values are
2309 unique. Messages encrypted using a reused nonce value are susceptible to
2310 serious confidentiality and integrity attacks. If an application changes
2311 the default nonce length to be longer than 12 bytes and then makes a
2312 change to the leading bytes of the nonce expecting the new value to be a
2313 new unique nonce then such an application could inadvertently encrypt
2314 messages with a reused nonce.
2316 Additionally the ignored bytes in a long nonce are not covered by the
2317 integrity guarantee of this cipher. Any application that relies on the
2318 integrity of these ignored leading bytes of a long nonce may be further
2319 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2320 is safe because no such use sets such a long nonce value. However user
2321 applications that use this cipher directly and set a non-default nonce
2322 length to be longer than 12 bytes may be vulnerable.
2324 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2330 * Added SCA hardening for modular field inversion in EC_GROUP through
2331 a new dedicated field_inv() pointer in EC_METHOD.
2332 This also addresses a leakage affecting conversions from projective
2333 to affine coordinates.
2335 *Billy Bob Brumley, Nicola Tuveri*
2337 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2338 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2342 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2346 * Remove the 'dist' target and add a tarball building script. The
2347 'dist' target has fallen out of use, and it shouldn't be
2348 necessary to configure just to create a source distribution.
2352 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2354 * Timing vulnerability in DSA signature generation
2356 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2357 timing side channel attack. An attacker could use variations in the signing
2358 algorithm to recover the private key.
2360 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2365 * Timing vulnerability in ECDSA signature generation
2367 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2368 timing side channel attack. An attacker could use variations in the signing
2369 algorithm to recover the private key.
2371 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2376 * Add coordinate blinding for EC_POINT and implement projective
2377 coordinate blinding for generic prime curves as a countermeasure to
2378 chosen point SCA attacks.
2380 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2382 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2384 * Client DoS due to large DH parameter
2386 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2387 malicious server can send a very large prime value to the client. This will
2388 cause the client to spend an unreasonably long period of time generating a
2389 key for this prime resulting in a hang until the client has finished. This
2390 could be exploited in a Denial Of Service attack.
2392 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2397 * Cache timing vulnerability in RSA Key Generation
2399 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2400 a cache timing side channel attack. An attacker with sufficient access to
2401 mount cache timing attacks during the RSA key generation process could
2402 recover the private key.
2404 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2405 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2410 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2411 parameter is no longer accepted, as it leads to a corrupt table. NULL
2412 pem_str is reserved for alias entries only.
2416 * Revert blinding in ECDSA sign and instead make problematic addition
2417 length-invariant. Switch even to fixed-length Montgomery multiplication.
2421 * Change generating and checking of primes so that the error rate of not
2422 being prime depends on the intended use based on the size of the input.
2423 For larger primes this will result in more rounds of Miller-Rabin.
2424 The maximal error rate for primes with more than 1080 bits is lowered
2427 *Kurt Roeckx, Annie Yousar*
2429 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2433 * Add blinding to ECDSA and DSA signatures to protect against side channel
2434 attacks discovered by Keegan Ryan (NCC Group).
2438 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2439 now allow empty (zero character) pass phrases.
2443 * Certificate time validation (X509_cmp_time) enforces stricter
2444 compliance with RFC 5280. Fractional seconds and timezone offsets
2445 are no longer allowed.
2449 * Fixed a text canonicalisation bug in CMS
2451 Where a CMS detached signature is used with text content the text goes
2452 through a canonicalisation process first prior to signing or verifying a
2453 signature. This process strips trailing space at the end of lines, converts
2454 line terminators to CRLF and removes additional trailing line terminators
2455 at the end of a file. A bug in the canonicalisation process meant that
2456 some characters, such as form-feed, were incorrectly treated as whitespace
2457 and removed. This is contrary to the specification (RFC5485). This fix
2458 could mean that detached text data signed with an earlier version of
2459 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2460 signed with a fixed OpenSSL may fail to verify with an earlier version of
2461 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2462 and use the "-binary" flag (for the "cms" command line application) or set
2463 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2467 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2469 * Constructed ASN.1 types with a recursive definition could exceed the stack
2471 Constructed ASN.1 types with a recursive definition (such as can be found
2472 in PKCS7) could eventually exceed the stack given malicious input with
2473 excessive recursion. This could result in a Denial Of Service attack. There
2474 are no such structures used within SSL/TLS that come from untrusted sources
2475 so this is considered safe.
2477 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2483 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2485 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2486 effectively reduced to only comparing the least significant bit of each
2487 byte. This allows an attacker to forge messages that would be considered as
2488 authenticated in an amount of tries lower than that guaranteed by the
2489 security claims of the scheme. The module can only be compiled by the
2490 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2492 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2498 * Add a build target 'build_all_generated', to build all generated files
2499 and only that. This can be used to prepare everything that requires
2500 things like perl for a system that lacks perl and then move everything
2501 to that system and do the rest of the build there.
2505 * Backport SSL_OP_NO_RENGOTIATION
2507 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2508 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2509 changes this is no longer possible in 1.1.0. Therefore the new
2510 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2511 1.1.0 to provide equivalent functionality.
2513 Note that if an application built against 1.1.0h headers (or above) is run
2514 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2515 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2519 * Removed the OS390-Unix config target. It relied on a script that doesn't
2524 * rsaz_1024_mul_avx2 overflow bug on x86_64
2526 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2527 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2528 Analysis suggests that attacks against RSA and DSA as a result of this
2529 defect would be very difficult to perform and are not believed likely.
2530 Attacks against DH1024 are considered just feasible, because most of the
2531 work necessary to deduce information about a private key may be performed
2532 offline. The amount of resources required for such an attack would be
2533 significant. However, for an attack on TLS to be meaningful, the server
2534 would have to share the DH1024 private key among multiple clients, which is
2535 no longer an option since CVE-2016-0701.
2537 This only affects processors that support the AVX2 but not ADX extensions
2538 like Intel Haswell (4th generation).
2540 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2541 was originally found via the OSS-Fuzz project.
2546 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2548 * bn_sqrx8x_internal carry bug on x86_64
2550 There is a carry propagating bug in the x86_64 Montgomery squaring
2551 procedure. No EC algorithms are affected. Analysis suggests that attacks
2552 against RSA and DSA as a result of this defect would be very difficult to
2553 perform and are not believed likely. Attacks against DH are considered just
2554 feasible (although very difficult) because most of the work necessary to
2555 deduce information about a private key may be performed offline. The amount
2556 of resources required for such an attack would be very significant and
2557 likely only accessible to a limited number of attackers. An attacker would
2558 additionally need online access to an unpatched system using the target
2559 private key in a scenario with persistent DH parameters and a private
2560 key that is shared between multiple clients.
2562 This only affects processors that support the BMI1, BMI2 and ADX extensions
2563 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2565 This issue was reported to OpenSSL by the OSS-Fuzz project.
2570 * Malformed X.509 IPAddressFamily could cause OOB read
2572 If an X.509 certificate has a malformed IPAddressFamily extension,
2573 OpenSSL could do a one-byte buffer overread. The most likely result
2574 would be an erroneous display of the certificate in text format.
2576 This issue was reported to OpenSSL by the OSS-Fuzz project.
2581 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2583 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2584 platform rather than 'mingw'.
2588 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2589 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2590 which is the minimum version we support.
2594 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2596 * Encrypt-Then-Mac renegotiation crash
2598 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2599 negotiated where it was not in the original handshake (or vice-versa) then
2600 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2601 and servers are affected.
2603 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2608 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2610 * Truncated packet could crash via OOB read
2612 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2613 cipher is being used, then a truncated packet can cause that host to
2614 perform an out-of-bounds read, usually resulting in a crash.
2616 This issue was reported to OpenSSL by Robert Święcki of Google.
2621 * Bad (EC)DHE parameters cause a client crash
2623 If a malicious server supplies bad parameters for a DHE or ECDHE key
2624 exchange then this can result in the client attempting to dereference a
2625 NULL pointer leading to a client crash. This could be exploited in a Denial
2628 This issue was reported to OpenSSL by Guido Vranken.
2633 * BN_mod_exp may produce incorrect results on x86_64
2635 There is a carry propagating bug in the x86_64 Montgomery squaring
2636 procedure. No EC algorithms are affected. Analysis suggests that attacks
2637 against RSA and DSA as a result of this defect would be very difficult to
2638 perform and are not believed likely. Attacks against DH are considered just
2639 feasible (although very difficult) because most of the work necessary to
2640 deduce information about a private key may be performed offline. The amount
2641 of resources required for such an attack would be very significant and
2642 likely only accessible to a limited number of attackers. An attacker would
2643 additionally need online access to an unpatched system using the target
2644 private key in a scenario with persistent DH parameters and a private
2645 key that is shared between multiple clients. For example this can occur by
2646 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2647 similar to CVE-2015-3193 but must be treated as a separate problem.
2649 This issue was reported to OpenSSL by the OSS-Fuzz project.
2654 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2656 * ChaCha20/Poly1305 heap-buffer-overflow
2658 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2659 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2660 crash. This issue is not considered to be exploitable beyond a DoS.
2662 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2667 * CMS Null dereference
2669 Applications parsing invalid CMS structures can crash with a NULL pointer
2670 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2671 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2672 structure callback if an attempt is made to free certain invalid encodings.
2673 Only CHOICE structures using a callback which do not handle NULL value are
2676 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2681 * Montgomery multiplication may produce incorrect results
2683 There is a carry propagating bug in the Broadwell-specific Montgomery
2684 multiplication procedure that handles input lengths divisible by, but
2685 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2686 and DH private keys are impossible. This is because the subroutine in
2687 question is not used in operations with the private key itself and an input
2688 of the attacker's direct choice. Otherwise the bug can manifest itself as
2689 transient authentication and key negotiation failures or reproducible
2690 erroneous outcome of public-key operations with specially crafted input.
2691 Among EC algorithms only Brainpool P-512 curves are affected and one
2692 presumably can attack ECDH key negotiation. Impact was not analyzed in
2693 detail, because pre-requisites for attack are considered unlikely. Namely
2694 multiple clients have to choose the curve in question and the server has to
2695 share the private key among them, neither of which is default behaviour.
2696 Even then only clients that chose the curve will be affected.
2698 This issue was publicly reported as transient failures and was not
2699 initially recognized as a security issue. Thanks to Richard Morgan for
2700 providing reproducible case.
2705 * Removed automatic addition of RPATH in shared libraries and executables,
2706 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2710 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2712 * Fix Use After Free for large message sizes
2714 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2715 message larger than approx 16k is received then the underlying buffer to
2716 store the incoming message is reallocated and moved. Unfortunately a
2717 dangling pointer to the old location is left which results in an attempt to
2718 write to the previously freed location. This is likely to result in a
2719 crash, however it could potentially lead to execution of arbitrary code.
2721 This issue only affects OpenSSL 1.1.0a.
2723 This issue was reported to OpenSSL by Robert Święcki.
2728 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2730 * OCSP Status Request extension unbounded memory growth
2732 A malicious client can send an excessively large OCSP Status Request
2733 extension. If that client continually requests renegotiation, sending a
2734 large OCSP Status Request extension each time, then there will be unbounded
2735 memory growth on the server. This will eventually lead to a Denial Of
2736 Service attack through memory exhaustion. Servers with a default
2737 configuration are vulnerable even if they do not support OCSP. Builds using
2738 the "no-ocsp" build time option are not affected.
2740 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2745 * SSL_peek() hang on empty record
2747 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2748 sends an empty record. This could be exploited by a malicious peer in a
2749 Denial Of Service attack.
2751 This issue was reported to OpenSSL by Alex Gaynor.
2756 * Excessive allocation of memory in tls_get_message_header() and
2757 dtls1_preprocess_fragment()
2759 A (D)TLS message includes 3 bytes for its length in the header for the
2760 message. This would allow for messages up to 16Mb in length. Messages of
2761 this length are excessive and OpenSSL includes a check to ensure that a
2762 peer is sending reasonably sized messages in order to avoid too much memory
2763 being consumed to service a connection. A flaw in the logic of version
2764 1.1.0 means that memory for the message is allocated too early, prior to
2765 the excessive message length check. Due to way memory is allocated in
2766 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2767 to service a connection. This could lead to a Denial of Service through
2768 memory exhaustion. However, the excessive message length check still takes
2769 place, and this would cause the connection to immediately fail. Assuming
2770 that the application calls SSL_free() on the failed connection in a timely
2771 manner then the 21Mb of allocated memory will then be immediately freed
2772 again. Therefore the excessive memory allocation will be transitory in
2773 nature. This then means that there is only a security impact if:
2775 1) The application does not call SSL_free() in a timely manner in the event
2776 that the connection fails
2778 2) The application is working in a constrained environment where there is
2779 very little free memory
2781 3) The attacker initiates multiple connection attempts such that there are
2782 multiple connections in a state where memory has been allocated for the
2783 connection; SSL_free() has not yet been called; and there is insufficient
2784 memory to service the multiple requests.
2786 Except in the instance of (1) above any Denial Of Service is likely to be
2787 transitory because as soon as the connection fails the memory is
2788 subsequently freed again in the SSL_free() call. However there is an
2789 increased risk during this period of application crashes due to the lack of
2790 memory - which would then mean a more serious Denial of Service.
2792 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2793 (CVE-2016-6307 and CVE-2016-6308)
2797 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2798 had to be removed. Primary reason is that vendor assembler can't
2799 assemble our modules with -KPIC flag. As result it, assembly
2800 support, was not even available as option. But its lack means
2801 lack of side-channel resistant code, which is incompatible with
2802 security by todays standards. Fortunately gcc is readily available
2803 prepackaged option, which we firmly point at...
2807 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2809 * Windows command-line tool supports UTF-8 opt-in option for arguments
2810 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2811 (to any value) allows Windows user to access PKCS#12 file generated
2812 with Windows CryptoAPI and protected with non-ASCII password, as well
2813 as files generated under UTF-8 locale on Linux also protected with
2818 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2819 have been disabled by default and removed from DEFAULT, just like RC4.
2820 See the RC4 item below to re-enable both.
2824 * The method for finding the storage location for the Windows RAND seed file
2825 has changed. First we check %RANDFILE%. If that is not set then we check
2826 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2827 all else fails we fall back to C:\.
2831 * The EVP_EncryptUpdate() function has had its return type changed from void
2832 to int. A return of 0 indicates and error while a return of 1 indicates
2837 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2838 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2839 off the constant time implementation for RSA, DSA and DH have been made
2840 no-ops and deprecated.
2844 * Windows RAND implementation was simplified to only get entropy by
2845 calling CryptGenRandom(). Various other RAND-related tickets
2848 *Joseph Wylie Yandle, Rich Salz*
2850 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2851 and `OPENSSL_LH_`, respectively. The old names are available
2852 with API compatibility. They new names are now completely documented.
2856 * Unify TYPE_up_ref(obj) methods signature.
2857 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2858 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2859 int (instead of void) like all others TYPE_up_ref() methods.
2860 So now these methods also check the return value of CRYPTO_atomic_add(),
2861 and the validity of object reference counter.
2863 *fdasilvayy@gmail.com*
2865 * With Windows Visual Studio builds, the .pdb files are installed
2866 alongside the installed libraries and executables. For a static
2867 library installation, ossl_static.pdb is the associate compiler
2868 generated .pdb file to be used when linking programs.
2872 * Remove openssl.spec. Packaging files belong with the packagers.
2876 * Automatic Darwin/OSX configuration has had a refresh, it will now
2877 recognise x86_64 architectures automatically. You can still decide
2878 to build for a different bitness with the environment variable
2879 KERNEL_BITS (can be 32 or 64), for example:
2881 KERNEL_BITS=32 ./config
2885 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2886 256 bit AES and HMAC with SHA256.
2890 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2894 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2898 * To enable users to have their own config files and build file templates,
2899 Configure looks in the directory indicated by the environment variable
2900 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2901 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2902 name and is used as is.
2906 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2907 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2908 X509_CERT_FILE_CTX was removed.
2912 * "shared" builds are now the default. To create only static libraries use
2913 the "no-shared" Configure option.
2917 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2918 All of these option have not worked for some while and are fundamental
2923 * Make various cleanup routines no-ops and mark them as deprecated. Most
2924 global cleanup functions are no longer required because they are handled
2925 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2926 Explicitly de-initing can cause problems (e.g. where a library that uses
2927 OpenSSL de-inits, but an application is still using it). The affected
2928 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2929 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2930 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2931 COMP_zlib_cleanup().
2935 * --strict-warnings no longer enables runtime debugging options
2936 such as REF_DEBUG. Instead, debug options are automatically
2937 enabled with '--debug' builds.
2939 *Andy Polyakov, Emilia Käsper*
2941 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2942 have been moved out of the public header files. New functions for managing
2943 these have been added.
2947 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2948 objects have been moved out of the public header files. New
2949 functions for managing these have been added.
2953 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2954 have been moved out of the public header files. New functions for managing
2955 these have been added.
2959 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2960 moved out of the public header files. New functions for managing these
2965 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2969 * Removed the mk1mf build scripts.
2973 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
2974 it is always safe to #include a header now.
2978 * Removed the aged BC-32 config and all its supporting scripts
2982 * Removed support for Ultrix, Netware, and OS/2.
2986 * Add support for HKDF.
2988 *Alessandro Ghedini*
2990 * Add support for blake2b and blake2s
2994 * Added support for "pipelining". Ciphers that have the
2995 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
2996 encryptions/decryptions simultaneously. There are currently no built-in
2997 ciphers with this property but the expectation is that engines will be able
2998 to offer it to significantly improve throughput. Support has been extended
2999 into libssl so that multiple records for a single connection can be
3000 processed in one go (for >=TLS 1.1).
3004 * Added the AFALG engine. This is an async capable engine which is able to
3005 offload work to the Linux kernel. In this initial version it only supports
3006 AES128-CBC. The kernel must be version 4.1.0 or greater.
3010 * OpenSSL now uses a new threading API. It is no longer necessary to
3011 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3012 are two supported threading models: pthreads and windows threads. It is
3013 also possible to configure OpenSSL at compile time for "no-threads". The
3014 old threading API should no longer be used. The functions have been
3015 replaced with "no-op" compatibility macros.
3017 *Alessandro Ghedini, Matt Caswell*
3019 * Modify behavior of ALPN to invoke callback after SNI/servername
3020 callback, such that updates to the SSL_CTX affect ALPN.
3024 * Add SSL_CIPHER queries for authentication and key-exchange.
3028 * Changes to the DEFAULT cipherlist:
3029 - Prefer (EC)DHE handshakes over plain RSA.
3030 - Prefer AEAD ciphers over legacy ciphers.
3031 - Prefer ECDSA over RSA when both certificates are available.
3032 - Prefer TLSv1.2 ciphers/PRF.
3033 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3038 * Change the ECC default curve list to be this, in order: x25519,
3039 secp256r1, secp521r1, secp384r1.
3043 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3044 disabled by default. They can be re-enabled using the
3045 enable-weak-ssl-ciphers option to Configure.
3049 * If the server has ALPN configured, but supports no protocols that the
3050 client advertises, send a fatal "no_application_protocol" alert.
3051 This behaviour is SHALL in RFC 7301, though it isn't universally
3052 implemented by other servers.
3056 * Add X25519 support.
3057 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3058 for public and private key encoding using the format documented in
3059 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3060 key generation and key derivation.
3062 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3067 * Deprecate SRP_VBASE_get_by_user.
3068 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3069 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3070 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3071 seed, even if the seed is configured.
3073 Users should use SRP_VBASE_get1_by_user instead. Note that in
3074 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3075 also that even though configuring the SRP seed attempts to hide
3076 invalid usernames by continuing the handshake with fake
3077 credentials, this behaviour is not constant time and no strong
3078 guarantees are made that the handshake is indistinguishable from
3079 that of a valid user.
3083 * Configuration change; it's now possible to build dynamic engines
3084 without having to build shared libraries and vice versa. This
3085 only applies to the engines in engines/, those in crypto/engine/
3086 will always be built into libcrypto (i.e. "static").
3088 Building dynamic engines is enabled by default; to disable, use
3089 the configuration option "disable-dynamic-engine".
3091 The only requirements for building dynamic engines are the
3092 presence of the DSO module and building with position independent
3093 code, so they will also automatically be disabled if configuring
3094 with "disable-dso" or "disable-pic".
3096 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3097 are also taken away from openssl/opensslconf.h, as they are
3102 * Configuration change; if there is a known flag to compile
3103 position independent code, it will always be applied on the
3104 libcrypto and libssl object files, and never on the application
3105 object files. This means other libraries that use routines from
3106 libcrypto / libssl can be made into shared libraries regardless
3107 of how OpenSSL was configured.
3109 If this isn't desirable, the configuration options "disable-pic"
3110 or "no-pic" can be used to disable the use of PIC. This will
3111 also disable building shared libraries and dynamic engines.
3115 * Removed JPAKE code. It was experimental and has no wide use.
3119 * The INSTALL_PREFIX Makefile variable has been renamed to
3120 DESTDIR. That makes for less confusion on what this variable
3121 is for. Also, the configuration option --install_prefix is
3126 * Heartbeat for TLS has been removed and is disabled by default
3127 for DTLS; configure with enable-heartbeats. Code that uses the
3128 old #define's might need to be updated.
3130 *Emilia Käsper, Rich Salz*
3132 * Rename REF_CHECK to REF_DEBUG.
3136 * New "unified" build system
3138 The "unified" build system is aimed to be a common system for all
3139 platforms we support. With it comes new support for VMS.
3141 This system builds supports building in a different directory tree
3142 than the source tree. It produces one Makefile (for unix family
3143 or lookalikes), or one descrip.mms (for VMS).
3145 The source of information to make the Makefile / descrip.mms is
3146 small files called 'build.info', holding the necessary
3147 information for each directory with source to compile, and a
3148 template in Configurations, like unix-Makefile.tmpl or
3151 With this change, the library names were also renamed on Windows
3152 and on VMS. They now have names that are closer to the standard
3153 on Unix, and include the major version number, and in certain
3154 cases, the architecture they are built for. See "Notes on shared
3155 libraries" in INSTALL.
3157 We rely heavily on the perl module Text::Template.
3161 * Added support for auto-initialisation and de-initialisation of the library.
3162 OpenSSL no longer requires explicit init or deinit routines to be called,
3163 except in certain circumstances. See the OPENSSL_init_crypto() and
3164 OPENSSL_init_ssl() man pages for further information.
3168 * The arguments to the DTLSv1_listen function have changed. Specifically the
3169 "peer" argument is now expected to be a BIO_ADDR object.
3171 * Rewrite of BIO networking library. The BIO library lacked consistent
3172 support of IPv6, and adding it required some more extensive
3173 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3174 which hold all types of addresses and chains of address information.
3175 It also introduces a new API, with functions like BIO_socket,
3176 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3177 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3178 have been adapted accordingly.
3182 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3187 * CRIME protection: disable compression by default, even if OpenSSL is
3188 compiled with zlib enabled. Applications can still enable compression
3189 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3190 using the SSL_CONF library to configure compression.
3194 * The signature of the session callback configured with
3195 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3196 was explicitly marked as `const unsigned char*` instead of
3201 * Always DPURIFY. Remove the use of uninitialized memory in the
3202 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3206 * Removed many obsolete configuration items, including
3207 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3208 MD2_CHAR, MD2_INT, MD2_LONG
3210 IDEA_SHORT, IDEA_LONG
3211 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3213 *Rich Salz, with advice from Andy Polyakov*
3215 * Many BN internals have been moved to an internal header file.
3217 *Rich Salz with help from Andy Polyakov*
3219 * Configuration and writing out the results from it has changed.
3220 Files such as Makefile include/openssl/opensslconf.h and are now
3221 produced through general templates, such as Makefile.in and
3222 crypto/opensslconf.h.in and some help from the perl module
3225 Also, the center of configuration information is no longer
3226 Makefile. Instead, Configure produces a perl module in
3227 configdata.pm which holds most of the config data (in the hash
3228 table %config), the target data that comes from the target
3229 configuration in one of the `Configurations/*.conf` files (in
3234 * To clarify their intended purposes, the Configure options
3235 --prefix and --openssldir change their semantics, and become more
3236 straightforward and less interdependent.
3238 --prefix shall be used exclusively to give the location INSTALLTOP
3239 where programs, scripts, libraries, include files and manuals are
3240 going to be installed. The default is now /usr/local.
3242 --openssldir shall be used exclusively to give the default
3243 location OPENSSLDIR where certificates, private keys, CRLs are
3244 managed. This is also where the default openssl.cnf gets
3246 If the directory given with this option is a relative path, the
3247 values of both the --prefix value and the --openssldir value will
3248 be combined to become OPENSSLDIR.
3249 The default for --openssldir is INSTALLTOP/ssl.
3251 Anyone who uses --openssldir to specify where OpenSSL is to be
3252 installed MUST change to use --prefix instead.
3256 * The GOST engine was out of date and therefore it has been removed. An up
3257 to date GOST engine is now being maintained in an external repository.
3258 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3259 support for GOST ciphersuites (these are only activated if a GOST engine
3264 * EGD is no longer supported by default; use enable-egd when
3267 *Ben Kaduk and Rich Salz*
3269 * The distribution now has Makefile.in files, which are used to
3270 create Makefile's when Configure is run. *Configure must be run
3271 before trying to build now.*
3275 * The return value for SSL_CIPHER_description() for error conditions
3280 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3282 Obtaining and performing DNSSEC validation of TLSA records is
3283 the application's responsibility. The application provides
3284 the TLSA records of its choice to OpenSSL, and these are then
3285 used to authenticate the peer.
3287 The TLSA records need not even come from DNS. They can, for
3288 example, be used to implement local end-entity certificate or
3289 trust-anchor "pinning", where the "pin" data takes the form
3290 of TLSA records, which can augment or replace verification
3291 based on the usual WebPKI public certification authorities.
3295 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3296 continues to support deprecated interfaces in default builds.
3297 However, applications are strongly advised to compile their
3298 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3299 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3300 or the 1.1.0 releases.
3302 In environments in which all applications have been ported to
3303 not use any deprecated interfaces OpenSSL's Configure script
3304 should be used with the --api=1.1.0 option to entirely remove
3305 support for the deprecated features from the library and
3306 unconditionally disable them in the installed headers.
3307 Essentially the same effect can be achieved with the "no-deprecated"
3308 argument to Configure, except that this will always restrict
3309 the build to just the latest API, rather than a fixed API
3312 As applications are ported to future revisions of the API,
3313 they should update their compile-time OPENSSL_API_COMPAT define
3314 accordingly, but in most cases should be able to continue to
3315 compile with later releases.
3317 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3318 0x10000000L and 0x00908000L, respectively. However those
3319 versions did not support the OPENSSL_API_COMPAT feature, and
3320 so applications are not typically tested for explicit support
3321 of just the undeprecated features of either release.
3325 * Add support for setting the minimum and maximum supported protocol.
3326 It can bet set via the SSL_set_min_proto_version() and
3327 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3328 MaxProtocol. It's recommended to use the new APIs to disable
3329 protocols instead of disabling individual protocols using
3330 SSL_set_options() or SSL_CONF's Protocol. This change also
3331 removes support for disabling TLS 1.2 in the OpenSSL TLS
3332 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3336 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3340 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3341 and integrates ECDSA and ECDH functionality into EC. Implementations can
3342 now redirect key generation and no longer need to convert to or from
3345 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3346 include the ec.h header file instead.
3350 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3351 ciphers who are no longer supported and drops support the ephemeral RSA key
3352 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3356 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3357 opaque. For HMAC_CTX, the following constructors and destructors
3360 HMAC_CTX *HMAC_CTX_new(void);
3361 void HMAC_CTX_free(HMAC_CTX *ctx);
3363 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3364 destroy such methods has been added. See EVP_MD_meth_new(3) and
3365 EVP_CIPHER_meth_new(3) for documentation.
3368 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3369 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3370 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3371 an already created structure.
3372 2) For consistency with the majority of our object creators and
3373 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3374 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3375 for deprecated builds.
3379 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3380 cryptographic operations to be performed asynchronously as long as an
3381 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3382 further details. Libssl has also had this capability integrated with the
3383 introduction of the new mode SSL_MODE_ASYNC and associated error
3384 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3385 pages. This work was developed in partnership with Intel Corp.
3389 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3390 always enabled now. If you want to disable the support you should
3391 exclude it using the list of supported ciphers. This also means that the
3392 "-no_ecdhe" option has been removed from s_server.
3396 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3397 SSL_{CTX_}set1_curves() which can set a list.
3401 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3402 curve you want to support using SSL_{CTX_}set1_curves().
3406 * State machine rewrite. The state machine code has been significantly
3407 refactored in order to remove much duplication of code and solve issues
3408 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3409 further details). This change does have some associated API changes.
3410 Notably the SSL_state() function has been removed and replaced by
3411 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3412 SSL_set_state() has been removed altogether. The previous handshake states
3413 defined in ssl.h and ssl3.h have also been removed.
3417 * All instances of the string "ssleay" in the public API were replaced
3418 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3419 Some error codes related to internal RSA_eay API's were renamed.
3423 * The demo files in crypto/threads were moved to demo/threads.
3427 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3430 *Matt Caswell, Rich Salz*
3432 * New ASN.1 embed macro.
3434 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3435 structure is not allocated: it is part of the parent. That is instead of
3443 This reduces memory fragmentation and make it impossible to accidentally
3444 set a mandatory field to NULL.
3446 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3447 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3448 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3453 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3457 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3458 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3459 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3460 DES and RC4 ciphersuites.
3464 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3465 This changes the decoding behaviour for some invalid messages,
3466 though the change is mostly in the more lenient direction, and
3467 legacy behaviour is preserved as much as possible.
3471 * Fix no-stdio build.
3472 *David Woodhouse <David.Woodhouse@intel.com> and also*
3473 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3475 * New testing framework
3476 The testing framework has been largely rewritten and is now using
3477 perl and the perl modules Test::Harness and an extended variant of
3478 Test::More called OpenSSL::Test to do its work. All test scripts in
3479 test/ have been rewritten into test recipes, and all direct calls to
3480 executables in test/Makefile have become individual recipes using the
3481 simplified testing OpenSSL::Test::Simple.
3483 For documentation on our testing modules, do:
3485 perldoc test/testlib/OpenSSL/Test/Simple.pm
3486 perldoc test/testlib/OpenSSL/Test.pm
3490 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3491 are used; the latter aborts on memory leaks (usually checked on exit).
3492 Some undocumented "set malloc, etc., hooks" functions were removed
3493 and others were changed. All are now documented.
3497 * In DSA_generate_parameters_ex, if the provided seed is too short,
3500 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3502 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3503 from RFC4279, RFC4785, RFC5487, RFC5489.
3505 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3506 original RSA_PSK patch.
3510 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3511 era flag was never set throughout the codebase (only read). Also removed
3512 SSL3_FLAGS_POP_BUFFER which was only used if
3513 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3517 * Changed the default name options in the "ca", "crl", "req" and "x509"
3518 to be "oneline" instead of "compat".
3522 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3523 not aware of clients that still exhibit this bug, and the workaround
3524 hasn't been working properly for a while.
3528 * The return type of BIO_number_read() and BIO_number_written() as well as
3529 the corresponding num_read and num_write members in the BIO structure has
3530 changed from unsigned long to uint64_t. On platforms where an unsigned
3531 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3536 * Given the pervasive nature of TLS extensions it is inadvisable to run
3537 OpenSSL without support for them. It also means that maintaining
3538 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3539 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3543 * Removed support for the two export grade static DH ciphersuites
3544 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3545 were newly added (along with a number of other static DH ciphersuites) to
3546 1.0.2. However the two export ones have *never* worked since they were
3547 introduced. It seems strange in any case to be adding new export
3548 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3552 * Version negotiation has been rewritten. In particular SSLv23_method(),
3553 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3554 and turned into macros which simply call the new preferred function names
3555 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3556 should use the new names instead. Also as part of this change the ssl23.h
3557 header file has been removed.
3561 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3562 code and the associated standard is no longer considered fit-for-purpose.
3566 * RT2547 was closed. When generating a private key, try to make the
3567 output file readable only by the owner. This behavior change might
3568 be noticeable when interacting with other software.
3570 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3575 * Added HTTP GET support to the ocsp command.
3579 * Changed default digest for the dgst and enc commands from MD5 to
3584 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3588 * Added support for TLS extended master secret from
3589 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3590 initial patch which was a great help during development.
3594 * All libssl internal structures have been removed from the public header
3595 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3596 now redundant). Users should not attempt to access internal structures
3597 directly. Instead they should use the provided API functions.
3601 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3602 Access to deprecated functions can be re-enabled by running config with
3603 "enable-deprecated". In addition applications wishing to use deprecated
3604 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3605 will, by default, disable some transitive includes that previously existed
3606 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3610 * Added support for OCB mode. OpenSSL has been granted a patent license
3611 compatible with the OpenSSL license for use of OCB. Details are available
3612 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3613 for OCB can be removed by calling config with no-ocb.
3617 * SSLv2 support has been removed. It still supports receiving a SSLv2
3618 compatible client hello.
3622 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3623 done while fixing the error code for the key-too-small case.
3625 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3627 * CA.sh has been removed; use CA.pl instead.
3631 * Removed old DES API.
3635 * Remove various unsupported platforms:
3641 Sinix/ReliantUNIX RM400
3646 16-bit platforms such as WIN16
3650 * Clean up OPENSSL_NO_xxx #define's
3651 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3652 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3653 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3654 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3655 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3656 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3657 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3658 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3659 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3660 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3664 * Cleaned up dead code
3665 Remove all but one '#ifdef undef' which is to be looked at.
3669 * Clean up calling of xxx_free routines.
3670 Just like free(), fix most of the xxx_free routines to accept
3671 NULL. Remove the non-null checks from callers. Save much code.
3675 * Add secure heap for storage of private keys (when possible).
3676 Add BIO_s_secmem(), CBIGNUM, etc.
3677 Contributed by Akamai Technologies under our Corporate CLA.
3681 * Experimental support for a new, fast, unbiased prime candidate generator,
3682 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3684 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3686 * New output format NSS in the sess_id command line tool. This allows
3687 exporting the session id and the master key in NSS keylog format.
3689 *Martin Kaiser <martin@kaiser.cx>*
3691 * Harmonize version and its documentation. -f flag is used to display
3694 *mancha <mancha1@zoho.com>*
3696 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3697 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3699 *mancha <mancha1@zoho.com>*
3701 * Fix some double frees. These are not thought to be exploitable.
3703 *mancha <mancha1@zoho.com>*
3705 * A missing bounds check in the handling of the TLS heartbeat extension
3706 can be used to reveal up to 64k of memory to a connected client or
3709 Thanks for Neel Mehta of Google Security for discovering this bug and to
3710 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3711 preparing the fix ([CVE-2014-0160])
3713 *Adam Langley, Bodo Moeller*
3715 * Fix for the attack described in the paper "Recovering OpenSSL
3716 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3717 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3718 <http://eprint.iacr.org/2014/140>
3720 Thanks to Yuval Yarom and Naomi Benger for discovering this
3721 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3723 *Yuval Yarom and Naomi Benger*
3725 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3726 this fixes a limitation in previous versions of OpenSSL.
3730 * Experimental encrypt-then-mac support.
3732 Experimental support for encrypt then mac from
3733 draft-gutmann-tls-encrypt-then-mac-02.txt
3735 To enable it set the appropriate extension number (0x42 for the test
3736 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3738 For non-compliant peers (i.e. just about everything) this should have no
3741 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3745 * Add EVP support for key wrapping algorithms, to avoid problems with
3746 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3747 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3748 algorithms and include tests cases.
3752 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3757 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3758 MGF1 digest and OAEP label.
3762 * Make openssl verify return errors.
3764 *Chris Palmer <palmer@google.com> and Ben Laurie*
3766 * New function ASN1_TIME_diff to calculate the difference between two
3767 ASN1_TIME structures or one structure and the current time.
3771 * Update fips_test_suite to support multiple command line options. New
3772 test to induce all self test errors in sequence and check expected
3777 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3778 sign or verify all in one operation.
3782 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3783 test programs and fips_test_suite. Includes functionality to parse
3784 the minimal script output of fipsalgest.pl directly.
3788 * Add authorisation parameter to FIPS_module_mode_set().
3792 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3796 * Use separate DRBG fields for internal and external flags. New function
3797 FIPS_drbg_health_check() to perform on demand health checking. Add
3798 generation tests to fips_test_suite with reduced health check interval to
3799 demonstrate periodic health checking. Add "nodh" option to
3800 fips_test_suite to skip very slow DH test.
3804 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3809 * More extensive health check for DRBG checking many more failure modes.
3810 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3811 combination: call this in fips_test_suite.
3815 * Add support for canonical generation of DSA parameter 'g'. See
3818 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3819 POST to handle HMAC cases.
3823 * Add functions FIPS_module_version() and FIPS_module_version_text()
3824 to return numerical and string versions of the FIPS module number.
3828 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3829 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3830 outside the validated module in the FIPS capable OpenSSL.
3834 * Minor change to DRBG entropy callback semantics. In some cases
3835 there is no multiple of the block length between min_len and
3836 max_len. Allow the callback to return more than max_len bytes
3837 of entropy but discard any extra: it is the callback's responsibility
3838 to ensure that the extra data discarded does not impact the
3839 requested amount of entropy.
3843 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3844 information in FIPS186-3, SP800-57 and SP800-131A.
3848 * CCM support via EVP. Interface is very similar to GCM case except we
3849 must supply all data in one chunk (i.e. no update, final) and the
3850 message length must be supplied if AAD is used. Add algorithm test
3855 * Initial version of POST overhaul. Add POST callback to allow the status
3856 of POST to be monitored and/or failures induced. Modify fips_test_suite
3857 to use callback. Always run all selftests even if one fails.
3861 * XTS support including algorithm test driver in the fips_gcmtest program.
3862 Note: this does increase the maximum key length from 32 to 64 bytes but
3863 there should be no binary compatibility issues as existing applications
3864 will never use XTS mode.
3868 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3869 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3870 performs algorithm blocking for unapproved PRNG types. Also do not
3871 set PRNG type in FIPS_mode_set(): leave this to the application.
3872 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3873 the standard OpenSSL PRNG: set additional data to a date time vector.
3877 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3878 This shouldn't present any incompatibility problems because applications
3879 shouldn't be using these directly and any that are will need to rethink
3880 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3884 * Extensive self tests and health checking required by SP800-90 DRBG.
3885 Remove strength parameter from FIPS_drbg_instantiate and always
3886 instantiate at maximum supported strength.
3890 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3894 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3898 * New function DH_compute_key_padded() to compute a DH key and pad with
3899 leading zeroes if needed: this complies with SP800-56A et al.
3903 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3904 anything, incomplete, subject to change and largely untested at present.
3908 * Modify fipscanisteronly build option to only build the necessary object
3909 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3913 * Add experimental option FIPSSYMS to give all symbols in
3914 fipscanister.o and FIPS or fips prefix. This will avoid
3915 conflicts with future versions of OpenSSL. Add perl script
3916 util/fipsas.pl to preprocess assembly language source files
3917 and rename any affected symbols.
3921 * Add selftest checks and algorithm block of non-fips algorithms in
3922 FIPS mode. Remove DES2 from selftests.
3926 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3927 return internal method without any ENGINE dependencies. Add new
3928 tiny fips sign and verify functions.
3932 * New build option no-ec2m to disable characteristic 2 code.
3936 * New build option "fipscanisteronly". This only builds fipscanister.o
3937 and (currently) associated fips utilities. Uses the file Makefile.fips
3938 instead of Makefile.org as the prototype.
3942 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3943 Update fips_gcmtest to use IV generator.
3947 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3948 setting output buffer to NULL. The `*Final` function must be
3949 called although it will not retrieve any additional data. The tag
3950 can be set or retrieved with a ctrl. The IV length is by default 12
3951 bytes (96 bits) but can be set to an alternative value. If the IV
3952 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3957 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3958 underlying do_cipher function handles all cipher semantics itself
3959 including padding and finalisation. This is useful if (for example)
3960 an ENGINE cipher handles block padding itself. The behaviour of
3961 do_cipher is subtly changed if this flag is set: the return value
3962 is the number of characters written to the output buffer (zero is
3963 no longer an error code) or a negative error code. Also if the
3964 input buffer is NULL and length 0 finalisation should be performed.
3968 * If a candidate issuer certificate is already part of the constructed
3969 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3973 * Improve forward-security support: add functions
3975 void SSL_CTX_set_not_resumable_session_callback(
3976 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
3977 void SSL_set_not_resumable_session_callback(
3978 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
3980 for use by SSL/TLS servers; the callback function will be called whenever a
3981 new session is created, and gets to decide whether the session may be
3982 cached to make it resumable (return 0) or not (return 1). (As by the
3983 SSL/TLS protocol specifications, the session_id sent by the server will be
3984 empty to indicate that the session is not resumable; also, the server will
3985 not generate RFC 4507 (RFC 5077) session tickets.)
3987 A simple reasonable callback implementation is to return is_forward_secure.
3988 This parameter will be set to 1 or 0 depending on the ciphersuite selected
3989 by the SSL/TLS server library, indicating whether it can provide forward
3992 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
3994 * New -verify_name option in command line utilities to set verification
3999 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4000 Add CMAC pkey methods.
4004 * Experimental renegotiation in s_server -www mode. If the client
4005 browses /reneg connection is renegotiated. If /renegcert it is
4006 renegotiated requesting a certificate.
4010 * Add an "external" session cache for debugging purposes to s_server. This
4011 should help trace issues which normally are only apparent in deployed
4012 multi-process servers.
4016 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4017 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4018 BIO_set_cipher() and some obscure PEM functions were changed so they
4019 can now return an error. The RAND changes required a change to the
4020 RAND_METHOD structure.
4024 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4025 a gcc attribute to warn if the result of a function is ignored. This
4026 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4027 whose return value is often ignored.
4031 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4032 These allow SCTs (signed certificate timestamps) to be requested and
4033 validated when establishing a connection.
4035 *Rob Percival <robpercival@google.com>*
4040 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4042 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4043 used even when parsing explicit parameters, when loading a encoded key
4044 or calling `EC_GROUP_new_from_ecpkparameters()`/
4045 `EC_GROUP_new_from_ecparameters()`.
4046 This prevents bypass of security hardening and performance gains,
4047 especially for curves with specialized EC_METHODs.
4048 By default, if a key encoded with explicit parameters is loaded and later
4049 encoded, the output is still encoded with explicit parameters, even if
4050 internally a "named" EC_GROUP is used for computation.
4054 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4055 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4056 NULL. After this change, only the cofactor parameter can be NULL. It also
4057 does some minimal sanity checks on the passed order.
4062 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4063 An attack is simple, if the first CMS_recipientInfo is valid but the
4064 second CMS_recipientInfo is chosen ciphertext. If the second
4065 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4066 encryption key will be replaced by garbage, and the message cannot be
4067 decoded, but if the RSA decryption fails, the correct encryption key is
4068 used and the recipient will not notice the attack.
4069 As a work around for this potential attack the length of the decrypted
4070 key must be equal to the cipher default key length, in case the
4071 certifiate is not given and all recipientInfo are tried out.
4072 The old behaviour can be re-enabled in the CMS code by setting the
4073 CMS_DEBUG_DECRYPT flag.
4078 * Document issue with installation paths in diverse Windows builds
4080 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4081 binaries and run-time config file.
4086 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4088 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4089 This changes the size when using the genpkey app when no size is given. It
4090 fixes an omission in earlier changes that changed all RSA, DSA and DH
4091 generation apps to use 2048 bits by default.
4095 * Add FIPS support for Android Arm 64-bit
4097 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4098 Module in Version 2.0.10. For some reason, the corresponding target
4099 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4100 built with FIPS support on Android Arm 64-bit. This omission has been
4103 *Matthias St. Pierre*
4105 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4107 * 0-byte record padding oracle
4109 If an application encounters a fatal protocol error and then calls
4110 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4111 then OpenSSL can respond differently to the calling application if a 0 byte
4112 record is received with invalid padding compared to if a 0 byte record is
4113 received with an invalid MAC. If the application then behaves differently
4114 based on that in a way that is detectable to the remote peer, then this
4115 amounts to a padding oracle that could be used to decrypt data.
4117 In order for this to be exploitable "non-stitched" ciphersuites must be in
4118 use. Stitched ciphersuites are optimised implementations of certain
4119 commonly used ciphersuites. Also the application must call SSL_shutdown()
4120 twice even if a protocol error has occurred (applications should not do
4121 this but some do anyway).
4123 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4124 Aviram, with additional investigation by Steven Collison and Andrew
4125 Hourselt. It was reported to OpenSSL on 10th December 2018.
4130 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4134 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4136 * Microarchitecture timing vulnerability in ECC scalar multiplication
4138 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4139 shown to be vulnerable to a microarchitecture timing side channel attack.
4140 An attacker with sufficient access to mount local timing attacks during
4141 ECDSA signature generation could recover the private key.
4143 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4144 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4150 * Timing vulnerability in DSA signature generation
4152 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4153 timing side channel attack. An attacker could use variations in the signing
4154 algorithm to recover the private key.
4156 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4161 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4162 Module, accidentally introduced while backporting security fixes from the
4163 development branch and hindering the use of ECC in FIPS mode.
4167 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4169 * Client DoS due to large DH parameter
4171 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4172 malicious server can send a very large prime value to the client. This will
4173 cause the client to spend an unreasonably long period of time generating a
4174 key for this prime resulting in a hang until the client has finished. This
4175 could be exploited in a Denial Of Service attack.
4177 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4182 * Cache timing vulnerability in RSA Key Generation
4184 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4185 a cache timing side channel attack. An attacker with sufficient access to
4186 mount cache timing attacks during the RSA key generation process could
4187 recover the private key.
4189 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4190 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4195 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4196 parameter is no longer accepted, as it leads to a corrupt table. NULL
4197 pem_str is reserved for alias entries only.
4201 * Revert blinding in ECDSA sign and instead make problematic addition
4202 length-invariant. Switch even to fixed-length Montgomery multiplication.
4206 * Change generating and checking of primes so that the error rate of not
4207 being prime depends on the intended use based on the size of the input.
4208 For larger primes this will result in more rounds of Miller-Rabin.
4209 The maximal error rate for primes with more than 1080 bits is lowered
4212 *Kurt Roeckx, Annie Yousar*
4214 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4218 * Add blinding to ECDSA and DSA signatures to protect against side channel
4219 attacks discovered by Keegan Ryan (NCC Group).
4223 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4224 now allow empty (zero character) pass phrases.
4228 * Certificate time validation (X509_cmp_time) enforces stricter
4229 compliance with RFC 5280. Fractional seconds and timezone offsets
4230 are no longer allowed.
4234 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4236 * Constructed ASN.1 types with a recursive definition could exceed the stack
4238 Constructed ASN.1 types with a recursive definition (such as can be found
4239 in PKCS7) could eventually exceed the stack given malicious input with
4240 excessive recursion. This could result in a Denial Of Service attack. There
4241 are no such structures used within SSL/TLS that come from untrusted sources
4242 so this is considered safe.
4244 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4250 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4252 * Read/write after SSL object in error state
4254 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4255 mechanism. The intent was that if a fatal error occurred during a handshake
4256 then OpenSSL would move into the error state and would immediately fail if
4257 you attempted to continue the handshake. This works as designed for the
4258 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4259 SSL_connect()), however due to a bug it does not work correctly if
4260 SSL_read() or SSL_write() is called directly. In that scenario, if the
4261 handshake fails then a fatal error will be returned in the initial function
4262 call. If SSL_read()/SSL_write() is subsequently called by the application
4263 for the same SSL object then it will succeed and the data is passed without
4264 being decrypted/encrypted directly from the SSL/TLS record layer.
4266 In order to exploit this issue an application bug would have to be present
4267 that resulted in a call to SSL_read()/SSL_write() being issued after having
4268 already received a fatal error.
4270 This issue was reported to OpenSSL by David Benjamin (Google).
4275 * rsaz_1024_mul_avx2 overflow bug on x86_64
4277 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4278 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4279 Analysis suggests that attacks against RSA and DSA as a result of this
4280 defect would be very difficult to perform and are not believed likely.
4281 Attacks against DH1024 are considered just feasible, because most of the
4282 work necessary to deduce information about a private key may be performed
4283 offline. The amount of resources required for such an attack would be
4284 significant. However, for an attack on TLS to be meaningful, the server
4285 would have to share the DH1024 private key among multiple clients, which is
4286 no longer an option since CVE-2016-0701.
4288 This only affects processors that support the AVX2 but not ADX extensions
4289 like Intel Haswell (4th generation).
4291 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4292 was originally found via the OSS-Fuzz project.
4297 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4299 * bn_sqrx8x_internal carry bug on x86_64
4301 There is a carry propagating bug in the x86_64 Montgomery squaring
4302 procedure. No EC algorithms are affected. Analysis suggests that attacks
4303 against RSA and DSA as a result of this defect would be very difficult to
4304 perform and are not believed likely. Attacks against DH are considered just
4305 feasible (although very difficult) because most of the work necessary to
4306 deduce information about a private key may be performed offline. The amount
4307 of resources required for such an attack would be very significant and
4308 likely only accessible to a limited number of attackers. An attacker would
4309 additionally need online access to an unpatched system using the target
4310 private key in a scenario with persistent DH parameters and a private
4311 key that is shared between multiple clients.
4313 This only affects processors that support the BMI1, BMI2 and ADX extensions
4314 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4316 This issue was reported to OpenSSL by the OSS-Fuzz project.
4321 * Malformed X.509 IPAddressFamily could cause OOB read
4323 If an X.509 certificate has a malformed IPAddressFamily extension,
4324 OpenSSL could do a one-byte buffer overread. The most likely result
4325 would be an erroneous display of the certificate in text format.
4327 This issue was reported to OpenSSL by the OSS-Fuzz project.
4332 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4334 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4335 platform rather than 'mingw'.
4339 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4341 * Truncated packet could crash via OOB read
4343 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4344 cipher is being used, then a truncated packet can cause that host to
4345 perform an out-of-bounds read, usually resulting in a crash.
4347 This issue was reported to OpenSSL by Robert Święcki of Google.
4352 * BN_mod_exp may produce incorrect results on x86_64
4354 There is a carry propagating bug in the x86_64 Montgomery squaring
4355 procedure. No EC algorithms are affected. Analysis suggests that attacks
4356 against RSA and DSA as a result of this defect would be very difficult to
4357 perform and are not believed likely. Attacks against DH are considered just
4358 feasible (although very difficult) because most of the work necessary to
4359 deduce information about a private key may be performed offline. The amount
4360 of resources required for such an attack would be very significant and
4361 likely only accessible to a limited number of attackers. An attacker would
4362 additionally need online access to an unpatched system using the target
4363 private key in a scenario with persistent DH parameters and a private
4364 key that is shared between multiple clients. For example this can occur by
4365 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4366 similar to CVE-2015-3193 but must be treated as a separate problem.
4368 This issue was reported to OpenSSL by the OSS-Fuzz project.
4373 * Montgomery multiplication may produce incorrect results
4375 There is a carry propagating bug in the Broadwell-specific Montgomery
4376 multiplication procedure that handles input lengths divisible by, but
4377 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4378 and DH private keys are impossible. This is because the subroutine in
4379 question is not used in operations with the private key itself and an input
4380 of the attacker's direct choice. Otherwise the bug can manifest itself as
4381 transient authentication and key negotiation failures or reproducible
4382 erroneous outcome of public-key operations with specially crafted input.
4383 Among EC algorithms only Brainpool P-512 curves are affected and one
4384 presumably can attack ECDH key negotiation. Impact was not analyzed in
4385 detail, because pre-requisites for attack are considered unlikely. Namely
4386 multiple clients have to choose the curve in question and the server has to
4387 share the private key among them, neither of which is default behaviour.
4388 Even then only clients that chose the curve will be affected.
4390 This issue was publicly reported as transient failures and was not
4391 initially recognized as a security issue. Thanks to Richard Morgan for
4392 providing reproducible case.
4397 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4398 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4399 prevent issues where no progress is being made and the peer continually
4400 sends unrecognised record types, using up resources processing them.
4404 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4406 * Missing CRL sanity check
4408 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4409 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4410 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4412 This issue only affects the OpenSSL 1.0.2i
4417 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4419 * OCSP Status Request extension unbounded memory growth
4421 A malicious client can send an excessively large OCSP Status Request
4422 extension. If that client continually requests renegotiation, sending a
4423 large OCSP Status Request extension each time, then there will be unbounded
4424 memory growth on the server. This will eventually lead to a Denial Of
4425 Service attack through memory exhaustion. Servers with a default
4426 configuration are vulnerable even if they do not support OCSP. Builds using
4427 the "no-ocsp" build time option are not affected.
4429 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4434 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4437 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4443 * OOB write in MDC2_Update()
4445 An overflow can occur in MDC2_Update() either if called directly or
4446 through the EVP_DigestUpdate() function using MDC2. If an attacker
4447 is able to supply very large amounts of input data after a previous
4448 call to EVP_EncryptUpdate() with a partial block then a length check
4449 can overflow resulting in a heap corruption.
4451 The amount of data needed is comparable to SIZE_MAX which is impractical
4454 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4459 * Malformed SHA512 ticket DoS
4461 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4462 DoS attack where a malformed ticket will result in an OOB read which will
4465 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4466 a custom server callback and ticket lookup mechanism.
4468 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4473 * OOB write in BN_bn2dec()
4475 The function BN_bn2dec() does not check the return value of BN_div_word().
4476 This can cause an OOB write if an application uses this function with an
4477 overly large BIGNUM. This could be a problem if an overly large certificate
4478 or CRL is printed out from an untrusted source. TLS is not affected because
4479 record limits will reject an oversized certificate before it is parsed.
4481 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4486 * OOB read in TS_OBJ_print_bio()
4488 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4489 the total length the OID text representation would use and not the amount
4490 of data written. This will result in OOB reads when large OIDs are
4493 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4498 * Pointer arithmetic undefined behaviour
4500 Avoid some undefined pointer arithmetic
4502 A common idiom in the codebase is to check limits in the following manner:
4505 Where "p" points to some malloc'd data of SIZE bytes and
4508 "len" here could be from some externally supplied data (e.g. from a TLS
4511 The rules of C pointer arithmetic are such that "p + len" is only well
4512 defined where len <= SIZE. Therefore the above idiom is actually
4513 undefined behaviour.
4515 For example this could cause problems if some malloc implementation
4516 provides an address for "p" such that "p + len" actually overflows for
4517 values of len that are too big and therefore p + len < limit.
4519 This issue was reported to OpenSSL by Guido Vranken
4524 * Constant time flag not preserved in DSA signing
4526 Operations in the DSA signing algorithm should run in constant time in
4527 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4528 implementation means that a non-constant time codepath is followed for
4529 certain operations. This has been demonstrated through a cache-timing
4530 attack to be sufficient for an attacker to recover the private DSA key.
4532 This issue was reported by César Pereida (Aalto University), Billy Brumley
4533 (Tampere University of Technology), and Yuval Yarom (The University of
4534 Adelaide and NICTA).
4539 * DTLS buffered message DoS
4541 In a DTLS connection where handshake messages are delivered out-of-order
4542 those messages that OpenSSL is not yet ready to process will be buffered
4543 for later use. Under certain circumstances, a flaw in the logic means that
4544 those messages do not get removed from the buffer even though the handshake
4545 has been completed. An attacker could force up to approx. 15 messages to
4546 remain in the buffer when they are no longer required. These messages will
4547 be cleared when the DTLS connection is closed. The default maximum size for
4548 a message is 100k. Therefore the attacker could force an additional 1500k
4549 to be consumed per connection. By opening many simulataneous connections an
4550 attacker could cause a DoS attack through memory exhaustion.
4552 This issue was reported to OpenSSL by Quan Luo.
4557 * DTLS replay protection DoS
4559 A flaw in the DTLS replay attack protection mechanism means that records
4560 that arrive for future epochs update the replay protection "window" before
4561 the MAC for the record has been validated. This could be exploited by an
4562 attacker by sending a record for the next epoch (which does not have to
4563 decrypt or have a valid MAC), with a very large sequence number. This means
4564 that all subsequent legitimate packets are dropped causing a denial of
4565 service for a specific DTLS connection.
4567 This issue was reported to OpenSSL by the OCAP audit team.
4572 * Certificate message OOB reads
4574 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4575 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4576 theoretical DoS risk but this has not been observed in practice on common
4579 The messages affected are client certificate, client certificate request
4580 and server certificate. As a result the attack can only be performed
4581 against a client or a server which enables client authentication.
4583 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4588 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4590 * Prevent padding oracle in AES-NI CBC MAC check
4592 A MITM attacker can use a padding oracle attack to decrypt traffic
4593 when the connection uses an AES CBC cipher and the server support
4596 This issue was introduced as part of the fix for Lucky 13 padding
4597 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4598 constant time by making sure that always the same bytes are read and
4599 compared against either the MAC or padding bytes. But it no longer
4600 checked that there was enough data to have both the MAC and padding
4603 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4608 * Fix EVP_EncodeUpdate overflow
4610 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4611 Base64 encoding of binary data. If an attacker is able to supply very large
4612 amounts of input data then a length check can overflow resulting in a heap
4615 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4616 the `PEM_write_bio*` family of functions. These are mainly used within the
4617 OpenSSL command line applications, so any application which processes data
4618 from an untrusted source and outputs it as a PEM file should be considered
4619 vulnerable to this issue. User applications that call these APIs directly
4620 with large amounts of untrusted data may also be vulnerable.
4622 This issue was reported by Guido Vranken.
4627 * Fix EVP_EncryptUpdate overflow
4629 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4630 is able to supply very large amounts of input data after a previous call to
4631 EVP_EncryptUpdate() with a partial block then a length check can overflow
4632 resulting in a heap corruption. Following an analysis of all OpenSSL
4633 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4634 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4635 the first called function after an EVP_EncryptInit(), and therefore that
4636 specific call must be safe. The second form is where the length passed to
4637 EVP_EncryptUpdate() can be seen from the code to be some small value and
4638 therefore there is no possibility of an overflow. Since all instances are
4639 one of these two forms, it is believed that there can be no overflows in
4640 internal code due to this problem. It should be noted that
4641 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4642 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4643 of these calls have also been analysed too and it is believed there are no
4644 instances in internal usage where an overflow could occur.
4646 This issue was reported by Guido Vranken.
4651 * Prevent ASN.1 BIO excessive memory allocation
4653 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4654 a short invalid encoding can cause allocation of large amounts of memory
4655 potentially consuming excessive resources or exhausting memory.
4657 Any application parsing untrusted data through d2i BIO functions is
4658 affected. The memory based functions such as d2i_X509() are *not* affected.
4659 Since the memory based functions are used by the TLS library, TLS
4660 applications are not affected.
4662 This issue was reported by Brian Carpenter.
4669 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4670 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4671 in arbitrary stack data being returned in the buffer.
4673 This issue was reported by Guido Vranken.
4678 * Modify behavior of ALPN to invoke callback after SNI/servername
4679 callback, such that updates to the SSL_CTX affect ALPN.
4683 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4688 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4689 methods are enabled and ssl2 is disabled the methods return NULL.
4693 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4695 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4696 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4697 provide any "EXPORT" or "LOW" strength ciphers.
4701 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4702 is by default disabled at build-time. Builds that are not configured with
4703 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4704 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4705 will need to explicitly call either of:
4707 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4709 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4711 as appropriate. Even if either of those is used, or the application
4712 explicitly uses the version-specific SSLv2_method() or its client and
4713 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4714 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4715 ciphers, and SSLv2 56-bit DES are no longer available.
4720 * Fix a double-free in DSA code
4722 A double free bug was discovered when OpenSSL parses malformed DSA private
4723 keys and could lead to a DoS attack or memory corruption for applications
4724 that receive DSA private keys from untrusted sources. This scenario is
4727 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4733 * Disable SRP fake user seed to address a server memory leak.
4735 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4737 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4738 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4739 was changed to ignore the "fake user" SRP seed, even if the seed
4742 Users should use SRP_VBASE_get1_by_user instead. Note that in
4743 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4744 also that even though configuring the SRP seed attempts to hide
4745 invalid usernames by continuing the handshake with fake
4746 credentials, this behaviour is not constant time and no strong
4747 guarantees are made that the handshake is indistinguishable from
4748 that of a valid user.
4753 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4755 In the BN_hex2bn function the number of hex digits is calculated using an
4756 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4757 large values of `i` this can result in `bn_expand` not allocating any
4758 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4759 field as NULL leading to a subsequent NULL ptr deref. For very large values
4760 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4761 In this case memory is allocated to the internal BIGNUM data field, but it
4762 is insufficiently sized leading to heap corruption. A similar issue exists
4763 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4764 is ever called by user applications with very large untrusted hex/dec data.
4765 This is anticipated to be a rare occurrence.
4767 All OpenSSL internal usage of these functions use data that is not expected
4768 to be untrusted, e.g. config file data or application command line
4769 arguments. If user developed applications generate config file data based
4770 on untrusted data then it is possible that this could also lead to security
4771 consequences. This is also anticipated to be rare.
4773 This issue was reported to OpenSSL by Guido Vranken.
4778 * Fix memory issues in `BIO_*printf` functions
4780 The internal `fmtstr` function used in processing a "%s" format string in
4781 the `BIO_*printf` functions could overflow while calculating the length of a
4782 string and cause an OOB read when printing very long strings.
4784 Additionally the internal `doapr_outch` function can attempt to write to an
4785 OOB memory location (at an offset from the NULL pointer) in the event of a
4786 memory allocation failure. In 1.0.2 and below this could be caused where
4787 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4788 could be in processing a very long "%s" format string. Memory leaks can
4791 The first issue may mask the second issue dependent on compiler behaviour.
4792 These problems could enable attacks where large amounts of untrusted data
4793 is passed to the `BIO_*printf` functions. If applications use these functions
4794 in this way then they could be vulnerable. OpenSSL itself uses these
4795 functions when printing out human-readable dumps of ASN.1 data. Therefore
4796 applications that print this data could be vulnerable if the data is from
4797 untrusted sources. OpenSSL command line applications could also be
4798 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4799 as command line arguments.
4801 Libssl is not considered directly vulnerable. Additionally certificates etc
4802 received via remote connections via libssl are also unlikely to be able to
4803 trigger these issues because of message size limits enforced within libssl.
4805 This issue was reported to OpenSSL Guido Vranken.
4810 * Side channel attack on modular exponentiation
4812 A side-channel attack was found which makes use of cache-bank conflicts on
4813 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4814 of RSA keys. The ability to exploit this issue is limited as it relies on
4815 an attacker who has control of code in a thread running on the same
4816 hyper-threaded core as the victim thread which is performing decryptions.
4818 This issue was reported to OpenSSL by Yuval Yarom, The University of
4819 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4820 Nadia Heninger, University of Pennsylvania with more information at
4821 <http://cachebleed.info>.
4826 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4827 if no keysize is specified with default_bits. This fixes an
4828 omission in an earlier change that changed all RSA/DSA key generation
4829 apps to use 2048 bits by default.
4833 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4835 * DH small subgroups
4837 Historically OpenSSL only ever generated DH parameters based on "safe"
4838 primes. More recently (in version 1.0.2) support was provided for
4839 generating X9.42 style parameter files such as those required for RFC 5114
4840 support. The primes used in such files may not be "safe". Where an
4841 application is using DH configured with parameters based on primes that are
4842 not "safe" then an attacker could use this fact to find a peer's private
4843 DH exponent. This attack requires that the attacker complete multiple
4844 handshakes in which the peer uses the same private DH exponent. For example
4845 this could be used to discover a TLS server's private DH exponent if it's
4846 reusing the private DH exponent or it's using a static DH ciphersuite.
4848 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4849 TLS. It is not on by default. If the option is not set then the server
4850 reuses the same private DH exponent for the life of the server process and
4851 would be vulnerable to this attack. It is believed that many popular
4852 applications do set this option and would therefore not be at risk.
4854 The fix for this issue adds an additional check where a "q" parameter is
4855 available (as is the case in X9.42 based parameters). This detects the
4856 only known attack, and is the only possible defense for static DH
4857 ciphersuites. This could have some performance impact.
4859 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4860 default and cannot be disabled. This could have some performance impact.
4862 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4867 * SSLv2 doesn't block disabled ciphers
4869 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4870 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4871 been disabled, provided that the SSLv2 protocol was not also disabled via
4874 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4875 and Sebastian Schinzel.
4880 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4882 * BN_mod_exp may produce incorrect results on x86_64
4884 There is a carry propagating bug in the x86_64 Montgomery squaring
4885 procedure. No EC algorithms are affected. Analysis suggests that attacks
4886 against RSA and DSA as a result of this defect would be very difficult to
4887 perform and are not believed likely. Attacks against DH are considered just
4888 feasible (although very difficult) because most of the work necessary to
4889 deduce information about a private key may be performed offline. The amount
4890 of resources required for such an attack would be very significant and
4891 likely only accessible to a limited number of attackers. An attacker would
4892 additionally need online access to an unpatched system using the target
4893 private key in a scenario with persistent DH parameters and a private
4894 key that is shared between multiple clients. For example this can occur by
4895 default in OpenSSL DHE based SSL/TLS ciphersuites.
4897 This issue was reported to OpenSSL by Hanno Böck.
4902 * Certificate verify crash with missing PSS parameter
4904 The signature verification routines will crash with a NULL pointer
4905 dereference if presented with an ASN.1 signature using the RSA PSS
4906 algorithm and absent mask generation function parameter. Since these
4907 routines are used to verify certificate signature algorithms this can be
4908 used to crash any certificate verification operation and exploited in a
4909 DoS attack. Any application which performs certificate verification is
4910 vulnerable including OpenSSL clients and servers which enable client
4913 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4918 * X509_ATTRIBUTE memory leak
4920 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4921 memory. This structure is used by the PKCS#7 and CMS routines so any
4922 application which reads PKCS#7 or CMS data from untrusted sources is
4923 affected. SSL/TLS is not affected.
4925 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4931 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4932 This changes the decoding behaviour for some invalid messages,
4933 though the change is mostly in the more lenient direction, and
4934 legacy behaviour is preserved as much as possible.
4938 * In DSA_generate_parameters_ex, if the provided seed is too short,
4941 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4943 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4945 * Alternate chains certificate forgery
4947 During certificate verification, OpenSSL will attempt to find an
4948 alternative certificate chain if the first attempt to build such a chain
4949 fails. An error in the implementation of this logic can mean that an
4950 attacker could cause certain checks on untrusted certificates to be
4951 bypassed, such as the CA flag, enabling them to use a valid leaf
4952 certificate to act as a CA and "issue" an invalid certificate.
4954 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4959 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4961 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4962 incompatibility in the handling of HMAC. The previous ABI has now been
4967 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4969 * Malformed ECParameters causes infinite loop
4971 When processing an ECParameters structure OpenSSL enters an infinite loop
4972 if the curve specified is over a specially malformed binary polynomial
4975 This can be used to perform denial of service against any
4976 system which processes public keys, certificate requests or
4977 certificates. This includes TLS clients and TLS servers with
4978 client authentication enabled.
4980 This issue was reported to OpenSSL by Joseph Barr-Pixton.
4985 * Exploitable out-of-bounds read in X509_cmp_time
4987 X509_cmp_time does not properly check the length of the ASN1_TIME
4988 string and can read a few bytes out of bounds. In addition,
4989 X509_cmp_time accepts an arbitrary number of fractional seconds in the
4992 An attacker can use this to craft malformed certificates and CRLs of
4993 various sizes and potentially cause a segmentation fault, resulting in
4994 a DoS on applications that verify certificates or CRLs. TLS clients
4995 that verify CRLs are affected. TLS clients and servers with client
4996 authentication enabled may be affected if they use custom verification
4999 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5000 independently by Hanno Böck.
5005 * PKCS7 crash with missing EnvelopedContent
5007 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5008 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5009 with missing content and trigger a NULL pointer dereference on parsing.
5011 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5012 structures from untrusted sources are affected. OpenSSL clients and
5013 servers are not affected.
5015 This issue was reported to OpenSSL by Michal Zalewski (Google).
5020 * CMS verify infinite loop with unknown hash function
5022 When verifying a signedData message the CMS code can enter an infinite loop
5023 if presented with an unknown hash function OID. This can be used to perform
5024 denial of service against any system which verifies signedData messages using
5026 This issue was reported to OpenSSL by Johannes Bauer.
5031 * Race condition handling NewSessionTicket
5033 If a NewSessionTicket is received by a multi-threaded client when attempting to
5034 reuse a previous ticket then a race condition can occur potentially leading to
5035 a double free of the ticket data.
5040 * Only support 256-bit or stronger elliptic curves with the
5041 'ecdh_auto' setting (server) or by default (client). Of supported
5042 curves, prefer P-256 (both).
5046 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5048 * ClientHello sigalgs DoS fix
5050 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5051 invalid signature algorithms extension a NULL pointer dereference will
5052 occur. This can be exploited in a DoS attack against the server.
5054 This issue was was reported to OpenSSL by David Ramos of Stanford
5058 *Stephen Henson and Matt Caswell*
5060 * Multiblock corrupted pointer fix
5062 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5063 feature only applies on 64 bit x86 architecture platforms that support AES
5064 NI instructions. A defect in the implementation of "multiblock" can cause
5065 OpenSSL's internal write buffer to become incorrectly set to NULL when
5066 using non-blocking IO. Typically, when the user application is using a
5067 socket BIO for writing, this will only result in a failed connection.
5068 However if some other BIO is used then it is likely that a segmentation
5069 fault will be triggered, thus enabling a potential DoS attack.
5071 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5076 * Segmentation fault in DTLSv1_listen fix
5078 The DTLSv1_listen function is intended to be stateless and processes the
5079 initial ClientHello from many peers. It is common for user code to loop
5080 over the call to DTLSv1_listen until a valid ClientHello is received with
5081 an associated cookie. A defect in the implementation of DTLSv1_listen means
5082 that state is preserved in the SSL object from one invocation to the next
5083 that can lead to a segmentation fault. Errors processing the initial
5084 ClientHello can trigger this scenario. An example of such an error could be
5085 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5088 This issue was reported to OpenSSL by Per Allansson.
5093 * Segmentation fault in ASN1_TYPE_cmp fix
5095 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5096 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5097 certificate signature algorithm consistency this can be used to crash any
5098 certificate verification operation and exploited in a DoS attack. Any
5099 application which performs certificate verification is vulnerable including
5100 OpenSSL clients and servers which enable client authentication.
5105 * Segmentation fault for invalid PSS parameters fix
5107 The signature verification routines will crash with a NULL pointer
5108 dereference if presented with an ASN.1 signature using the RSA PSS
5109 algorithm and invalid parameters. Since these routines are used to verify
5110 certificate signature algorithms this can be used to crash any
5111 certificate verification operation and exploited in a DoS attack. Any
5112 application which performs certificate verification is vulnerable including
5113 OpenSSL clients and servers which enable client authentication.
5115 This issue was was reported to OpenSSL by Brian Carpenter.
5120 * ASN.1 structure reuse memory corruption fix
5122 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5123 memory corruption via an invalid write. Such reuse is and has been
5124 strongly discouraged and is believed to be rare.
5126 Applications that parse structures containing CHOICE or ANY DEFINED BY
5127 components may be affected. Certificate parsing (d2i_X509 and related
5128 functions) are however not affected. OpenSSL clients and servers are
5134 * PKCS7 NULL pointer dereferences fix
5136 The PKCS#7 parsing code does not handle missing outer ContentInfo
5137 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5138 missing content and trigger a NULL pointer dereference on parsing.
5140 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5141 otherwise parse PKCS#7 structures from untrusted sources are
5142 affected. OpenSSL clients and servers are not affected.
5144 This issue was reported to OpenSSL by Michal Zalewski (Google).
5149 * DoS via reachable assert in SSLv2 servers fix
5151 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5152 servers that both support SSLv2 and enable export cipher suites by sending
5153 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5155 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5156 (OpenSSL development team).
5161 * Empty CKE with client auth and DHE fix
5163 If client auth is used then a server can seg fault in the event of a DHE
5164 ciphersuite being selected and a zero length ClientKeyExchange message
5165 being sent by the client. This could be exploited in a DoS attack.
5170 * Handshake with unseeded PRNG fix
5172 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5173 with an unseeded PRNG. The conditions are:
5174 - The client is on a platform where the PRNG has not been seeded
5175 automatically, and the user has not seeded manually
5176 - A protocol specific client method version has been used (i.e. not
5177 SSL_client_methodv23)
5178 - A ciphersuite is used that does not require additional random data from
5179 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5181 If the handshake succeeds then the client random that has been used will
5182 have been generated from a PRNG with insufficient entropy and therefore the
5183 output may be predictable.
5185 For example using the following command with an unseeded openssl will
5186 succeed on an unpatched platform:
5188 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5193 * Use After Free following d2i_ECPrivatekey error fix
5195 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5196 could cause a use after free condition. This, in turn, could cause a double
5197 free in several private key parsing functions (such as d2i_PrivateKey
5198 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5199 for applications that receive EC private keys from untrusted
5200 sources. This scenario is considered rare.
5202 This issue was discovered by the BoringSSL project and fixed in their
5208 * X509_to_X509_REQ NULL pointer deref fix
5210 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5211 the certificate key is invalid. This function is rarely used in practice.
5213 This issue was discovered by Brian Carpenter.
5218 * Removed the export ciphers from the DEFAULT ciphers
5222 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5224 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5225 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5226 So far those who have to target multiple platforms would compromise
5227 and argue that binary targeting say ARMv5 would still execute on
5228 ARMv8. "Universal" build resolves this compromise by providing
5229 near-optimal performance even on newer platforms.
5233 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5234 (other platforms pending).
5236 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5238 * Add support for the SignedCertificateTimestampList certificate and
5239 OCSP response extensions from RFC6962.
5243 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5244 for corner cases. (Certain input points at infinity could lead to
5245 bogus results, with non-infinity inputs mapped to infinity too.)
5249 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5250 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5251 common cases are optimized and there still is room for further
5252 improvements. Vector Permutation AES for Altivec is also added.
5256 * Add support for little-endian ppc64 Linux target.
5258 *Marcelo Cerri (IBM)*
5260 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5261 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5262 are optimized and there still is room for further improvements.
5263 Both 32- and 64-bit modes are supported.
5265 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5267 * Improved ARMv7 NEON support.
5271 * Support for SPARC Architecture 2011 crypto extensions, first
5272 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5273 SHA256/512, MD5, GHASH and modular exponentiation.
5275 *Andy Polyakov, David Miller*
5277 * Accelerated modular exponentiation for Intel processors, a.k.a.
5280 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5282 * Support for new and upcoming Intel processors, including AVX2,
5283 BMI and SHA ISA extensions. This includes additional "stitched"
5284 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5287 This work was sponsored by Intel Corp.
5291 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5292 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5293 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5297 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5298 this fixes a limitation in previous versions of OpenSSL.
5302 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5303 MGF1 digest and OAEP label.
5307 * Add EVP support for key wrapping algorithms, to avoid problems with
5308 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5309 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5310 algorithms and include tests cases.
5314 * Add functions to allocate and set the fields of an ECDSA_METHOD
5317 *Douglas E. Engert, Steve Henson*
5319 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5320 difference in days and seconds between two tm or ASN1_TIME structures.
5324 * Add -rev test option to s_server to just reverse order of characters
5325 received by client and send back to server. Also prints an abbreviated
5326 summary of the connection parameters.
5330 * New option -brief for s_client and s_server to print out a brief summary
5331 of connection parameters.
5335 * Add callbacks for arbitrary TLS extensions.
5337 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5339 * New option -crl_download in several openssl utilities to download CRLs
5340 from CRLDP extension in certificates.
5344 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5348 * New function X509_CRL_diff to generate a delta CRL from the difference
5349 of two full CRLs. Add support to "crl" utility.
5353 * New functions to set lookup_crls function and to retrieve
5354 X509_STORE from X509_STORE_CTX.
5358 * Print out deprecated issuer and subject unique ID fields in
5363 * Extend OCSP I/O functions so they can be used for simple general purpose
5364 HTTP as well as OCSP. New wrapper function which can be used to download
5365 CRLs using the OCSP API.
5369 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5373 * `SSL_CONF*` functions. These provide a common framework for application
5374 configuration using configuration files or command lines.
5378 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5379 message callback and prints the results. Needs compile time option
5380 "enable-ssl-trace". New options to s_client and s_server to enable
5385 * New ctrl and macro to retrieve supported points extensions.
5386 Print out extension in s_server and s_client.
5390 * New functions to retrieve certificate signature and signature
5395 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5400 * New Suite B modes for TLS code. These use and enforce the requirements
5401 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5402 only use Suite B curves. The Suite B modes can be set by using the
5403 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5407 * New chain verification flags for Suite B levels of security. Check
5408 algorithms are acceptable when flags are set in X509_verify_cert.
5412 * Make tls1_check_chain return a set of flags indicating checks passed
5413 by a certificate chain. Add additional tests to handle client
5414 certificates: checks for matching certificate type and issuer name
5419 * If an attempt is made to use a signature algorithm not in the peer
5420 preference list abort the handshake. If client has no suitable
5421 signature algorithms in response to a certificate request do not
5422 use the certificate.
5426 * If server EC tmp key is not in client preference list abort handshake.
5430 * Add support for certificate stores in CERT structure. This makes it
5431 possible to have different stores per SSL structure or one store in
5432 the parent SSL_CTX. Include distinct stores for certificate chain
5433 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5434 to build and store a certificate chain in CERT structure: returning
5435 an error if the chain cannot be built: this will allow applications
5436 to test if a chain is correctly configured.
5438 Note: if the CERT based stores are not set then the parent SSL_CTX
5439 store is used to retain compatibility with existing behaviour.
5443 * New function ssl_set_client_disabled to set a ciphersuite disabled
5444 mask based on the current session, check mask when sending client
5445 hello and checking the requested ciphersuite.
5449 * New ctrls to retrieve and set certificate types in a certificate
5450 request message. Print out received values in s_client. If certificate
5451 types is not set with custom values set sensible values based on
5452 supported signature algorithms.
5456 * Support for distinct client and server supported signature algorithms.
5460 * Add certificate callback. If set this is called whenever a certificate
5461 is required by client or server. An application can decide which
5462 certificate chain to present based on arbitrary criteria: for example
5463 supported signature algorithms. Add very simple example to s_server.
5464 This fixes many of the problems and restrictions of the existing client
5465 certificate callback: for example you can now clear an existing
5466 certificate and specify the whole chain.
5470 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5471 the certificate can be used for (if anything). Set valid_flags field
5472 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5473 to have similar checks in it.
5475 Add new "cert_flags" field to CERT structure and include a "strict mode".
5476 This enforces some TLS certificate requirements (such as only permitting
5477 certificate signature algorithms contained in the supported algorithms
5478 extension) which some implementations ignore: this option should be used
5479 with caution as it could cause interoperability issues.
5483 * Update and tidy signature algorithm extension processing. Work out
5484 shared signature algorithms based on preferences and peer algorithms
5485 and print them out in s_client and s_server. Abort handshake if no
5486 shared signature algorithms.
5490 * Add new functions to allow customised supported signature algorithms
5491 for SSL and SSL_CTX structures. Add options to s_client and s_server
5496 * New function SSL_certs_clear() to delete all references to certificates
5497 from an SSL structure. Before this once a certificate had been added
5498 it couldn't be removed.
5502 * Integrate hostname, email address and IP address checking with certificate
5503 verification. New verify options supporting checking in openssl utility.
5507 * Fixes and wildcard matching support to hostname and email checking
5508 functions. Add manual page.
5510 *Florian Weimer (Red Hat Product Security Team)*
5512 * New functions to check a hostname email or IP address against a
5513 certificate. Add options x509 utility to print results of checks against
5518 * Fix OCSP checking.
5520 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5522 * Initial experimental support for explicitly trusted non-root CAs.
5523 OpenSSL still tries to build a complete chain to a root but if an
5524 intermediate CA has a trust setting included that is used. The first
5525 setting is used: whether to trust (e.g., -addtrust option to the x509
5530 * Add -trusted_first option which attempts to find certificates in the
5531 trusted store even if an untrusted chain is also supplied.
5535 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5536 platform support for Linux and Android.
5540 * Support for linux-x32, ILP32 environment in x86_64 framework.
5544 * Experimental multi-implementation support for FIPS capable OpenSSL.
5545 When in FIPS mode the approved implementations are used as normal,
5546 when not in FIPS mode the internal unapproved versions are used instead.
5547 This means that the FIPS capable OpenSSL isn't forced to use the
5548 (often lower performance) FIPS implementations outside FIPS mode.
5552 * Transparently support X9.42 DH parameters when calling
5553 PEM_read_bio_DHparameters. This means existing applications can handle
5554 the new parameter format automatically.
5558 * Initial experimental support for X9.42 DH parameter format: mainly
5559 to support use of 'q' parameter for RFC5114 parameters.
5563 * Add DH parameters from RFC5114 including test data to dhtest.
5567 * Support for automatic EC temporary key parameter selection. If enabled
5568 the most preferred EC parameters are automatically used instead of
5569 hardcoded fixed parameters. Now a server just has to call:
5570 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5571 support ECDH and use the most appropriate parameters.
5575 * Enhance and tidy EC curve and point format TLS extension code. Use
5576 static structures instead of allocation if default values are used.
5577 New ctrls to set curves we wish to support and to retrieve shared curves.
5578 Print out shared curves in s_server. New options to s_server and s_client
5579 to set list of supported curves.
5583 * New ctrls to retrieve supported signature algorithms and
5584 supported curve values as an array of NIDs. Extend openssl utility
5585 to print out received values.
5589 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5590 between NIDs and the more common NIST names such as "P-256". Enhance
5591 ecparam utility and ECC method to recognise the NIST names for curves.
5595 * Enhance SSL/TLS certificate chain handling to support different
5596 chains for each certificate instead of one chain in the parent SSL_CTX.
5600 * Support for fixed DH ciphersuite client authentication: where both
5601 server and client use DH certificates with common parameters.
5605 * Support for fixed DH ciphersuites: those requiring DH server
5610 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5612 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5613 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5614 X509_CINF_get_signature were reverted post internal team review.
5619 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5621 * OCSP Status Request extension unbounded memory growth
5623 A malicious client can send an excessively large OCSP Status Request
5624 extension. If that client continually requests renegotiation, sending a
5625 large OCSP Status Request extension each time, then there will be unbounded
5626 memory growth on the server. This will eventually lead to a Denial Of
5627 Service attack through memory exhaustion. Servers with a default
5628 configuration are vulnerable even if they do not support OCSP. Builds using
5629 the "no-ocsp" build time option are not affected.
5631 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5636 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5639 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5645 * OOB write in MDC2_Update()
5647 An overflow can occur in MDC2_Update() either if called directly or
5648 through the EVP_DigestUpdate() function using MDC2. If an attacker
5649 is able to supply very large amounts of input data after a previous
5650 call to EVP_EncryptUpdate() with a partial block then a length check
5651 can overflow resulting in a heap corruption.
5653 The amount of data needed is comparable to SIZE_MAX which is impractical
5656 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5661 * Malformed SHA512 ticket DoS
5663 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5664 DoS attack where a malformed ticket will result in an OOB read which will
5667 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5668 a custom server callback and ticket lookup mechanism.
5670 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5675 * OOB write in BN_bn2dec()
5677 The function BN_bn2dec() does not check the return value of BN_div_word().
5678 This can cause an OOB write if an application uses this function with an
5679 overly large BIGNUM. This could be a problem if an overly large certificate
5680 or CRL is printed out from an untrusted source. TLS is not affected because
5681 record limits will reject an oversized certificate before it is parsed.
5683 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5688 * OOB read in TS_OBJ_print_bio()
5690 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5691 the total length the OID text representation would use and not the amount
5692 of data written. This will result in OOB reads when large OIDs are
5695 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5700 * Pointer arithmetic undefined behaviour
5702 Avoid some undefined pointer arithmetic
5704 A common idiom in the codebase is to check limits in the following manner:
5707 Where "p" points to some malloc'd data of SIZE bytes and
5710 "len" here could be from some externally supplied data (e.g. from a TLS
5713 The rules of C pointer arithmetic are such that "p + len" is only well
5714 defined where len <= SIZE. Therefore the above idiom is actually
5715 undefined behaviour.
5717 For example this could cause problems if some malloc implementation
5718 provides an address for "p" such that "p + len" actually overflows for
5719 values of len that are too big and therefore p + len < limit.
5721 This issue was reported to OpenSSL by Guido Vranken
5726 * Constant time flag not preserved in DSA signing
5728 Operations in the DSA signing algorithm should run in constant time in
5729 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5730 implementation means that a non-constant time codepath is followed for
5731 certain operations. This has been demonstrated through a cache-timing
5732 attack to be sufficient for an attacker to recover the private DSA key.
5734 This issue was reported by César Pereida (Aalto University), Billy Brumley
5735 (Tampere University of Technology), and Yuval Yarom (The University of
5736 Adelaide and NICTA).
5741 * DTLS buffered message DoS
5743 In a DTLS connection where handshake messages are delivered out-of-order
5744 those messages that OpenSSL is not yet ready to process will be buffered
5745 for later use. Under certain circumstances, a flaw in the logic means that
5746 those messages do not get removed from the buffer even though the handshake
5747 has been completed. An attacker could force up to approx. 15 messages to
5748 remain in the buffer when they are no longer required. These messages will
5749 be cleared when the DTLS connection is closed. The default maximum size for
5750 a message is 100k. Therefore the attacker could force an additional 1500k
5751 to be consumed per connection. By opening many simulataneous connections an
5752 attacker could cause a DoS attack through memory exhaustion.
5754 This issue was reported to OpenSSL by Quan Luo.
5759 * DTLS replay protection DoS
5761 A flaw in the DTLS replay attack protection mechanism means that records
5762 that arrive for future epochs update the replay protection "window" before
5763 the MAC for the record has been validated. This could be exploited by an
5764 attacker by sending a record for the next epoch (which does not have to
5765 decrypt or have a valid MAC), with a very large sequence number. This means
5766 that all subsequent legitimate packets are dropped causing a denial of
5767 service for a specific DTLS connection.
5769 This issue was reported to OpenSSL by the OCAP audit team.
5774 * Certificate message OOB reads
5776 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5777 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5778 theoretical DoS risk but this has not been observed in practice on common
5781 The messages affected are client certificate, client certificate request
5782 and server certificate. As a result the attack can only be performed
5783 against a client or a server which enables client authentication.
5785 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5790 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5792 * Prevent padding oracle in AES-NI CBC MAC check
5794 A MITM attacker can use a padding oracle attack to decrypt traffic
5795 when the connection uses an AES CBC cipher and the server support
5798 This issue was introduced as part of the fix for Lucky 13 padding
5799 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5800 constant time by making sure that always the same bytes are read and
5801 compared against either the MAC or padding bytes. But it no longer
5802 checked that there was enough data to have both the MAC and padding
5805 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5810 * Fix EVP_EncodeUpdate overflow
5812 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5813 Base64 encoding of binary data. If an attacker is able to supply very large
5814 amounts of input data then a length check can overflow resulting in a heap
5817 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5818 the `PEM_write_bio*` family of functions. These are mainly used within the
5819 OpenSSL command line applications, so any application which processes data
5820 from an untrusted source and outputs it as a PEM file should be considered
5821 vulnerable to this issue. User applications that call these APIs directly
5822 with large amounts of untrusted data may also be vulnerable.
5824 This issue was reported by Guido Vranken.
5829 * Fix EVP_EncryptUpdate overflow
5831 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5832 is able to supply very large amounts of input data after a previous call to
5833 EVP_EncryptUpdate() with a partial block then a length check can overflow
5834 resulting in a heap corruption. Following an analysis of all OpenSSL
5835 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5836 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5837 the first called function after an EVP_EncryptInit(), and therefore that
5838 specific call must be safe. The second form is where the length passed to
5839 EVP_EncryptUpdate() can be seen from the code to be some small value and
5840 therefore there is no possibility of an overflow. Since all instances are
5841 one of these two forms, it is believed that there can be no overflows in
5842 internal code due to this problem. It should be noted that
5843 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5844 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5845 of these calls have also been analysed too and it is believed there are no
5846 instances in internal usage where an overflow could occur.
5848 This issue was reported by Guido Vranken.
5853 * Prevent ASN.1 BIO excessive memory allocation
5855 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5856 a short invalid encoding can casuse allocation of large amounts of memory
5857 potentially consuming excessive resources or exhausting memory.
5859 Any application parsing untrusted data through d2i BIO functions is
5860 affected. The memory based functions such as d2i_X509() are *not* affected.
5861 Since the memory based functions are used by the TLS library, TLS
5862 applications are not affected.
5864 This issue was reported by Brian Carpenter.
5871 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5872 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5873 in arbitrary stack data being returned in the buffer.
5875 This issue was reported by Guido Vranken.
5880 * Modify behavior of ALPN to invoke callback after SNI/servername
5881 callback, such that updates to the SSL_CTX affect ALPN.
5885 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5890 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5891 methods are enabled and ssl2 is disabled the methods return NULL.
5895 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5897 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5898 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5899 provide any "EXPORT" or "LOW" strength ciphers.
5903 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5904 is by default disabled at build-time. Builds that are not configured with
5905 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5906 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5907 will need to explicitly call either of:
5909 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5911 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5913 as appropriate. Even if either of those is used, or the application
5914 explicitly uses the version-specific SSLv2_method() or its client and
5915 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5916 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5917 ciphers, and SSLv2 56-bit DES are no longer available.
5922 * Fix a double-free in DSA code
5924 A double free bug was discovered when OpenSSL parses malformed DSA private
5925 keys and could lead to a DoS attack or memory corruption for applications
5926 that receive DSA private keys from untrusted sources. This scenario is
5929 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5935 * Disable SRP fake user seed to address a server memory leak.
5937 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5939 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5940 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5941 was changed to ignore the "fake user" SRP seed, even if the seed
5944 Users should use SRP_VBASE_get1_by_user instead. Note that in
5945 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5946 also that even though configuring the SRP seed attempts to hide
5947 invalid usernames by continuing the handshake with fake
5948 credentials, this behaviour is not constant time and no strong
5949 guarantees are made that the handshake is indistinguishable from
5950 that of a valid user.
5955 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5957 In the BN_hex2bn function the number of hex digits is calculated using an
5958 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5959 large values of `i` this can result in `bn_expand` not allocating any
5960 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5961 field as NULL leading to a subsequent NULL ptr deref. For very large values
5962 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5963 In this case memory is allocated to the internal BIGNUM data field, but it
5964 is insufficiently sized leading to heap corruption. A similar issue exists
5965 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5966 is ever called by user applications with very large untrusted hex/dec data.
5967 This is anticipated to be a rare occurrence.
5969 All OpenSSL internal usage of these functions use data that is not expected
5970 to be untrusted, e.g. config file data or application command line
5971 arguments. If user developed applications generate config file data based
5972 on untrusted data then it is possible that this could also lead to security
5973 consequences. This is also anticipated to be rare.
5975 This issue was reported to OpenSSL by Guido Vranken.
5980 * Fix memory issues in `BIO_*printf` functions
5982 The internal `fmtstr` function used in processing a "%s" format string in
5983 the `BIO_*printf` functions could overflow while calculating the length of a
5984 string and cause an OOB read when printing very long strings.
5986 Additionally the internal `doapr_outch` function can attempt to write to an
5987 OOB memory location (at an offset from the NULL pointer) in the event of a
5988 memory allocation failure. In 1.0.2 and below this could be caused where
5989 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
5990 could be in processing a very long "%s" format string. Memory leaks can
5993 The first issue may mask the second issue dependent on compiler behaviour.
5994 These problems could enable attacks where large amounts of untrusted data
5995 is passed to the `BIO_*printf` functions. If applications use these functions
5996 in this way then they could be vulnerable. OpenSSL itself uses these
5997 functions when printing out human-readable dumps of ASN.1 data. Therefore
5998 applications that print this data could be vulnerable if the data is from
5999 untrusted sources. OpenSSL command line applications could also be
6000 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6001 as command line arguments.
6003 Libssl is not considered directly vulnerable. Additionally certificates etc
6004 received via remote connections via libssl are also unlikely to be able to
6005 trigger these issues because of message size limits enforced within libssl.
6007 This issue was reported to OpenSSL Guido Vranken.
6012 * Side channel attack on modular exponentiation
6014 A side-channel attack was found which makes use of cache-bank conflicts on
6015 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6016 of RSA keys. The ability to exploit this issue is limited as it relies on
6017 an attacker who has control of code in a thread running on the same
6018 hyper-threaded core as the victim thread which is performing decryptions.
6020 This issue was reported to OpenSSL by Yuval Yarom, The University of
6021 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6022 Nadia Heninger, University of Pennsylvania with more information at
6023 <http://cachebleed.info>.
6028 * Change the req app to generate a 2048-bit RSA/DSA key by default,
6029 if no keysize is specified with default_bits. This fixes an
6030 omission in an earlier change that changed all RSA/DSA key generation
6031 apps to use 2048 bits by default.
6035 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6037 * Protection for DH small subgroup attacks
6039 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6040 switched on by default and cannot be disabled. This could have some
6045 * SSLv2 doesn't block disabled ciphers
6047 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6048 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6049 been disabled, provided that the SSLv2 protocol was not also disabled via
6052 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6053 and Sebastian Schinzel.
6058 * Reject DH handshakes with parameters shorter than 1024 bits.
6062 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6064 * Certificate verify crash with missing PSS parameter
6066 The signature verification routines will crash with a NULL pointer
6067 dereference if presented with an ASN.1 signature using the RSA PSS
6068 algorithm and absent mask generation function parameter. Since these
6069 routines are used to verify certificate signature algorithms this can be
6070 used to crash any certificate verification operation and exploited in a
6071 DoS attack. Any application which performs certificate verification is
6072 vulnerable including OpenSSL clients and servers which enable client
6075 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6080 * X509_ATTRIBUTE memory leak
6082 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6083 memory. This structure is used by the PKCS#7 and CMS routines so any
6084 application which reads PKCS#7 or CMS data from untrusted sources is
6085 affected. SSL/TLS is not affected.
6087 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6093 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6094 This changes the decoding behaviour for some invalid messages,
6095 though the change is mostly in the more lenient direction, and
6096 legacy behaviour is preserved as much as possible.
6100 * In DSA_generate_parameters_ex, if the provided seed is too short,
6101 use a random seed, as already documented.
6103 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6105 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6107 * Alternate chains certificate forgery
6109 During certificate verfification, OpenSSL will attempt to find an
6110 alternative certificate chain if the first attempt to build such a chain
6111 fails. An error in the implementation of this logic can mean that an
6112 attacker could cause certain checks on untrusted certificates to be
6113 bypassed, such as the CA flag, enabling them to use a valid leaf
6114 certificate to act as a CA and "issue" an invalid certificate.
6116 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6122 * Race condition handling PSK identify hint
6124 If PSK identity hints are received by a multi-threaded client then
6125 the values are wrongly updated in the parent SSL_CTX structure. This can
6126 result in a race condition potentially leading to a double free of the
6132 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6134 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6135 incompatibility in the handling of HMAC. The previous ABI has now been
6138 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6140 * Malformed ECParameters causes infinite loop
6142 When processing an ECParameters structure OpenSSL enters an infinite loop
6143 if the curve specified is over a specially malformed binary polynomial
6146 This can be used to perform denial of service against any
6147 system which processes public keys, certificate requests or
6148 certificates. This includes TLS clients and TLS servers with
6149 client authentication enabled.
6151 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6156 * Exploitable out-of-bounds read in X509_cmp_time
6158 X509_cmp_time does not properly check the length of the ASN1_TIME
6159 string and can read a few bytes out of bounds. In addition,
6160 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6163 An attacker can use this to craft malformed certificates and CRLs of
6164 various sizes and potentially cause a segmentation fault, resulting in
6165 a DoS on applications that verify certificates or CRLs. TLS clients
6166 that verify CRLs are affected. TLS clients and servers with client
6167 authentication enabled may be affected if they use custom verification
6170 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6171 independently by Hanno Böck.
6176 * PKCS7 crash with missing EnvelopedContent
6178 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6179 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6180 with missing content and trigger a NULL pointer dereference on parsing.
6182 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6183 structures from untrusted sources are affected. OpenSSL clients and
6184 servers are not affected.
6186 This issue was reported to OpenSSL by Michal Zalewski (Google).
6191 * CMS verify infinite loop with unknown hash function
6193 When verifying a signedData message the CMS code can enter an infinite loop
6194 if presented with an unknown hash function OID. This can be used to perform
6195 denial of service against any system which verifies signedData messages using
6197 This issue was reported to OpenSSL by Johannes Bauer.
6202 * Race condition handling NewSessionTicket
6204 If a NewSessionTicket is received by a multi-threaded client when attempting to
6205 reuse a previous ticket then a race condition can occur potentially leading to
6206 a double free of the ticket data.
6211 * Reject DH handshakes with parameters shorter than 768 bits.
6213 *Kurt Roeckx and Emilia Kasper*
6215 * dhparam: generate 2048-bit parameters by default.
6217 *Kurt Roeckx and Emilia Kasper*
6219 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6221 * Segmentation fault in ASN1_TYPE_cmp fix
6223 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6224 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6225 certificate signature algorithm consistency this can be used to crash any
6226 certificate verification operation and exploited in a DoS attack. Any
6227 application which performs certificate verification is vulnerable including
6228 OpenSSL clients and servers which enable client authentication.
6233 * ASN.1 structure reuse memory corruption fix
6235 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6236 memory corruption via an invalid write. Such reuse is and has been
6237 strongly discouraged and is believed to be rare.
6239 Applications that parse structures containing CHOICE or ANY DEFINED BY
6240 components may be affected. Certificate parsing (d2i_X509 and related
6241 functions) are however not affected. OpenSSL clients and servers are
6247 * PKCS7 NULL pointer dereferences fix
6249 The PKCS#7 parsing code does not handle missing outer ContentInfo
6250 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6251 missing content and trigger a NULL pointer dereference on parsing.
6253 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6254 otherwise parse PKCS#7 structures from untrusted sources are
6255 affected. OpenSSL clients and servers are not affected.
6257 This issue was reported to OpenSSL by Michal Zalewski (Google).
6262 * DoS via reachable assert in SSLv2 servers fix
6264 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6265 servers that both support SSLv2 and enable export cipher suites by sending
6266 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6268 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6269 (OpenSSL development team).
6274 * Use After Free following d2i_ECPrivatekey error fix
6276 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6277 could cause a use after free condition. This, in turn, could cause a double
6278 free in several private key parsing functions (such as d2i_PrivateKey
6279 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6280 for applications that receive EC private keys from untrusted
6281 sources. This scenario is considered rare.
6283 This issue was discovered by the BoringSSL project and fixed in their
6289 * X509_to_X509_REQ NULL pointer deref fix
6291 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6292 the certificate key is invalid. This function is rarely used in practice.
6294 This issue was discovered by Brian Carpenter.
6299 * Removed the export ciphers from the DEFAULT ciphers
6303 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6305 * Build fixes for the Windows and OpenVMS platforms
6307 *Matt Caswell and Richard Levitte*
6309 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6311 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6312 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6313 dereference. This could lead to a Denial Of Service attack. Thanks to
6314 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6319 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6320 dtls1_buffer_record function under certain conditions. In particular this
6321 could occur if an attacker sent repeated DTLS records with the same
6322 sequence number but for the next epoch. The memory leak could be exploited
6323 by an attacker in a Denial of Service attack through memory exhaustion.
6324 Thanks to Chris Mueller for reporting this issue.
6329 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6330 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6331 method would be set to NULL which could later result in a NULL pointer
6332 dereference. Thanks to Frank Schmirler for reporting this issue.
6337 * Abort handshake if server key exchange message is omitted for ephemeral
6340 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6341 reporting this issue.
6346 * Remove non-export ephemeral RSA code on client and server. This code
6347 violated the TLS standard by allowing the use of temporary RSA keys in
6348 non-export ciphersuites and could be used by a server to effectively
6349 downgrade the RSA key length used to a value smaller than the server
6350 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6351 INRIA or reporting this issue.
6356 * Fixed issue where DH client certificates are accepted without verification.
6357 An OpenSSL server will accept a DH certificate for client authentication
6358 without the certificate verify message. This effectively allows a client to
6359 authenticate without the use of a private key. This only affects servers
6360 which trust a client certificate authority which issues certificates
6361 containing DH keys: these are extremely rare and hardly ever encountered.
6362 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6368 * Ensure that the session ID context of an SSL is updated when its
6369 SSL_CTX is updated via SSL_set_SSL_CTX.
6371 The session ID context is typically set from the parent SSL_CTX,
6372 and can vary with the CTX.
6376 * Fix various certificate fingerprint issues.
6378 By using non-DER or invalid encodings outside the signed portion of a
6379 certificate the fingerprint can be changed without breaking the signature.
6380 Although no details of the signed portion of the certificate can be changed
6381 this can cause problems with some applications: e.g. those using the
6382 certificate fingerprint for blacklists.
6384 1. Reject signatures with non zero unused bits.
6386 If the BIT STRING containing the signature has non zero unused bits reject
6387 the signature. All current signature algorithms require zero unused bits.
6389 2. Check certificate algorithm consistency.
6391 Check the AlgorithmIdentifier inside TBS matches the one in the
6392 certificate signature. NB: this will result in signature failure
6393 errors for some broken certificates.
6395 Thanks to Konrad Kraszewski from Google for reporting this issue.
6397 3. Check DSA/ECDSA signatures use DER.
6399 Re-encode DSA/ECDSA signatures and compare with the original received
6400 signature. Return an error if there is a mismatch.
6402 This will reject various cases including garbage after signature
6403 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6404 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6405 (negative or with leading zeroes).
6407 Further analysis was conducted and fixes were developed by Stephen Henson
6408 of the OpenSSL core team.
6414 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6415 results on some platforms, including x86_64. This bug occurs at random
6416 with a very low probability, and is not known to be exploitable in any
6417 way, though its exact impact is difficult to determine. Thanks to Pieter
6418 Wuille (Blockstream) who reported this issue and also suggested an initial
6419 fix. Further analysis was conducted by the OpenSSL development team and
6420 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6421 the OpenSSL core team.
6426 * Do not resume sessions on the server if the negotiated protocol
6427 version does not match the session's version. Resuming with a different
6428 version, while not strictly forbidden by the RFC, is of questionable
6429 sanity and breaks all known clients.
6431 *David Benjamin, Emilia Käsper*
6433 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6434 early CCS messages during renegotiation. (Note that because
6435 renegotiation is encrypted, this early CCS was not exploitable.)
6439 * Tighten client-side session ticket handling during renegotiation:
6440 ensure that the client only accepts a session ticket if the server sends
6441 the extension anew in the ServerHello. Previously, a TLS client would
6442 reuse the old extension state and thus accept a session ticket if one was
6443 announced in the initial ServerHello.
6445 Similarly, ensure that the client requires a session ticket if one
6446 was advertised in the ServerHello. Previously, a TLS client would
6447 ignore a missing NewSessionTicket message.
6451 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6455 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6456 sends a carefully crafted handshake message, to cause OpenSSL to fail
6457 to free up to 64k of memory causing a memory leak. This could be
6458 exploited in a Denial Of Service attack. This issue affects OpenSSL
6459 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6460 whether SRTP is used or configured. Implementations of OpenSSL that
6461 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6463 The fix was developed by the OpenSSL team.
6468 * Session Ticket Memory Leak.
6470 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6471 integrity of that ticket is first verified. In the event of a session
6472 ticket integrity check failing, OpenSSL will fail to free memory
6473 causing a memory leak. By sending a large number of invalid session
6474 tickets an attacker could exploit this issue in a Denial Of Service
6480 * Build option no-ssl3 is incomplete.
6482 When OpenSSL is configured with "no-ssl3" as a build option, servers
6483 could accept and complete a SSL 3.0 handshake, and clients could be
6484 configured to send them.
6487 *Akamai and the OpenSSL team*
6489 * Add support for TLS_FALLBACK_SCSV.
6490 Client applications doing fallback retries should call
6491 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6494 *Adam Langley, Bodo Moeller*
6496 * Add additional DigestInfo checks.
6498 Re-encode DigestInto in DER and check against the original when
6499 verifying RSA signature: this will reject any improperly encoded
6500 DigestInfo structures.
6502 Note: this is a precautionary measure and no attacks are currently known.
6506 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6508 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6509 SRP code can be overrun an internal buffer. Add sanity check that
6510 g, A, B < N to SRP code.
6512 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6513 Group for discovering this issue.
6518 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6519 TLS 1.0 instead of higher protocol versions when the ClientHello message
6520 is badly fragmented. This allows a man-in-the-middle attacker to force a
6521 downgrade to TLS 1.0 even if both the server and the client support a
6522 higher protocol version, by modifying the client's TLS records.
6524 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6525 researching this issue.
6530 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6531 to a denial of service attack. A malicious server can crash the client
6532 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6533 ciphersuite and sending carefully crafted handshake messages.
6535 Thanks to Felix Gröbert (Google) for discovering and researching this
6541 * By sending carefully crafted DTLS packets an attacker could cause openssl
6542 to leak memory. This can be exploited through a Denial of Service attack.
6543 Thanks to Adam Langley for discovering and researching this issue.
6548 * An attacker can force openssl to consume large amounts of memory whilst
6549 processing DTLS handshake messages. This can be exploited through a
6550 Denial of Service attack.
6551 Thanks to Adam Langley for discovering and researching this issue.
6556 * An attacker can force an error condition which causes openssl to crash
6557 whilst processing DTLS packets due to memory being freed twice. This
6558 can be exploited through a Denial of Service attack.
6559 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6565 * If a multithreaded client connects to a malicious server using a resumed
6566 session and the server sends an ec point format extension it could write
6567 up to 255 bytes to freed memory.
6569 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6575 * A malicious server can crash an OpenSSL client with a null pointer
6576 dereference (read) by specifying an SRP ciphersuite even though it was not
6577 properly negotiated with the client. This can be exploited through a
6578 Denial of Service attack.
6580 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6581 discovering and researching this issue.
6586 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6587 X509_name_oneline, X509_name_print_ex et al. to leak some information
6588 from the stack. Applications may be affected if they echo pretty printing
6589 output to the attacker.
6591 Thanks to Ivan Fratric (Google) for discovering this issue.
6594 *Emilia Käsper, and Steve Henson*
6596 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6597 for corner cases. (Certain input points at infinity could lead to
6598 bogus results, with non-infinity inputs mapped to infinity too.)
6602 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6604 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6605 handshake can force the use of weak keying material in OpenSSL
6606 SSL/TLS clients and servers.
6608 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6609 researching this issue. ([CVE-2014-0224])
6611 *KIKUCHI Masashi, Steve Henson*
6613 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6614 OpenSSL DTLS client the code can be made to recurse eventually crashing
6617 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6620 *Imre Rad, Steve Henson*
6622 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6623 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6624 client or server. This is potentially exploitable to run arbitrary
6625 code on a vulnerable client or server.
6627 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6629 *Jüri Aedla, Steve Henson*
6631 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6632 are subject to a denial of service attack.
6634 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6635 this issue. ([CVE-2014-3470])
6637 *Felix Gröbert, Ivan Fratric, Steve Henson*
6639 * Harmonize version and its documentation. -f flag is used to display
6642 *mancha <mancha1@zoho.com>*
6644 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6645 in i2d_ECPrivateKey.
6647 *mancha <mancha1@zoho.com>*
6649 * Fix some double frees. These are not thought to be exploitable.
6651 *mancha <mancha1@zoho.com>*
6653 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6655 * A missing bounds check in the handling of the TLS heartbeat extension
6656 can be used to reveal up to 64k of memory to a connected client or
6659 Thanks for Neel Mehta of Google Security for discovering this bug and to
6660 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6661 preparing the fix ([CVE-2014-0160])
6663 *Adam Langley, Bodo Moeller*
6665 * Fix for the attack described in the paper "Recovering OpenSSL
6666 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6667 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6668 <http://eprint.iacr.org/2014/140>
6670 Thanks to Yuval Yarom and Naomi Benger for discovering this
6671 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6673 *Yuval Yarom and Naomi Benger*
6675 * TLS pad extension: draft-agl-tls-padding-03
6677 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6678 TLS client Hello record length value would otherwise be > 255 and
6679 less that 512 pad with a dummy extension containing zeroes so it
6680 is at least 512 bytes long.
6682 *Adam Langley, Steve Henson*
6684 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6686 * Fix for TLS record tampering bug. A carefully crafted invalid
6687 handshake could crash OpenSSL with a NULL pointer exception.
6688 Thanks to Anton Johansson for reporting this issues.
6691 * Keep original DTLS digest and encryption contexts in retransmission
6692 structures so we can use the previous session parameters if they need
6693 to be resent. ([CVE-2013-6450])
6697 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6698 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6699 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6700 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6701 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6702 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6704 *Rob Stradling, Adam Langley*
6706 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6708 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6709 supporting platforms or when small records were transferred.
6711 *Andy Polyakov, Steve Henson*
6713 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6715 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6717 This addresses the flaw in CBC record processing discovered by
6718 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6719 at: <http://www.isg.rhul.ac.uk/tls/>
6721 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6722 Security Group at Royal Holloway, University of London
6723 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6724 Emilia Käsper for the initial patch.
6727 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6729 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6730 ciphersuites which can be exploited in a denial of service attack.
6731 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6732 and detecting this bug and to Wolfgang Ettlinger
6733 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6738 * Return an error when checking OCSP signatures when key is NULL.
6739 This fixes a DoS attack. ([CVE-2013-0166])
6743 * Make openssl verify return errors.
6745 *Chris Palmer <palmer@google.com> and Ben Laurie*
6747 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6748 the right response is stapled. Also change SSL_get_certificate()
6749 so it returns the certificate actually sent.
6750 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6752 *Rob Stradling <rob.stradling@comodo.com>*
6754 * Fix possible deadlock when decoding public keys.
6758 * Don't use TLS 1.0 record version number in initial client hello
6763 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6765 * Sanity check record length before skipping explicit IV in TLS
6766 1.2, 1.1 and DTLS to fix DoS attack.
6768 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6769 fuzzing as a service testing platform.
6774 * Initialise tkeylen properly when encrypting CMS messages.
6775 Thanks to Solar Designer of Openwall for reporting this issue.
6779 * In FIPS mode don't try to use composite ciphers as they are not
6784 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6786 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6787 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6788 mean any application compiled against OpenSSL 1.0.0 headers setting
6789 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6790 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6791 0x10000000L Any application which was previously compiled against
6792 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6793 will need to be recompiled as a result. Letting be results in
6794 inability to disable specifically TLS 1.1 and in client context,
6795 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6799 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6800 disable just protocol X, but all protocols above X *if* there are
6801 protocols *below* X still enabled. In more practical terms it means
6802 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6803 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6804 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6809 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6811 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6812 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6813 in CRYPTO_realloc_clean.
6815 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6816 issue and to Adam Langley <agl@chromium.org> for fixing it.
6819 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6821 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6825 * Workarounds for some broken servers that "hang" if a client hello
6826 record length exceeds 255 bytes.
6828 1. Do not use record version number > TLS 1.0 in initial client
6829 hello: some (but not all) hanging servers will now work.
6830 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6831 the number of ciphers sent in the client hello. This should be
6832 set to an even number, such as 50, for example by passing:
6833 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6834 Most broken servers should now work.
6835 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6836 TLS 1.2 client support entirely.
6840 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6844 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6846 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6847 STRING form instead of a DigestInfo.
6851 * The format used for MDC2 RSA signatures is inconsistent between EVP
6852 and the RSA_sign/RSA_verify functions. This was made more apparent when
6853 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6854 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6855 the correct format in RSA_verify so both forms transparently work.
6859 * Some servers which support TLS 1.0 can choke if we initially indicate
6860 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6861 encrypted premaster secret. As a workaround use the maximum permitted
6862 client version in client hello, this should keep such servers happy
6863 and still work with previous versions of OpenSSL.
6867 * Add support for TLS/DTLS heartbeats.
6869 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6871 * Add support for SCTP.
6873 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6875 * Improved PRNG seeding for VOS.
6877 *Paul Green <Paul.Green@stratus.com>*
6879 * Extensive assembler packs updates, most notably:
6881 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6882 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6883 - x86_64: bit-sliced AES implementation;
6884 - ARM: NEON support, contemporary platforms optimizations;
6885 - s390x: z196 support;
6886 - `*`: GHASH and GF(2^m) multiplication implementations;
6890 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6891 (removal of unnecessary code)
6893 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6895 * Add TLS key material exporter from RFC 5705.
6899 * Add DTLS-SRTP negotiation from RFC 5764.
6903 * Add Next Protocol Negotiation,
6904 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6905 disabled with a no-npn flag to config or Configure. Code donated
6908 *Adam Langley <agl@google.com> and Ben Laurie*
6910 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6911 NIST-P256, NIST-P521, with constant-time single point multiplication on
6912 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6913 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6914 Code made available under Apache License version 2.0.
6916 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6917 line to include this in your build of OpenSSL, and run "make depend" (or
6918 "make update"). This enables the following EC_METHODs:
6920 EC_GFp_nistp224_method()
6921 EC_GFp_nistp256_method()
6922 EC_GFp_nistp521_method()
6924 EC_GROUP_new_by_curve_name() will automatically use these (while
6925 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6928 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6930 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6931 all platforms. Move ssize_t definition from e_os.h to the public
6932 header file e_os2.h as it now appears in public header file cms.h
6936 * New -sigopt option to the ca, req and x509 utilities. Additional
6937 signature parameters can be passed using this option and in
6942 * Add RSA PSS signing function. This will generate and set the
6943 appropriate AlgorithmIdentifiers for PSS based on those in the
6944 corresponding EVP_MD_CTX structure. No application support yet.
6948 * Support for companion algorithm specific ASN1 signing routines.
6949 New function ASN1_item_sign_ctx() signs a pre-initialised
6950 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6951 the appropriate parameters.
6955 * Add new algorithm specific ASN1 verification initialisation function
6956 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6957 handling will be the same no matter what EVP_PKEY_METHOD is used.
6958 Add a PSS handler to support verification of PSS signatures: checked
6959 against a number of sample certificates.
6963 * Add signature printing for PSS. Add PSS OIDs.
6965 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6967 * Add algorithm specific signature printing. An individual ASN1 method
6968 can now print out signatures instead of the standard hex dump.
6970 More complex signatures (e.g. PSS) can print out more meaningful
6971 information. Include DSA version that prints out the signature
6976 * Password based recipient info support for CMS library: implementing
6981 * Split password based encryption into PBES2 and PBKDF2 functions. This
6982 neatly separates the code into cipher and PBE sections and is required
6983 for some algorithms that split PBES2 into separate pieces (such as
6984 password based CMS).
6988 * Session-handling fixes:
6989 - Fix handling of connections that are resuming with a session ID,
6990 but also support Session Tickets.
6991 - Fix a bug that suppressed issuing of a new ticket if the client
6992 presented a ticket with an expired session.
6993 - Try to set the ticket lifetime hint to something reasonable.
6994 - Make tickets shorter by excluding irrelevant information.
6995 - On the client side, don't ignore renewed tickets.
6997 *Adam Langley, Bodo Moeller (Google)*
6999 * Fix PSK session representation.
7003 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7005 This work was sponsored by Intel.
7009 * Add GCM support to TLS library. Some custom code is needed to split
7010 the IV between the fixed (from PRF) and explicit (from TLS record)
7011 portions. This adds all GCM ciphersuites supported by RFC5288 and
7012 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7013 add a special AESGCM string for GCM only.
7017 * Expand range of ctrls for AES GCM. Permit setting invocation
7018 field on decrypt and retrieval of invocation field only on encrypt.
7022 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7023 As required by RFC5289 these ciphersuites cannot be used if for
7024 versions of TLS earlier than 1.2.
7028 * For FIPS capable OpenSSL interpret a NULL default public key method
7029 as unset and return the appropriate default but do *not* set the default.
7030 This means we can return the appropriate method in applications that
7031 switch between FIPS and non-FIPS modes.
7035 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7036 ENGINE is used then we cannot handle that in the FIPS module so we
7037 keep original code iff non-FIPS operations are allowed.
7041 * Add -attime option to openssl utilities.
7043 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7045 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7049 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7050 FIPS EC methods unconditionally for now.
7054 * New build option no-ec2m to disable characteristic 2 code.
7058 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7059 all cases can be covered as some introduce binary incompatibilities.
7063 * Redirect RSA operations to FIPS module including keygen,
7064 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7068 * Add similar low level API blocking to ciphers.
7072 * Low level digest APIs are not approved in FIPS mode: any attempt
7073 to use these will cause a fatal error. Applications that *really* want
7074 to use them can use the `private_*` version instead.
7078 * Redirect cipher operations to FIPS module for FIPS builds.
7082 * Redirect digest operations to FIPS module for FIPS builds.
7086 * Update build system to add "fips" flag which will link in fipscanister.o
7087 for static and shared library builds embedding a signature if needed.
7091 * Output TLS supported curves in preference order instead of numerical
7092 order. This is currently hardcoded for the highest order curves first.
7093 This should be configurable so applications can judge speed vs strength.
7097 * Add TLS v1.2 server support for client authentication.
7101 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7106 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7107 FIPS modules versions.
7111 * Add TLS v1.2 client side support for client authentication. Keep cache
7112 of handshake records longer as we don't know the hash algorithm to use
7113 until after the certificate request message is received.
7117 * Initial TLS v1.2 client support. Add a default signature algorithms
7118 extension including all the algorithms we support. Parse new signature
7119 format in client key exchange. Relax some ECC signing restrictions for
7120 TLS v1.2 as indicated in RFC5246.
7124 * Add server support for TLS v1.2 signature algorithms extension. Switch
7125 to new signature format when needed using client digest preference.
7126 All server ciphersuites should now work correctly in TLS v1.2. No client
7127 support yet and no support for client certificates.
7131 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7132 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7133 ciphersuites. At present only RSA key exchange ciphersuites work with
7134 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7135 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7136 and version checking.
7140 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7141 with this defined it will not be affected by any changes to ssl internal
7142 structures. Add several utility functions to allow openssl application
7143 to work with OPENSSL_NO_SSL_INTERN defined.
7147 * A long standing patch to add support for SRP from EdelWeb (Peter
7148 Sylvester and Christophe Renou) was integrated.
7149 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7150 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7153 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7157 * Permit abbreviated handshakes when renegotiating using the function
7158 SSL_renegotiate_abbreviated().
7160 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7162 * Add call to ENGINE_register_all_complete() to
7163 ENGINE_load_builtin_engines(), so some implementations get used
7164 automatically instead of needing explicit application support.
7168 * Add support for TLS key exporter as described in RFC5705.
7170 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7172 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7173 a few changes are required:
7175 Add SSL_OP_NO_TLSv1_1 flag.
7176 Add TLSv1_1 methods.
7177 Update version checking logic to handle version 1.1.
7178 Add explicit IV handling (ported from DTLS code).
7179 Add command line options to s_client/s_server.
7186 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7188 * X509_ATTRIBUTE memory leak
7190 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7191 memory. This structure is used by the PKCS#7 and CMS routines so any
7192 application which reads PKCS#7 or CMS data from untrusted sources is
7193 affected. SSL/TLS is not affected.
7195 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7201 * Race condition handling PSK identify hint
7203 If PSK identity hints are received by a multi-threaded client then
7204 the values are wrongly updated in the parent SSL_CTX structure. This can
7205 result in a race condition potentially leading to a double free of the
7211 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7213 * Malformed ECParameters causes infinite loop
7215 When processing an ECParameters structure OpenSSL enters an infinite loop
7216 if the curve specified is over a specially malformed binary polynomial
7219 This can be used to perform denial of service against any
7220 system which processes public keys, certificate requests or
7221 certificates. This includes TLS clients and TLS servers with
7222 client authentication enabled.
7224 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7229 * Exploitable out-of-bounds read in X509_cmp_time
7231 X509_cmp_time does not properly check the length of the ASN1_TIME
7232 string and can read a few bytes out of bounds. In addition,
7233 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7236 An attacker can use this to craft malformed certificates and CRLs of
7237 various sizes and potentially cause a segmentation fault, resulting in
7238 a DoS on applications that verify certificates or CRLs. TLS clients
7239 that verify CRLs are affected. TLS clients and servers with client
7240 authentication enabled may be affected if they use custom verification
7243 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7244 independently by Hanno Böck.
7249 * PKCS7 crash with missing EnvelopedContent
7251 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7252 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7253 with missing content and trigger a NULL pointer dereference on parsing.
7255 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7256 structures from untrusted sources are affected. OpenSSL clients and
7257 servers are not affected.
7259 This issue was reported to OpenSSL by Michal Zalewski (Google).
7264 * CMS verify infinite loop with unknown hash function
7266 When verifying a signedData message the CMS code can enter an infinite loop
7267 if presented with an unknown hash function OID. This can be used to perform
7268 denial of service against any system which verifies signedData messages using
7270 This issue was reported to OpenSSL by Johannes Bauer.
7275 * Race condition handling NewSessionTicket
7277 If a NewSessionTicket is received by a multi-threaded client when attempting to
7278 reuse a previous ticket then a race condition can occur potentially leading to
7279 a double free of the ticket data.
7284 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7286 * Segmentation fault in ASN1_TYPE_cmp fix
7288 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7289 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7290 certificate signature algorithm consistency this can be used to crash any
7291 certificate verification operation and exploited in a DoS attack. Any
7292 application which performs certificate verification is vulnerable including
7293 OpenSSL clients and servers which enable client authentication.
7298 * ASN.1 structure reuse memory corruption fix
7300 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7301 memory corruption via an invalid write. Such reuse is and has been
7302 strongly discouraged and is believed to be rare.
7304 Applications that parse structures containing CHOICE or ANY DEFINED BY
7305 components may be affected. Certificate parsing (d2i_X509 and related
7306 functions) are however not affected. OpenSSL clients and servers are
7312 * PKCS7 NULL pointer dereferences fix
7314 The PKCS#7 parsing code does not handle missing outer ContentInfo
7315 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7316 missing content and trigger a NULL pointer dereference on parsing.
7318 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7319 otherwise parse PKCS#7 structures from untrusted sources are
7320 affected. OpenSSL clients and servers are not affected.
7322 This issue was reported to OpenSSL by Michal Zalewski (Google).
7327 * DoS via reachable assert in SSLv2 servers fix
7329 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7330 servers that both support SSLv2 and enable export cipher suites by sending
7331 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7333 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7334 (OpenSSL development team).
7339 * Use After Free following d2i_ECPrivatekey error fix
7341 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7342 could cause a use after free condition. This, in turn, could cause a double
7343 free in several private key parsing functions (such as d2i_PrivateKey
7344 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7345 for applications that receive EC private keys from untrusted
7346 sources. This scenario is considered rare.
7348 This issue was discovered by the BoringSSL project and fixed in their
7354 * X509_to_X509_REQ NULL pointer deref fix
7356 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7357 the certificate key is invalid. This function is rarely used in practice.
7359 This issue was discovered by Brian Carpenter.
7364 * Removed the export ciphers from the DEFAULT ciphers
7368 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7370 * Build fixes for the Windows and OpenVMS platforms
7372 *Matt Caswell and Richard Levitte*
7374 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7376 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7377 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7378 dereference. This could lead to a Denial Of Service attack. Thanks to
7379 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7384 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7385 dtls1_buffer_record function under certain conditions. In particular this
7386 could occur if an attacker sent repeated DTLS records with the same
7387 sequence number but for the next epoch. The memory leak could be exploited
7388 by an attacker in a Denial of Service attack through memory exhaustion.
7389 Thanks to Chris Mueller for reporting this issue.
7394 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7395 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7396 method would be set to NULL which could later result in a NULL pointer
7397 dereference. Thanks to Frank Schmirler for reporting this issue.
7402 * Abort handshake if server key exchange message is omitted for ephemeral
7405 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7406 reporting this issue.
7411 * Remove non-export ephemeral RSA code on client and server. This code
7412 violated the TLS standard by allowing the use of temporary RSA keys in
7413 non-export ciphersuites and could be used by a server to effectively
7414 downgrade the RSA key length used to a value smaller than the server
7415 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7416 INRIA or reporting this issue.
7421 * Fixed issue where DH client certificates are accepted without verification.
7422 An OpenSSL server will accept a DH certificate for client authentication
7423 without the certificate verify message. This effectively allows a client to
7424 authenticate without the use of a private key. This only affects servers
7425 which trust a client certificate authority which issues certificates
7426 containing DH keys: these are extremely rare and hardly ever encountered.
7427 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7433 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7434 results on some platforms, including x86_64. This bug occurs at random
7435 with a very low probability, and is not known to be exploitable in any
7436 way, though its exact impact is difficult to determine. Thanks to Pieter
7437 Wuille (Blockstream) who reported this issue and also suggested an initial
7438 fix. Further analysis was conducted by the OpenSSL development team and
7439 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7440 the OpenSSL core team.
7445 * Fix various certificate fingerprint issues.
7447 By using non-DER or invalid encodings outside the signed portion of a
7448 certificate the fingerprint can be changed without breaking the signature.
7449 Although no details of the signed portion of the certificate can be changed
7450 this can cause problems with some applications: e.g. those using the
7451 certificate fingerprint for blacklists.
7453 1. Reject signatures with non zero unused bits.
7455 If the BIT STRING containing the signature has non zero unused bits reject
7456 the signature. All current signature algorithms require zero unused bits.
7458 2. Check certificate algorithm consistency.
7460 Check the AlgorithmIdentifier inside TBS matches the one in the
7461 certificate signature. NB: this will result in signature failure
7462 errors for some broken certificates.
7464 Thanks to Konrad Kraszewski from Google for reporting this issue.
7466 3. Check DSA/ECDSA signatures use DER.
7468 Reencode DSA/ECDSA signatures and compare with the original received
7469 signature. Return an error if there is a mismatch.
7471 This will reject various cases including garbage after signature
7472 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7473 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7474 (negative or with leading zeroes).
7476 Further analysis was conducted and fixes were developed by Stephen Henson
7477 of the OpenSSL core team.
7483 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7485 * Session Ticket Memory Leak.
7487 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7488 integrity of that ticket is first verified. In the event of a session
7489 ticket integrity check failing, OpenSSL will fail to free memory
7490 causing a memory leak. By sending a large number of invalid session
7491 tickets an attacker could exploit this issue in a Denial Of Service
7497 * Build option no-ssl3 is incomplete.
7499 When OpenSSL is configured with "no-ssl3" as a build option, servers
7500 could accept and complete a SSL 3.0 handshake, and clients could be
7501 configured to send them.
7504 *Akamai and the OpenSSL team*
7506 * Add support for TLS_FALLBACK_SCSV.
7507 Client applications doing fallback retries should call
7508 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7511 *Adam Langley, Bodo Moeller*
7513 * Add additional DigestInfo checks.
7515 Reencode DigestInto in DER and check against the original when
7516 verifying RSA signature: this will reject any improperly encoded
7517 DigestInfo structures.
7519 Note: this is a precautionary measure and no attacks are currently known.
7523 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7525 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7526 to a denial of service attack. A malicious server can crash the client
7527 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7528 ciphersuite and sending carefully crafted handshake messages.
7530 Thanks to Felix Gröbert (Google) for discovering and researching this
7536 * By sending carefully crafted DTLS packets an attacker could cause openssl
7537 to leak memory. This can be exploited through a Denial of Service attack.
7538 Thanks to Adam Langley for discovering and researching this issue.
7543 * An attacker can force openssl to consume large amounts of memory whilst
7544 processing DTLS handshake messages. This can be exploited through a
7545 Denial of Service attack.
7546 Thanks to Adam Langley for discovering and researching this issue.
7551 * An attacker can force an error condition which causes openssl to crash
7552 whilst processing DTLS packets due to memory being freed twice. This
7553 can be exploited through a Denial of Service attack.
7554 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7560 * If a multithreaded client connects to a malicious server using a resumed
7561 session and the server sends an ec point format extension it could write
7562 up to 255 bytes to freed memory.
7564 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7570 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7571 X509_name_oneline, X509_name_print_ex et al. to leak some information
7572 from the stack. Applications may be affected if they echo pretty printing
7573 output to the attacker.
7575 Thanks to Ivan Fratric (Google) for discovering this issue.
7578 *Emilia Käsper, and Steve Henson*
7580 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7581 for corner cases. (Certain input points at infinity could lead to
7582 bogus results, with non-infinity inputs mapped to infinity too.)
7586 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7588 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7589 handshake can force the use of weak keying material in OpenSSL
7590 SSL/TLS clients and servers.
7592 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7593 researching this issue. ([CVE-2014-0224])
7595 *KIKUCHI Masashi, Steve Henson*
7597 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7598 OpenSSL DTLS client the code can be made to recurse eventually crashing
7601 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7604 *Imre Rad, Steve Henson*
7606 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7607 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7608 client or server. This is potentially exploitable to run arbitrary
7609 code on a vulnerable client or server.
7611 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7613 *Jüri Aedla, Steve Henson*
7615 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7616 are subject to a denial of service attack.
7618 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7619 this issue. ([CVE-2014-3470])
7621 *Felix Gröbert, Ivan Fratric, Steve Henson*
7623 * Harmonize version and its documentation. -f flag is used to display
7626 *mancha <mancha1@zoho.com>*
7628 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7629 in i2d_ECPrivateKey.
7631 *mancha <mancha1@zoho.com>*
7633 * Fix some double frees. These are not thought to be exploitable.
7635 *mancha <mancha1@zoho.com>*
7637 * Fix for the attack described in the paper "Recovering OpenSSL
7638 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7639 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7640 <http://eprint.iacr.org/2014/140>
7642 Thanks to Yuval Yarom and Naomi Benger for discovering this
7643 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7645 *Yuval Yarom and Naomi Benger*
7647 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7649 * Keep original DTLS digest and encryption contexts in retransmission
7650 structures so we can use the previous session parameters if they need
7651 to be resent. ([CVE-2013-6450])
7655 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7656 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7657 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7658 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7659 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7660 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7662 *Rob Stradling, Adam Langley*
7664 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7666 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7668 This addresses the flaw in CBC record processing discovered by
7669 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7670 at: <http://www.isg.rhul.ac.uk/tls/>
7672 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7673 Security Group at Royal Holloway, University of London
7674 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7675 Emilia Käsper for the initial patch.
7678 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7680 * Return an error when checking OCSP signatures when key is NULL.
7681 This fixes a DoS attack. ([CVE-2013-0166])
7685 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7686 the right response is stapled. Also change SSL_get_certificate()
7687 so it returns the certificate actually sent.
7688 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7689 (This is a backport)
7691 *Rob Stradling <rob.stradling@comodo.com>*
7693 * Fix possible deadlock when decoding public keys.
7697 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7699 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7702 * Sanity check record length before skipping explicit IV in DTLS
7705 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7706 fuzzing as a service testing platform.
7711 * Initialise tkeylen properly when encrypting CMS messages.
7712 Thanks to Solar Designer of Openwall for reporting this issue.
7716 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7718 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7719 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7720 in CRYPTO_realloc_clean.
7722 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7723 issue and to Adam Langley <agl@chromium.org> for fixing it.
7726 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7728 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7730 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7731 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7732 content decryption and always return the same error. Note: this attack
7733 needs on average 2^20 messages so it only affects automated senders. The
7734 old behaviour can be re-enabled in the CMS code by setting the
7735 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7736 an MMA defence is not necessary.
7737 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7738 this issue. ([CVE-2012-0884])
7742 * Fix CVE-2011-4619: make sure we really are receiving a
7743 client hello before rejecting multiple SGC restarts. Thanks to
7744 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7748 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7750 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7751 Thanks to Antonio Martin, Enterprise Secure Access Research and
7752 Development, Cisco Systems, Inc. for discovering this bug and
7753 preparing a fix. ([CVE-2012-0050])
7757 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7759 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7760 of the Vaudenay padding oracle attack on CBC mode encryption
7761 which enables an efficient plaintext recovery attack against
7762 the OpenSSL implementation of DTLS. Their attack exploits timing
7763 differences arising during decryption processing. A research
7764 paper describing this attack can be found at:
7765 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7766 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7767 Security Group at Royal Holloway, University of London
7768 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7769 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7770 for preparing the fix. ([CVE-2011-4108])
7772 *Robin Seggelmann, Michael Tuexen*
7774 * Clear bytes used for block padding of SSL 3.0 records.
7777 *Adam Langley (Google)*
7779 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7780 Kadianakis <desnacked@gmail.com> for discovering this issue and
7781 Adam Langley for preparing the fix. ([CVE-2011-4619])
7783 *Adam Langley (Google)*
7785 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7787 *Andrey Kulikov <amdeich@gmail.com>*
7789 * Prevent malformed RFC3779 data triggering an assertion failure.
7790 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7791 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7793 *Rob Austein <sra@hactrn.net>*
7795 * Improved PRNG seeding for VOS.
7797 *Paul Green <Paul.Green@stratus.com>*
7799 * Fix ssl_ciph.c set-up race.
7801 *Adam Langley (Google)*
7803 * Fix spurious failures in ecdsatest.c.
7805 *Emilia Käsper (Google)*
7807 * Fix the BIO_f_buffer() implementation (which was mixing different
7808 interpretations of the `..._len` fields).
7810 *Adam Langley (Google)*
7812 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7813 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7814 threads won't reuse the same blinding coefficients.
7816 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7817 lock to call BN_BLINDING_invert_ex, and avoids one use of
7818 BN_BLINDING_update for each BN_BLINDING structure (previously,
7819 the last update always remained unused).
7821 *Emilia Käsper (Google)*
7823 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7825 *Bob Buckholz (Google)*
7827 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7829 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7830 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7832 *Kaspar Brand <ossl@velox.ch>*
7834 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7835 for multi-threaded use of ECDH. ([CVE-2011-3210])
7837 *Adam Langley (Google)*
7839 * Fix x509_name_ex_d2i memory leak on bad inputs.
7843 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7844 signature public key algorithm by using OID xref utilities instead.
7845 Before this you could only use some ECC ciphersuites with SHA1 only.
7849 * Add protection against ECDSA timing attacks as mentioned in the paper
7850 by Billy Bob Brumley and Nicola Tuveri, see:
7851 <http://eprint.iacr.org/2011/232.pdf>
7853 *Billy Bob Brumley and Nicola Tuveri*
7855 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7857 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7859 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7861 * Fix bug in string printing code: if *any* escaping is enabled we must
7862 escape the escape character (backslash) or the resulting string is
7867 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7869 * Disable code workaround for ancient and obsolete Netscape browsers
7870 and servers: an attacker can use it in a ciphersuite downgrade attack.
7871 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7875 * Fixed J-PAKE implementation error, originally discovered by
7876 Sebastien Martini, further info and confirmation from Stefan
7877 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7881 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7883 * Fix extension code to avoid race conditions which can result in a buffer
7884 overrun vulnerability: resumed sessions must not be modified as they can
7885 be shared by multiple threads. CVE-2010-3864
7889 * Fix WIN32 build system to correctly link an ENGINE directory into
7894 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7896 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7899 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7901 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7903 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7904 context. The operation can be customised via the ctrl mechanism in
7905 case ENGINEs want to include additional functionality.
7909 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7913 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7914 output hashes compatible with older versions of OpenSSL.
7916 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7918 * Fix compression algorithm handling: if resuming a session use the
7919 compression algorithm of the resumed session instead of determining
7920 it from client hello again. Don't allow server to change algorithm.
7924 * Add load_crls() function to apps tidying load_certs() too. Add option
7925 to verify utility to allow additional CRLs to be included.
7929 * Update OCSP request code to permit adding custom headers to the request:
7930 some responders need this.
7934 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7937 *Julia Lawall <julia@diku.dk>*
7939 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7940 needlessly dereferenced structures, used obsolete functions and
7941 didn't handle all updated verify codes correctly.
7945 * Disable MD2 in the default configuration.
7949 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7950 indicate the initial BIO being pushed or popped. This makes it possible
7951 to determine whether the BIO is the one explicitly called or as a result
7952 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7953 it handles reference counts correctly and doesn't zero out the I/O bio
7954 when it is not being explicitly popped. WARNING: applications which
7955 included workarounds for the old buggy behaviour will need to be modified
7956 or they could free up already freed BIOs.
7960 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7961 renaming to all platforms (within the 0.9.8 branch, this was
7962 done conditionally on Netware platforms to avoid a name clash).
7964 *Guenter <lists@gknw.net>*
7966 * Add ECDHE and PSK support to DTLS.
7968 *Michael Tuexen <tuexen@fh-muenster.de>*
7970 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
7975 * Add "missing" function EVP_MD_flags() (without this the only way to
7976 retrieve a digest flags is by accessing the structure directly. Update
7977 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
7978 or cipher is registered as in the "from" argument. Print out all
7979 registered digests in the dgst usage message instead of manually
7980 attempting to work them out.
7984 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
7985 this allows the use of compression and extensions. Change default cipher
7986 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
7987 by default unless an application cipher string requests it.
7991 * Alter match criteria in PKCS12_parse(). It used to try to use local
7992 key ids to find matching certificates and keys but some PKCS#12 files
7993 don't follow the (somewhat unwritten) rules and this strategy fails.
7994 Now just gather all certificates together and the first private key
7995 then look for the first certificate that matches the key.
7999 * Support use of registered digest and cipher names for dgst and cipher
8000 commands instead of having to add each one as a special case. So now
8007 openssl dgst -sha256 foo
8009 and this works for ENGINE based algorithms too.
8013 * Update Gost ENGINE to support parameter files.
8015 *Victor B. Wagner <vitus@cryptocom.ru>*
8017 * Support GeneralizedTime in ca utility.
8019 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8021 * Enhance the hash format used for certificate directory links. The new
8022 form uses the canonical encoding (meaning equivalent names will work
8023 even if they aren't identical) and uses SHA1 instead of MD5. This form
8024 is incompatible with the older format and as a result c_rehash should
8025 be used to rebuild symbolic links.
8029 * Make PKCS#8 the default write format for private keys, replacing the
8030 traditional format. This form is standardised, more secure and doesn't
8031 include an implicit MD5 dependency.
8035 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8036 committed to OpenSSL should pass this lot as a minimum.
8040 * Add session ticket override functionality for use by EAP-FAST.
8042 *Jouni Malinen <j@w1.fi>*
8044 * Modify HMAC functions to return a value. Since these can be implemented
8045 in an ENGINE errors can occur.
8049 * Type-checked OBJ_bsearch_ex.
8053 * Type-checked OBJ_bsearch. Also some constification necessitated
8054 by type-checking. Still to come: TXT_DB, bsearch(?),
8055 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8060 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8061 seconds to a tm structure directly, instead of going through OS
8062 specific date routines. This avoids any issues with OS routines such
8063 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8064 and X509_time_adj_ex() to cover the extended range. The existing
8065 X509_time_adj() is still usable and will no longer have any date issues.
8069 * Delta CRL support. New use deltas option which will attempt to locate
8070 and search any appropriate delta CRLs available.
8072 This work was sponsored by Google.
8076 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8077 code and add additional score elements. Validate alternate CRL paths
8078 as part of the CRL checking and indicate a new error "CRL path validation
8079 error" in this case. Applications wanting additional details can use
8080 the verify callback and check the new "parent" field. If this is not
8081 NULL CRL path validation is taking place. Existing applications won't
8082 see this because it requires extended CRL support which is off by
8085 This work was sponsored by Google.
8089 * Support for freshest CRL extension.
8091 This work was sponsored by Google.
8095 * Initial indirect CRL support. Currently only supported in the CRLs
8096 passed directly and not via lookup. Process certificate issuer
8097 CRL entry extension and lookup CRL entries by bother issuer name
8098 and serial number. Check and process CRL issuer entry in IDP extension.
8100 This work was sponsored by Google.
8104 * Add support for distinct certificate and CRL paths. The CRL issuer
8105 certificate is validated separately in this case. Only enabled if
8106 an extended CRL support flag is set: this flag will enable additional
8107 CRL functionality in future.
8109 This work was sponsored by Google.
8113 * Add support for policy mappings extension.
8115 This work was sponsored by Google.
8119 * Fixes to pathlength constraint, self issued certificate handling,
8120 policy processing to align with RFC3280 and PKITS tests.
8122 This work was sponsored by Google.
8126 * Support for name constraints certificate extension. DN, email, DNS
8127 and URI types are currently supported.
8129 This work was sponsored by Google.
8133 * To cater for systems that provide a pointer-based thread ID rather
8134 than numeric, deprecate the current numeric thread ID mechanism and
8135 replace it with a structure and associated callback type. This
8136 mechanism allows a numeric "hash" to be extracted from a thread ID in
8137 either case, and on platforms where pointers are larger than 'long',
8138 mixing is done to help ensure the numeric 'hash' is usable even if it
8139 can't be guaranteed unique. The default mechanism is to use "&errno"
8140 as a pointer-based thread ID to distinguish between threads.
8142 Applications that want to provide their own thread IDs should now use
8143 CRYPTO_THREADID_set_callback() to register a callback that will call
8144 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8146 Note that ERR_remove_state() is now deprecated, because it is tied
8147 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8148 to free the current thread's error state should be replaced by
8149 ERR_remove_thread_state(NULL).
8151 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8152 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8153 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8154 application was previously providing a numeric thread callback that
8155 was inappropriate for distinguishing threads, then uniqueness might
8156 have been obtained with &errno that happened immediately in the
8157 intermediate development versions of OpenSSL; this is no longer the
8158 case, the numeric thread callback will now override the automatic use
8161 *Geoff Thorpe, with help from Bodo Moeller*
8163 * Initial support for different CRL issuing certificates. This covers a
8164 simple case where the self issued certificates in the chain exist and
8165 the real CRL issuer is higher in the existing chain.
8167 This work was sponsored by Google.
8171 * Removed effectively defunct crypto/store from the build.
8175 * Revamp of STACK to provide stronger type-checking. Still to come:
8176 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8177 ASN1_STRING, CONF_VALUE.
8181 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8182 RAM on SSL connections. This option can save about 34k per idle SSL.
8186 * Revamp of LHASH to provide stronger type-checking. Still to come:
8187 STACK, TXT_DB, bsearch, qsort.
8191 * Initial support for Cryptographic Message Syntax (aka CMS) based
8192 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8193 support for data, signedData, compressedData, digestedData and
8194 encryptedData, envelopedData types included. Scripts to check against
8195 RFC4134 examples draft and interop and consistency checks of many
8196 content types and variants.
8200 * Add options to enc utility to support use of zlib compression BIO.
8204 * Extend mk1mf to support importing of options and assembly language
8205 files from Configure script, currently only included in VC-WIN32.
8206 The assembly language rules can now optionally generate the source
8207 files from the associated perl scripts.
8211 * Implement remaining functionality needed to support GOST ciphersuites.
8212 Interop testing has been performed using CryptoPro implementations.
8214 *Victor B. Wagner <vitus@cryptocom.ru>*
8216 * s390x assembler pack.
8220 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8225 * Implement Opaque PRF Input TLS extension as specified in
8226 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8227 official specification yet and no extension type assignment by
8228 IANA exists, this extension (for now) will have to be explicitly
8229 enabled when building OpenSSL by providing the extension number
8230 to use. For example, specify an option
8232 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8234 to the "config" or "Configure" script to enable the extension,
8235 assuming extension number 0x9527 (which is a completely arbitrary
8236 and unofficial assignment based on the MD5 hash of the Internet
8237 Draft). Note that by doing so, you potentially lose
8238 interoperability with other TLS implementations since these might
8239 be using the same extension number for other purposes.
8241 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8242 opaque PRF input value to use in the handshake. This will create
8243 an internal copy of the length-'len' string at 'src', and will
8244 return non-zero for success.
8246 To get more control and flexibility, provide a callback function
8249 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8250 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8254 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8257 Callback function 'cb' will be called in handshakes, and is
8258 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8259 Argument 'arg' is for application purposes (the value as given to
8260 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8261 be provided to the callback function). The callback function
8262 has to return non-zero to report success: usually 1 to use opaque
8263 PRF input just if possible, or 2 to enforce use of the opaque PRF
8264 input. In the latter case, the library will abort the handshake
8265 if opaque PRF input is not successfully negotiated.
8267 Arguments 'peerinput' and 'len' given to the callback function
8268 will always be NULL and 0 in the case of a client. A server will
8269 see the client's opaque PRF input through these variables if
8270 available (NULL and 0 otherwise). Note that if the server
8271 provides an opaque PRF input, the length must be the same as the
8272 length of the client's opaque PRF input.
8274 Note that the callback function will only be called when creating
8275 a new session (session resumption can resume whatever was
8276 previously negotiated), and will not be called in SSL 2.0
8277 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8278 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8279 for applications that need to enforce opaque PRF input.
8283 * Update ssl code to support digests other than SHA1+MD5 for handshake
8286 *Victor B. Wagner <vitus@cryptocom.ru>*
8288 * Add RFC4507 support to OpenSSL. This includes the corrections in
8289 RFC4507bis. The encrypted ticket format is an encrypted encoded
8290 SSL_SESSION structure, that way new session features are automatically
8293 If a client application caches session in an SSL_SESSION structure
8294 support is transparent because tickets are now stored in the encoded
8297 The SSL_CTX structure automatically generates keys for ticket
8298 protection in servers so again support should be possible
8299 with no application modification.
8301 If a client or server wishes to disable RFC4507 support then the option
8302 SSL_OP_NO_TICKET can be set.
8304 Add a TLS extension debugging callback to allow the contents of any client
8305 or server extensions to be examined.
8307 This work was sponsored by Google.
8311 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8312 OpenSSL should now compile cleanly on gcc 4.2
8314 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8316 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8317 support including streaming MAC support: this is required for GOST
8318 ciphersuite support.
8320 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8322 * Add option -stream to use PKCS#7 streaming in smime utility. New
8323 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8324 to output in BER and PEM format.
8328 * Experimental support for use of HMAC via EVP_PKEY interface. This
8329 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8330 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8331 ENGINE support for HMAC keys which are unextractable. New -mac and
8332 -macopt options to dgst utility.
8336 * New option -sigopt to dgst utility. Update dgst to use
8337 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8338 alternative signing parameters such as X9.31 or PSS in the dgst
8343 * Change ssl_cipher_apply_rule(), the internal function that does
8344 the work each time a ciphersuite string requests enabling
8345 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8346 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8347 the order of disabled ciphersuites such that those ciphersuites
8348 that most recently went from enabled to disabled not only stay
8349 in order with respect to each other, but also have higher priority
8350 than other disabled ciphersuites the next time ciphersuites are
8353 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8354 the same ciphersuites as with "HIGH" alone, but in a specific
8355 order where the PSK ciphersuites come first (since they are the
8356 most recently disabled ciphersuites when "HIGH" is parsed).
8358 Also, change ssl_create_cipher_list() (using this new
8359 functionality) such that between otherwise identical
8360 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8365 * Change ssl_create_cipher_list() so that it automatically
8366 arranges the ciphersuites in reasonable order before starting
8367 to process the rule string. Thus, the definition for "DEFAULT"
8368 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8369 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8370 This makes it much easier to arrive at a reasonable default order
8371 in applications for which anonymous ciphers are OK (meaning
8372 that you can't actually use DEFAULT).
8374 *Bodo Moeller; suggested by Victor Duchovni*
8376 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8377 processing) into multiple integers instead of setting
8378 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8379 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8380 (These masks as well as the individual bit definitions are hidden
8381 away into the non-exported interface ssl/ssl_locl.h, so this
8382 change to the definition of the SSL_CIPHER structure shouldn't
8383 affect applications.) This give us more bits for each of these
8384 categories, so there is no longer a need to coagulate AES128 and
8385 AES256 into a single algorithm bit, and to coagulate Camellia128
8386 and Camellia256 into a single algorithm bit, which has led to all
8389 Thus, among other things, the kludge introduced in 0.9.7m and
8390 0.9.8e for masking out AES256 independently of AES128 or masking
8391 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8393 With the change, we also introduce new ciphersuite aliases that
8394 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8399 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8400 Use the leftmost N bytes of the signature input if the input is
8401 larger than the prime q (with N being the size in bytes of q).
8405 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8406 it yet and it is largely untested.
8410 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8414 * Initial incomplete changes to avoid need for function casts in OpenSSL
8415 some compilers (gcc 4.2 and later) reject their use. Safestack is
8416 reimplemented. Update ASN1 to avoid use of legacy functions.
8420 * Win32/64 targets are linked with Winsock2.
8424 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8425 to external functions. This can be used to increase CRL handling
8426 efficiency especially when CRLs are very large by (for example) storing
8427 the CRL revoked certificates in a database.
8431 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8432 new CRLs added to a directory can be used. New command line option
8433 -verify_return_error to s_client and s_server. This causes real errors
8434 to be returned by the verify callback instead of carrying on no matter
8435 what. This reflects the way a "real world" verify callback would behave.
8439 * GOST engine, supporting several GOST algorithms and public key formats.
8440 Kindly donated by Cryptocom.
8444 * Partial support for Issuing Distribution Point CRL extension. CRLs
8445 partitioned by DP are handled but no indirect CRL or reason partitioning
8446 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8447 selected via a scoring technique which handles IDP and AKID in CRLs.
8451 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8452 will ultimately be used for all verify operations: this will remove the
8453 X509_STORE dependency on certificate verification and allow alternative
8454 lookup methods. X509_STORE based implementations of these two callbacks.
8458 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8459 Modify get_crl() to find a valid (unexpired) CRL if possible.
8463 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8464 this would be called X509_CRL_cmp() but that name is already used by
8465 a function that just compares CRL issuer names. Cache several CRL
8466 extensions in X509_CRL structure and cache CRLDP in X509.
8470 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8471 this maps equivalent X509_NAME structures into a consistent structure.
8472 Name comparison can then be performed rapidly using memcmp().
8476 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8481 * Allow digests to supply their own micalg string for S/MIME type using
8482 the ctrl EVP_MD_CTRL_MICALG.
8486 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8487 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8488 ctrl. It can then customise the structure before and/or after signing
8493 * New function OBJ_add_sigid() to allow application defined signature OIDs
8494 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8495 to free up any added signature OIDs.
8499 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8500 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8501 digest and cipher tables. New options added to openssl utility:
8502 list-message-digest-algorithms and list-cipher-algorithms.
8506 * Change the array representation of binary polynomials: the list
8507 of degrees of non-zero coefficients is now terminated with -1.
8508 Previously it was terminated with 0, which was also part of the
8509 value; thus, the array representation was not applicable to
8510 polynomials where t^0 has coefficient zero. This change makes
8511 the array representation useful in a more general context.
8515 * Various modifications and fixes to SSL/TLS cipher string
8516 handling. For ECC, the code now distinguishes between fixed ECDH
8517 with RSA certificates on the one hand and with ECDSA certificates
8518 on the other hand, since these are separate ciphersuites. The
8519 unused code for Fortezza ciphersuites has been removed.
8521 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8522 (not "ECDHE"). For consistency with the code for DH
8523 certificates, use of ECDH certificates is now considered ECDH
8524 authentication, not RSA or ECDSA authentication (the latter is
8525 merely the CA's signing algorithm and not actively used in the
8528 The temporary ciphersuite alias "ECCdraft" is no longer
8529 available, and ECC ciphersuites are no longer excluded from "ALL"
8530 and "DEFAULT". The following aliases now exist for RFC 4492
8531 ciphersuites, most of these by analogy with the DH case:
8533 kECDHr - ECDH cert, signed with RSA
8534 kECDHe - ECDH cert, signed with ECDSA
8535 kECDH - ECDH cert (signed with either RSA or ECDSA)
8536 kEECDH - ephemeral ECDH
8537 ECDH - ECDH cert or ephemeral ECDH
8543 AECDH - anonymous ECDH
8544 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8548 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8549 Use correct micalg parameters depending on digest(s) in signed message.
8553 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8554 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8558 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8559 an engine to register a method. Add ENGINE lookups for methods and
8560 functional reference processing.
8564 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8565 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8570 * New -resign option to smime utility. This adds one or more signers
8571 to an existing PKCS#7 signedData structure. Also -md option to use an
8572 alternative message digest algorithm for signing.
8576 * Tidy up PKCS#7 routines and add new functions to make it easier to
8577 create PKCS7 structures containing multiple signers. Update smime
8578 application to support multiple signers.
8582 * New -macalg option to pkcs12 utility to allow setting of an alternative
8587 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8588 Reorganize PBE internals to lookup from a static table using NIDs,
8589 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8590 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8591 PRF which will be automatically used with PBES2.
8595 * Replace the algorithm specific calls to generate keys in "req" with the
8600 * Update PKCS#7 enveloped data routines to use new API. This is now
8601 supported by any public key method supporting the encrypt operation. A
8602 ctrl is added to allow the public key algorithm to examine or modify
8603 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8608 * Add a ctrl to asn1 method to allow a public key algorithm to express
8609 a default digest type to use. In most cases this will be SHA1 but some
8610 algorithms (such as GOST) need to specify an alternative digest. The
8611 return value indicates how strong the preference is 1 means optional and
8612 2 is mandatory (that is it is the only supported type). Modify
8613 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8614 use the default md. Update openssl utilities to use the default digest
8615 type for signing if it is not explicitly indicated.
8619 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8620 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8621 signing method from the key type. This effectively removes the link
8622 between digests and public key types.
8626 * Add an OID cross reference table and utility functions. Its purpose is to
8627 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8628 rsaEncryption. This will allow some of the algorithm specific hackery
8629 needed to use the correct OID to be removed.
8633 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8634 structures for PKCS7_sign(). They are now set up by the relevant public
8639 * Add provisional EC pkey method with support for ECDSA and ECDH.
8643 * Add support for key derivation (agreement) in the API, DH method and
8648 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8649 public and private key formats. As a side effect these add additional
8650 command line functionality not previously available: DSA signatures can be
8651 generated and verified using pkeyutl and DH key support and generation in
8658 *Oliver Tappe <zooey@hirschkaefer.de>*
8660 * New make target "install_html_docs" installs HTML renditions of the
8663 *Oliver Tappe <zooey@hirschkaefer.de>*
8665 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8666 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8667 support key and parameter generation and add initial key generation
8668 functionality for RSA.
8672 * Add functions for main EVP_PKEY_method operations. The undocumented
8673 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8674 `EVP_PKEY_{encrypt,decrypt}_old`.
8678 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8679 key API, doesn't do much yet.
8683 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8684 public key algorithms. New option to openssl utility:
8685 "list-public-key-algorithms" to print out info.
8689 * Implement the Supported Elliptic Curves Extension for
8690 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8694 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8695 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8699 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8700 utilities such as rsa, dsa, dsaparam etc except they process any key
8705 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8706 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8707 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8712 * Initial support for pluggable public key ASN1.
8713 De-spaghettify the public key ASN1 handling. Move public and private
8714 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8715 algorithm specific handling to a single module within the relevant
8716 algorithm directory. Add functions to allow (near) opaque processing
8717 of public and private key structures.
8721 * Implement the Supported Point Formats Extension for
8722 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8726 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8727 for the psk identity [hint] and the psk callback functions to the
8728 SSL_SESSION, SSL and SSL_CTX structure.
8731 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8735 SSL_CTX_use_psk_identity_hint
8736 SSL_get_psk_identity_hint
8737 SSL_get_psk_identity
8738 SSL_use_psk_identity_hint
8740 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8742 * Add RFC 3161 compliant time stamp request creation, response generation
8743 and response verification functionality.
8745 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8747 * Add initial support for TLS extensions, specifically for the server_name
8748 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8749 have new members for a host name. The SSL data structure has an
8750 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8751 stored in that context to allow for session resumption, even after the
8752 SSL has been switched to a new SSL_CTX in reaction to a client's
8753 server_name extension.
8755 New functions (subject to change):
8757 SSL_get_servername()
8758 SSL_get_servername_type()
8761 New CTRL codes and macros (subject to change):
8763 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8764 - SSL_CTX_set_tlsext_servername_callback()
8765 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8766 - SSL_CTX_set_tlsext_servername_arg()
8767 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8769 openssl s_client has a new '-servername ...' option.
8771 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8772 '-key2 ...', '-servername_fatal' (subject to change). This allows
8773 testing the HostName extension for a specific single host name ('-cert'
8774 and '-key' remain fallbacks for handshakes without HostName
8775 negotiation). If the unrecognized_name alert has to be sent, this by
8776 default is a warning; it becomes fatal with the '-servername_fatal'
8779 *Peter Sylvester, Remy Allais, Christophe Renou*
8781 * Whirlpool hash implementation is added.
8785 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8786 bn(64,32). Because of instruction set limitations it doesn't have
8787 any negative impact on performance. This was done mostly in order
8788 to make it possible to share assembler modules, such as bn_mul_mont
8789 implementations, between 32- and 64-bit builds without hassle.
8793 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8794 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8799 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8800 dedicated Montgomery multiplication procedure, is introduced.
8801 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8802 "64-bit" performance on certain 32-bit targets.
8806 * New option SSL_OP_NO_COMP to disable use of compression selectively
8807 in SSL structures. New SSL ctrl to set maximum send fragment size.
8808 Save memory by setting the I/O buffer sizes dynamically instead of
8809 using the maximum available value.
8813 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8814 in addition to the text details.
8818 * Very, very preliminary EXPERIMENTAL support for printing of general
8819 ASN1 structures. This currently produces rather ugly output and doesn't
8820 handle several customised structures at all.
8824 * Integrated support for PVK file format and some related formats such
8825 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8826 these in the 'rsa' and 'dsa' utilities.
8830 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8834 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8835 place for the (very old) "NETSCAPE" format certificates which are now
8836 handled using new ASN1 code equivalents.
8840 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8841 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8842 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8846 * Modify CRL distribution points extension code to print out previously
8847 unsupported fields. Enhance extension setting code to allow setting of
8852 * Add print and set support for Issuing Distribution Point CRL extension.
8856 * Change 'Configure' script to enable Camellia by default.
8863 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8865 * When rejecting SSL/TLS records due to an incorrect version number, never
8866 update s->server with a new major version number. As of
8867 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8868 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8869 the previous behavior could result in a read attempt at NULL when
8870 receiving specific incorrect SSL/TLS records once record payload
8871 protection is active. ([CVE-2010-0740])
8873 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8875 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8876 could be crashed if the relevant tables were not present (e.g. chrooted).
8878 *Tomas Hoger <thoger@redhat.com>*
8880 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8882 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8884 *Martin Olsson, Neel Mehta*
8886 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8887 accommodate for stack sorting, always a write lock!).
8891 * On some versions of WIN32 Heap32Next is very slow. This can cause
8892 excessive delays in the RAND_poll(): over a minute. As a workaround
8893 include a time check in the inner Heap32Next loop too.
8897 * The code that handled flushing of data in SSL/TLS originally used the
8898 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8899 the problem outlined in PR#1949. The fix suggested there however can
8900 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8901 of Apache). So instead simplify the code to flush unconditionally.
8902 This should be fine since flushing with no data to flush is a no op.
8906 * Handle TLS versions 2.0 and later properly and correctly use the
8907 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8908 off ancient servers have a habit of sticking around for a while...
8912 * Modify compression code so it frees up structures without using the
8913 ex_data callbacks. This works around a problem where some applications
8914 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8915 restarting) then use compression (e.g. SSL with compression) later.
8916 This results in significant per-connection memory leaks and
8917 has caused some security issues including CVE-2008-1678 and
8922 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8923 change when encrypting or decrypting.
8927 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8928 connect and renegotiate with servers which do not support RI.
8929 Until RI is more widely deployed this option is enabled by default.
8933 * Add "missing" ssl ctrls to clear options and mode.
8937 * If client attempts to renegotiate and doesn't support RI respond with
8938 a no_renegotiation alert as required by RFC5746. Some renegotiating
8939 TLS clients will continue a connection gracefully when they receive
8940 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8941 waiting for a server hello which it will never receive. Now we treat a
8942 received no_renegotiation alert as a fatal error. This is because
8943 applications requesting a renegotiation might well expect it to succeed
8944 and would have no code in place to handle the server denying it so the
8945 only safe thing to do is to terminate the connection.
8949 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8950 peer supports secure renegotiation and 0 otherwise. Print out peer
8951 renegotiation support in s_client/s_server.
8955 * Replace the highly broken and deprecated SPKAC certification method with
8956 the updated NID creation version. This should correctly handle UTF8.
8960 * Implement RFC5746. Re-enable renegotiation but require the extension
8961 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8962 turns out to be a bad idea. It has been replaced by
8963 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8964 SSL_CTX_set_options(). This is really not recommended unless you
8965 know what you are doing.
8967 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8969 * Fixes to stateless session resumption handling. Use initial_ctx when
8970 issuing and attempting to decrypt tickets in case it has changed during
8971 servername handling. Use a non-zero length session ID when attempting
8972 stateless session resumption: this makes it possible to determine if
8973 a resumption has occurred immediately after receiving server hello
8974 (several places in OpenSSL subtly assume this) instead of later in
8979 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
8980 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
8981 fixes for a few places where the return code is not checked
8984 *Julia Lawall <julia@diku.dk>*
8986 * Add --strict-warnings option to Configure script to include devteam
8987 warnings in other configurations.
8991 * Add support for --libdir option and LIBDIR variable in makefiles. This
8992 makes it possible to install openssl libraries in locations which
8993 have names other than "lib", for example "/usr/lib64" which some
8996 *Steve Henson, based on patch from Jeremy Utley*
8998 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
8999 X690 8.9.12 and can produce some misleading textual output of OIDs.
9001 *Steve Henson, reported by Dan Kaminsky*
9003 * Delete MD2 from algorithm tables. This follows the recommendation in
9004 several standards that it is not used in new applications due to
9005 several cryptographic weaknesses. For binary compatibility reasons
9006 the MD2 API is still compiled in by default.
9010 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9015 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9016 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9019 *Guenter <lists@gknw.net>*
9021 * Fix the server certificate chain building code to use X509_verify_cert(),
9022 it used to have an ad-hoc builder which was unable to cope with anything
9023 other than a simple chain.
9025 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9027 * Don't check self signed certificate signatures in X509_verify_cert()
9028 by default (a flag can override this): it just wastes time without
9029 adding any security. As a useful side effect self signed root CAs
9030 with non-FIPS digests are now usable in FIPS mode.
9034 * In dtls1_process_out_of_seq_message() the check if the current message
9035 is already buffered was missing. For every new message was memory
9036 allocated, allowing an attacker to perform an denial of service attack
9037 with sending out of seq handshake messages until there is no memory
9038 left. Additionally every future message was buffered, even if the
9039 sequence number made no sense and would be part of another handshake.
9040 So only messages with sequence numbers less than 10 in advance will be
9041 buffered. ([CVE-2009-1378])
9043 *Robin Seggelmann, discovered by Daniel Mentz*
9045 * Records are buffered if they arrive with a future epoch to be
9046 processed after finishing the corresponding handshake. There is
9047 currently no limitation to this buffer allowing an attacker to perform
9048 a DOS attack with sending records with future epochs until there is no
9049 memory left. This patch adds the pqueue_size() function to determine
9050 the size of a buffer and limits the record buffer to 100 entries.
9053 *Robin Seggelmann, discovered by Daniel Mentz*
9055 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9056 parent structure is freed. ([CVE-2009-1379])
9060 * Handle non-blocking I/O properly in SSL_shutdown() call.
9062 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9064 * Add `2.5.4.*` OIDs
9066 *Ilya O. <vrghost@gmail.com>*
9068 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9070 * Disable renegotiation completely - this fixes a severe security
9071 problem ([CVE-2009-3555]) at the cost of breaking all
9072 renegotiation. Renegotiation can be re-enabled by setting
9073 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9074 run-time. This is really not recommended unless you know what
9079 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9081 * Don't set val to NULL when freeing up structures, it is freed up by
9082 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9083 zeroing past the valid field. ([CVE-2009-0789])
9085 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9087 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9088 checked correctly. This would allow some invalid signed attributes to
9089 appear to verify correctly. ([CVE-2009-0591])
9091 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9093 * Reject UniversalString and BMPString types with invalid lengths. This
9094 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9095 a legal length. ([CVE-2009-0590])
9099 * Set S/MIME signing as the default purpose rather than setting it
9100 unconditionally. This allows applications to override it at the store
9105 * Permit restricted recursion of ASN1 strings. This is needed in practice
9106 to handle some structures.
9110 * Improve efficiency of mem_gets: don't search whole buffer each time
9113 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9115 * New -hex option for openssl rand.
9119 * Print out UTF8String and NumericString when parsing ASN1.
9123 * Support NumericString type for name components.
9127 * Allow CC in the environment to override the automatically chosen
9128 compiler. Note that nothing is done to ensure flags work with the
9133 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9135 * Properly check EVP_VerifyFinal() and similar return values
9138 *Ben Laurie, Bodo Moeller, Google Security Team*
9140 * Enable TLS extensions by default.
9144 * Allow the CHIL engine to be loaded, whether the application is
9145 multithreaded or not. (This does not release the developer from the
9146 obligation to set up the dynamic locking callbacks.)
9148 *Sander Temme <sander@temme.net>*
9150 * Use correct exit code if there is an error in dgst command.
9152 *Steve Henson; problem pointed out by Roland Dirlewanger*
9154 * Tweak Configure so that you need to say "experimental-jpake" to enable
9155 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9159 * Add experimental JPAKE support, including demo authentication in
9160 s_client and s_server.
9164 * Set the comparison function in v3_addr_canonize().
9166 *Rob Austein <sra@hactrn.net>*
9168 * Add support for XMPP STARTTLS in s_client.
9170 *Philip Paeps <philip@freebsd.org>*
9172 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9173 to ensure that even with this option, only ciphersuites in the
9174 server's preference list will be accepted. (Note that the option
9175 applies only when resuming a session, so the earlier behavior was
9176 just about the algorithm choice for symmetric cryptography.)
9180 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9182 * Fix NULL pointer dereference if a DTLS server received
9183 ChangeCipherSpec as first record ([CVE-2009-1386]).
9187 * Fix a state transition in s3_srvr.c and d1_srvr.c
9188 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9192 * The fix in 0.9.8c that supposedly got rid of unsafe
9193 double-checked locking was incomplete for RSA blinding,
9194 addressing just one layer of what turns out to have been
9195 doubly unsafe triple-checked locking.
9197 So now fix this for real by retiring the MONT_HELPER macro
9198 in crypto/rsa/rsa_eay.c.
9200 *Bodo Moeller; problem pointed out by Marius Schilder*
9202 * Various precautionary measures:
9204 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9206 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9207 (NB: This would require knowledge of the secret session ticket key
9208 to exploit, in which case you'd be SOL either way.)
9210 - Change bn_nist.c so that it will properly handle input BIGNUMs
9211 outside the expected range.
9213 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9216 *Neel Mehta, Bodo Moeller*
9218 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9219 the load fails. Useful for distros.
9221 *Ben Laurie and the FreeBSD team*
9223 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9227 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9231 * Expand ENGINE to support engine supplied SSL client certificate functions.
9233 This work was sponsored by Logica.
9237 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9238 keystores. Support for SSL/TLS client authentication too.
9239 Not compiled unless enable-capieng specified to Configure.
9241 This work was sponsored by Logica.
9245 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9246 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9247 attribute creation routines such as certificate requests and PKCS#12
9252 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9254 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9255 handshake which could lead to a client crash as found using the
9256 Codenomicon TLS test suite ([CVE-2008-1672])
9258 *Steve Henson, Mark Cox*
9260 * Fix double free in TLS server name extensions which could lead to
9261 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9265 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9267 Clear the error queue to ensure that error entries left from
9268 older function calls do not interfere with the correct operation.
9270 *Lutz Jaenicke, Erik de Castro Lopo*
9272 * Remove root CA certificates of commercial CAs:
9274 The OpenSSL project does not recommend any specific CA and does not
9275 have any policy with respect to including or excluding any CA.
9276 Therefore it does not make any sense to ship an arbitrary selection
9277 of root CA certificates with the OpenSSL software.
9281 * RSA OAEP patches to fix two separate invalid memory reads.
9282 The first one involves inputs when 'lzero' is greater than
9283 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9284 before the beginning of from). The second one involves inputs where
9285 the 'db' section contains nothing but zeroes (there is a one-byte
9286 invalid read after the end of 'db').
9288 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9290 * Partial backport from 0.9.9-dev:
9292 Introduce bn_mul_mont (dedicated Montgomery multiplication
9293 procedure) as a candidate for BIGNUM assembler implementation.
9294 While 0.9.9-dev uses assembler for various architectures, only
9295 x86_64 is available by default here in the 0.9.8 branch, and
9296 32-bit x86 is available through a compile-time setting.
9298 To try the 32-bit x86 assembler implementation, use Configure
9299 option "enable-montasm" (which exists only for this backport).
9301 As "enable-montasm" for 32-bit x86 disclaims code stability
9302 anyway, in this constellation we activate additional code
9303 backported from 0.9.9-dev for further performance improvements,
9304 namely BN_from_montgomery_word. (To enable this otherwise,
9305 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9307 *Andy Polyakov (backport partially by Bodo Moeller)*
9309 * Add TLS session ticket callback. This allows an application to set
9310 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9311 values. This is useful for key rollover for example where several key
9312 sets may exist with different names.
9316 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9317 This was broken until now in 0.9.8 releases, such that the only way
9318 a registered ENGINE could be used (assuming it initialises
9319 successfully on the host) was to explicitly set it as the default
9320 for the relevant algorithms. This is in contradiction with 0.9.7
9321 behaviour and the documentation. With this fix, when an ENGINE is
9322 registered into a given algorithm's table of implementations, the
9323 'uptodate' flag is reset so that auto-discovery will be used next
9324 time a new context for that algorithm attempts to select an
9327 *Ian Lister (tweaked by Geoff Thorpe)*
9329 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9330 implementation in the following ways:
9332 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9335 Lack of BER streaming support means one pass streaming processing is
9336 only supported if data is detached: setting the streaming flag is
9337 ignored for embedded content.
9339 CMS support is disabled by default and must be explicitly enabled
9340 with the enable-cms configuration option.
9344 * Update the GMP engine glue to do direct copies between BIGNUM and
9345 mpz_t when openssl and GMP use the same limb size. Otherwise the
9346 existing "conversion via a text string export" trick is still used.
9348 *Paul Sheer <paulsheer@gmail.com>*
9350 * Zlib compression BIO. This is a filter BIO which compressed and
9351 uncompresses any data passed through it.
9355 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9356 RFC3394 compatible AES key wrapping.
9360 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9361 sets string data without copying. X509_ALGOR_set0() and
9362 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9363 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9364 from an X509_ATTRIBUTE structure optionally checking it occurs only
9365 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9370 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9371 to get the expected BN_FLG_CONSTTIME behavior.
9373 *Bodo Moeller (Google)*
9377 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9378 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9379 - added some more tests to do_tests.pl
9380 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9381 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9382 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9383 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9384 - various changes to netware.pl to enable gcc-cross builds on Win32
9386 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9387 - various changes to fix missing prototype warnings
9388 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9389 - added AES, WHIRLPOOL and CPUID assembler code to build files
9390 - added missing AES assembler make rules to mk1mf.pl
9391 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9393 *Guenter Knauf <eflash@gmx.net>*
9395 * Implement certificate status request TLS extension defined in RFC3546.
9396 A client can set the appropriate parameters and receive the encoded
9397 OCSP response via a callback. A server can query the supplied parameters
9398 and set the encoded OCSP response in the callback. Add simplified examples
9399 to s_client and s_server.
9403 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9406 + Binary incompatibility of ssl_ctx_st structure
9407 + DTLS interoperation with non-compliant servers
9408 + Don't call get_session_cb() without proposed session
9409 + Fix ia64 assembler code
9411 *Andy Polyakov, Steve Henson*
9413 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9415 * DTLS Handshake overhaul. There were longstanding issues with
9416 OpenSSL DTLS implementation, which were making it impossible for
9417 RFC 4347 compliant client to communicate with OpenSSL server.
9418 Unfortunately just fixing these incompatibilities would "cut off"
9419 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9420 server keeps tolerating non RFC compliant syntax. The opposite is
9421 not true, 0.9.8f client can not communicate with earlier server.
9422 This update even addresses CVE-2007-4995.
9426 * Changes to avoid need for function casts in OpenSSL: some compilers
9427 (gcc 4.2 and later) reject their use.
9428 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9431 * Add RFC4507 support to OpenSSL. This includes the corrections in
9432 RFC4507bis. The encrypted ticket format is an encrypted encoded
9433 SSL_SESSION structure, that way new session features are automatically
9436 If a client application caches session in an SSL_SESSION structure
9437 support is transparent because tickets are now stored in the encoded
9440 The SSL_CTX structure automatically generates keys for ticket
9441 protection in servers so again support should be possible
9442 with no application modification.
9444 If a client or server wishes to disable RFC4507 support then the option
9445 SSL_OP_NO_TICKET can be set.
9447 Add a TLS extension debugging callback to allow the contents of any client
9448 or server extensions to be examined.
9450 This work was sponsored by Google.
9454 * Add initial support for TLS extensions, specifically for the server_name
9455 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9456 have new members for a host name. The SSL data structure has an
9457 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9458 stored in that context to allow for session resumption, even after the
9459 SSL has been switched to a new SSL_CTX in reaction to a client's
9460 server_name extension.
9462 New functions (subject to change):
9464 SSL_get_servername()
9465 SSL_get_servername_type()
9468 New CTRL codes and macros (subject to change):
9470 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9471 - SSL_CTX_set_tlsext_servername_callback()
9472 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9473 - SSL_CTX_set_tlsext_servername_arg()
9474 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9476 openssl s_client has a new '-servername ...' option.
9478 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9479 '-key2 ...', '-servername_fatal' (subject to change). This allows
9480 testing the HostName extension for a specific single host name ('-cert'
9481 and '-key' remain fallbacks for handshakes without HostName
9482 negotiation). If the unrecognized_name alert has to be sent, this by
9483 default is a warning; it becomes fatal with the '-servername_fatal'
9486 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9488 * Add AES and SSE2 assembly language support to VC++ build.
9492 * Mitigate attack on final subtraction in Montgomery reduction.
9496 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9497 (which previously caused an internal error).
9501 * Squeeze another 10% out of IGE mode when in != out.
9505 * AES IGE mode speedup.
9507 *Dean Gaudet (Google)*
9509 * Add the Korean symmetric 128-bit cipher SEED (see
9510 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9511 add SEED ciphersuites from RFC 4162:
9513 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9514 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9515 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9516 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9518 To minimize changes between patchlevels in the OpenSSL 0.9.8
9519 series, SEED remains excluded from compilation unless OpenSSL
9520 is configured with 'enable-seed'.
9522 *KISA, Bodo Moeller*
9524 * Mitigate branch prediction attacks, which can be practical if a
9525 single processor is shared, allowing a spy process to extract
9526 information. For detailed background information, see
9527 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9528 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9529 and Necessary Software Countermeasures"). The core of the change
9530 are new versions BN_div_no_branch() and
9531 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9532 respectively, which are slower, but avoid the security-relevant
9533 conditional branches. These are automatically called by BN_div()
9534 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9535 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9536 remove a conditional branch.
9538 BN_FLG_CONSTTIME is the new name for the previous
9539 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9540 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9541 in the exponent causes BN_mod_exp_mont() to use the alternative
9542 implementation in BN_mod_exp_mont_consttime().) The old name
9543 remains as a deprecated alias.
9545 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9546 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9547 constant-time implementations for more than just exponentiation.
9548 Here too the old name is kept as a deprecated alias.
9550 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9551 the BN_BLINDING structure gets an independent copy of the
9552 modulus. This means that the previous `BIGNUM *m` argument to
9553 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9554 essentially becomes `const BIGNUM *m`, although we can't actually
9555 change this in the header file before 0.9.9. It allows
9556 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9557 enable BN_FLG_CONSTTIME.
9559 *Matthew D Wood (Intel Corp)*
9561 * In the SSL/TLS server implementation, be strict about session ID
9562 context matching (which matters if an application uses a single
9563 external cache for different purposes). Previously,
9564 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9565 set. This did ensure strict client verification, but meant that,
9566 with applications using a single external cache for quite
9567 different requirements, clients could circumvent ciphersuite
9568 restrictions for a given session ID context by starting a session
9569 in a different context.
9573 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9574 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9575 authentication-only ciphersuites.
9579 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9580 not complete and could lead to a possible single byte overflow
9581 ([CVE-2007-5135]) [Ben Laurie]
9583 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9585 * Since AES128 and AES256 (and similarly Camellia128 and
9586 Camellia256) share a single mask bit in the logic of
9587 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9588 kludge to work properly if AES128 is available and AES256 isn't
9589 (or if Camellia128 is available and Camellia256 isn't).
9593 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9594 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9595 When a point or a seed is encoded in a BIT STRING, we need to
9596 prevent the removal of trailing zero bits to get the proper DER
9597 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9598 of a NamedBitList, for which trailing 0 bits need to be removed.)
9602 * Have SSL/TLS server implementation tolerate "mismatched" record
9603 protocol version while receiving ClientHello even if the
9604 ClientHello is fragmented. (The server can't insist on the
9605 particular protocol version it has chosen before the ServerHello
9606 message has informed the client about his choice.)
9610 * Add RFC 3779 support.
9612 *Rob Austein for ARIN, Ben Laurie*
9614 * Load error codes if they are not already present instead of using a
9615 static variable. This allows them to be cleanly unloaded and reloaded.
9616 Improve header file function name parsing.
9620 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9621 or CAPABILITY handshake as required by RFCs.
9625 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9627 * Introduce limits to prevent malicious keys being able to
9628 cause a denial of service. ([CVE-2006-2940])
9630 *Steve Henson, Bodo Moeller*
9632 * Fix ASN.1 parsing of certain invalid structures that can result
9633 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9635 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9636 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9638 * Fix SSL client code which could crash if connecting to a
9639 malicious SSLv2 server. ([CVE-2006-4343])
9641 *Tavis Ormandy and Will Drewry, Google Security Team*
9643 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9644 match only those. Before that, "AES256-SHA" would be interpreted
9645 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9646 the same strength classification in 0.9.7h) as we currently only
9647 have a single AES bit in the ciphersuite description bitmap.
9648 That change, however, also applied to ciphersuite strings such as
9649 "RC4-MD5" that intentionally matched multiple ciphersuites --
9650 namely, SSL 2.0 ciphersuites in addition to the more common ones
9651 from SSL 3.0/TLS 1.0.
9653 So we change the selection algorithm again: Naming an explicit
9654 ciphersuite selects this one ciphersuite, and any other similar
9655 ciphersuite (same bitmap) from *other* protocol versions.
9656 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9657 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9659 Since SSL 2.0 does not have any ciphersuites for which the
9660 128/256 bit distinction would be relevant, this works for now.
9661 The proper fix will be to use different bits for AES128 and
9662 AES256, which would have avoided the problems from the beginning;
9663 however, bits are scarce, so we can only do this in a new release
9664 (not just a patchlevel) when we can change the SSL_CIPHER
9665 definition to split the single 'unsigned long mask' bitmap into
9666 multiple values to extend the available space.
9670 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9672 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9673 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9675 * Add AES IGE and biIGE modes.
9679 * Change the Unix randomness entropy gathering to use poll() when
9680 possible instead of select(), since the latter has some
9681 undesirable limitations.
9683 *Darryl Miles via Richard Levitte and Bodo Moeller*
9685 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9686 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9687 cannot be implicitly activated as part of, e.g., the "AES" alias.
9688 However, please upgrade to OpenSSL 0.9.9[-dev] for
9689 non-experimental use of the ECC ciphersuites to get TLS extension
9690 support, which is required for curve and point format negotiation
9691 to avoid potential handshake problems.
9695 * Disable rogue ciphersuites:
9697 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9698 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9699 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9701 The latter two were purportedly from
9702 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9705 Also deactivate the remaining ciphersuites from
9706 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9707 unofficial, and the ID has long expired.
9711 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9712 dual-core machines) and other potential thread-safety issues.
9716 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9717 versions), which is now available for royalty-free use
9718 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9719 Also, add Camellia TLS ciphersuites from RFC 4132.
9721 To minimize changes between patchlevels in the OpenSSL 0.9.8
9722 series, Camellia remains excluded from compilation unless OpenSSL
9723 is configured with 'enable-camellia'.
9727 * Disable the padding bug check when compression is in use. The padding
9728 bug check assumes the first packet is of even length, this is not
9729 necessarily true if compression is enabled and can result in false
9730 positives causing handshake failure. The actual bug test is ancient
9731 code so it is hoped that implementations will either have fixed it by
9732 now or any which still have the bug do not support compression.
9736 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9738 * When applying a cipher rule check to see if string match is an explicit
9739 cipher suite and only match that one cipher suite if it is.
9743 * Link in manifests for VC++ if needed.
9745 *Austin Ziegler <halostatue@gmail.com>*
9747 * Update support for ECC-based TLS ciphersuites according to
9748 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9749 TLS extensions, which are supported starting with the 0.9.9
9750 branch, not in the OpenSSL 0.9.8 branch).
9754 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9755 opaque EVP_CIPHER_CTX handling.
9759 * Fixes and enhancements to zlib compression code. We now only use
9760 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9761 to conform with the standards mentioned here:
9762 <http://www.zlib.net/DLL_FAQ.txt>
9763 Static zlib linking now works on Windows and the new --with-zlib-include
9764 --with-zlib-lib options to Configure can be used to supply the location
9765 of the headers and library. Gracefully handle case where zlib library
9770 * Several fixes and enhancements to the OID generation code. The old code
9771 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9772 handle numbers larger than ULONG_MAX, truncated printing and had a
9773 non standard OBJ_obj2txt() behaviour.
9777 * Add support for building of engines under engine/ as shared libraries
9778 under VC++ build system.
9782 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9783 Hopefully, we will not see any false combination of paths any more.
9787 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9789 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9790 (part of SSL_OP_ALL). This option used to disable the
9791 countermeasure against man-in-the-middle protocol-version
9792 rollback in the SSL 2.0 server implementation, which is a bad
9793 idea. ([CVE-2005-2969])
9795 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9796 for Information Security, National Institute of Advanced Industrial
9797 Science and Technology [AIST], Japan)*
9799 * Add two function to clear and return the verify parameter flags.
9803 * Keep cipherlists sorted in the source instead of sorting them at
9804 runtime, thus removing the need for a lock.
9808 * Avoid some small subgroup attacks in Diffie-Hellman.
9810 *Nick Mathewson and Ben Laurie*
9812 * Add functions for well-known primes.
9816 * Extended Windows CE support.
9818 *Satoshi Nakamura and Andy Polyakov*
9820 * Initialize SSL_METHOD structures at compile time instead of during
9821 runtime, thus removing the need for a lock.
9825 * Make PKCS7_decrypt() work even if no certificate is supplied by
9826 attempting to decrypt each encrypted key in turn. Add support to
9831 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9833 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9836 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9840 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9841 key into the same file any more.
9845 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9849 * Add -utf8 command line and config file option to 'ca'.
9851 *Stefan <stf@udoma.org*
9853 * Removed the macro des_crypt(), as it seems to conflict with some
9854 libraries. Use DES_crypt().
9858 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9859 involves renaming the source and generated shared-libs for
9860 both. The engines will accept the corrected or legacy ids
9861 ('ncipher' and '4758_cca' respectively) when binding. NB,
9862 this only applies when building 'shared'.
9864 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9866 * Add attribute functions to EVP_PKEY structure. Modify
9867 PKCS12_create() to recognize a CSP name attribute and
9868 use it. Make -CSP option work again in pkcs12 utility.
9872 * Add new functionality to the bn blinding code:
9873 - automatic re-creation of the BN_BLINDING parameters after
9874 a fixed number of uses (currently 32)
9875 - add new function for parameter creation
9876 - introduce flags to control the update behaviour of the
9877 BN_BLINDING parameters
9878 - hide BN_BLINDING structure
9879 Add a second BN_BLINDING slot to the RSA structure to improve
9880 performance when a single RSA object is shared among several
9885 * Add support for DTLS.
9887 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9889 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9890 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9894 * Remove buggy and incomplete DH cert support from
9895 ssl/ssl_rsa.c and ssl/s3_both.c
9899 * Use SHA-1 instead of MD5 as the default digest algorithm for
9900 the apps/openssl applications.
9904 * Compile clean with "-Wall -Wmissing-prototypes
9905 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9906 DEBUG_SAFESTACK must also be set.
9910 * Change ./Configure so that certain algorithms can be disabled by default.
9911 The new counterpiece to "no-xxx" is "enable-xxx".
9913 The patented RC5 and MDC2 algorithms will now be disabled unless
9914 "enable-rc5" and "enable-mdc2", respectively, are specified.
9916 (IDEA remains enabled despite being patented. This is because IDEA
9917 is frequently required for interoperability, and there is no license
9918 fee for non-commercial use. As before, "no-idea" can be used to
9919 avoid this algorithm.)
9923 * Add processing of proxy certificates (see RFC 3820). This work was
9924 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9925 EGEE (Enabling Grids for E-science in Europe).
9929 * RC4 performance overhaul on modern architectures/implementations, such
9930 as Intel P4, IA-64 and AMD64.
9934 * New utility extract-section.pl. This can be used specify an alternative
9935 section number in a pod file instead of having to treat each file as
9936 a separate case in Makefile. This can be done by adding two lines to the
9939 =for comment openssl_section:XXX
9941 The blank line is mandatory.
9945 * New arguments -certform, -keyform and -pass for s_client and s_server
9946 to allow alternative format key and certificate files and passphrase
9951 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9952 update associated structures and add various utility functions.
9954 Add new policy related verify parameters, include policy checking in
9955 standard verify code. Enhance 'smime' application with extra parameters
9956 to support policy checking and print out.
9960 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9961 Nehemiah processors. These extensions support AES encryption in hardware
9962 as well as RNG (though RNG support is currently disabled).
9964 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9966 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9970 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9972 *Andy Polyakov and a number of other people*
9974 * Improved PowerPC platform support. Most notably BIGNUM assembler
9975 implementation contributed by IBM.
9977 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
9979 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
9980 exponent rather than 'unsigned long'. There is a corresponding change to
9981 the new 'rsa_keygen' element of the RSA_METHOD structure.
9983 *Jelte Jansen, Geoff Thorpe*
9985 * Functionality for creating the initial serial number file is now
9986 moved from CA.pl to the 'ca' utility with a new option -create_serial.
9988 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
9989 number file to 1, which is bound to cause problems. To avoid
9990 the problems while respecting compatibility between different 0.9.7
9991 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
9992 CA.pl for serial number initialization. With the new release 0.9.8,
9993 we can fix the problem directly in the 'ca' utility.)
9997 * Reduced header interdependencies by declaring more opaque objects in
9998 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
9999 give fewer recursive includes, which could break lazy source code - so
10000 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10001 developers should define this symbol when building and using openssl to
10002 ensure they track the recommended behaviour, interfaces, [etc], but
10003 backwards-compatible behaviour prevails when this isn't defined.
10007 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10011 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10012 This will generate a random key of the appropriate length based on the
10013 cipher context. The EVP_CIPHER can provide its own random key generation
10014 routine to support keys of a specific form. This is used in the des and
10015 3des routines to generate a key of the correct parity. Update S/MIME
10016 code to use new functions and hence generate correct parity DES keys.
10017 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10018 valid (weak or incorrect parity).
10022 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10023 as looking them up. This is useful when the verified structure may contain
10024 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10025 present unless the new PKCS7_NO_CRL flag is asserted.
10029 * Extend ASN1 oid configuration module. It now additionally accepts the
10032 shortName = some long name, 1.2.3.4
10036 * Reimplemented the BN_CTX implementation. There is now no more static
10037 limitation on the number of variables it can handle nor the depth of the
10038 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10039 information can now expand as required, and rather than having a single
10040 static array of bignums, BN_CTX now uses a linked-list of such arrays
10041 allowing it to expand on demand whilst maintaining the usefulness of
10042 BN_CTX's "bundling".
10046 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10047 to allow all RSA operations to function using a single BN_CTX.
10051 * Preliminary support for certificate policy evaluation and checking. This
10052 is initially intended to pass the tests outlined in "Conformance Testing
10053 of Relying Party Client Certificate Path Processing Logic" v1.07.
10057 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10058 remained unused and not that useful. A variety of other little bignum
10059 tweaks and fixes have also been made continuing on from the audit (see
10064 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10065 associated ASN1, EVP and SSL functions and old ASN1 macros.
10069 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10070 and this should never fail. So the return value from the use of
10071 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10072 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10076 * BN_CTX_get() should return zero-valued bignums, providing the same
10077 initialised value as BN_new().
10079 *Geoff Thorpe, suggested by Ulf Möller*
10081 * Support for inhibitAnyPolicy certificate extension.
10085 * An audit of the BIGNUM code is underway, for which debugging code is
10086 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10087 is considered valid when processing BIGNUMs, and causes execution to
10088 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10089 further steps are taken to deliberately pollute unused data in BIGNUM
10090 structures to try and expose faulty code further on. For now, openssl will
10091 (in its default mode of operation) continue to tolerate the inconsistent
10092 forms that it has tolerated in the past, but authors and packagers should
10093 consider trying openssl and their own applications when compiled with
10094 these debugging symbols defined. It will help highlight potential bugs in
10095 their own code, and will improve the test coverage for OpenSSL itself. At
10096 some point, these tighter rules will become openssl's default to improve
10097 maintainability, though the assert()s and other overheads will remain only
10098 in debugging configurations. See bn.h for more details.
10100 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10102 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10103 that can only be obtained through BN_CTX_new() (which implicitly
10104 initialises it). The presence of this function only made it possible
10105 to overwrite an existing structure (and cause memory leaks).
10109 * Because of the callback-based approach for implementing LHASH as a
10110 template type, lh_insert() adds opaque objects to hash-tables and
10111 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10112 to clean up those corresponding objects before destroying the hash table
10113 (and losing the object pointers). So some over-zealous constifications in
10114 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10115 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10116 prototyped to have "const" restrictions on the object pointers they are
10117 given (and so aren't required to cast them away any more).
10121 * The tmdiff.h API was so ugly and minimal that our own timing utility
10122 (speed) prefers to use its own implementation. The two implementations
10123 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10124 its object type properly exposed (MS_TM) instead of casting to/from
10125 `char *`. This may still change yet if someone realises MS_TM and
10127 aren't necessarily the greatest nomenclatures - but this is what was used
10128 internally to the implementation so I've used that for now.
10132 * Ensure that deprecated functions do not get compiled when
10133 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10134 the self-tests were still using deprecated key-generation functions so
10135 these have been updated also.
10139 * Reorganise PKCS#7 code to separate the digest location functionality
10140 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10141 New function PKCS7_set_digest() to set the digest type for PKCS#7
10142 digestedData type. Add additional code to correctly generate the
10143 digestedData type and add support for this type in PKCS7 initialization
10148 * New function PKCS7_set0_type_other() this initializes a PKCS7
10149 structure of type "other".
10153 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10154 sure the loop does correctly stop and breaking ("division by zero")
10155 modulus operations are not performed. The (pre-generated) prime
10156 table crypto/bn/bn_prime.h was already correct, but it could not be
10157 re-generated on some platforms because of the "division by zero"
10158 situation in the script.
10160 *Ralf S. Engelschall*
10162 * Update support for ECC-based TLS ciphersuites according to
10163 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10164 SHA-1 now is only used for "small" curves (where the
10165 representation of a field element takes up to 24 bytes); for
10166 larger curves, the field element resulting from ECDH is directly
10167 used as premaster secret.
10169 *Douglas Stebila (Sun Microsystems Laboratories)*
10171 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10172 curve secp160r1 to the tests.
10174 *Douglas Stebila (Sun Microsystems Laboratories)*
10176 * Add the possibility to load symbols globally with DSO.
10178 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10180 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10181 control of the error stack.
10185 * Add support for STORE in ENGINE.
10189 * Add the STORE type. The intention is to provide a common interface
10190 to certificate and key stores, be they simple file-based stores, or
10191 HSM-type store, or LDAP stores, or...
10192 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10196 * Add a generic structure called OPENSSL_ITEM. This can be used to
10197 pass a list of arguments to any function as well as provide a way
10198 for a function to pass data back to the caller.
10202 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10203 works like BUF_strdup() but can be used to duplicate a portion of
10204 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10209 * Add the function sk_find_ex() which works like sk_find(), but will
10210 return an index to an element even if an exact match couldn't be
10211 found. The index is guaranteed to point at the element where the
10212 searched-for key would be inserted to preserve sorting order.
10216 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10217 takes an extra flags argument for optional functionality. Currently,
10218 the following flags are defined:
10220 OBJ_BSEARCH_VALUE_ON_NOMATCH
10221 This one gets OBJ_bsearch_ex() to return a pointer to the first
10222 element where the comparing function returns a negative or zero
10225 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10226 This one gets OBJ_bsearch_ex() to return a pointer to the first
10227 element where the comparing function returns zero. This is useful
10228 if there are more than one element where the comparing function
10233 * Make it possible to create self-signed certificates with 'openssl ca'
10234 in such a way that the self-signed certificate becomes part of the
10235 CA database and uses the same mechanisms for serial number generation
10236 as all other certificate signing. The new flag '-selfsign' enables
10237 this functionality. Adapt CA.sh and CA.pl.in.
10241 * Add functionality to check the public key of a certificate request
10242 against a given private. This is useful to check that a certificate
10243 request can be signed by that key (self-signing).
10247 * Make it possible to have multiple active certificates with the same
10248 subject in the CA index file. This is done only if the keyword
10249 'unique_subject' is set to 'no' in the main CA section (default
10250 if 'CA_default') of the configuration file. The value is saved
10251 with the database itself in a separate index attribute file,
10252 named like the index file with '.attr' appended to the name.
10256 * Generate multi-valued AVAs using '+' notation in config files for
10261 * Support for nameConstraints certificate extension.
10265 * Support for policyConstraints certificate extension.
10269 * Support for policyMappings certificate extension.
10273 * Make sure the default DSA_METHOD implementation only uses its
10274 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10275 and change its own handlers to be NULL so as to remove unnecessary
10276 indirection. This lets alternative implementations fallback to the
10277 default implementation more easily.
10281 * Support for directoryName in GeneralName related extensions
10286 * Make it possible to link applications using Makefile.shared.
10287 Make that possible even when linking against static libraries!
10291 * Support for single pass processing for S/MIME signing. This now
10292 means that S/MIME signing can be done from a pipe, in addition
10293 cleartext signing (multipart/signed type) is effectively streaming
10294 and the signed data does not need to be all held in memory.
10296 This is done with a new flag PKCS7_STREAM. When this flag is set
10297 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10298 is done after the data is output (and digests calculated) in
10299 SMIME_write_PKCS7().
10303 * Add full support for -rpath/-R, both in shared libraries and
10304 applications, at least on the platforms where it's known how
10309 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10310 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10311 will now compute a table of multiples of the generator that
10312 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10313 faster (notably in the case of a single point multiplication,
10314 scalar * generator).
10316 *Nils Larsch, Bodo Moeller*
10318 * IPv6 support for certificate extensions. The various extensions
10319 which use the IP:a.b.c.d can now take IPv6 addresses using the
10320 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10325 * Added an ENGINE that implements RSA by performing private key
10326 exponentiations with the GMP library. The conversions to and from
10327 GMP's mpz_t format aren't optimised nor are any montgomery forms
10328 cached, and on x86 it appears OpenSSL's own performance has caught up.
10329 However there are likely to be other architectures where GMP could
10330 provide a boost. This ENGINE is not built in by default, but it can be
10331 specified at Configure time and should be accompanied by the necessary
10332 linker additions, eg;
10333 ./config -DOPENSSL_USE_GMP -lgmp
10337 * "openssl engine" will not display ENGINE/DSO load failure errors when
10338 testing availability of engines with "-t" - the old behaviour is
10339 produced by increasing the feature's verbosity with "-tt".
10343 * ECDSA routines: under certain error conditions uninitialized BN objects
10344 could be freed. Solution: make sure initialization is performed early
10345 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10350 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10351 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10352 software implementations. For DSA and DH, parameter generation can
10353 also be overridden by providing the appropriate method callbacks.
10357 * Change the "progress" mechanism used in key-generation and
10358 primality testing to functions that take a new BN_GENCB pointer in
10359 place of callback/argument pairs. The new API functions have `_ex`
10360 postfixes and the older functions are reimplemented as wrappers for
10361 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10362 declarations of the old functions to help (graceful) attempts to
10363 migrate to the new functions. Also, the new key-generation API
10364 functions operate on a caller-supplied key-structure and return
10365 success/failure rather than returning a key or NULL - this is to
10366 help make "keygen" another member function of RSA_METHOD etc.
10368 Example for using the new callback interface:
10370 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10371 void *my_arg = ...;
10374 BN_GENCB_set(&my_cb, my_callback, my_arg);
10376 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10377 /* For the meaning of a, b in calls to my_callback(), see the
10378 * documentation of the function that calls the callback.
10379 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10380 * my_callback should return 1 if it wants BN_is_prime_ex()
10381 * to continue, or 0 to stop.
10386 * Change the ZLIB compression method to be stateful, and make it
10387 available to TLS with the number defined in
10388 draft-ietf-tls-compression-04.txt.
10392 * Add the ASN.1 structures and functions for CertificatePair, which
10393 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10395 CertificatePair ::= SEQUENCE {
10396 forward [0] Certificate OPTIONAL,
10397 reverse [1] Certificate OPTIONAL,
10398 -- at least one of the pair shall be present -- }
10400 Also implement the PEM functions to read and write certificate
10401 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10403 This needed to be defined, mostly for the sake of the LDAP
10404 attribute crossCertificatePair, but may prove useful elsewhere as
10409 * Make it possible to inhibit symlinking of shared libraries in
10410 Makefile.shared, for Cygwin's sake.
10414 * Extend the BIGNUM API by creating a function
10415 void BN_set_negative(BIGNUM *a, int neg);
10416 and a macro that behave like
10417 int BN_is_negative(const BIGNUM *a);
10419 to avoid the need to access 'a->neg' directly in applications.
10423 * Implement fast modular reduction for pseudo-Mersenne primes
10424 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10425 EC_GROUP_new_curve_GFp() will now automatically use this
10428 *Nils Larsch <nla@trustcenter.de>*
10430 * Add new lock type (CRYPTO_LOCK_BN).
10434 * Change the ENGINE framework to automatically load engines
10435 dynamically from specific directories unless they could be
10436 found to already be built in or loaded. Move all the
10437 current engines except for the cryptodev one to a new
10438 directory engines/.
10439 The engines in engines/ are built as shared libraries if
10440 the "shared" options was given to ./Configure or ./config.
10441 Otherwise, they are inserted in libcrypto.a.
10442 /usr/local/ssl/engines is the default directory for dynamic
10443 engines, but that can be overridden at configure time through
10444 the usual use of --prefix and/or --openssldir, and at run
10445 time with the environment variable OPENSSL_ENGINES.
10447 *Geoff Thorpe and Richard Levitte*
10449 * Add Makefile.shared, a helper makefile to build shared
10450 libraries. Adapt Makefile.org.
10454 * Add version info to Win32 DLLs.
10456 *Peter 'Luna' Runestig" <peter@runestig.com>*
10458 * Add new 'medium level' PKCS#12 API. Certificates and keys
10459 can be added using this API to created arbitrary PKCS#12
10460 files while avoiding the low level API.
10462 New options to PKCS12_create(), key or cert can be NULL and
10463 will then be omitted from the output file. The encryption
10464 algorithm NIDs can be set to -1 for no encryption, the mac
10465 iteration count can be set to 0 to omit the mac.
10467 Enhance pkcs12 utility by making the -nokeys and -nocerts
10468 options work when creating a PKCS#12 file. New option -nomac
10469 to omit the mac, NONE can be set for an encryption algorithm.
10470 New code is modified to use the enhanced PKCS12_create()
10471 instead of the low level API.
10475 * Extend ASN1 encoder to support indefinite length constructed
10476 encoding. This can output sequences tags and octet strings in
10477 this form. Modify pk7_asn1.c to support indefinite length
10478 encoding. This is experimental and needs additional code to
10479 be useful, such as an ASN1 bio and some enhanced streaming
10482 Extend template encode functionality so that tagging is passed
10483 down to the template encoder.
10487 * Let 'openssl req' fail if an argument to '-newkey' is not
10488 recognized instead of using RSA as a default.
10492 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10493 As these are not official, they are not included in "ALL";
10494 the "ECCdraft" ciphersuite group alias can be used to select them.
10496 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10498 * Add ECDH engine support.
10500 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10502 * Add ECDH in new directory crypto/ecdh/.
10504 *Douglas Stebila (Sun Microsystems Laboratories)*
10506 * Let BN_rand_range() abort with an error after 100 iterations
10507 without success (which indicates a broken PRNG).
10511 * Change BN_mod_sqrt() so that it verifies that the input value
10512 is really the square of the return value. (Previously,
10513 BN_mod_sqrt would show GIGO behaviour.)
10517 * Add named elliptic curves over binary fields from X9.62, SECG,
10518 and WAP/WTLS; add OIDs that were still missing.
10520 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10522 * Extend the EC library for elliptic curves over binary fields
10523 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10526 EC_GF2m_simple_method
10530 EC_GROUP_new_curve_GF2m
10531 EC_GROUP_set_curve_GF2m
10532 EC_GROUP_get_curve_GF2m
10533 EC_POINT_set_affine_coordinates_GF2m
10534 EC_POINT_get_affine_coordinates_GF2m
10535 EC_POINT_set_compressed_coordinates_GF2m
10537 Point compression for binary fields is disabled by default for
10538 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10541 As binary polynomials are represented as BIGNUMs, various members
10542 of the EC_GROUP and EC_POINT data structures can be shared
10543 between the implementations for prime fields and binary fields;
10544 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10545 are essentially identical to their `..._GFp` counterparts.
10546 (For simplicity, the `..._GFp` prefix has been dropped from
10547 various internal method names.)
10549 An internal 'field_div' method (similar to 'field_mul' and
10550 'field_sqr') has been added; this is used only for binary fields.
10552 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10554 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10555 through methods ('mul', 'precompute_mult').
10557 The generic implementations (now internally called 'ec_wNAF_mul'
10558 and 'ec_wNAF_precomputed_mult') remain the default if these
10559 methods are undefined.
10561 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10563 * New function EC_GROUP_get_degree, which is defined through
10564 EC_METHOD. For curves over prime fields, this returns the bit
10565 length of the modulus.
10567 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10569 * New functions EC_GROUP_dup, EC_POINT_dup.
10570 (These simply call ..._new and ..._copy).
10572 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10574 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10575 Polynomials are represented as BIGNUMs (where the sign bit is not
10576 used) in the following functions [macros]:
10579 BN_GF2m_sub [= BN_GF2m_add]
10580 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10581 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10582 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10584 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10585 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10586 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10587 BN_GF2m_cmp [= BN_ucmp]
10589 (Note that only the 'mod' functions are actually for fields GF(2^m).
10590 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10592 For some functions, an the irreducible polynomial defining a
10593 field can be given as an 'unsigned int[]' with strictly
10594 decreasing elements giving the indices of those bits that are set;
10595 i.e., p[] represents the polynomial
10596 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10598 p[0] > p[1] > ... > p[k] = 0.
10599 This applies to the following functions:
10602 BN_GF2m_mod_mul_arr
10603 BN_GF2m_mod_sqr_arr
10604 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10605 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10606 BN_GF2m_mod_exp_arr
10607 BN_GF2m_mod_sqrt_arr
10608 BN_GF2m_mod_solve_quad_arr
10612 Conversion can be performed by the following functions:
10617 bntest.c has additional tests for binary polynomial arithmetic.
10619 Two implementations for BN_GF2m_mod_div() are available.
10620 The default algorithm simply uses BN_GF2m_mod_inv() and
10621 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10622 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10623 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10625 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10627 * Add new error code 'ERR_R_DISABLED' that can be used when some
10628 functionality is disabled at compile-time.
10630 *Douglas Stebila <douglas.stebila@sun.com>*
10632 * Change default behaviour of 'openssl asn1parse' so that more
10633 information is visible when viewing, e.g., a certificate:
10635 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10636 mode the content of non-printable OCTET STRINGs is output in a
10637 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10638 avoid the appearance of a printable string.
10640 *Nils Larsch <nla@trustcenter.de>*
10642 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10644 EC_GROUP_set_asn1_flag()
10645 EC_GROUP_get_asn1_flag()
10646 EC_GROUP_set_point_conversion_form()
10647 EC_GROUP_get_point_conversion_form()
10648 These control ASN1 encoding details:
10649 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10650 has been set to OPENSSL_EC_NAMED_CURVE.
10651 - Points are encoded in uncompressed form by default; options for
10652 asn1_for are as for point2oct, namely
10653 POINT_CONVERSION_COMPRESSED
10654 POINT_CONVERSION_UNCOMPRESSED
10655 POINT_CONVERSION_HYBRID
10657 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10659 EC_GROUP_set_seed()
10660 EC_GROUP_get0_seed()
10661 EC_GROUP_get_seed_len()
10662 This is used only for ASN1 purposes (so far).
10664 *Nils Larsch <nla@trustcenter.de>*
10666 * Add 'field_type' member to EC_METHOD, which holds the NID
10667 of the appropriate field type OID. The new function
10668 EC_METHOD_get_field_type() returns this value.
10670 *Nils Larsch <nla@trustcenter.de>*
10673 EC_POINT_point2bn()
10674 EC_POINT_bn2point()
10675 EC_POINT_point2hex()
10676 EC_POINT_hex2point()
10677 providing useful interfaces to EC_POINT_point2oct() and
10678 EC_POINT_oct2point().
10680 *Nils Larsch <nla@trustcenter.de>*
10682 * Change internals of the EC library so that the functions
10683 EC_GROUP_set_generator()
10684 EC_GROUP_get_generator()
10685 EC_GROUP_get_order()
10686 EC_GROUP_get_cofactor()
10687 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10688 to methods, which would lead to unnecessary code duplication when
10689 adding different types of curves.
10691 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10693 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10694 arithmetic, and such that modified wNAFs are generated
10695 (which avoid length expansion in many cases).
10699 * Add a function EC_GROUP_check_discriminant() (defined via
10700 EC_METHOD) that verifies that the curve discriminant is non-zero.
10702 Add a function EC_GROUP_check() that makes some sanity tests
10703 on a EC_GROUP, its generator and order. This includes
10704 EC_GROUP_check_discriminant().
10706 *Nils Larsch <nla@trustcenter.de>*
10708 * Add ECDSA in new directory crypto/ecdsa/.
10710 Add applications 'openssl ecparam' and 'openssl ecdsa'
10711 (these are based on 'openssl dsaparam' and 'openssl dsa').
10713 ECDSA support is also included in various other files across the
10714 library. Most notably,
10715 - 'openssl req' now has a '-newkey ecdsa:file' option;
10716 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10717 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10718 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10719 them suitable for ECDSA where domain parameters must be
10720 extracted before the specific public key;
10721 - ECDSA engine support has been added.
10723 *Nils Larsch <nla@trustcenter.de>*
10725 * Include some named elliptic curves, and add OIDs from X9.62,
10726 SECG, and WAP/WTLS. Each curve can be obtained from the new
10728 EC_GROUP_new_by_curve_name(),
10729 and the list of available named curves can be obtained with
10730 EC_get_builtin_curves().
10731 Also add a 'curve_name' member to EC_GROUP objects, which can be
10733 EC_GROUP_set_curve_name()
10734 EC_GROUP_get_curve_name()
10736 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10738 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10739 was actually never needed) and in BN_mul(). The removal in BN_mul()
10740 required a small change in bn_mul_part_recursive() and the addition
10741 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10742 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10743 bn_sub_words() and bn_add_words() except they take arrays with
10748 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10750 * Cleanse PEM buffers before freeing them since they may contain
10753 *Benjamin Bennett <ben@psc.edu>*
10755 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10756 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10757 authentication-only ciphersuites.
10761 * Since AES128 and AES256 share a single mask bit in the logic of
10762 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10763 kludge to work properly if AES128 is available and AES256 isn't.
10767 * Expand security boundary to match 1.1.1 module.
10771 * Remove redundant features: hash file source, editing of test vectors
10772 modify fipsld to use external fips_premain.c signature.
10776 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10777 run algorithm test programs.
10781 * Make algorithm test programs more tolerant of whitespace.
10785 * Have SSL/TLS server implementation tolerate "mismatched" record
10786 protocol version while receiving ClientHello even if the
10787 ClientHello is fragmented. (The server can't insist on the
10788 particular protocol version it has chosen before the ServerHello
10789 message has informed the client about his choice.)
10793 * Load error codes if they are not already present instead of using a
10794 static variable. This allows them to be cleanly unloaded and reloaded.
10798 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10800 * Introduce limits to prevent malicious keys being able to
10801 cause a denial of service. ([CVE-2006-2940])
10803 *Steve Henson, Bodo Moeller*
10805 * Fix ASN.1 parsing of certain invalid structures that can result
10806 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10808 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10809 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10811 * Fix SSL client code which could crash if connecting to a
10812 malicious SSLv2 server. ([CVE-2006-4343])
10814 *Tavis Ormandy and Will Drewry, Google Security Team*
10816 * Change ciphersuite string processing so that an explicit
10817 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10818 will no longer include "AES128-SHA"), and any other similar
10819 ciphersuite (same bitmap) from *other* protocol versions (so that
10820 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10821 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10822 changes from 0.9.8b and 0.9.8d.
10826 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10828 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10829 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10831 * Change the Unix randomness entropy gathering to use poll() when
10832 possible instead of select(), since the latter has some
10833 undesirable limitations.
10835 *Darryl Miles via Richard Levitte and Bodo Moeller*
10837 * Disable rogue ciphersuites:
10839 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10840 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10841 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10843 The latter two were purportedly from
10844 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10847 Also deactivate the remaining ciphersuites from
10848 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10849 unofficial, and the ID has long expired.
10853 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10854 dual-core machines) and other potential thread-safety issues.
10858 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10860 * Adapt fipsld and the build system to link against the validated FIPS
10861 module in FIPS mode.
10865 * Fixes for VC++ 2005 build under Windows.
10869 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10870 from a Windows bash shell such as MSYS. It is autodetected from the
10871 "config" script when run from a VC++ environment. Modify standard VC++
10872 build to use fipscanister.o from the GNU make build.
10876 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10878 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10879 The value now differs depending on if you build for FIPS or not.
10880 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10881 safely run with a non-FIPSed libcrypto, as it may crash because of
10882 the difference induced by this change.
10886 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10888 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10889 (part of SSL_OP_ALL). This option used to disable the
10890 countermeasure against man-in-the-middle protocol-version
10891 rollback in the SSL 2.0 server implementation, which is a bad
10892 idea. ([CVE-2005-2969])
10894 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10895 for Information Security, National Institute of Advanced Industrial
10896 Science and Technology [AIST, Japan)]*
10898 * Minimal support for X9.31 signatures and PSS padding modes. This is
10899 mainly for FIPS compliance and not fully integrated at this stage.
10903 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10904 the exponentiation using a fixed-length exponent. (Otherwise,
10905 the information leaked through timing could expose the secret key
10906 after many signatures; cf. Bleichenbacher's attack on DSA with
10911 * Make a new fixed-window mod_exp implementation the default for
10912 RSA, DSA, and DH private-key operations so that the sequence of
10913 squares and multiplies and the memory access pattern are
10914 independent of the particular secret key. This will mitigate
10915 cache-timing and potential related attacks.
10917 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10918 and this is automatically used by BN_mod_exp_mont() if the new flag
10919 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10920 will use this BN flag for private exponents unless the flag
10921 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10922 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10924 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10926 * Change the client implementation for SSLv23_method() and
10927 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10928 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10929 (Previously, the SSL 2.0 backwards compatible Client Hello
10930 message format would be used even with SSL_OP_NO_SSLv2.)
10934 * Add support for smime-type MIME parameter in S/MIME messages which some
10939 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10940 a threadsafe manner. Modify rsa code to use new function and add calls
10941 to dsa and dh code (which had race conditions before).
10945 * Include the fixed error library code in the C error file definitions
10946 instead of fixing them up at runtime. This keeps the error code
10947 structures constant.
10951 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10953 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10956 * Fixes for newer kerberos headers. NB: the casts are needed because
10957 the 'length' field is signed on one version and unsigned on another
10958 with no (?) obvious way to tell the difference, without these VC++
10959 complains. Also the "definition" of FAR (blank) is no longer included
10960 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10961 some needed definitions.
10965 * Undo Cygwin change.
10969 * Added support for proxy certificates according to RFC 3820.
10970 Because they may be a security thread to unaware applications,
10971 they must be explicitly allowed in run-time. See
10972 docs/HOWTO/proxy_certificates.txt for further information.
10976 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
10978 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
10979 server and client random values. Previously
10980 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
10981 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
10983 This change has negligible security impact because:
10985 1. Server and client random values still have 24 bytes of pseudo random
10988 2. Server and client random values are sent in the clear in the initial
10991 3. The master secret is derived using the premaster secret (48 bytes in
10992 size for static RSA ciphersuites) as well as client server and random
10995 The OpenSSL team would like to thank the UK NISCC for bringing this issue
10998 *Stephen Henson, reported by UK NISCC*
11000 * Use Windows randomness collection on Cygwin.
11004 * Fix hang in EGD/PRNGD query when communication socket is closed
11005 prematurely by EGD/PRNGD.
11007 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11009 * Prompt for pass phrases when appropriate for PKCS12 input format.
11013 * Back-port of selected performance improvements from development
11014 branch, as well as improved support for PowerPC platforms.
11018 * Add lots of checks for memory allocation failure, error codes to indicate
11019 failure and freeing up memory if a failure occurs.
11021 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11023 * Add new -passin argument to dgst.
11027 * Perform some character comparisons of different types in X509_NAME_cmp:
11028 this is needed for some certificates that re-encode DNs into UTF8Strings
11029 (in violation of RFC3280) and can't or won't issue name rollover
11034 * Make an explicit check during certificate validation to see that
11035 the CA setting in each certificate on the chain is correct. As a
11036 side effect always do the following basic checks on extensions,
11037 not just when there's an associated purpose to the check:
11039 - if there is an unhandled critical extension (unless the user
11040 has chosen to ignore this fault)
11041 - if the path length has been exceeded (if one is set at all)
11042 - that certain extensions fit the associated purpose (if one has
11047 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11049 * Avoid a race condition when CRLs are checked in a multi threaded
11050 environment. This would happen due to the reordering of the revoked
11051 entries during signature checking and serial number lookup. Now the
11052 encoding is cached and the serial number sort performed under a lock.
11053 Add new STACK function sk_is_sorted().
11057 * Add Delta CRL to the extension code.
11061 * Various fixes to s3_pkt.c so alerts are sent properly.
11063 *David Holmes <d.holmes@f5.com>*
11065 * Reduce the chances of duplicate issuer name and serial numbers (in
11066 violation of RFC3280) using the OpenSSL certificate creation utilities.
11067 This is done by creating a random 64 bit value for the initial serial
11068 number when a serial number file is created or when a self signed
11069 certificate is created using 'openssl req -x509'. The initial serial
11070 number file is created using 'openssl x509 -next_serial' in CA.pl
11071 rather than being initialized to 1.
11075 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11077 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11078 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11080 *Joe Orton, Steve Henson*
11082 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11085 *Joe Orton, Steve Henson*
11087 * Make it possible to have multiple active certificates with the same
11088 subject in the CA index file. This is done only if the keyword
11089 'unique_subject' is set to 'no' in the main CA section (default
11090 if 'CA_default') of the configuration file. The value is saved
11091 with the database itself in a separate index attribute file,
11092 named like the index file with '.attr' appended to the name.
11096 * X509 verify fixes. Disable broken certificate workarounds when
11097 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11098 keyUsage extension present. Don't accept CRLs with unhandled critical
11099 extensions: since verify currently doesn't process CRL extensions this
11100 rejects a CRL with *any* critical extensions. Add new verify error codes
11105 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11106 A clarification of RFC2560 will require the use of OCTET STRINGs and
11107 some implementations cannot handle the current raw format. Since OpenSSL
11108 copies and compares OCSP nonces as opaque blobs without any attempt at
11109 parsing them this should not create any compatibility issues.
11113 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11114 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11115 this HMAC (and other) operations are several times slower than OpenSSL
11120 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11122 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11124 * Use the correct content when signing type "other".
11128 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11130 * Fix various bugs revealed by running the NISCC test suite:
11132 Stop out of bounds reads in the ASN1 code when presented with
11133 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11135 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11137 If verify callback ignores invalid public key errors don't try to check
11138 certificate signature with the NULL public key.
11142 * New -ignore_err option in ocsp application to stop the server
11143 exiting on the first error in a request.
11147 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11148 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11153 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11154 extra data after the compression methods not only for TLS 1.0
11155 but also for SSL 3.0 (as required by the specification).
11157 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11159 * Change X509_certificate_type() to mark the key as exported/exportable
11160 when it's 512 *bits* long, not 512 bytes.
11164 * Change AES_cbc_encrypt() so it outputs exact multiple of
11165 blocks during encryption.
11169 * Various fixes to base64 BIO and non blocking I/O. On write
11170 flushes were not handled properly if the BIO retried. On read
11171 data was not being buffered properly and had various logic bugs.
11172 This also affects blocking I/O when the data being decoded is a
11177 * Various S/MIME bugfixes and compatibility changes:
11178 output correct application/pkcs7 MIME type if
11179 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11180 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11181 of files as .eml work). Correctly handle very long lines in MIME
11186 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11188 * Countermeasure against the Klima-Pokorny-Rosa extension of
11189 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11190 a protocol version number mismatch like a decryption error
11191 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11195 * Turn on RSA blinding by default in the default implementation
11196 to avoid a timing attack. Applications that don't want it can call
11197 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11198 They would be ill-advised to do so in most cases.
11200 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11202 * Change RSA blinding code so that it works when the PRNG is not
11203 seeded (in this case, the secret RSA exponent is abused as
11204 an unpredictable seed -- if it is not unpredictable, there
11205 is no point in blinding anyway). Make RSA blinding thread-safe
11206 by remembering the creator's thread ID in rsa->blinding and
11207 having all other threads use local one-time blinding factors
11208 (this requires more computation than sharing rsa->blinding, but
11209 avoids excessive locking; and if an RSA object is not shared
11210 between threads, blinding will still be very fast).
11214 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11215 ENGINE as defaults for all supported algorithms irrespective of
11216 the 'flags' parameter. 'flags' is now honoured, so applications
11217 should make sure they are passing it correctly.
11221 * Target "mingw" now allows native Windows code to be generated in
11222 the Cygwin environment as well as with the MinGW compiler.
11226 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11228 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11229 via timing by performing a MAC computation even if incorrect
11230 block cipher padding has been found. This is a countermeasure
11231 against active attacks where the attacker has to distinguish
11232 between bad padding and a MAC verification error. ([CVE-2003-0078])
11234 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11235 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11236 Martin Vuagnoux (EPFL, Ilion)*
11238 * Make the no-err option work as intended. The intention with no-err
11239 is not to have the whole error stack handling routines removed from
11240 libcrypto, it's only intended to remove all the function name and
11241 reason texts, thereby removing some of the footprint that may not
11242 be interesting if those errors aren't displayed anyway.
11244 NOTE: it's still possible for any application or module to have its
11245 own set of error texts inserted. The routines are there, just not
11246 used by default when no-err is given.
11250 * Add support for FreeBSD on IA64.
11252 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11254 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11255 Kerberos function mit_des_cbc_cksum(). Before this change,
11256 the value returned by DES_cbc_cksum() was like the one from
11257 mit_des_cbc_cksum(), except the bytes were swapped.
11259 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11261 * Allow an application to disable the automatic SSL chain building.
11262 Before this a rather primitive chain build was always performed in
11263 ssl3_output_cert_chain(): an application had no way to send the
11264 correct chain if the automatic operation produced an incorrect result.
11266 Now the chain builder is disabled if either:
11268 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11270 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11272 The reasoning behind this is that an application would not want the
11273 auto chain building to take place if extra chain certificates are
11274 present and it might also want a means of sending no additional
11275 certificates (for example the chain has two certificates and the
11280 * Add the possibility to build without the ENGINE framework.
11282 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11284 * Under Win32 gmtime() can return NULL: check return value in
11285 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11289 * DSA routines: under certain error conditions uninitialized BN objects
11290 could be freed. Solution: make sure initialization is performed early
11291 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11292 Nils Larsch <nla@trustcenter.de> via PR#459)
11296 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11297 checked on reconnect on the client side, therefore session resumption
11298 could still fail with a "ssl session id is different" error. This
11299 behaviour is masked when SSL_OP_ALL is used due to
11300 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11301 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11302 followup to PR #377.
11306 * IA-32 assembler support enhancements: unified ELF targets, support
11307 for SCO/Caldera platforms, fix for Cygwin shared build.
11311 * Add support for FreeBSD on sparc64. As a consequence, support for
11312 FreeBSD on non-x86 processors is separate from x86 processors on
11313 the config script, much like the NetBSD support.
11315 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11317 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11319 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11322 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11323 code (06) was taken as the first octet of the session ID and the last
11324 octet was ignored consequently. As a result SSLv2 client side session
11325 caching could not have worked due to the session ID mismatch between
11327 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11332 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11333 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11338 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11339 seems that in spite of existing for more than a year, many application
11340 author have done nothing to provide the necessary callbacks, which
11341 means that this particular engine will not work properly anywhere.
11342 This is a very unfortunate situation which forces us, in the name
11343 of usability, to give the hw_ncipher.c a static lock, which is part
11345 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11346 appear in 0.9.8 or later. We EXPECT application authors to have
11347 dealt properly with this when 0.9.8 is released (unless we actually
11348 make such changes in the libcrypto locking code that changes will
11349 have to be made anyway).
11353 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11354 octets have been read, EOF or an error occurs. Without this change
11355 some truncated ASN1 structures will not produce an error.
11359 * Disable Heimdal support, since it hasn't been fully implemented.
11360 Still give the possibility to force the use of Heimdal, but with
11361 warnings and a request that patches get sent to openssl-dev.
11365 * Add the VC-CE target, introduce the WINCE sysname, and add
11366 INSTALL.WCE and appropriate conditionals to make it build.
11368 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11370 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11371 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11372 edit numbers of the version.
11374 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11376 * Introduce safe string copy and catenation functions
11377 (BUF_strlcpy() and BUF_strlcat()).
11379 *Ben Laurie (CHATS) and Richard Levitte*
11381 * Avoid using fixed-size buffers for one-line DNs.
11383 *Ben Laurie (CHATS)*
11385 * Add BUF_MEM_grow_clean() to avoid information leakage when
11386 resizing buffers containing secrets, and use where appropriate.
11388 *Ben Laurie (CHATS)*
11390 * Avoid using fixed size buffers for configuration file location.
11392 *Ben Laurie (CHATS)*
11394 * Avoid filename truncation for various CA files.
11396 *Ben Laurie (CHATS)*
11398 * Use sizeof in preference to magic numbers.
11400 *Ben Laurie (CHATS)*
11402 * Avoid filename truncation in cert requests.
11404 *Ben Laurie (CHATS)*
11406 * Add assertions to check for (supposedly impossible) buffer
11409 *Ben Laurie (CHATS)*
11411 * Don't cache truncated DNS entries in the local cache (this could
11412 potentially lead to a spoofing attack).
11414 *Ben Laurie (CHATS)*
11416 * Fix various buffers to be large enough for hex/decimal
11417 representations in a platform independent manner.
11419 *Ben Laurie (CHATS)*
11421 * Add CRYPTO_realloc_clean() to avoid information leakage when
11422 resizing buffers containing secrets, and use where appropriate.
11424 *Ben Laurie (CHATS)*
11426 * Add BIO_indent() to avoid much slightly worrying code to do
11429 *Ben Laurie (CHATS)*
11431 * Convert sprintf()/BIO_puts() to BIO_printf().
11433 *Ben Laurie (CHATS)*
11435 * buffer_gets() could terminate with the buffer only half
11438 *Ben Laurie (CHATS)*
11440 * Add assertions to prevent user-supplied crypto functions from
11441 overflowing internal buffers by having large block sizes, etc.
11443 *Ben Laurie (CHATS)*
11445 * New OPENSSL_assert() macro (similar to assert(), but enabled
11448 *Ben Laurie (CHATS)*
11450 * Eliminate unused copy of key in RC4.
11452 *Ben Laurie (CHATS)*
11454 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11456 *Ben Laurie (CHATS)*
11458 * Fix off-by-one error in EGD path.
11460 *Ben Laurie (CHATS)*
11462 * If RANDFILE path is too long, ignore instead of truncating.
11464 *Ben Laurie (CHATS)*
11466 * Eliminate unused and incorrectly sized X.509 structure
11469 *Ben Laurie (CHATS)*
11471 * Eliminate unused and dangerous function knumber().
11473 *Ben Laurie (CHATS)*
11475 * Eliminate unused and dangerous structure, KSSL_ERR.
11477 *Ben Laurie (CHATS)*
11479 * Protect against overlong session ID context length in an encoded
11480 session object. Since these are local, this does not appear to be
11483 *Ben Laurie (CHATS)*
11485 * Change from security patch (see 0.9.6e below) that did not affect
11486 the 0.9.6 release series:
11488 Remote buffer overflow in SSL3 protocol - an attacker could
11489 supply an oversized master key in Kerberos-enabled versions.
11492 *Ben Laurie (CHATS)*
11494 * Change the SSL kerb5 codes to match RFC 2712.
11498 * Make -nameopt work fully for req and add -reqopt switch.
11500 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11502 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11504 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11506 * Make sure tests can be performed even if the corresponding algorithms
11507 have been removed entirely. This was also the last step to make
11508 OpenSSL compilable with DJGPP under all reasonable conditions.
11510 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11512 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11513 to allow version independent disabling of normally unselected ciphers,
11514 which may be activated as a side-effect of selecting a single cipher.
11516 (E.g., cipher list string "RSA" enables ciphersuites that are left
11517 out of "ALL" because they do not provide symmetric encryption.
11518 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11520 *Lutz Jaenicke, Bodo Moeller*
11522 * Add appropriate support for separate platform-dependent build
11523 directories. The recommended way to make a platform-dependent
11524 build directory is the following (tested on Linux), maybe with
11527 # Place yourself outside of the OpenSSL source tree. In
11528 # this example, the environment variable OPENSSL_SOURCE
11529 # is assumed to contain the absolute OpenSSL source directory.
11530 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11531 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11532 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11533 mkdir -p `dirname $F`
11534 ln -s $OPENSSL_SOURCE/$F $F
11537 To be absolutely sure not to disturb the source tree, a "make clean"
11538 is a good thing. If it isn't successful, don't worry about it,
11539 it probably means the source directory is very clean.
11543 * Make sure any ENGINE control commands make local copies of string
11544 pointers passed to them whenever necessary. Otherwise it is possible
11545 the caller may have overwritten (or deallocated) the original string
11546 data when a later ENGINE operation tries to use the stored values.
11548 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11550 * Improve diagnostics in file reading and command-line digests.
11552 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11554 * Add AES modes CFB and OFB to the object database. Correct an
11555 error in AES-CFB decryption.
11559 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11560 allows existing EVP_CIPHER_CTX structures to be reused after
11561 calling `EVP_*Final()`. This behaviour is used by encryption
11562 BIOs and some applications. This has the side effect that
11563 applications must explicitly clean up cipher contexts with
11564 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11568 * Check the values of dna and dnb in bn_mul_recursive before calling
11569 bn_mul_comba (a non zero value means the a or b arrays do not contain
11570 n2 elements) and fallback to bn_mul_normal if either is not zero.
11574 * Fix escaping of non-ASCII characters when using the -subj option
11575 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11579 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11580 form for "surname", serialNumber has no short form.
11581 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11582 therefore remove "mail" short name for "internet 7".
11583 The OID for unique identifiers in X509 certificates is
11584 x500UniqueIdentifier, not uniqueIdentifier.
11585 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11589 * Add an "init" command to the ENGINE config module and auto initialize
11590 ENGINEs. Without any "init" command the ENGINE will be initialized
11591 after all ctrl commands have been executed on it. If init=1 the
11592 ENGINE is initialized at that point (ctrls before that point are run
11593 on the uninitialized ENGINE and after on the initialized one). If
11594 init=0 then the ENGINE will not be initialized at all.
11598 * Fix the 'app_verify_callback' interface so that the user-defined
11599 argument is actually passed to the callback: In the
11600 SSL_CTX_set_cert_verify_callback() prototype, the callback
11601 declaration has been changed from
11604 int (*cb)(X509_STORE_CTX *,void *);
11605 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11606 i=s->ctx->app_verify_callback(&ctx)
11607 has been changed into
11608 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11610 To update applications using SSL_CTX_set_cert_verify_callback(),
11611 a dummy argument can be added to their callback functions.
11613 *D. K. Smetters <smetters@parc.xerox.com>*
11615 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11617 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11619 * Add and OPENSSL_LOAD_CONF define which will cause
11620 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11621 This allows older applications to transparently support certain
11622 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11623 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11624 load the config file and OPENSSL_add_all_algorithms_conf() which will
11625 always load it have also been added.
11629 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11630 Adjust NIDs and EVP layer.
11632 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11634 * Config modules support in openssl utility.
11636 Most commands now load modules from the config file,
11637 though in a few (such as version) this isn't done
11638 because it couldn't be used for anything.
11640 In the case of ca and req the config file used is
11641 the same as the utility itself: that is the -config
11642 command line option can be used to specify an
11647 * Move default behaviour from OPENSSL_config(). If appname is NULL
11648 use "openssl_conf" if filename is NULL use default openssl config file.
11652 * Add an argument to OPENSSL_config() to allow the use of an alternative
11653 config section name. Add a new flag to tolerate a missing config file
11654 and move code to CONF_modules_load_file().
11658 * Support for crypto accelerator cards from Accelerated Encryption
11659 Processing, www.aep.ie. (Use engine 'aep')
11660 The support was copied from 0.9.6c [engine] and adapted/corrected
11661 to work with the new engine framework.
11663 *AEP Inc. and Richard Levitte*
11665 * Support for SureWare crypto accelerator cards from Baltimore
11666 Technologies. (Use engine 'sureware')
11667 The support was copied from 0.9.6c [engine] and adapted
11668 to work with the new engine framework.
11672 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11673 make the newer ENGINE framework commands for the CHIL engine work.
11675 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11677 * Make it possible to produce shared libraries on ReliantUNIX.
11679 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11681 * Add the configuration target debug-linux-ppro.
11682 Make 'openssl rsa' use the general key loading routines
11683 implemented in apps.c, and make those routines able to
11684 handle the key format FORMAT_NETSCAPE and the variant
11687 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11689 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11691 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11693 * Add -keyform to rsautl, and document -engine.
11695 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11697 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11698 BIO_R_NO_SUCH_FILE error code rather than the generic
11699 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11703 * Add new functions
11704 ERR_peek_last_error
11705 ERR_peek_last_error_line
11706 ERR_peek_last_error_line_data.
11707 These are similar to
11709 ERR_peek_error_line
11710 ERR_peek_error_line_data,
11711 but report on the latest error recorded rather than the first one
11712 still in the error queue.
11714 *Ben Laurie, Bodo Moeller*
11716 * default_algorithms option in ENGINE config module. This allows things
11718 default_algorithms = ALL
11719 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11723 * Preliminary ENGINE config module.
11727 * New experimental application configuration code.
11731 * Change the AES code to follow the same name structure as all other
11732 symmetric ciphers, and behave the same way. Move everything to
11733 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11735 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11737 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11739 *Ben Laurie and Theo de Raadt*
11741 * Add option to output public keys in req command.
11743 *Massimiliano Pala madwolf@openca.org*
11745 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11746 (up to about 10% better than before for P-192 and P-224).
11750 * New functions/macros
11752 SSL_CTX_set_msg_callback(ctx, cb)
11753 SSL_CTX_set_msg_callback_arg(ctx, arg)
11754 SSL_set_msg_callback(ssl, cb)
11755 SSL_set_msg_callback_arg(ssl, arg)
11757 to request calling a callback function
11759 void cb(int write_p, int version, int content_type,
11760 const void *buf, size_t len, SSL *ssl, void *arg)
11762 whenever a protocol message has been completely received
11763 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11764 protocol version according to which the SSL library interprets
11765 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11766 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11767 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11768 specification (change_cipher_spec(20), alert(21), handshake(22)).
11769 'buf' and 'len' point to the actual message, 'ssl' to the
11770 SSL object, and 'arg' is the application-defined value set by
11771 SSL[_CTX]_set_msg_callback_arg().
11773 'openssl s_client' and 'openssl s_server' have new '-msg' options
11774 to enable a callback that displays all protocol messages.
11778 * Change the shared library support so shared libraries are built as
11779 soon as the corresponding static library is finished, and thereby get
11780 openssl and the test programs linked against the shared library.
11781 This still only happens when the keyword "shard" has been given to
11782 the configuration scripts.
11784 NOTE: shared library support is still an experimental thing, and
11785 backward binary compatibility is still not guaranteed.
11787 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11789 * Add support for Subject Information Access extension.
11791 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11793 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11794 additional bytes when new memory had to be allocated, not just
11795 when reusing an existing buffer.
11799 * New command line and configuration option 'utf8' for the req command.
11800 This allows field values to be specified as UTF8 strings.
11804 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11805 runs for the former and machine-readable output for the latter.
11809 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11810 of the e-mail address in the DN (i.e., it will go into a certificate
11811 extension only). The new configuration file option 'email_in_dn = no'
11812 has the same effect.
11814 *Massimiliano Pala madwolf@openca.org*
11816 * Change all functions with names starting with `des_` to be starting
11817 with `DES_` instead. Add wrappers that are compatible with libdes,
11818 but are named `_ossl_old_des_*`. Finally, add macros that map the
11819 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11820 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11821 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11824 Since we provide two compatibility mappings, the user needs to
11825 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11826 compatibility is desired. The default (i.e., when that macro
11827 isn't defined) is OpenSSL 0.9.6c compatibility.
11829 There are also macros that enable and disable the support of old
11830 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11831 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11832 are defined, the default will apply: to support the old des routines.
11834 In either case, one must include openssl/des.h to get the correct
11835 definitions. Do not try to just include openssl/des_old.h, that
11838 NOTE: This is a major break of an old API into a new one. Software
11839 authors are encouraged to switch to the `DES_` style functions. Some
11840 time in the future, des_old.h and the libdes compatibility functions
11841 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11842 default), and then completely removed.
11846 * Test for certificates which contain unsupported critical extensions.
11847 If such a certificate is found during a verify operation it is
11848 rejected by default: this behaviour can be overridden by either
11849 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11850 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11851 X509_supported_extension() has also been added which returns 1 if a
11852 particular extension is supported.
11856 * Modify the behaviour of EVP cipher functions in similar way to digests
11857 to retain compatibility with existing code.
11861 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11862 compatibility with existing code. In particular the 'ctx' parameter does
11863 not have to be to be initialized before the call to EVP_DigestInit() and
11864 it is tidied up after a call to EVP_DigestFinal(). New function
11865 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11866 EVP_MD_CTX_copy() changed to not require the destination to be
11867 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11868 requires the destination to be valid.
11870 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11871 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11875 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11876 so that complete 'Handshake' protocol structures are kept in memory
11877 instead of overwriting 'msg_type' and 'length' with 'body' data.
11881 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11883 *Massimo Santin via Richard Levitte*
11885 * Major restructuring to the underlying ENGINE code. This includes
11886 reduction of linker bloat, separation of pure "ENGINE" manipulation
11887 (initialisation, etc) from functionality dealing with implementations
11888 of specific crypto interfaces. This change also introduces integrated
11889 support for symmetric ciphers and digest implementations - so ENGINEs
11890 can now accelerate these by providing EVP_CIPHER and EVP_MD
11891 implementations of their own. This is detailed in
11892 [crypto/engine/README.md](crypto/engine/README.md)
11893 as it couldn't be adequately described here. However, there are a few
11894 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11895 were changed in the original introduction of ENGINE code have now
11896 reverted back - the hooking from this code to ENGINE is now a good
11897 deal more passive and at run-time, operations deal directly with
11898 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11899 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11900 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11901 they were not being used by the framework as there is no concept of a
11902 BIGNUM_METHOD and they could not be generalised to the new
11903 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11904 ENGINE_cpy() has been removed as it cannot be consistently defined in
11909 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11913 * Change mkdef.pl to sort symbols that get the same entry number,
11914 and make sure the automatically generated functions `ERR_load_*`
11915 become part of libeay.num as well.
11919 * New function SSL_renegotiate_pending(). This returns true once
11920 renegotiation has been requested (either SSL_renegotiate() call
11921 or HelloRequest/ClientHello received from the peer) and becomes
11922 false once a handshake has been completed.
11923 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11924 sends a HelloRequest, but does not ensure that a handshake takes
11925 place. SSL_renegotiate_pending() is useful for checking if the
11926 client has followed the request.)
11930 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11931 By default, clients may request session resumption even during
11932 renegotiation (if session ID contexts permit); with this option,
11933 session resumption is possible only in the first handshake.
11935 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11936 more bits available for options that should not be part of
11937 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11941 * Add some demos for certificate and certificate request creation.
11945 * Make maximum certificate chain size accepted from the peer application
11946 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11947 "Douglas E. Engert" <deengert@anl.gov>.
11951 * Add support for shared libraries for Unixware-7
11952 (Boyd Lynn Gerber <gerberb@zenez.com>).
11956 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11957 be done prior to destruction. Use this to unload error strings from
11958 ENGINEs that load their own error strings. NB: This adds two new API
11959 functions to "get" and "set" this destroy handler in an ENGINE.
11963 * Alter all existing ENGINE implementations (except "openssl" and
11964 "openbsd") to dynamically instantiate their own error strings. This
11965 makes them more flexible to be built both as statically-linked ENGINEs
11966 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11967 Also, add stub code to each that makes building them as self-contained
11968 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11972 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11973 implementations into applications that are completely implemented in
11974 self-contained shared-libraries. The "dynamic" ENGINE exposes control
11975 commands that can be used to configure what shared-library to load and
11976 to control aspects of the way it is handled. Also, made an update to
11977 the [README-Engine.md](README-Engine.md) file
11978 that brings its information up-to-date and
11979 provides some information and instructions on the "dynamic" ENGINE
11980 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
11984 * Make it possible to unload ranges of ERR strings with a new
11985 "ERR_unload_strings" function.
11989 * Add a copy() function to EVP_MD.
11993 * Make EVP_MD routines take a context pointer instead of just the
11994 md_data void pointer.
11998 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
11999 that the digest can only process a single chunk of data
12000 (typically because it is provided by a piece of
12001 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12002 is only going to provide a single chunk of data, and hence the
12003 framework needn't accumulate the data for oneshot drivers.
12007 * As with "ERR", make it possible to replace the underlying "ex_data"
12008 functions. This change also alters the storage and management of global
12009 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12010 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12011 index counters. The API functions that use this state have been changed
12012 to take a "class_index" rather than pointers to the class's local STACK
12013 and counter, and there is now an API function to dynamically create new
12014 classes. This centralisation allows us to (a) plug a lot of the
12015 thread-safety problems that existed, and (b) makes it possible to clean
12016 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12017 such data would previously have always leaked in application code and
12018 workarounds were in place to make the memory debugging turn a blind eye
12019 to it. Application code that doesn't use this new function will still
12020 leak as before, but their memory debugging output will announce it now
12021 rather than letting it slide.
12023 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12024 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12025 has a return value to indicate success or failure.
12029 * Make it possible to replace the underlying "ERR" functions such that the
12030 global state (2 LHASH tables and 2 locks) is only used by the "default"
12031 implementation. This change also adds two functions to "get" and "set"
12032 the implementation prior to it being automatically set the first time
12033 any other ERR function takes place. Ie. an application can call "get",
12034 pass the return value to a module it has just loaded, and that module
12035 can call its own "set" function using that value. This means the
12036 module's "ERR" operations will use (and modify) the error state in the
12037 application and not in its own statically linked copy of OpenSSL code.
12041 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12042 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12043 the operation, and provides a more encapsulated way for external code
12044 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12045 to use these functions rather than manually incrementing the counts.
12047 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12051 * Add EVP test program.
12055 * Add symmetric cipher support to ENGINE. Expect the API to change!
12059 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12060 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12061 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12062 These allow a CRL to be built without having to access X509_CRL fields
12063 directly. Modify 'ca' application to use new functions.
12067 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12068 bug workarounds. Rollback attack detection is a security feature.
12069 The problem will only arise on OpenSSL servers when TLSv1 is not
12070 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12071 Software authors not wanting to support TLSv1 will have special reasons
12072 for their choice and can explicitly enable this option.
12074 *Bodo Moeller, Lutz Jaenicke*
12076 * Rationalise EVP so it can be extended: don't include a union of
12077 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12078 (similar to those existing for EVP_CIPHER_CTX).
12083 EVP_MD_CTX_init(&md); /* new function call */
12084 EVP_DigestInit(&md, EVP_sha1());
12085 EVP_DigestUpdate(&md, in, len);
12086 EVP_DigestFinal(&md, out, NULL);
12087 EVP_MD_CTX_cleanup(&md); /* new function call */
12091 * Make DES key schedule conform to the usual scheme, as well as
12092 correcting its structure. This means that calls to DES functions
12093 now have to pass a pointer to a des_key_schedule instead of a
12094 plain des_key_schedule (which was actually always a pointer
12097 des_key_schedule ks;
12099 des_set_key_checked(..., &ks);
12100 des_ncbc_encrypt(..., &ks, ...);
12102 (Note that a later change renames 'des_...' into 'DES_...'.)
12106 * Initial reduction of linker bloat: the use of some functions, such as
12107 PEM causes large amounts of unused functions to be linked in due to
12108 poor organisation. For example pem_all.c contains every PEM function
12109 which has a knock on effect of linking in large amounts of (unused)
12110 ASN1 code. Grouping together similar functions and splitting unrelated
12111 functions prevents this.
12115 * Cleanup of EVP macros.
12119 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12120 correct `_ecb suffix`.
12124 * Add initial OCSP responder support to ocsp application. The
12125 revocation information is handled using the text based index
12126 use by the ca application. The responder can either handle
12127 requests generated internally, supplied in files (for example
12128 via a CGI script) or using an internal minimal server.
12132 * Add configuration choices to get zlib compression for TLS.
12136 * Changes to Kerberos SSL for RFC 2712 compliance:
12137 1. Implemented real KerberosWrapper, instead of just using
12138 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12139 2. Implemented optional authenticator field of KerberosWrapper.
12141 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12142 and authenticator structs; see crypto/krb5/.
12144 Generalized Kerberos calls to support multiple Kerberos libraries.
12145 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12146 via Richard Levitte*
12148 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12149 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12150 values for each of the key sizes rather than having just
12151 parameters (and 'speed' generating keys each time).
12155 * Speed up EVP routines.
12158 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12159 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12160 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12161 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12163 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12164 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12165 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12168 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12170 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12174 * Added the OS2-EMX target.
12176 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12178 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12179 to support NCONF routines in extension code. New function CONF_set_nconf()
12180 to allow functions which take an NCONF to also handle the old LHASH
12181 structure: this means that the old CONF compatible routines can be
12182 retained (in particular wrt extensions) without having to duplicate the
12183 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12187 * Enhance the general user interface with mechanisms for inner control
12188 and with possibilities to have yes/no kind of prompts.
12192 * Change all calls to low level digest routines in the library and
12193 applications to use EVP. Add missing calls to HMAC_cleanup() and
12194 don't assume HMAC_CTX can be copied using memcpy().
12196 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12198 * Add the possibility to control engines through control names but with
12199 arbitrary arguments instead of just a string.
12200 Change the key loaders to take a UI_METHOD instead of a callback
12201 function pointer. NOTE: this breaks binary compatibility with earlier
12202 versions of OpenSSL [engine].
12203 Adapt the nCipher code for these new conditions and add a card insertion
12208 * Enhance the general user interface with mechanisms to better support
12209 dialog box interfaces, application-defined prompts, the possibility
12210 to use defaults (for example default passwords from somewhere else)
12211 and interrupts/cancellations.
12215 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12216 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12220 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12221 tidy up some unnecessarily weird code in 'sk_new()').
12223 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12225 * Change the key loading routines for ENGINEs to use the same kind
12226 callback (pem_password_cb) as all other routines that need this
12231 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12232 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12233 than this minimum value is recommended.
12237 * New random seeder for OpenVMS, using the system process statistics
12238 that are easily reachable.
12242 * Windows apparently can't transparently handle global
12243 variables defined in DLLs. Initialisations such as:
12245 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12247 won't compile. This is used by the any applications that need to
12248 declare their own ASN1 modules. This was fixed by adding the option
12249 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12250 needed for static libraries under Win32.
12254 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12255 setting of purpose and trust fields. New X509_STORE trust and
12256 purpose functions and tidy up setting in other SSL functions.
12260 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12261 structure. These are inherited by X509_STORE_CTX when it is
12262 initialised. This allows various defaults to be set in the
12263 X509_STORE structure (such as flags for CRL checking and custom
12264 purpose or trust settings) for functions which only use X509_STORE_CTX
12265 internally such as S/MIME.
12267 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12268 trust settings if they are not set in X509_STORE. This allows X509_STORE
12269 purposes and trust (in S/MIME for example) to override any set by default.
12271 Add command line options for CRL checking to smime, s_client and s_server
12276 * Initial CRL based revocation checking. If the CRL checking flag(s)
12277 are set then the CRL is looked up in the X509_STORE structure and
12278 its validity and signature checked, then if the certificate is found
12279 in the CRL the verify fails with a revoked error.
12281 Various new CRL related callbacks added to X509_STORE_CTX structure.
12283 Command line options added to 'verify' application to support this.
12285 This needs some additional work, such as being able to handle multiple
12286 CRLs with different times, extension based lookup (rather than just
12287 by subject name) and ultimately more complete V2 CRL extension
12292 * Add a general user interface API (crypto/ui/). This is designed
12293 to replace things like des_read_password and friends (backward
12294 compatibility functions using this new API are provided).
12295 The purpose is to remove prompting functions from the DES code
12296 section as well as provide for prompting through dialog boxes in
12297 a window system and the like.
12301 * Add "ex_data" support to ENGINE so implementations can add state at a
12302 per-structure level rather than having to store it globally.
12306 * Make it possible for ENGINE structures to be copied when retrieved by
12307 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12308 This causes the "original" ENGINE structure to act like a template,
12309 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12310 operational state can be localised to each ENGINE structure, despite the
12311 fact they all share the same "methods". New ENGINE structures returned in
12312 this case have no functional references and the return value is the single
12313 structural reference. This matches the single structural reference returned
12314 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12319 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12320 needs to match any other type at all we need to manually clear the
12325 * Changes to the "openssl engine" utility to include;
12326 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12327 about an ENGINE's available control commands.
12328 - executing control commands from command line arguments using the
12329 '-pre' and '-post' switches. '-post' is only used if '-t' is
12330 specified and the ENGINE is successfully initialised. The syntax for
12331 the individual commands are colon-separated, for example;
12332 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12336 * New dynamic control command support for ENGINEs. ENGINEs can now
12337 declare their own commands (numbers), names (strings), descriptions,
12338 and input types for run-time discovery by calling applications. A
12339 subset of these commands are implicitly classed as "executable"
12340 depending on their input type, and only these can be invoked through
12341 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12342 can be based on user input, config files, etc). The distinction is
12343 that "executable" commands cannot return anything other than a boolean
12344 result and can only support numeric or string input, whereas some
12345 discoverable commands may only be for direct use through
12346 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12347 pointers, or other custom uses. The "executable" commands are to
12348 support parameterisations of ENGINE behaviour that can be
12349 unambiguously defined by ENGINEs and used consistently across any
12350 OpenSSL-based application. Commands have been added to all the
12351 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12352 control over shared-library paths without source code alterations.
12356 * Changed all ENGINE implementations to dynamically allocate their
12357 ENGINEs rather than declaring them statically. Apart from this being
12358 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12359 this also allows the implementations to compile without using the
12360 internal engine_int.h header.
12364 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12365 'const' value. Any code that should be able to modify a RAND_METHOD
12366 should already have non-const pointers to it (ie. they should only
12367 modify their own ones).
12371 * Made a variety of little tweaks to the ENGINE code.
12372 - "atalla" and "ubsec" string definitions were moved from header files
12373 to C code. "nuron" string definitions were placed in variables
12374 rather than hard-coded - allowing parameterisation of these values
12375 later on via ctrl() commands.
12376 - Removed unused "#if 0"'d code.
12377 - Fixed engine list iteration code so it uses ENGINE_free() to release
12378 structural references.
12379 - Constified the RAND_METHOD element of ENGINE structures.
12380 - Constified various get/set functions as appropriate and added
12381 missing functions (including a catch-all ENGINE_cpy that duplicates
12382 all ENGINE values onto a new ENGINE except reference counts/state).
12383 - Removed NULL parameter checks in get/set functions. Setting a method
12384 or function to NULL is a way of cancelling out a previously set
12385 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12386 and doesn't justify the extra error symbols and code.
12387 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12388 flags from engine_int.h to engine.h.
12389 - Changed prototypes for ENGINE handler functions (init(), finish(),
12390 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12394 * Implement binary inversion algorithm for BN_mod_inverse in addition
12395 to the algorithm using long division. The binary algorithm can be
12396 used only if the modulus is odd. On 32-bit systems, it is faster
12397 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12398 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12399 up to 450 bits. In 64-bit environments, the binary algorithm
12400 appears to be advantageous for much longer moduli; here we use it
12401 for moduli up to 2048 bits.
12405 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12406 could not support the combine flag in choice fields.
12410 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12411 extensions from a certificate request to the certificate.
12415 * Allow multiple 'certopt' and 'nameopt' options to be separated
12416 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12417 file: this allows the display of the certificate about to be
12418 signed to be customised, to allow certain fields to be included
12419 or excluded and extension details. The old system didn't display
12420 multicharacter strings properly, omitted fields not in the policy
12421 and couldn't display additional details such as extensions.
12425 * Function EC_POINTs_mul for multiple scalar multiplication
12426 of an arbitrary number of elliptic curve points
12427 \sum scalars[i]*points[i],
12428 optionally including the generator defined for the EC_GROUP:
12429 scalar*generator + \sum scalars[i]*points[i].
12431 EC_POINT_mul is a simple wrapper function for the typical case
12432 that the point list has just one item (besides the optional
12437 * First EC_METHODs for curves over GF(p):
12439 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12440 operations and provides various method functions that can also
12441 operate with faster implementations of modular arithmetic.
12443 EC_GFp_mont_method() reuses most functions that are part of
12444 EC_GFp_simple_method, but uses Montgomery arithmetic.
12446 *Bodo Moeller; point addition and point doubling
12447 implementation directly derived from source code provided by
12448 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12450 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12451 crypto/ec/ec_lib.c):
12453 Curves are EC_GROUP objects (with an optional group generator)
12454 based on EC_METHODs that are built into the library.
12456 Points are EC_POINT objects based on EC_GROUP objects.
12458 Most of the framework would be able to handle curves over arbitrary
12459 finite fields, but as there are no obvious types for fields other
12460 than GF(p), some functions are limited to that for now.
12464 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12465 that the file contains a complete HTTP response.
12469 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12470 change the def and num file printf format specifier from "%-40sXXX"
12471 to "%-39s XXX". The latter will always guarantee a space after the
12472 field while the former will cause them to run together if the field
12473 is 40 of more characters long.
12477 * Constify the cipher and digest 'method' functions and structures
12478 and modify related functions to take constant EVP_MD and EVP_CIPHER
12483 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12484 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12488 * Modify `EVP_Digest*()` routines so they now return values. Although the
12489 internal software routines can never fail additional hardware versions
12494 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12496 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12497 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12500 ERR_R_NESTED_ASN1_ERROR
12502 ERR_R_MISSING_ASN1_EOS
12503 were 4 .. 9, conflicting with
12504 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12506 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12507 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12509 Add new error code 'ERR_R_INTERNAL_ERROR'.
12513 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12518 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12519 sets the subject name for a new request or supersedes the
12520 subject name in a given request. Formats that can be parsed are
12521 'CN=Some Name, OU=myOU, C=IT'
12523 'CN=Some Name/OU=myOU/C=IT'.
12525 Add options '-batch' and '-verbose' to 'openssl req'.
12527 *Massimiliano Pala <madwolf@hackmasters.net>*
12529 * Introduce the possibility to access global variables through
12530 functions on platform were that's the best way to handle exporting
12531 global variables in shared libraries. To enable this functionality,
12532 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12533 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12534 is normally done by Configure or something similar).
12536 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12537 in the source file (foo.c) like this:
12539 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12540 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12542 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12543 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12545 OPENSSL_DECLARE_GLOBAL(int,foo);
12546 #define foo OPENSSL_GLOBAL_REF(foo)
12547 OPENSSL_DECLARE_GLOBAL(double,bar);
12548 #define bar OPENSSL_GLOBAL_REF(bar)
12550 The #defines are very important, and therefore so is including the
12551 header file everywhere where the defined globals are used.
12553 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12554 of ASN.1 items, but that structure is a bit different.
12556 The largest change is in util/mkdef.pl which has been enhanced with
12557 better and easier to understand logic to choose which symbols should
12558 go into the Windows .def files as well as a number of fixes and code
12559 cleanup (among others, algorithm keywords are now sorted
12560 lexicographically to avoid constant rewrites).
12564 * In BN_div() keep a copy of the sign of 'num' before writing the
12565 result to 'rm' because if rm==num the value will be overwritten
12566 and produce the wrong result if 'num' is negative: this caused
12567 problems with BN_mod() and BN_nnmod().
12571 * Function OCSP_request_verify(). This checks the signature on an
12572 OCSP request and verifies the signer certificate. The signer
12573 certificate is just checked for a generic purpose and OCSP request
12578 * Add OCSP_check_validity() function to check the validity of OCSP
12579 responses. OCSP responses are prepared in real time and may only
12580 be a few seconds old. Simply checking that the current time lies
12581 between thisUpdate and nextUpdate max reject otherwise valid responses
12582 caused by either OCSP responder or client clock inaccuracy. Instead
12583 we allow thisUpdate and nextUpdate to fall within a certain period of
12584 the current time. The age of the response can also optionally be
12585 checked. Two new options -validity_period and -status_age added to
12590 * If signature or public key algorithm is unrecognized print out its
12591 OID rather that just UNKNOWN.
12595 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12596 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12597 ID to be generated from the issuer certificate alone which can then be
12598 passed to OCSP_id_issuer_cmp().
12602 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12603 ASN1 modules to export functions returning ASN1_ITEM pointers
12604 instead of the ASN1_ITEM structures themselves. This adds several
12605 new macros which allow the underlying ASN1 function/structure to
12606 be accessed transparently. As a result code should not use ASN1_ITEM
12607 references directly (such as &X509_it) but instead use the relevant
12608 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12609 use of the new ASN1 code on platforms where exporting structures
12610 is problematical (for example in shared libraries) but exporting
12611 functions returning pointers to structures is not.
12615 * Add support for overriding the generation of SSL/TLS session IDs.
12616 These callbacks can be registered either in an SSL_CTX or per SSL.
12617 The purpose of this is to allow applications to control, if they wish,
12618 the arbitrary values chosen for use as session IDs, particularly as it
12619 can be useful for session caching in multiple-server environments. A
12620 command-line switch for testing this (and any client code that wishes
12621 to use such a feature) has been added to "s_server".
12623 *Geoff Thorpe, Lutz Jaenicke*
12625 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12626 of the form `#if defined(...) || defined(...) || ...` and
12627 `#if !defined(...) && !defined(...) && ...`. This also avoids
12628 the growing number of special cases it was previously handling.
12632 * Make all configuration macros available for application by making
12633 sure they are available in opensslconf.h, by giving them names starting
12634 with `OPENSSL_` to avoid conflicts with other packages and by making
12635 sure e_os2.h will cover all platform-specific cases together with
12637 Additionally, it is now possible to define configuration/platform-
12638 specific names (called "system identities"). In the C code, these
12639 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12640 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12641 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12646 * New option -set_serial to 'req' and 'x509' this allows the serial
12647 number to use to be specified on the command line. Previously self
12648 signed certificates were hard coded with serial number 0 and the
12649 CA options of 'x509' had to use a serial number in a file which was
12654 * New options to 'ca' utility to support V2 CRL entry extensions.
12655 Currently CRL reason, invalidity date and hold instruction are
12656 supported. Add new CRL extensions to V3 code and some new objects.
12660 * New function EVP_CIPHER_CTX_set_padding() this is used to
12661 disable standard block padding (aka PKCS#5 padding) in the EVP
12662 API, which was previously mandatory. This means that the data is
12663 not padded in any way and so the total length much be a multiple
12664 of the block size, otherwise an error occurs.
12668 * Initial (incomplete) OCSP SSL support.
12672 * New function OCSP_parse_url(). This splits up a URL into its host,
12673 port and path components: primarily to parse OCSP URLs. New -url
12674 option to ocsp utility.
12678 * New nonce behavior. The return value of OCSP_check_nonce() now
12679 reflects the various checks performed. Applications can decide
12680 whether to tolerate certain situations such as an absent nonce
12681 in a response when one was present in a request: the ocsp application
12682 just prints out a warning. New function OCSP_add1_basic_nonce()
12683 this is to allow responders to include a nonce in a response even if
12684 the request is nonce-less.
12688 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12689 skipped when using openssl x509 multiple times on a single input file,
12690 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12694 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12695 set string type: to handle setting ASN1_TIME structures. Fix ca
12696 utility to correctly initialize revocation date of CRLs.
12700 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12701 the clients preferred ciphersuites and rather use its own preferences.
12702 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12703 Internet Explorer by ensuring unchanged hash method during stepup.
12704 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12708 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12709 to aes and add a new 'exist' option to print out symbols that don't
12714 * Additional options to ocsp utility to allow flags to be set and
12715 additional certificates supplied.
12719 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12720 OCSP client a number of certificate to only verify the response
12725 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12726 handle the new API. Currently only ECB, CBC modes supported. Add new
12729 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12730 Encryption Standard (AES) Ciphersuites for Transport Layer
12731 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12732 not enabled by default and were not part of the "ALL" ciphersuite
12733 alias because they were not yet official; they could be
12734 explicitly requested by specifying the "AESdraft" ciphersuite
12735 group alias. In the final release of OpenSSL 0.9.7, the group
12736 alias is called "AES" and is part of "ALL".)
12738 *Ben Laurie, Steve Henson, Bodo Moeller*
12740 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12741 request to response.
12745 * Functions for OCSP responders. OCSP_request_onereq_count(),
12746 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12747 extract information from a certificate request. OCSP_response_create()
12748 creates a response and optionally adds a basic response structure.
12749 OCSP_basic_add1_status() adds a complete single response to a basic
12750 response and returns the OCSP_SINGLERESP structure just added (to allow
12751 extensions to be included for example). OCSP_basic_add1_cert() adds a
12752 certificate to a basic response and OCSP_basic_sign() signs a basic
12753 response with various flags. New helper functions ASN1_TIME_check()
12754 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12755 (converts ASN1_TIME to GeneralizedTime).
12759 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12760 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12761 structure from a certificate. X509_pubkey_digest() digests the public_key
12762 contents: this is used in various key identifiers.
12766 * Make sk_sort() tolerate a NULL argument.
12768 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12770 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12771 passed by the function are trusted implicitly. If any of them signed the
12772 response then it is assumed to be valid and is not verified.
12776 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12777 to data. This was previously part of the PKCS7 ASN1 code. This
12778 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12779 *Steve Henson, reported by Kenneth R. Robinette
12780 <support@securenetterm.com>*
12782 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12783 routines: without these tracing memory leaks is very painful.
12784 Fix leaks in PKCS12 and PKCS7 routines.
12788 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12789 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12790 effectively meant GeneralizedTime would never be used. Now it
12791 is initialised to -1 but X509_time_adj() now has to check the value
12792 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12793 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12794 *Steve Henson, reported by Kenneth R. Robinette
12795 <support@securenetterm.com>*
12797 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12798 result in a zero length in the ASN1_INTEGER structure which was
12799 not consistent with the structure when d2i_ASN1_INTEGER() was used
12800 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12801 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12802 where it did not print out a minus for negative ASN1_INTEGER.
12806 * Add summary printout to ocsp utility. The various functions which
12807 convert status values to strings have been renamed to:
12808 OCSP_response_status_str(), OCSP_cert_status_str() and
12809 OCSP_crl_reason_str() and are no longer static. New options
12810 to verify nonce values and to disable verification. OCSP response
12811 printout format cleaned up.
12815 * Add additional OCSP certificate checks. These are those specified
12816 in RFC2560. This consists of two separate checks: the CA of the
12817 certificate being checked must either be the OCSP signer certificate
12818 or the issuer of the OCSP signer certificate. In the latter case the
12819 OCSP signer certificate must contain the OCSP signing extended key
12820 usage. This check is performed by attempting to match the OCSP
12821 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12822 in the OCSP_CERTID structures of the response.
12826 * Initial OCSP certificate verification added to OCSP_basic_verify()
12827 and related routines. This uses the standard OpenSSL certificate
12828 verify routines to perform initial checks (just CA validity) and
12829 to obtain the certificate chain. Then additional checks will be
12830 performed on the chain. Currently the root CA is checked to see
12831 if it is explicitly trusted for OCSP signing. This is used to set
12832 a root CA as a global signing root: that is any certificate that
12833 chains to that CA is an acceptable OCSP signing certificate.
12837 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12838 extensions from a separate configuration file.
12839 As when reading extensions from the main configuration file,
12840 the '-extensions ...' option may be used for specifying the
12843 *Massimiliano Pala <madwolf@comune.modena.it>*
12845 * New OCSP utility. Allows OCSP requests to be generated or
12846 read. The request can be sent to a responder and the output
12847 parsed, outputted or printed in text form. Not complete yet:
12848 still needs to check the OCSP response validity.
12852 * New subcommands for 'openssl ca':
12853 `openssl ca -status <serial>` prints the status of the cert with
12854 the given serial number (according to the index file).
12855 `openssl ca -updatedb` updates the expiry status of certificates
12858 *Massimiliano Pala <madwolf@comune.modena.it>*
12860 * New '-newreq-nodes' command option to CA.pl. This is like
12861 '-newreq', but calls 'openssl req' with the '-nodes' option
12862 so that the resulting key is not encrypted.
12864 *Damien Miller <djm@mindrot.org>*
12866 * New configuration for the GNU Hurd.
12868 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12870 * Initial code to implement OCSP basic response verify. This
12871 is currently incomplete. Currently just finds the signer's
12872 certificate and verifies the signature on the response.
12876 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12877 value of OPENSSLDIR. This is available via the new '-d' option
12878 to 'openssl version', and is also included in 'openssl version -a'.
12882 * Allowing defining memory allocation callbacks that will be given
12883 file name and line number information in additional arguments
12884 (a `const char*` and an int). The basic functionality remains, as
12885 well as the original possibility to just replace malloc(),
12886 realloc() and free() by functions that do not know about these
12887 additional arguments. To register and find out the current
12888 settings for extended allocation functions, the following
12889 functions are provided:
12891 CRYPTO_set_mem_ex_functions
12892 CRYPTO_set_locked_mem_ex_functions
12893 CRYPTO_get_mem_ex_functions
12894 CRYPTO_get_locked_mem_ex_functions
12896 These work the same way as CRYPTO_set_mem_functions and friends.
12897 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12898 extended allocation function is enabled.
12899 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12900 a conventional allocation function is enabled.
12902 *Richard Levitte, Bodo Moeller*
12904 * Finish off removing the remaining LHASH function pointer casts.
12905 There should no longer be any prototype-casting required when using
12906 the LHASH abstraction, and any casts that remain are "bugs". See
12907 the callback types and macros at the head of lhash.h for details
12908 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12912 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12913 If /dev/[u]random devices are not available or do not return enough
12914 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12916 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12917 /etc/entropy will be queried once each in this sequence, querying stops
12918 when enough entropy was collected without querying more sockets.
12922 * Change the Unix RAND_poll() variant to be able to poll several
12923 random devices, as specified by DEVRANDOM, until a sufficient amount
12924 of data has been collected. We spend at most 10 ms on each file
12925 (select timeout) and read in non-blocking mode. DEVRANDOM now
12926 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12927 (previously it was just the string "/dev/urandom"), so on typical
12928 platforms the 10 ms delay will never occur.
12929 Also separate out the Unix variant to its own file, rand_unix.c.
12930 For VMS, there's a currently-empty rand_vms.c.
12934 * Move OCSP client related routines to ocsp_cl.c. These
12935 provide utility functions which an application needing
12936 to issue a request to an OCSP responder and analyse the
12937 response will typically need: as opposed to those which an
12938 OCSP responder itself would need which will be added later.
12940 OCSP_request_sign() signs an OCSP request with an API similar
12941 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12942 response. OCSP_response_get1_basic() extracts basic response
12943 from response. OCSP_resp_find_status(): finds and extracts status
12944 information from an OCSP_CERTID structure (which will be created
12945 when the request structure is built). These are built from lower
12946 level functions which work on OCSP_SINGLERESP structures but
12947 won't normally be used unless the application wishes to examine
12948 extensions in the OCSP response for example.
12950 Replace nonce routines with a pair of functions.
12951 OCSP_request_add1_nonce() adds a nonce value and optionally
12952 generates a random value. OCSP_check_nonce() checks the
12953 validity of the nonce in an OCSP response.
12957 * Change function OCSP_request_add() to OCSP_request_add0_id().
12958 This doesn't copy the supplied OCSP_CERTID and avoids the
12959 need to free up the newly created id. Change return type
12960 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12961 This can then be used to add extensions to the request.
12962 Deleted OCSP_request_new(), since most of its functionality
12963 is now in OCSP_REQUEST_new() (and the case insensitive name
12964 clash) apart from the ability to set the request name which
12965 will be added elsewhere.
12969 * Update OCSP API. Remove obsolete extensions argument from
12970 various functions. Extensions are now handled using the new
12971 OCSP extension code. New simple OCSP HTTP function which
12972 can be used to send requests and parse the response.
12976 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
12977 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
12978 uses the special reorder version of SET OF to sort the attributes
12979 and reorder them to match the encoded order. This resolves a long
12980 standing problem: a verify on a PKCS7 structure just after signing
12981 it used to fail because the attribute order did not match the
12982 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
12983 it uses the received order. This is necessary to tolerate some broken
12984 software that does not order SET OF. This is handled by encoding
12985 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
12986 to produce the required SET OF.
12990 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
12991 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
12992 files to get correct declarations of the ASN.1 item variables.
12996 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
12997 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
12998 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
12999 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13000 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13001 ASN1_ITEM and no wrapper functions.
13005 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13006 replace the old function pointer based I/O routines. Change most of
13007 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13011 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13012 lines, recognize more "algorithms" that can be deselected, and make
13013 it complain about algorithm deselection that isn't recognised.
13017 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13018 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13019 to use new functions. Add NO_ASN1_OLD which can be set to remove
13020 some old style ASN1 functions: this can be used to determine if old
13021 code will still work when these eventually go away.
13025 * New extension functions for OCSP structures, these follow the
13026 same conventions as certificates and CRLs.
13030 * New function X509V3_add1_i2d(). This automatically encodes and
13031 adds an extension. Its behaviour can be customised with various
13032 flags to append, replace or delete. Various wrappers added for
13033 certificates and CRLs.
13037 * Fix to avoid calling the underlying ASN1 print routine when
13038 an extension cannot be parsed. Correct a typo in the
13039 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13043 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13044 entries for variables.
13048 * Add functionality to apps/openssl.c for detecting locking
13049 problems: As the program is single-threaded, all we have
13050 to do is register a locking callback using an array for
13051 storing which locks are currently held by the program.
13055 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13056 SSL_get_ex_data_X509_STORE_idx(), which is used in
13057 ssl_verify_cert_chain() and thus can be called at any time
13058 during TLS/SSL handshakes so that thread-safety is essential.
13059 Unfortunately, the ex_data design is not at all suited
13060 for multi-threaded use, so it probably should be abolished.
13064 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13066 *Broadcom, tweaked and integrated by Geoff Thorpe*
13068 * Move common extension printing code to new function
13069 X509V3_print_extensions(). Reorganise OCSP print routines and
13070 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13074 * New function X509_signature_print() to remove duplication in some
13079 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13080 set (this was treated exactly the same as SET OF previously). This
13081 is used to reorder the STACK representing the structure to match the
13082 encoding. This will be used to get round a problem where a PKCS7
13083 structure which was signed could not be verified because the STACK
13084 order did not reflect the encoded order.
13088 * Reimplement the OCSP ASN1 module using the new code.
13092 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13093 for its ASN1 operations. The old style function pointers still exist
13094 for now but they will eventually go away.
13098 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13099 completely replaces the old ASN1 functionality with a table driven
13100 encoder and decoder which interprets an ASN1_ITEM structure describing
13101 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13102 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13103 has also been converted to the new form.
13107 * Change BN_mod_exp_recp so that negative moduli are tolerated
13108 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13109 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13110 for negative moduli.
13114 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13115 of not touching the result's sign bit.
13119 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13124 * Changed the LHASH code to use prototypes for callbacks, and created
13125 macros to declare and implement thin (optionally static) functions
13126 that provide type-safety and avoid function pointer casting for the
13127 type-specific callbacks.
13131 * Added Kerberos Cipher Suites to be used with TLS, as written in
13133 *Veers Staats <staatsvr@asc.hpc.mil>,
13134 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13136 * Reformat the FAQ so the different questions and answers can be divided
13137 in sections depending on the subject.
13141 * Have the zlib compression code load ZLIB.DLL dynamically under
13146 * New function BN_mod_sqrt for computing square roots modulo a prime
13147 (using the probabilistic Tonelli-Shanks algorithm unless
13148 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13149 be handled deterministically).
13151 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13153 * Make BN_mod_inverse faster by explicitly handling small quotients
13154 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13155 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13159 * New function BN_kronecker.
13163 * Fix BN_gcd so that it works on negative inputs; the result is
13164 positive unless both parameters are zero.
13165 Previously something reasonably close to an infinite loop was
13166 possible because numbers could be growing instead of shrinking
13167 in the implementation of Euclid's algorithm.
13171 * Fix BN_is_word() and BN_is_one() macros to take into account the
13172 sign of the number in question.
13174 Fix BN_is_word(a,w) to work correctly for w == 0.
13176 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13177 because its test if the absolute value of 'a' equals 'w'.
13178 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13179 it exists mostly for use in the implementations of BN_is_zero(),
13180 BN_is_one(), and BN_is_word().
13184 * New function BN_swap.
13188 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13189 the exponentiation functions are more likely to produce reasonable
13190 results on negative inputs.
13194 * Change BN_mod_mul so that the result is always non-negative.
13195 Previously, it could be negative if one of the factors was negative;
13196 I don't think anyone really wanted that behaviour.
13200 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13201 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13202 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13203 and add new functions:
13212 BN_mod_lshift1_quick
13214 BN_mod_lshift_quick
13216 These functions always generate non-negative results.
13218 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13219 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13221 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13222 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13223 be reduced modulo `m`.
13225 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13228 The following entry accidentally appeared in the CHANGES file
13229 distributed with OpenSSL 0.9.7. The modifications described in
13230 it do *not* apply to OpenSSL 0.9.7.
13232 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13233 was actually never needed) and in BN_mul(). The removal in BN_mul()
13234 required a small change in bn_mul_part_recursive() and the addition
13235 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13236 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13237 bn_sub_words() and bn_add_words() except they take arrays with
13243 * In 'openssl passwd', verify passwords read from the terminal
13244 unless the '-salt' option is used (which usually means that
13245 verification would just waste user's time since the resulting
13246 hash is going to be compared with some given password hash)
13247 or the new '-noverify' option is used.
13249 This is an incompatible change, but it does not affect
13250 non-interactive use of 'openssl passwd' (passwords on the command
13251 line, '-stdin' option, '-in ...' option) and thus should not
13252 cause any problems.
13256 * Remove all references to RSAref, since there's no more need for it.
13260 * Make DSO load along a path given through an environment variable
13261 (SHLIB_PATH) with shl_load().
13265 * Constify the ENGINE code as a result of BIGNUM constification.
13266 Also constify the RSA code and most things related to it. In a
13267 few places, most notable in the depth of the ASN.1 code, ugly
13268 casts back to non-const were required (to be solved at a later
13273 * Make it so the openssl application has all engines loaded by default.
13277 * Constify the BIGNUM routines a little more.
13281 * Add the following functions:
13283 ENGINE_load_cswift()
13285 ENGINE_load_atalla()
13286 ENGINE_load_nuron()
13287 ENGINE_load_builtin_engines()
13289 That way, an application can itself choose if external engines that
13290 are built-in in OpenSSL shall ever be used or not. The benefit is
13291 that applications won't have to be linked with libdl or other dso
13292 libraries unless it's really needed.
13294 Changed 'openssl engine' to load all engines on demand.
13295 Changed the engine header files to avoid the duplication of some
13296 declarations (they differed!).
13300 * 'openssl engine' can now list capabilities.
13304 * Better error reporting in 'openssl engine'.
13308 * Never call load_dh_param(NULL) in s_server.
13312 * Add engine application. It can currently list engines by name and
13313 identity, and test if they are actually available.
13317 * Improve RPM specification file by forcing symbolic linking and making
13318 sure the installed documentation is also owned by root.root.
13320 *Damien Miller <djm@mindrot.org>*
13322 * Give the OpenSSL applications more possibilities to make use of
13323 keys (public as well as private) handled by engines.
13327 * Add OCSP code that comes from CertCo.
13331 * Add VMS support for the Rijndael code.
13335 * Added untested support for Nuron crypto accelerator.
13339 * Add support for external cryptographic devices. This code was
13340 previously distributed separately as the "engine" branch.
13342 *Geoff Thorpe, Richard Levitte*
13344 * Rework the filename-translation in the DSO code. It is now possible to
13345 have far greater control over how a "name" is turned into a filename
13346 depending on the operating environment and any oddities about the
13347 different shared library filenames on each system.
13351 * Support threads on FreeBSD-elf in Configure.
13355 * Fix for SHA1 assembly problem with MASM: it produces
13356 warnings about corrupt line number information when assembling
13357 with debugging information. This is caused by the overlapping
13360 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13363 NCONF_get_number() has no error checking at all. As a replacement,
13364 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13365 promoted strongly. The old NCONF_get_number is kept around for
13366 binary backward compatibility.
13367 Make it possible for methods to load from something other than a BIO,
13368 by providing a function pointer that is given a name instead of a BIO.
13369 For example, this could be used to load configuration data from an
13374 * Fix for non blocking accept BIOs. Added new I/O special reason
13375 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13376 with non blocking I/O was not possible because no retry code was
13377 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13382 * Added the beginnings of Rijndael support.
13386 * Fix for bug in DirectoryString mask setting. Add support for
13387 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13388 to allow certificate printing to more controllable, additional
13389 'certopt' option to 'x509' to allow new printing options to be
13394 * Clean old EAY MD5 hack from e_os.h.
13398 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13400 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13401 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13403 *Joe Orton, Steve Henson*
13405 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13407 * Fix additional bug revealed by the NISCC test suite:
13409 Stop bug triggering large recursion when presented with
13410 certain ASN.1 tags ([CVE-2003-0851])
13414 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13416 * Fix various bugs revealed by running the NISCC test suite:
13418 Stop out of bounds reads in the ASN1 code when presented with
13419 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13421 If verify callback ignores invalid public key errors don't try to check
13422 certificate signature with the NULL public key.
13426 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13427 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13432 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13433 extra data after the compression methods not only for TLS 1.0
13434 but also for SSL 3.0 (as required by the specification).
13436 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13438 * Change X509_certificate_type() to mark the key as exported/exportable
13439 when it's 512 *bits* long, not 512 bytes.
13443 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13445 * Countermeasure against the Klima-Pokorny-Rosa extension of
13446 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13447 a protocol version number mismatch like a decryption error
13448 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13452 * Turn on RSA blinding by default in the default implementation
13453 to avoid a timing attack. Applications that don't want it can call
13454 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13455 They would be ill-advised to do so in most cases.
13457 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13459 * Change RSA blinding code so that it works when the PRNG is not
13460 seeded (in this case, the secret RSA exponent is abused as
13461 an unpredictable seed -- if it is not unpredictable, there
13462 is no point in blinding anyway). Make RSA blinding thread-safe
13463 by remembering the creator's thread ID in rsa->blinding and
13464 having all other threads use local one-time blinding factors
13465 (this requires more computation than sharing rsa->blinding, but
13466 avoids excessive locking; and if an RSA object is not shared
13467 between threads, blinding will still be very fast).
13471 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13473 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13474 via timing by performing a MAC computation even if incorrect
13475 block cipher padding has been found. This is a countermeasure
13476 against active attacks where the attacker has to distinguish
13477 between bad padding and a MAC verification error. ([CVE-2003-0078])
13479 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13480 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13481 Martin Vuagnoux (EPFL, Ilion)*
13483 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13485 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13486 memory from its contents. This is done with a counter that will
13487 place alternating values in each byte. This can be used to solve
13488 two issues: 1) the removal of calls to memset() by highly optimizing
13489 compilers, and 2) cleansing with other values than 0, since those can
13490 be read through on certain media, for example a swap space on disk.
13494 * Bugfix: client side session caching did not work with external caching,
13495 because the session->cipher setting was not restored when reloading
13496 from the external cache. This problem was masked, when
13497 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13498 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13502 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13503 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13505 *Zeev Lieber <zeev-l@yahoo.com>*
13507 * Undo an undocumented change introduced in 0.9.6e which caused
13508 repeated calls to OpenSSL_add_all_ciphers() and
13509 OpenSSL_add_all_digests() to be ignored, even after calling
13514 * Change the default configuration reader to deal with last line not
13515 being properly terminated.
13519 * Change X509_NAME_cmp() so it applies the special rules on handling
13520 DN values that are of type PrintableString, as well as RDNs of type
13521 emailAddress where the value has the type ia5String.
13523 *stefank@valicert.com via Richard Levitte*
13525 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13526 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13527 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13528 the bitwise-OR of the two for use by the majority of applications
13529 wanting this behaviour, and update the docs. The documented
13530 behaviour and actual behaviour were inconsistent and had been
13531 changing anyway, so this is more a bug-fix than a behavioural
13534 *Geoff Thorpe, diagnosed by Nadav Har'El*
13536 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13537 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13541 * Fix initialization code race conditions in
13542 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13543 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13544 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13545 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13546 ssl2_get_cipher_by_char(),
13547 ssl3_get_cipher_by_char().
13549 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13551 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13552 the cached sessions are flushed, as the remove_cb() might use ex_data
13553 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13554 (see [openssl.org #212]).
13556 *Geoff Thorpe, Lutz Jaenicke*
13558 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13559 length, instead of the encoding length to d2i_ASN1_OBJECT.
13563 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13565 * [In 0.9.6g-engine release:]
13566 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13568 *Lynn Gazis <lgazis@rainbow.com>*
13570 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13572 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13573 and get fix the header length calculation.
13574 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13575 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13577 * Use proper error handling instead of 'assertions' in buffer
13578 overflow checks added in 0.9.6e. This prevents DoS (the
13579 assertions could call abort()).
13581 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13583 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13585 * Add various sanity checks to asn1_get_length() to reject
13586 the ASN1 length bytes if they exceed sizeof(long), will appear
13587 negative or the content length exceeds the length of the
13590 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13592 * Fix cipher selection routines: ciphers without encryption had no flags
13593 for the cipher strength set and where therefore not handled correctly
13594 by the selection routines (PR #130).
13598 * Fix EVP_dsa_sha macro.
13603 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13604 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13605 that was added in OpenSSL 0.9.6d.
13607 As the countermeasure turned out to be incompatible with some
13608 broken SSL implementations, the new option is part of SSL_OP_ALL.
13609 SSL_OP_ALL is usually employed when compatibility with weird SSL
13610 implementations is desired (e.g. '-bugs' option to 's_client' and
13611 's_server'), so the new option is automatically set in many
13616 * Changes in security patch:
13618 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13619 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13620 Air Force Materiel Command, USAF, under agreement number
13623 * Add various sanity checks to asn1_get_length() to reject
13624 the ASN1 length bytes if they exceed sizeof(long), will appear
13625 negative or the content length exceeds the length of the
13626 supplied buffer. ([CVE-2002-0659])
13628 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13630 * Assertions for various potential buffer overflows, not known to
13631 happen in practice.
13633 *Ben Laurie (CHATS)*
13635 * Various temporary buffers to hold ASCII versions of integers were
13636 too small for 64 bit platforms. ([CVE-2002-0655])
13637 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13639 * Remote buffer overflow in SSL3 protocol - an attacker could
13640 supply an oversized session ID to a client. ([CVE-2002-0656])
13642 *Ben Laurie (CHATS)*
13644 * Remote buffer overflow in SSL2 protocol - an attacker could
13645 supply an oversized client master key. ([CVE-2002-0656])
13647 *Ben Laurie (CHATS)*
13649 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13651 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13652 encoded as NULL) with id-dsa-with-sha1.
13654 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13656 * Check various `X509_...()` return values in apps/req.c.
13658 *Nils Larsch <nla@trustcenter.de>*
13660 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13661 an end-of-file condition would erroneously be flagged, when the CRLF
13662 was just at the end of a processed block. The bug was discovered when
13663 processing data through a buffering memory BIO handing the data to a
13664 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13665 <ptsekov@syntrex.com> and Nedelcho Stanev.
13669 * Implement a countermeasure against a vulnerability recently found
13670 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13671 before application data chunks to avoid the use of known IVs
13672 with data potentially chosen by the attacker.
13676 * Fix length checks in ssl3_get_client_hello().
13680 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13681 to prevent ssl3_read_internal() from incorrectly assuming that
13682 ssl3_read_bytes() found application data while handshake
13683 processing was enabled when in fact s->s3->in_read_app_data was
13684 merely automatically cleared during the initial handshake.
13686 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13688 * Fix object definitions for Private and Enterprise: they were not
13689 recognized in their shortname (=lowercase) representation. Extend
13690 obj_dat.pl to issue an error when using undefined keywords instead
13691 of silently ignoring the problem (Svenning Sorensen
13692 <sss@sss.dnsalias.net>).
13696 * Fix DH_generate_parameters() so that it works for 'non-standard'
13697 generators, i.e. generators other than 2 and 5. (Previously, the
13698 code did not properly initialise the 'add' and 'rem' values to
13699 BN_generate_prime().)
13701 In the new general case, we do not insist that 'generator' is
13702 actually a primitive root: This requirement is rather pointless;
13703 a generator of the order-q subgroup is just as good, if not
13708 * Map new X509 verification errors to alerts. Discovered and submitted by
13709 Tom Wu <tom@arcot.com>.
13713 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13714 returning non-zero before the data has been completely received
13715 when using non-blocking I/O.
13717 *Bodo Moeller; problem pointed out by John Hughes*
13719 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13721 *Ben Laurie, Lutz Jaenicke*
13723 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13724 Yoram Zahavi <YoramZ@gilian.com>).
13728 * Add information about CygWin 1.3 and on, and preserve proper
13729 configuration for the versions before that.
13731 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13733 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13734 check whether we deal with a copy of a session and do not delete from
13735 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13736 <izhar@checkpoint.com>.
13740 * Do not store session data into the internal session cache, if it
13741 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13742 flag is set). Proposed by Aslam <aslam@funk.com>.
13746 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13751 * [In 0.9.6d-engine release:]
13752 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13754 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13756 * Add the configuration target linux-s390x.
13758 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13760 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13761 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13762 variable as an indication that a ClientHello message has been
13763 received. As the flag value will be lost between multiple
13764 invocations of ssl3_accept when using non-blocking I/O, the
13765 function may not be aware that a handshake has actually taken
13766 place, thus preventing a new session from being added to the
13769 To avoid this problem, we now set s->new_session to 2 instead of
13770 using a local variable.
13772 *Lutz Jaenicke, Bodo Moeller*
13774 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13775 if the SSL_R_LENGTH_MISMATCH error is detected.
13777 *Geoff Thorpe, Bodo Moeller*
13779 * New 'shared_ldflag' column in Configure platform table.
13783 * Fix EVP_CIPHER_mode macro.
13785 *"Dan S. Camper" <dan@bti.net>*
13787 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13788 type, we must throw them away by setting rr->length to 0.
13790 *D P Chang <dpc@qualys.com>*
13792 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13794 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13795 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13796 worked incorrectly for those cases where range = `10..._2` and
13797 `3*range` is two bits longer than range.)
13801 * Only add signing time to PKCS7 structures if it is not already
13806 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13807 OBJ_ld_ce should be OBJ_id_ce.
13808 Also some ip-pda OIDs in crypto/objects/objects.txt were
13809 incorrect (cf. RFC 3039).
13811 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13813 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13814 returns early because it has nothing to do.
13816 *Andy Schneider <andy.schneider@bjss.co.uk>*
13818 * [In 0.9.6c-engine release:]
13819 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13821 *Andy Schneider <andy.schneider@bjss.co.uk>*
13823 * [In 0.9.6c-engine release:]
13824 Add support for Cryptographic Appliance's keyserver technology.
13825 (Use engine 'keyclient')
13827 *Cryptographic Appliances and Geoff Thorpe*
13829 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13830 is called via tools/c89.sh because arguments have to be
13831 rearranged (all '-L' options must appear before the first object
13834 *Richard Shapiro <rshapiro@abinitio.com>*
13836 * [In 0.9.6c-engine release:]
13837 Add support for Broadcom crypto accelerator cards, backported
13840 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13842 * [In 0.9.6c-engine release:]
13843 Add support for SureWare crypto accelerator cards from
13844 Baltimore Technologies. (Use engine 'sureware')
13846 *Baltimore Technologies and Mark Cox*
13848 * [In 0.9.6c-engine release:]
13849 Add support for crypto accelerator cards from Accelerated
13850 Encryption Processing, www.aep.ie. (Use engine 'aep')
13852 *AEP Inc. and Mark Cox*
13854 * Add a configuration entry for gcc on UnixWare.
13856 *Gary Benson <gbenson@redhat.com>*
13858 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13859 messages are stored in a single piece (fixed-length part and
13860 variable-length part combined) and fix various bugs found on the way.
13864 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13865 instead. BIO_gethostbyname() does not know what timeouts are
13866 appropriate, so entries would stay in cache even when they have
13868 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13870 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13871 faced with a pathologically small ClientHello fragment that does
13872 not contain client_version: Instead of aborting with an error,
13873 simply choose the highest available protocol version (i.e.,
13874 TLS 1.0 unless it is disabled). In practice, ClientHello
13875 messages are never sent like this, but this change gives us
13876 strictly correct behaviour at least for TLS.
13880 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13881 never resets s->method to s->ctx->method when called from within
13882 one of the SSL handshake functions.
13884 *Bodo Moeller; problem pointed out by Niko Baric*
13886 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13887 (sent using the client's version number) if client_version is
13888 smaller than the protocol version in use. Also change
13889 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13890 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13891 the client will at least see that alert.
13895 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13900 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13901 client receives HelloRequest while in a handshake.
13903 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13905 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13906 should end in 'break', not 'goto end' which circumvents various
13907 cleanups done in state SSL_ST_OK. But session related stuff
13908 must be disabled for SSL_ST_OK in the case that we just sent a
13911 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13912 before just sending a HelloRequest.
13914 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13916 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13917 reveal whether illegal block cipher padding was found or a MAC
13918 verification error occurred. (Neither SSLerr() codes nor alerts
13919 are directly visible to potential attackers, but the information
13920 may leak via logfiles.)
13922 Similar changes are not required for the SSL 2.0 implementation
13923 because the number of padding bytes is sent in clear for SSL 2.0,
13924 and the extra bytes are just ignored. However ssl/s2_pkt.c
13925 failed to verify that the purported number of padding bytes is in
13930 * Add OpenUNIX-8 support including shared libraries
13931 (Boyd Lynn Gerber <gerberb@zenez.com>).
13935 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13936 'wristwatch attack' using huge encoding parameters (cf.
13937 James H. Manger's CRYPTO 2001 paper). Note that the
13938 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13939 encoding parameters and hence was not vulnerable.
13943 * BN_sqr() bug fix.
13945 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13947 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13948 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13949 followed by modular reduction.
13951 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13953 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13954 equivalent based on BN_pseudo_rand() instead of BN_rand().
13958 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13959 This function was broken, as the check for a new client hello message
13960 to handle SGC did not allow these large messages.
13961 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13965 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13969 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13970 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
13974 * Rework the configuration and shared library support for Tru64 Unix.
13975 The configuration part makes use of modern compiler features and
13976 still retains old compiler behavior for those that run older versions
13977 of the OS. The shared library support part includes a variant that
13978 uses the RPATH feature, and is available through the special
13979 configuration target "alpha-cc-rpath", which will never be selected
13982 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
13984 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
13985 with the same message size as in ssl3_get_certificate_request().
13986 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
13987 messages might inadvertently be reject as too long.
13989 *Petr Lampa <lampa@fee.vutbr.cz>*
13991 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
13995 * Modified SSL library such that the verify_callback that has been set
13996 specifically for an SSL object with SSL_set_verify() is actually being
13997 used. Before the change, a verify_callback set with this function was
13998 ignored and the verify_callback() set in the SSL_CTX at the time of
13999 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14000 to allow the necessary settings.
14004 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14005 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14006 done automatically (in contradiction to the requirements of the C
14007 standard). This made problems when used from OpenSSH.
14011 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14012 dh->length and always used
14014 BN_rand_range(priv_key, dh->p).
14016 BN_rand_range() is not necessary for Diffie-Hellman, and this
14017 specific range makes Diffie-Hellman unnecessarily inefficient if
14018 dh->length (recommended exponent length) is much smaller than the
14019 length of dh->p. We could use BN_rand_range() if the order of
14020 the subgroup was stored in the DH structure, but we only have
14025 BN_rand(priv_key, l, ...)
14027 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14034 RSA_eay_public_encrypt
14035 RSA_eay_private_decrypt
14036 RSA_eay_private_encrypt (signing)
14037 RSA_eay_public_decrypt (signature verification)
14039 (default implementations for RSA_public_encrypt,
14040 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14041 always reject numbers >= n.
14045 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14046 to synchronize access to 'locking_thread'. This is necessary on
14047 systems where access to 'locking_thread' (an 'unsigned long'
14048 variable) is not atomic.
14052 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14053 *before* setting the 'crypto_lock_rand' flag. The previous code had
14054 a race condition if 0 is a valid thread ID.
14056 *Travis Vitek <vitek@roguewave.com>*
14058 * Add support for shared libraries under Irix.
14060 *Albert Chin-A-Young <china@thewrittenword.com>*
14062 * Add configuration option to build on Linux on both big-endian and
14063 little-endian MIPS.
14065 *Ralf Baechle <ralf@uni-koblenz.de>*
14067 * Add the possibility to create shared libraries on HP-UX.
14071 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14073 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14074 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14075 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14076 PRNG state recovery was possible based on the output of
14077 one PRNG request appropriately sized to gain knowledge on
14078 'md' followed by enough consecutive 1-byte PRNG requests
14079 to traverse all of 'state'.
14081 1. When updating 'md_local' (the current thread's copy of 'md')
14082 during PRNG output generation, hash all of the previous
14083 'md_local' value, not just the half used for PRNG output.
14085 2. Make the number of bytes from 'state' included into the hash
14086 independent from the number of PRNG bytes requested.
14088 The first measure alone would be sufficient to avoid
14089 Markku-Juhani's attack. (Actually it had never occurred
14090 to me that the half of 'md_local' used for chaining was the
14091 half from which PRNG output bytes were taken -- I had always
14092 assumed that the secret half would be used.) The second
14093 measure makes sure that additional data from 'state' is never
14094 mixed into 'md_local' in small portions; this heuristically
14095 further strengthens the PRNG.
14099 * Fix crypto/bn/asm/mips3.s.
14103 * When only the key is given to "enc", the IV is undefined. Print out
14104 an error message in this case.
14108 * Handle special case when X509_NAME is empty in X509 printing routines.
14112 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14113 positive and less than q.
14117 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14118 used: it isn't thread safe and the add_lock_callback should handle
14121 *Paul Rose <Paul.Rose@bridge.com>*
14123 * Verify that incoming data obeys the block size in
14124 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14130 *Ulf Möller, Bodo Möller*
14132 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14133 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14134 when fixing the server behaviour for backwards-compatible 'client
14135 hello' messages. (Note that the attack is impractical against
14136 SSL 3.0 and TLS 1.0 anyway because length and version checking
14137 means that the probability of guessing a valid ciphertext is
14138 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14141 Before 0.9.5, the countermeasure (hide the error by generating a
14142 random 'decryption result') did not work properly because
14143 ERR_clear_error() was missing, meaning that SSL_get_error() would
14144 detect the supposedly ignored error.
14146 Both problems are now fixed.
14150 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14151 (previously it was 1024).
14155 * Fix for compatibility mode trust settings: ignore trust settings
14156 unless some valid trust or reject settings are present.
14160 * Fix for blowfish EVP: its a variable length cipher.
14164 * Fix various bugs related to DSA S/MIME verification. Handle missing
14165 parameters in DSA public key structures and return an error in the
14166 DSA routines if parameters are absent.
14170 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14171 in the current directory if neither $RANDFILE nor $HOME was set.
14172 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14173 caused some confusion to Windows users who haven't defined $HOME.
14174 Thus RAND_file_name() is changed again: e_os.h can define a
14175 DEFAULT_HOME, which will be used if $HOME is not set.
14176 For Windows, we use "C:"; on other platforms, we still require
14177 environment variables.
14179 * Move 'if (!initialized) RAND_poll()' into regions protected by
14180 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14181 having multiple threads call RAND_poll() concurrently.
14185 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14186 combination of a flag and a thread ID variable.
14187 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14188 flag), *other* threads can enter ssleay_add_bytes without obeying
14189 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14190 that they do not hold after the first thread unsets add_do_not_lock).
14194 * Change bctest again: '-x' expressions are not available in all
14195 versions of 'test'.
14199 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14201 * Fix a couple of memory leaks in PKCS7_dataDecode()
14203 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14205 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14206 the default extension for executables, if any. Also, make the perl
14207 scripts that use symlink() to test if it really exists and use "cp"
14208 if it doesn't. All this made OpenSSL compilable and installable in
14213 * Fix for asn1_GetSequence() for indefinite length constructed data.
14214 If SEQUENCE is length is indefinite just set c->slen to the total
14215 amount of data available.
14217 *Steve Henson, reported by shige@FreeBSD.org*
14219 *This change does not apply to 0.9.7.*
14221 * Change bctest to avoid here-documents inside command substitution
14222 (workaround for FreeBSD /bin/sh bug).
14223 For compatibility with Ultrix, avoid shell functions (introduced
14224 in the bctest version that searches along $PATH).
14228 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14229 with des_encrypt() defined on some operating systems, like Solaris
14234 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14235 On the Importance of Eliminating Errors in Cryptographic
14236 Computations, J. Cryptology 14 (2001) 2, 101-119,
14237 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14241 * MIPS assembler BIGNUM division bug fix.
14245 * Disabled incorrect Alpha assembler code.
14249 * Fix PKCS#7 decode routines so they correctly update the length
14250 after reading an EOC for the EXPLICIT tag.
14254 *This change does not apply to 0.9.7.*
14256 * Fix bug in PKCS#12 key generation routines. This was triggered
14257 if a 3DES key was generated with a 0 initial byte. Include
14258 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14259 (but broken) behaviour.
14263 * Enhance bctest to search for a working bc along $PATH and print
14266 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14268 * Fix memory leaks in err.c: free err_data string if necessary;
14269 don't write to the wrong index in ERR_set_error_data.
14273 * Implement ssl23_peek (analogous to ssl23_read), which previously
14278 * Replace rdtsc with `_emit` statements for VC++ version 5.
14280 *Jeremy Cooper <jeremy@baymoo.org>*
14282 * Make it possible to reuse SSLv2 sessions.
14286 * In copy_email() check for >= 0 as a return value for
14287 X509_NAME_get_index_by_NID() since 0 is a valid index.
14289 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14291 * Avoid coredump with unsupported or invalid public keys by checking if
14292 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14293 PKCS7_verify() fails with non detached data.
14297 * Don't use getenv in library functions when run as setuid/setgid.
14298 New function OPENSSL_issetugid().
14302 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14303 due to incorrect handling of multi-threading:
14305 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14307 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14309 3. Count how many times MemCheck_off() has been called so that
14310 nested use can be treated correctly. This also avoids
14311 inband-signalling in the previous code (which relied on the
14312 assumption that thread ID 0 is impossible).
14316 * Add "-rand" option also to s_client and s_server.
14320 * Fix CPU detection on Irix 6.x.
14321 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14322 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14324 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14329 *This change does not apply to 0.9.7.*
14331 * Use the cached encoding of an X509_NAME structure rather than
14332 copying it. This is apparently the reason for the libsafe "errors"
14333 but the code is actually correct.
14337 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14338 Bleichenbacher's DSA attack.
14339 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14340 to be set and top=0 forces the highest bit to be set; top=-1 is new
14341 and leaves the highest bit random.
14343 *Ulf Moeller, Bodo Moeller*
14345 * In the `NCONF_...`-based implementations for `CONF_...` queries
14346 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14347 a temporary CONF structure with the data component set to NULL
14348 (which gives segmentation faults in lh_retrieve).
14349 Instead, use NULL for the CONF pointer in CONF_get_string and
14350 CONF_get_number (which may use environment variables) and directly
14351 return NULL from CONF_get_section.
14355 * Fix potential buffer overrun for EBCDIC.
14359 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14360 keyUsage if basicConstraints absent for a CA.
14364 * Make SMIME_write_PKCS7() write mail header values with a format that
14365 is more generally accepted (no spaces before the semicolon), since
14366 some programs can't parse those values properly otherwise. Also make
14367 sure BIO's that break lines after each write do not create invalid
14372 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14373 macros previously used would not encode an empty SEQUENCE OF
14374 and break the signature.
14378 *This change does not apply to 0.9.7.*
14380 * Zero the premaster secret after deriving the master secret in
14385 * Add some EVP_add_digest_alias registrations (as found in
14386 OpenSSL_add_all_digests()) to SSL_library_init()
14387 aka OpenSSL_add_ssl_algorithms(). This provides improved
14388 compatibility with peers using X.509 certificates
14389 with unconventional AlgorithmIdentifier OIDs.
14393 * Fix for Irix with NO_ASM.
14395 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14397 * ./config script fixes.
14399 *Ulf Moeller, Richard Levitte*
14401 * Fix 'openssl passwd -1'.
14405 * Change PKCS12_key_gen_asc() so it can cope with non null
14406 terminated strings whose length is passed in the passlen
14407 parameter, for example from PEM callbacks. This was done
14408 by adding an extra length parameter to asc2uni().
14410 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14412 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14413 call failed, free the DSA structure.
14417 * Fix to uni2asc() to cope with zero length Unicode strings.
14418 These are present in some PKCS#12 files.
14422 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14423 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14424 when writing a 32767 byte record.
14426 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14428 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14429 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14431 (RSA objects have a reference count access to which is protected
14432 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14433 so they are meant to be shared between threads.)
14434 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14435 "Reddie, Steven" <Steven.Reddie@ca.com>*
14437 * Fix a deadlock in CRYPTO_mem_leaks().
14441 * Use better test patterns in bntest.
14445 * rand_win.c fix for Borland C.
14449 * BN_rshift bugfix for n == 0.
14453 * Add a 'bctest' script that checks for some known 'bc' bugs
14454 so that 'make test' does not abort just because 'bc' is broken.
14458 * Store verify_result within SSL_SESSION also for client side to
14459 avoid potential security hole. (Re-used sessions on the client side
14460 always resulted in verify_result==X509_V_OK, not using the original
14461 result of the server certificate verification.)
14465 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14466 SSL3_RT_APPLICATION_DATA, return 0.
14467 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14472 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14473 releases, have been re-implemented by renaming the previous
14474 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14475 and ssl3_read_internal, respectively, and adding 'peek' parameters
14476 to them. The new ssl[23]_{read,peek} functions are calls to
14477 ssl[23]_read_internal with the 'peek' flag set appropriately.
14478 A 'peek' parameter has also been added to ssl3_read_bytes, which
14479 does the actual work for ssl3_read_internal.
14483 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14484 the method-specific "init()" handler. Also clean up ex_data after
14485 calling the method-specific "finish()" handler. Previously, this was
14486 happening the other way round.
14490 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14491 The previous value, 12, was not always sufficient for BN_mod_exp().
14495 * Make sure that shared libraries get the internal name engine with
14496 the full version number and not just 0. This should mark the
14497 shared libraries as not backward compatible. Of course, this should
14498 be changed again when we can guarantee backward binary compatibility.
14502 * Fix typo in get_cert_by_subject() in by_dir.c
14504 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14506 * Rework the system to generate shared libraries:
14508 - Make note of the expected extension for the shared libraries and
14509 if there is a need for symbolic links from for example libcrypto.so.0
14510 to libcrypto.so.0.9.7. There is extended info in Configure for
14513 - Make as few rebuilds of the shared libraries as possible.
14515 - Still avoid linking the OpenSSL programs with the shared libraries.
14517 - When installing, install the shared libraries separately from the
14522 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14524 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14525 and not in SSL_clear because the latter is also used by the
14526 accept/connect functions; previously, the settings made by
14527 SSL_set_read_ahead would be lost during the handshake.
14529 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14531 * Correct util/mkdef.pl to be selective about disabled algorithms.
14532 Previously, it would create entries for disabled algorithms no
14537 * Added several new manual pages for SSL_* function.
14541 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14543 * In ssl23_get_client_hello, generate an error message when faced
14544 with an initial SSL 3.0/TLS record that is too small to contain the
14545 first two bytes of the ClientHello message, i.e. client_version.
14546 (Note that this is a pathologic case that probably has never happened
14547 in real life.) The previous approach was to use the version number
14548 from the record header as a substitute; but our protocol choice
14549 should not depend on that one because it is not authenticated
14550 by the Finished messages.
14554 * More robust randomness gathering functions for Windows.
14556 *Jeffrey Altman <jaltman@columbia.edu>*
14558 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14559 not set then we don't setup the error code for issuer check errors
14560 to avoid possibly overwriting other errors which the callback does
14561 handle. If an application does set the flag then we assume it knows
14562 what it is doing and can handle the new informational codes
14567 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14568 a general "ANY" type, as such it should be able to decode anything
14569 including tagged types. However it didn't check the class so it would
14570 wrongly interpret tagged types in the same way as their universal
14571 counterpart and unknown types were just rejected. Changed so that the
14572 tagged and unknown types are handled in the same way as a SEQUENCE:
14573 that is the encoding is stored intact. There is also a new type
14574 "V_ASN1_OTHER" which is used when the class is not universal, in this
14575 case we have no idea what the actual type is so we just lump them all
14580 * On VMS, stdout may very well lead to a file that is written to
14581 in a record-oriented fashion. That means that every write() will
14582 write a separate record, which will be read separately by the
14583 programs trying to read from it. This can be very confusing.
14585 The solution is to put a BIO filter in the way that will buffer
14586 text until a linefeed is reached, and then write everything a
14587 line at a time, so every record written will be an actual line,
14588 not chunks of lines and not (usually doesn't happen, but I've
14589 seen it once) several lines in one record. BIO_f_linebuffer() is
14592 Currently, it's a VMS-only method, because that's where it has
14593 been tested well enough.
14597 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14598 it can return incorrect results.
14599 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14600 but it was in 0.9.6-beta[12].)
14604 * Disable the check for content being present when verifying detached
14605 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14606 include zero length content when signing messages.
14610 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14611 BIO_ctrl (for BIO pairs).
14615 * Add DSO method for VMS.
14619 * Bug fix: Montgomery multiplication could produce results with the
14624 * Add RPM specification openssl.spec and modify it to build three
14625 packages. The default package contains applications, application
14626 documentation and run-time libraries. The devel package contains
14627 include files, static libraries and function documentation. The
14628 doc package contains the contents of the doc directory. The original
14629 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14633 * Add a large number of documentation files for many SSL routines.
14635 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14637 * Add a configuration entry for Sony News 4.
14639 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14641 * Don't set the two most significant bits to one when generating a
14642 random number < q in the DSA library.
14646 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14647 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14648 the underlying transport is blocking) if a handshake took place.
14649 (The default behaviour is needed by applications such as s_client
14650 and s_server that use select() to determine when to use SSL_read;
14651 but for applications that know in advance when to expect data, it
14652 just makes things more complicated.)
14656 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14661 * Add a few more EBCDIC conditionals that make `req` and `x509`
14662 work better on such systems.
14664 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14666 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14667 Update PKCS12_parse() so it copies the friendlyName and the
14668 keyid to the certificates aux info.
14672 * Fix bug in PKCS7_verify() which caused an infinite loop
14673 if there was more than one signature.
14675 *Sven Uszpelkat <su@celocom.de>*
14677 * Major change in util/mkdef.pl to include extra information
14678 about each symbol, as well as presenting variables as well
14679 as functions. This change means that there's n more need
14680 to rebuild the .num files when some algorithms are excluded.
14684 * Allow the verify time to be set by an application,
14685 rather than always using the current time.
14689 * Phase 2 verify code reorganisation. The certificate
14690 verify code now looks up an issuer certificate by a
14691 number of criteria: subject name, authority key id
14692 and key usage. It also verifies self signed certificates
14693 by the same criteria. The main comparison function is
14694 X509_check_issued() which performs these checks.
14696 Lot of changes were necessary in order to support this
14697 without completely rewriting the lookup code.
14699 Authority and subject key identifier are now cached.
14701 The LHASH 'certs' is X509_STORE has now been replaced
14702 by a STACK_OF(X509_OBJECT). This is mainly because an
14703 LHASH can't store or retrieve multiple objects with
14704 the same hash value.
14706 As a result various functions (which were all internal
14707 use only) have changed to handle the new X509_STORE
14708 structure. This will break anything that messed round
14709 with X509_STORE internally.
14711 The functions X509_STORE_add_cert() now checks for an
14712 exact match, rather than just subject name.
14714 The X509_STORE API doesn't directly support the retrieval
14715 of multiple certificates matching a given criteria, however
14716 this can be worked round by performing a lookup first
14717 (which will fill the cache with candidate certificates)
14718 and then examining the cache for matches. This is probably
14719 the best we can do without throwing out X509_LOOKUP
14720 entirely (maybe later...).
14722 The X509_VERIFY_CTX structure has been enhanced considerably.
14724 All certificate lookup operations now go via a get_issuer()
14725 callback. Although this currently uses an X509_STORE it
14726 can be replaced by custom lookups. This is a simple way
14727 to bypass the X509_STORE hackery necessary to make this
14728 work and makes it possible to use more efficient techniques
14729 in future. A very simple version which uses a simple
14730 STACK for its trusted certificate store is also provided
14731 using X509_STORE_CTX_trusted_stack().
14733 The verify_cb() and verify() callbacks now have equivalents
14734 in the X509_STORE_CTX structure.
14736 X509_STORE_CTX also has a 'flags' field which can be used
14737 to customise the verify behaviour.
14741 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14742 excludes S/MIME capabilities.
14746 * When a certificate request is read in keep a copy of the
14747 original encoding of the signed data and use it when outputting
14748 again. Signatures then use the original encoding rather than
14749 a decoded, encoded version which may cause problems if the
14750 request is improperly encoded.
14754 * For consistency with other BIO_puts implementations, call
14755 buffer_write(b, ...) directly in buffer_puts instead of calling
14758 In BIO_puts, increment b->num_write as in BIO_write.
14760 *Peter.Sylvester@EdelWeb.fr*
14762 * Fix BN_mul_word for the case where the word is 0. (We have to use
14763 BN_zero, we may not return a BIGNUM with an array consisting of
14764 words set to zero.)
14768 * Avoid calling abort() from within the library when problems are
14769 detected, except if preprocessor symbols have been defined
14770 (such as REF_CHECK, BN_DEBUG etc.).
14774 * New openssl application 'rsautl'. This utility can be
14775 used for low level RSA operations. DER public key
14776 BIO/fp routines also added.
14780 * New Configure entry and patches for compiling on QNX 4.
14782 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14784 * A demo state-machine implementation was sponsored by
14785 Nuron (<http://www.nuron.com/>) and is now available in
14786 demos/state_machine.
14790 * New options added to the 'dgst' utility for signature
14791 generation and verification.
14795 * Unrecognized PKCS#7 content types are now handled via a
14796 catch all ASN1_TYPE structure. This allows unsupported
14797 types to be stored as a "blob" and an application can
14798 encode and decode it manually.
14802 * Fix various signed/unsigned issues to make a_strex.c
14803 compile under VC++.
14805 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14807 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14808 length if passed a buffer. ASN1_INTEGER_to_BN failed
14809 if passed a NULL BN and its argument was negative.
14811 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14813 * Modification to PKCS#7 encoding routines to output definite
14814 length encoding. Since currently the whole structures are in
14815 memory there's not real point in using indefinite length
14816 constructed encoding. However if OpenSSL is compiled with
14817 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14821 * Added BIO_vprintf() and BIO_vsnprintf().
14825 * Added more prefixes to parse for in the strings written
14826 through a logging bio, to cover all the levels that are available
14827 through syslog. The prefixes are now:
14829 PANIC, EMERG, EMR => LOG_EMERG
14830 ALERT, ALR => LOG_ALERT
14831 CRIT, CRI => LOG_CRIT
14832 ERROR, ERR => LOG_ERR
14833 WARNING, WARN, WAR => LOG_WARNING
14834 NOTICE, NOTE, NOT => LOG_NOTICE
14835 INFO, INF => LOG_INFO
14836 DEBUG, DBG => LOG_DEBUG
14838 and as before, if none of those prefixes are present at the
14839 beginning of the string, LOG_ERR is chosen.
14841 On Win32, the `LOG_*` levels are mapped according to this:
14843 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14844 LOG_WARNING => EVENTLOG_WARNING_TYPE
14845 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14849 * Made it possible to reconfigure with just the configuration
14850 argument "reconf" or "reconfigure". The command line arguments
14851 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14852 and are retrieved from there when reconfiguring.
14858 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14860 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14864 * The obj_dat.pl script was messing up the sorting of object
14865 names. The reason was that it compared the quoted version
14866 of strings as a result "OCSP" > "OCSP Signing" because
14867 " > SPACE. Changed script to store unquoted versions of
14868 names and add quotes on output. It was also omitting some
14869 names from the lookup table if they were given a default
14870 value (that is if SN is missing it is given the same
14871 value as LN and vice versa), these are now added on the
14872 grounds that if an object has a name we should be able to
14873 look it up. Finally added warning output when duplicate
14874 short or long names are found.
14878 * Changes needed for Tandem NSK.
14880 *Scott Uroff <scott@xypro.com>*
14882 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14883 RSA_padding_check_SSLv23(), special padding was never detected
14884 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14885 version rollback attacks was not effective.
14887 In s23_clnt.c, don't use special rollback-attack detection padding
14888 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14889 client; similarly, in s23_srvr.c, don't do the rollback check if
14890 SSL 2.0 is the only protocol enabled in the server.
14894 * Make it possible to get hexdumps of unprintable data with 'openssl
14895 asn1parse'. By implication, the functions ASN1_parse_dump() and
14896 BIO_dump_indent() are added.
14900 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14901 these print out strings and name structures based on various
14902 flags including RFC2253 support and proper handling of
14903 multibyte characters. Added options to the 'x509' utility
14904 to allow the various flags to be set.
14908 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14909 Also change the functions X509_cmp_current_time() and
14910 X509_gmtime_adj() work with an ASN1_TIME structure,
14911 this will enable certificates using GeneralizedTime in validity
14912 dates to be checked.
14916 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14917 negative public key encodings) on by default,
14918 NO_NEG_PUBKEY_BUG can be set to disable it.
14922 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14923 content octets. An i2c_ASN1_OBJECT is unnecessary because
14924 the encoding can be trivially obtained from the structure.
14928 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14929 not read locks (`CRYPTO_r_[un]lock`).
14933 * A first attempt at creating official support for shared
14934 libraries through configuration. I've kept it so the
14935 default is static libraries only, and the OpenSSL programs
14936 are always statically linked for now, but there are
14937 preparations for dynamic linking in place.
14938 This has been tested on Linux and Tru64.
14942 * Randomness polling function for Win9x, as described in:
14943 Peter Gutmann, Software Generation of Practically Strong
14948 * Fix so PRNG is seeded in req if using an already existing
14953 * New options to smime application. -inform and -outform
14954 allow alternative formats for the S/MIME message including
14955 PEM and DER. The -content option allows the content to be
14956 specified separately. This should allow things like Netscape
14957 form signing output easier to verify.
14961 * Fix the ASN1 encoding of tags using the 'long form'.
14965 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14966 STRING types. These convert content octets to and from the
14967 underlying type. The actual tag and length octets are
14968 already assumed to have been read in and checked. These
14969 are needed because all other string types have virtually
14970 identical handling apart from the tag. By having versions
14971 of the ASN1 functions that just operate on content octets
14972 IMPLICIT tagging can be handled properly. It also allows
14973 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
14974 and ASN1_INTEGER are identical apart from the tag.
14978 * Change the handling of OID objects as follows:
14980 - New object identifiers are inserted in objects.txt, following
14981 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
14982 - objects.pl is used to process obj_mac.num and create a new
14984 - obj_dat.pl is used to create a new obj_dat.h, using the data in
14987 This is currently kind of a hack, and the perl code in objects.pl
14988 isn't very elegant, but it works as I intended. The simplest way
14989 to check that it worked correctly is to look in obj_dat.h and
14990 check the array nid_objs and make sure the objects haven't moved
14991 around (this is important!). Additions are OK, as well as
14992 consistent name changes.
14996 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15000 * Addition of the command line parameter '-rand file' to 'openssl req'.
15001 The given file adds to whatever has already been seeded into the
15002 random pool through the RANDFILE configuration file option or
15003 environment variable, or the default random state file.
15007 * mkstack.pl now sorts each macro group into lexical order.
15008 Previously the output order depended on the order the files
15009 appeared in the directory, resulting in needless rewriting
15014 * Patches to make OpenSSL compile under Win32 again. Mostly
15015 work arounds for the VC++ problem that it treats func() as
15016 func(void). Also stripped out the parts of mkdef.pl that
15017 added extra typesafe functions: these no longer exist.
15021 * Reorganisation of the stack code. The macros are now all
15022 collected in safestack.h . Each macro is defined in terms of
15023 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15024 DEBUG_SAFESTACK is now handled in terms of function casts,
15025 this has the advantage of retaining type safety without the
15026 use of additional functions. If DEBUG_SAFESTACK is not defined
15027 then the non typesafe macros are used instead. Also modified the
15028 mkstack.pl script to handle the new form. Needs testing to see
15029 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15030 the default if no major problems. Similar behaviour for ASN1_SET_OF
15031 and PKCS12_STACK_OF.
15035 * When some versions of IIS use the 'NET' form of private key the
15036 key derivation algorithm is different. Normally MD5(password) is
15037 used as a 128 bit RC4 key. In the modified case
15038 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15039 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15040 as the old Netscape_RSA functions except they have an additional
15041 'sgckey' parameter which uses the modified algorithm. Also added
15042 an -sgckey command line option to the rsa utility. Thanks to
15043 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15044 algorithm to openssl-dev.
15048 * The evp_local.h macros were using 'c.##kname' which resulted in
15049 invalid expansion on some systems (SCO 5.0.5 for example).
15050 Corrected to 'c.kname'.
15052 *Phillip Porch <root@theporch.com>*
15054 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15055 a STACK of email addresses from a certificate or request, these look
15056 in the subject name and the subject alternative name extensions and
15057 omit any duplicate addresses.
15061 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15062 This makes DSA verification about 2 % faster.
15066 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15067 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15068 plus overhead for 1024 bit moduli).
15069 This makes exponentiations about 0.5 % faster for 1024 bit
15070 exponents (as measured by "openssl speed rsa2048").
15074 * Rename memory handling macros to avoid conflicts with other
15076 Malloc => OPENSSL_malloc
15077 Malloc_locked => OPENSSL_malloc_locked
15078 Realloc => OPENSSL_realloc
15079 Free => OPENSSL_free
15083 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15084 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15088 * CygWin32 support.
15090 *John Jarvie <jjarvie@newsguy.com>*
15092 * The type-safe stack code has been rejigged. It is now only compiled
15093 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15094 by default all type-specific stack functions are "#define"d back to
15095 standard stack functions. This results in more streamlined output
15096 but retains the type-safety checking possibilities of the original
15101 * The STACK code has been cleaned up, and certain type declarations
15102 that didn't make a lot of sense have been brought in line. This has
15103 also involved a cleanup of sorts in safestack.h to more correctly
15104 map type-safe stack functions onto their plain stack counterparts.
15105 This work has also resulted in a variety of "const"ifications of
15106 lots of the code, especially `_cmp` operations which should normally
15107 be prototyped with "const" parameters anyway.
15111 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15112 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15113 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15114 where all of 'md' is used each time the PRNG is used, but 'state'
15115 is used only indexed by a cyclic counter. As entropy may not be
15116 well distributed from the beginning, 'md' is important as a
15117 chaining variable. However, the output function chains only half
15118 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15119 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15120 in all of 'state' being rewritten, with the new values depending
15121 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15125 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15126 the handshake is continued after ssl_verify_cert_chain();
15127 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15128 can lead to 'unexplainable' connection aborts later.
15130 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15132 * Major EVP API cipher revision.
15133 Add hooks for extra EVP features. This allows various cipher
15134 parameters to be set in the EVP interface. Support added for variable
15135 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15136 setting of RC2 and RC5 parameters.
15138 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15141 Remove lots of duplicated code from the EVP library. For example *every*
15142 cipher init() function handles the 'iv' in the same way according to the
15143 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15144 for CFB and OFB modes they zero ctx->num.
15146 New functionality allows removal of S/MIME code RC2 hack.
15148 Most of the routines have the same form and so can be declared in terms
15151 By shifting this to the top level EVP_CipherInit() it can be removed from
15152 all individual ciphers. If the cipher wants to handle IVs or keys
15153 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15156 Change lots of functions like EVP_EncryptUpdate() to now return a
15157 value: although software versions of the algorithms cannot fail
15158 any installed hardware versions can.
15162 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15163 this option is set, tolerate broken clients that send the negotiated
15164 protocol version number instead of the requested protocol version
15169 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15170 i.e. non-zero for export ciphersuites, zero otherwise.
15171 Previous versions had this flag inverted, inconsistent with
15172 rsa_tmp_cb (..._TMP_RSA_CB).
15174 *Bodo Moeller; problem reported by Amit Chopra*
15176 * Add missing DSA library text string. Work around for some IIS
15177 key files with invalid SEQUENCE encoding.
15181 * Add a document (doc/standards.txt) that list all kinds of standards
15182 and so on that are implemented in OpenSSL.
15186 * Enhance c_rehash script. Old version would mishandle certificates
15187 with the same subject name hash and wouldn't handle CRLs at all.
15188 Added -fingerprint option to crl utility, to support new c_rehash
15193 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15197 * Fix for SSL server purpose checking. Server checking was
15198 rejecting certificates which had extended key usage present
15199 but no ssl client purpose.
15201 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15203 * Make PKCS#12 code work with no password. The PKCS#12 spec
15204 is a little unclear about how a blank password is handled.
15205 Since the password in encoded as a BMPString with terminating
15206 double NULL a zero length password would end up as just the
15207 double NULL. However no password at all is different and is
15208 handled differently in the PKCS#12 key generation code. NS
15209 treats a blank password as zero length. MSIE treats it as no
15210 password on export: but it will try both on import. We now do
15211 the same: PKCS12_parse() tries zero length and no password if
15212 the password is set to "" or NULL (NULL is now a valid password:
15213 it wasn't before) as does the pkcs12 application.
15217 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15218 perror when PEM_read_bio_X509_REQ fails, the error message must
15219 be obtained from the error queue.
15223 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15224 it in ERR_remove_state if appropriate, and change ERR_get_state
15225 accordingly to avoid race conditions (this is necessary because
15226 thread_hash is no longer constant once set).
15230 * Bugfix for linux-elf makefile.one.
15234 * RSA_get_default_method() will now cause a default
15235 RSA_METHOD to be chosen if one doesn't exist already.
15236 Previously this was only set during a call to RSA_new()
15237 or RSA_new_method(NULL) meaning it was possible for
15238 RSA_get_default_method() to return NULL.
15242 * Added native name translation to the existing DSO code
15243 that will convert (if the flag to do so is set) filenames
15244 that are sufficiently small and have no path information
15245 into a canonical native form. Eg. "blah" converted to
15246 "libblah.so" or "blah.dll" etc.
15250 * New function ERR_error_string_n(e, buf, len) which is like
15251 ERR_error_string(e, buf), but writes at most 'len' bytes
15252 including the 0 terminator. For ERR_error_string_n, 'buf'
15255 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15257 * CONF library reworked to become more general. A new CONF
15258 configuration file reader "class" is implemented as well as a
15259 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15260 old `CONF_*` functions are still there, but are reimplemented to
15261 work in terms of the new functions. Also, a set of functions
15262 to handle the internal storage of the configuration data is
15263 provided to make it easier to write new configuration file
15264 reader "classes" (I can definitely see something reading a
15265 configuration file in XML format, for example), called `_CONF_*`,
15266 or "the configuration storage API"...
15268 The new configuration file reading functions are:
15270 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15271 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15273 NCONF_default, NCONF_WIN32
15275 NCONF_dump_fp, NCONF_dump_bio
15277 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15278 NCONF_new creates a new CONF object. This works in the same way
15279 as other interfaces in OpenSSL, like the BIO interface.
15280 `NCONF_dump_*` dump the internal storage of the configuration file,
15281 which is useful for debugging. All other functions take the same
15282 arguments as the old `CONF_*` functions with the exception of the
15283 first that must be a `CONF *` instead of a `LHASH *`.
15285 To make it easier to use the new classes with the old `CONF_*` functions,
15286 the function CONF_set_default_method is provided.
15290 * Add '-tls1' option to 'openssl ciphers', which was already
15291 mentioned in the documentation but had not been implemented.
15292 (This option is not yet really useful because even the additional
15293 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15297 * Initial DSO code added into libcrypto for letting OpenSSL (and
15298 OpenSSL-based applications) load shared libraries and bind to
15299 them in a portable way.
15301 *Geoff Thorpe, with contributions from Richard Levitte*
15303 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15305 * Make sure _lrotl and _lrotr are only used with MSVC.
15307 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15308 (the default implementation of RAND_status).
15310 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15311 to '-clrext' (= clear extensions), as intended and documented.
15312 *Bodo Moeller; inconsistency pointed out by Michael Attili
15313 <attili@amaxo.com>*
15315 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15316 was larger than the MD block size.
15318 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15320 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15321 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15322 using the passed key: if the passed key was a private key the result
15323 of X509_print(), for example, would be to print out all the private key
15328 * des_quad_cksum() byte order bug fix.
15329 *Ulf Möller, using the problem description in krb4-0.9.7, where
15330 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15332 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15335 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15337 * For easily testing in shell scripts whether some command
15338 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15339 returns with exit code 0 iff no command of the given name is available.
15340 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15341 the output goes to stdout and nothing is printed to stderr.
15342 Additional arguments are always ignored.
15344 Since for each cipher there is a command of the same name,
15345 the 'no-cipher' compilation switches can be tested this way.
15347 ('openssl no-XXX' is not able to detect pseudo-commands such
15348 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15352 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15356 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15357 is set; it will be thrown away anyway because each handshake creates
15359 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15360 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15361 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15362 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15366 * New s_client option -ign_eof: EOF at stdin is ignored, and
15367 'Q' and 'R' lose their special meanings (quit/renegotiate).
15368 This is part of what -quiet does; unlike -quiet, -ign_eof
15369 does not suppress any output.
15373 * Add compatibility options to the purpose and trust code. The
15374 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15375 accepts a certificate or CA, this was the previous behaviour,
15376 with all the associated security issues.
15378 X509_TRUST_COMPAT is the old trust behaviour: only and
15379 automatically trust self signed roots in certificate store. A
15380 new trust setting X509_TRUST_DEFAULT is used to specify that
15381 a purpose has no associated trust setting and it should instead
15382 use the value in the default purpose.
15386 * Fix the PKCS#8 DSA private key code so it decodes keys again
15387 and fix a memory leak.
15391 * In util/mkerr.pl (which implements 'make errors'), preserve
15392 reason strings from the previous version of the .c file, as
15393 the default to have only downcase letters (and digits) in
15394 automatically generated reasons codes is not always appropriate.
15398 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15399 using strerror. Previously, ERR_reason_error_string() returned
15400 library names as reason strings for SYSerr; but SYSerr is a special
15401 case where small numbers are errno values, not library numbers.
15405 * Add '-dsaparam' option to 'openssl dhparam' application. This
15406 converts DSA parameters into DH parameters. (When creating parameters,
15407 DSA_generate_parameters is used.)
15411 * Include 'length' (recommended exponent length) in C code generated
15412 by 'openssl dhparam -C'.
15416 * The second argument to set_label in perlasm was already being used
15417 so couldn't be used as a "file scope" flag. Moved to third argument
15422 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15423 instead of RAND_bytes for encryption IVs and salts.
15427 * Include RAND_status() into RAND_METHOD instead of implementing
15428 it only for md_rand.c Otherwise replacing the PRNG by calling
15429 RAND_set_rand_method would be impossible.
15433 * Don't let DSA_generate_key() enter an infinite loop if the random
15434 number generation fails.
15438 * New 'rand' application for creating pseudo-random output.
15442 * Added configuration support for Linux/IA64
15444 *Rolf Haberrecker <rolf@suse.de>*
15446 * Assembler module support for Mingw32.
15450 * Shared library support for HPUX (in shlib/).
15452 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15454 * Shared library support for Solaris gcc.
15456 *Lutz Behnke <behnke@trustcenter.de>*
15458 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15460 * PKCS7_encrypt() was adding text MIME headers twice because they
15461 were added manually and by SMIME_crlf_copy().
15465 * In bntest.c don't call BN_rand with zero bits argument.
15467 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15469 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15470 case was implemented. This caused BN_div_recp() to fail occasionally.
15474 * Add an optional second argument to the set_label() in the perl
15475 assembly language builder. If this argument exists and is set
15476 to 1 it signals that the assembler should use a symbol whose
15477 scope is the entire file, not just the current function. This
15478 is needed with MASM which uses the format label:: for this scope.
15480 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15482 * Change the ASN1 types so they are typedefs by default. Before
15483 almost all types were #define'd to ASN1_STRING which was causing
15484 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15489 * Change names of new functions to the new get1/get0 naming
15490 convention: After 'get1', the caller owns a reference count
15491 and has to call `..._free`; 'get0' returns a pointer to some
15492 data structure without incrementing reference counters.
15493 (Some of the existing 'get' functions increment a reference
15494 counter, some don't.)
15495 Similarly, 'set1' and 'add1' functions increase reference
15496 counters or duplicate objects.
15500 * Allow for the possibility of temp RSA key generation failure:
15501 the code used to assume it always worked and crashed on failure.
15505 * Fix potential buffer overrun problem in BIO_printf().
15506 *Ulf Möller, using public domain code by Patrick Powell; problem
15507 pointed out by David Sacerdote <das33@cornell.edu>*
15509 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15510 RAND_egd() and RAND_status(). In the command line application,
15511 the EGD socket can be specified like a seed file using RANDFILE
15516 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15517 Some CAs (e.g. Verisign) distribute certificates in this form.
15521 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15522 list to exclude them. This means that no special compilation option
15523 is needed to use anonymous DH: it just needs to be included in the
15528 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15529 EVP_MD_type. The old functionality is available in a new macro called
15530 EVP_MD_md(). Change code that uses it and update docs.
15534 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15535 where the `void *` argument is replaced by a function pointer argument.
15536 Previously `void *` was abused to point to functions, which works on
15537 many platforms, but is not correct. As these functions are usually
15538 called by macros defined in OpenSSL header files, most source code
15539 should work without changes.
15543 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15544 sections with information on -D... compiler switches used for
15545 compiling the library so that applications can see them. To enable
15546 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15547 must be defined. E.g.,
15548 #define OPENSSL_ALGORITHM_DEFINES
15549 #include <openssl/opensslconf.h>
15550 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15552 *Richard Levitte, Ulf and Bodo Möller*
15554 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15559 * Change the 'other' type in certificate aux info to a STACK_OF
15560 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15561 the required ASN1 format: arbitrary types determined by an OID.
15565 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15566 argument to 'req'. This is not because the function is newer or
15567 better than others it just uses the work 'NEW' in the certificate
15568 request header lines. Some software needs this.
15572 * Reorganise password command line arguments: now passwords can be
15573 obtained from various sources. Delete the PEM_cb function and make
15574 it the default behaviour: i.e. if the callback is NULL and the
15575 usrdata argument is not NULL interpret it as a null terminated pass
15576 phrase. If usrdata and the callback are NULL then the pass phrase
15577 is prompted for as usual.
15581 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15582 the support is automatically enabled. The resulting binaries will
15583 autodetect the card and use it if present.
15585 *Ben Laurie and Compaq Inc.*
15587 * Work around for Netscape hang bug. This sends certificate request
15588 and server done in one record. Since this is perfectly legal in the
15589 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15590 the bugs/SSLv3 entry for more info.
15594 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15598 * Add -rand argument to smime and pkcs12 applications and read/write
15603 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15607 * Add command line password options to the remaining applications.
15611 * Bug fix for BN_div_recp() for numerators with an even number of
15616 * More tests in bntest.c, and changed test_bn output.
15620 * ./config recognizes MacOS X now.
15624 * Bug fix for BN_div() when the first words of num and divisor are
15625 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15629 * Add support for various broken PKCS#8 formats, and command line
15630 options to produce them.
15634 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15635 get temporary BIGNUMs from a BN_CTX.
15639 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15644 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15645 include a #define from the old name to the new. The original intent
15646 was that statically linked binaries could for example just call
15647 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15648 link with digests. This never worked because SSLeay_add_all_digests()
15649 and SSLeay_add_all_ciphers() were in the same source file so calling
15650 one would link with the other. They are now in separate source files.
15654 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15658 * Use a less unusual form of the Miller-Rabin primality test (it used
15659 a binary algorithm for exponentiation integrated into the Miller-Rabin
15660 loop, our standard modexp algorithms are faster).
15664 * Support for the EBCDIC character set completed.
15666 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15668 * Source code cleanups: use const where appropriate, eliminate casts,
15669 use `void *` instead of `char *` in lhash.
15673 * Bugfix: ssl3_send_server_key_exchange was not restartable
15674 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15675 this the server could overwrite ephemeral keys that the client
15680 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15681 using 50 iterations of the Rabin-Miller test.
15683 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15684 iterations of the Rabin-Miller test as required by the appendix
15685 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15686 As BN_is_prime_fasttest includes trial division, DSA parameter
15687 generation becomes much faster.
15689 This implies a change for the callback functions in DSA_is_prime
15690 and DSA_generate_parameters: The callback function is called once
15691 for each positive witness in the Rabin-Miller test, not just
15692 occasionally in the inner loop; and the parameters to the
15693 callback function now provide an iteration count for the outer
15694 loop rather than for the current invocation of the inner loop.
15695 DSA_generate_parameters additionally can call the callback
15696 function with an 'iteration count' of -1, meaning that a
15697 candidate has passed the trial division test (when q is generated
15698 from an application-provided seed, trial division is skipped).
15702 * New function BN_is_prime_fasttest that optionally does trial
15703 division before starting the Rabin-Miller test and has
15704 an additional BN_CTX * argument (whereas BN_is_prime always
15705 has to allocate at least one BN_CTX).
15706 'callback(1, -1, cb_arg)' is called when a number has passed the
15707 trial division stage.
15711 * Fix for bug in CRL encoding. The validity dates weren't being handled
15716 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15720 * New function BN_pseudo_rand().
15724 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15725 bignum version of BN_from_montgomery() with the working code from
15726 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15731 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15732 made it impossible to use the same SSL_SESSION data structure in
15733 SSL2 clients in multiple threads.
15737 * The return value of RAND_load_file() no longer counts bytes obtained
15738 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15739 to seed the PRNG (previously an explicit byte count was required).
15741 *Ulf Möller, Bodo Möller*
15743 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15744 used `char *` instead of `void *` and had casts all over the place.
15748 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15752 * Retain source code compatibility for BN_prime_checks macro:
15753 BN_is_prime(..., BN_prime_checks, ...) now uses
15754 BN_prime_checks_for_size to determine the appropriate number of
15755 Rabin-Miller iterations.
15759 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15760 DH_CHECK_P_NOT_SAFE_PRIME.
15761 (Check if this is true? OpenPGP calls them "strong".)
15765 * Merge the functionality of "dh" and "gendh" programs into a new program
15766 "dhparam". The old programs are retained for now but will handle DH keys
15767 (instead of parameters) in future.
15771 * Make the ciphers, s_server and s_client programs check the return values
15772 when a new cipher list is set.
15776 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15777 ciphers. Before when the 56bit ciphers were enabled the sorting was
15780 The syntax for the cipher sorting has been extended to support sorting by
15781 cipher-strength (using the strength_bits hard coded in the tables).
15782 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15784 Fix a bug in the cipher-command parser: when supplying a cipher command
15785 string with an "undefined" symbol (neither command nor alphanumeric
15786 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15787 an error is flagged.
15789 Due to the strength-sorting extension, the code of the
15790 ssl_create_cipher_list() function was completely rearranged. I hope that
15791 the readability was also increased :-)
15793 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15795 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15796 for the first serial number and places 2 in the serial number file. This
15797 avoids problems when the root CA is created with serial number zero and
15798 the first user certificate has the same issuer name and serial number
15803 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15804 the new code. Add documentation for this stuff.
15808 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15809 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15810 structures and behave in an analogous way to the X509v3 functions:
15811 they shouldn't be called directly but wrapper functions should be used
15814 So we also now have some wrapper functions that call the X509at functions
15815 when passed certificate requests. (TO DO: similar things can be done with
15816 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15817 things. Some of these need some d2i or i2d and print functionality
15818 because they handle more complex structures.)
15822 * Add missing #ifndefs that caused missing symbols when building libssl
15823 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15824 NO_RSA in `ssl/s2*.c`.
15826 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15828 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15829 has a return value which indicates the quality of the random data
15830 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15831 error queue. New function RAND_pseudo_bytes() generates output that is
15832 guaranteed to be unique but not unpredictable. RAND_add is like
15833 RAND_seed, but takes an extra argument for an entropy estimate
15834 (RAND_seed always assumes full entropy).
15838 * Do more iterations of Rabin-Miller probable prime test (specifically,
15839 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15840 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15841 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15842 false-positive rate of at most 2^-80 for random input.
15846 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15850 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15851 in the 0.9.5 release), this returns the chain
15852 from an X509_CTX structure with a dup of the stack and all
15853 the X509 reference counts upped: so the stack will exist
15854 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15857 Also make SSL_SESSION_print() print out the verify return
15862 * Add manpage for the pkcs12 command. Also change the default
15863 behaviour so MAC iteration counts are used unless the new
15864 -nomaciter option is used. This improves file security and
15865 only older versions of MSIE (4.0 for example) need it.
15869 * Honor the no-xxx Configure options when creating .DEF files.
15873 * Add PKCS#10 attributes to field table: challengePassword,
15874 unstructuredName and unstructuredAddress. These are taken from
15875 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15876 international characters are used.
15878 More changes to X509_ATTRIBUTE code: allow the setting of types
15879 based on strings. Remove the 'loc' parameter when adding
15880 attributes because these will be a SET OF encoding which is sorted
15885 * Initial changes to the 'req' utility to allow request generation
15886 automation. This will allow an application to just generate a template
15887 file containing all the field values and have req construct the
15890 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15891 used all over the place including certificate requests and PKCS#7
15892 structures. They are currently handled manually where necessary with
15893 some primitive wrappers for PKCS#7. The new functions behave in a
15894 manner analogous to the X509 extension functions: they allow
15895 attributes to be looked up by NID and added.
15897 Later something similar to the X509V3 code would be desirable to
15898 automatically handle the encoding, decoding and printing of the
15899 more complex types. The string types like challengePassword can
15900 be handled by the string table functions.
15902 Also modified the multi byte string table handling. Now there is
15903 a 'global mask' which masks out certain types. The table itself
15904 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15905 is useful when for example there is only one permissible type
15906 (as in countryName) and using the mask might result in no valid
15911 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15912 SSL_get_peer_finished to allow applications to obtain the latest
15913 Finished messages sent to the peer or expected from the peer,
15914 respectively. (SSL_get_peer_finished is usually the Finished message
15915 actually received from the peer, otherwise the protocol will be aborted.)
15917 As the Finished message are message digests of the complete handshake
15918 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15919 be used for external authentication procedures when the authentication
15920 provided by SSL/TLS is not desired or is not enough.
15924 * Enhanced support for Alpha Linux is added. Now ./config checks if
15925 the host supports BWX extension and if Compaq C is present on the
15926 $PATH. Just exploiting of the BWX extension results in 20-30%
15927 performance kick for some algorithms, e.g. DES and RC4 to mention
15928 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15933 * Add support for MS "fast SGC". This is arguably a violation of the
15934 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15935 weak crypto and after checking the certificate is SGC a second one
15936 with strong crypto. MS SGC stops the first handshake after receiving
15937 the server certificate message and sends a second client hello. Since
15938 a server will typically do all the time consuming operations before
15939 expecting any further messages from the client (server key exchange
15940 is the most expensive) there is little difference between the two.
15942 To get OpenSSL to support MS SGC we have to permit a second client
15943 hello message after we have sent server done. In addition we have to
15944 reset the MAC if we do get this second client hello.
15948 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15949 if a DER encoded private key is RSA or DSA traditional format. Changed
15950 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15951 format DER encoded private key. Newer code should use PKCS#8 format which
15952 has the key type encoded in the ASN1 structure. Added DER private key
15953 support to pkcs8 application.
15957 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15958 ciphersuites has been selected (as required by the SSL 3/TLS 1
15959 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15960 is set, we interpret this as a request to violate the specification
15961 (the worst that can happen is a handshake failure, and 'correct'
15962 behaviour would result in a handshake failure anyway).
15966 * In SSL_CTX_add_session, take into account that there might be multiple
15967 SSL_SESSION structures with the same session ID (e.g. when two threads
15968 concurrently obtain them from an external cache).
15969 The internal cache can handle only one SSL_SESSION with a given ID,
15970 so if there's a conflict, we now throw out the old one to achieve
15975 * Add OIDs for idea and blowfish in CBC mode. This will allow both
15976 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
15977 some routines that use cipher OIDs: some ciphers do not have OIDs
15978 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
15983 * Simplify the trust setting structure and code. Now we just have
15984 two sequences of OIDs for trusted and rejected settings. These will
15985 typically have values the same as the extended key usage extension
15986 and any application specific purposes.
15988 The trust checking code now has a default behaviour: it will just
15989 check for an object with the same NID as the passed id. Functions can
15990 be provided to override either the default behaviour or the behaviour
15991 for a given id. SSL client, server and email already have functions
15992 in place for compatibility: they check the NID and also return "trusted"
15993 if the certificate is self signed.
15997 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
15998 traditional format into an EVP_PKEY structure.
16002 * Add a password callback function PEM_cb() which either prompts for
16003 a password if usr_data is NULL or otherwise assumes it is a null
16004 terminated password. Allow passwords to be passed on command line
16005 environment or config files in a few more utilities.
16009 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16010 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16011 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16012 Update documentation.
16016 * Support for ASN1 "NULL" type. This could be handled before by using
16017 ASN1_TYPE but there wasn't any function that would try to read a NULL
16018 and produce an error if it couldn't. For compatibility we also have
16019 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16020 don't allocate anything because they don't need to.
16024 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16027 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16029 * Rebuild of the memory allocation routines used by OpenSSL code and
16030 possibly others as well. The purpose is to make an interface that
16031 provide hooks so anyone can build a separate set of allocation and
16032 deallocation routines to be used by OpenSSL, for example memory
16033 pool implementations, or something else, which was previously hard
16034 since Malloc(), Realloc() and Free() were defined as macros having
16035 the values malloc, realloc and free, respectively (except for Win32
16036 compilations). The same is provided for memory debugging code.
16037 OpenSSL already comes with functionality to find memory leaks, but
16038 this gives people a chance to debug other memory problems.
16040 With these changes, a new set of functions and macros have appeared:
16042 CRYPTO_set_mem_debug_functions() [F]
16043 CRYPTO_get_mem_debug_functions() [F]
16044 CRYPTO_dbg_set_options() [F]
16045 CRYPTO_dbg_get_options() [F]
16046 CRYPTO_malloc_debug_init() [M]
16048 The memory debug functions are NULL by default, unless the library
16049 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16050 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16051 gives the standard debugging functions that come with OpenSSL) or
16052 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16053 provided by the library user) must be used. When the standard
16054 debugging functions are used, CRYPTO_dbg_set_options can be used to
16055 request additional information:
16056 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16057 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16059 Also, things like CRYPTO_set_mem_functions will always give the
16060 expected result (the new set of functions is used for allocation
16061 and deallocation) at all times, regardless of platform and compiler
16064 To finish it up, some functions that were never use in any other
16065 way than through macros have a new API and new semantic:
16067 CRYPTO_dbg_malloc()
16068 CRYPTO_dbg_realloc()
16071 All macros of value have retained their old syntax.
16073 *Richard Levitte and Bodo Moeller*
16075 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16076 ordering of SMIMECapabilities wasn't in "strength order" and there
16077 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16082 * Some ASN1 types with illegal zero length encoding (INTEGER,
16083 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16085 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16087 * Merge in my S/MIME library for OpenSSL. This provides a simple
16088 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16089 functionality to handle multipart/signed properly) and a utility
16090 called 'smime' to call all this stuff. This is based on code I
16091 originally wrote for Celo who have kindly allowed it to be
16092 included in OpenSSL.
16096 * Add variants des_set_key_checked and des_set_key_unchecked of
16097 des_set_key (aka des_key_sched). Global variable des_check_key
16098 decides which of these is called by des_set_key; this way
16099 des_check_key behaves as it always did, but applications and
16100 the library itself, which was buggy for des_check_key == 1,
16101 have a cleaner way to pick the version they need.
16105 * New function PKCS12_newpass() which changes the password of a
16110 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16111 dynamic mix. In both cases the ids can be used as an index into the
16112 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16113 functions so they accept a list of the field values and the
16114 application doesn't need to directly manipulate the X509_TRUST
16119 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16124 * Modify the way the V3 extension code looks up extensions. This now
16125 works in a similar way to the object code: we have some "standard"
16126 extensions in a static table which is searched with OBJ_bsearch()
16127 and the application can add dynamic ones if needed. The file
16128 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16129 updated whenever a new extension is added to the core code and kept
16130 in ext_nid order. There is a simple program 'tabtest.c' which checks
16131 this. New extensions are not added too often so this file can readily
16132 be maintained manually.
16134 There are two big advantages in doing things this way. The extensions
16135 can be looked up immediately and no longer need to be "added" using
16136 X509V3_add_standard_extensions(): this function now does nothing.
16137 Side note: I get *lots* of email saying the extension code doesn't
16138 work because people forget to call this function.
16139 Also no dynamic allocation is done unless new extensions are added:
16140 so if we don't add custom extensions there is no need to call
16141 X509V3_EXT_cleanup().
16145 * Modify enc utility's salting as follows: make salting the default. Add a
16146 magic header, so unsalted files fail gracefully instead of just decrypting
16147 to garbage. This is because not salting is a big security hole, so people
16148 should be discouraged from doing it.
16152 * Fixes and enhancements to the 'x509' utility. It allowed a message
16153 digest to be passed on the command line but it only used this
16154 parameter when signing a certificate. Modified so all relevant
16155 operations are affected by the digest parameter including the
16156 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16157 DSA key was used because it didn't fix the digest.
16161 * Initial certificate chain verify code. Currently tests the untrusted
16162 certificates for consistency with the verify purpose (which is set
16163 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16165 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16166 this is because it will reject chains with invalid extensions whereas
16167 every previous version of OpenSSL and SSLeay made no checks at all.
16169 Trust code: checks the root CA for the relevant trust settings. Trust
16170 settings have an initial value consistent with the verify purpose: e.g.
16171 if the verify purpose is for SSL client use it expects the CA to be
16172 trusted for SSL client use. However the default value can be changed to
16173 permit custom trust settings: one example of this would be to only trust
16174 certificates from a specific "secure" set of CAs.
16176 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16177 which should be used for version portability: especially since the
16178 verify structure is likely to change more often now.
16180 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16181 to set them. If not set then assume SSL clients will verify SSL servers
16184 Two new options to the verify program: -untrusted allows a set of
16185 untrusted certificates to be passed in and -purpose which sets the
16186 intended purpose of the certificate. If a purpose is set then the
16187 new chain verify code is used to check extension consistency.
16191 * Support for the authority information access extension.
16195 * Modify RSA and DSA PEM read routines to transparently handle
16196 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16197 public keys in a format compatible with certificate
16198 SubjectPublicKeyInfo structures. Unfortunately there were already
16199 functions called *_PublicKey_* which used various odd formats so
16200 these are retained for compatibility: however the DSA variants were
16201 never in a public release so they have been deleted. Changed dsa/rsa
16202 utilities to handle the new format: note no releases ever handled public
16203 keys so we should be OK.
16205 The primary motivation for this change is to avoid the same fiasco
16206 that dogs private keys: there are several incompatible private key
16207 formats some of which are standard and some OpenSSL specific and
16208 require various evil hacks to allow partial transparent handling and
16209 even then it doesn't work with DER formats. Given the option anything
16210 other than PKCS#8 should be dumped: but the other formats have to
16211 stay in the name of compatibility.
16213 With public keys and the benefit of hindsight one standard format
16214 is used which works with EVP_PKEY, RSA or DSA structures: though
16215 it clearly returns an error if you try to read the wrong kind of key.
16217 Added a -pubkey option to the 'x509' utility to output the public key.
16218 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16219 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16220 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16221 that do the same as the `EVP_PKEY_assign_*()` except they up the
16222 reference count of the added key (they don't "swallow" the
16227 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16228 CRLs would fail if the file contained no certificates or no CRLs:
16229 added a new function to read in both types and return the number
16230 read: this means that if none are read it will be an error. The
16231 DER versions of the certificate and CRL reader would always fail
16232 because it isn't possible to mix certificates and CRLs in DER format
16233 without choking one or the other routine. Changed this to just read
16234 a certificate: this is the best we can do. Also modified the code
16235 in apps/verify.c to take notice of return codes: it was previously
16236 attempting to read in certificates from NULL pointers and ignoring
16237 any errors: this is one reason why the cert and CRL reader seemed
16238 to work. It doesn't check return codes from the default certificate
16239 routines: these may well fail if the certificates aren't installed.
16243 * Code to support otherName option in GeneralName.
16247 * First update to verify code. Change the verify utility
16248 so it warns if it is passed a self signed certificate:
16249 for consistency with the normal behaviour. X509_verify
16250 has been modified to it will now verify a self signed
16251 certificate if *exactly* the same certificate appears
16252 in the store: it was previously impossible to trust a
16253 single self signed certificate. This means that:
16254 openssl verify ss.pem
16255 now gives a warning about a self signed certificate but
16256 openssl verify -CAfile ss.pem ss.pem
16261 * For servers, store verify_result in SSL_SESSION data structure
16262 (and add it to external session representation).
16263 This is needed when client certificate verifications fails,
16264 but an application-provided verification callback (set by
16265 SSL_CTX_set_cert_verify_callback) allows accepting the session
16266 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16267 but returns 1): When the session is reused, we have to set
16268 ssl->verify_result to the appropriate error code to avoid
16271 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16273 * Fix a bug in the new PKCS#7 code: it didn't consider the
16274 case in PKCS7_dataInit() where the signed PKCS7 structure
16275 didn't contain any existing data because it was being created.
16277 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16279 * Add a salt to the key derivation routines in enc.c. This
16280 forms the first 8 bytes of the encrypted file. Also add a
16281 -S option to allow a salt to be input on the command line.
16285 * New function X509_cmp(). Oddly enough there wasn't a function
16286 to compare two certificates. We do this by working out the SHA1
16287 hash and comparing that. X509_cmp() will be needed by the trust
16292 * SSL_get1_session() is like SSL_get_session(), but increments
16293 the reference count in the SSL_SESSION returned.
16295 *Geoff Thorpe <geoff@eu.c2.net>*
16297 * Fix for 'req': it was adding a null to request attributes.
16298 Also change the X509_LOOKUP and X509_INFO code to handle
16299 certificate auxiliary information.
16303 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16308 * Add the possibility to add extra information to the memory leak
16309 detecting output, to form tracebacks, showing from where each
16310 allocation was originated: CRYPTO_push_info("constant string") adds
16311 the string plus current file name and line number to a per-thread
16312 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16313 is like calling CYRPTO_pop_info() until the stack is empty.
16314 Also updated memory leak detection code to be multi-thread-safe.
16318 * Add options -text and -noout to pkcs7 utility and delete the
16319 encryption options which never did anything. Update docs.
16323 * Add options to some of the utilities to allow the pass phrase
16324 to be included on either the command line (not recommended on
16325 OSes like Unix) or read from the environment. Update the
16326 manpages and fix a few bugs.
16330 * Add a few manpages for some of the openssl commands.
16334 * Fix the -revoke option in ca. It was freeing up memory twice,
16335 leaking and not finding already revoked certificates.
16339 * Extensive changes to support certificate auxiliary information.
16340 This involves the use of X509_CERT_AUX structure and X509_AUX
16341 functions. An X509_AUX function such as PEM_read_X509_AUX()
16342 can still read in a certificate file in the usual way but it
16343 will also read in any additional "auxiliary information". By
16344 doing things this way a fair degree of compatibility can be
16345 retained: existing certificates can have this information added
16346 using the new 'x509' options.
16348 Current auxiliary information includes an "alias" and some trust
16349 settings. The trust settings will ultimately be used in enhanced
16350 certificate chain verification routines: currently a certificate
16351 can only be trusted if it is self signed and then it is trusted
16356 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16357 The problem was that one of the replacement routines had not been working
16358 since SSLeay releases. For now the offending routine has been replaced
16359 with non-optimised assembler. Even so, this now gives around 95%
16360 performance improvement for 1024 bit RSA signs.
16364 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16365 handling. Most clients have the effective key size in bits equal to
16366 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16367 A few however don't do this and instead use the size of the decrypted key
16368 to determine the RC2 key length and the AlgorithmIdentifier to determine
16369 the effective key length. In this case the effective key length can still
16370 be 40 bits but the key length can be 168 bits for example. This is fixed
16371 by manually forcing an RC2 key into the EVP_PKEY structure because the
16372 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16373 the key length and effective key length are equal.
16377 * Add a bunch of functions that should simplify the creation of
16378 X509_NAME structures. Now you should be able to do:
16379 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16380 and have it automatically work out the correct field type and fill in
16381 the structures. The more adventurous can try:
16382 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16383 and it will (hopefully) work out the correct multibyte encoding.
16387 * Change the 'req' utility to use the new field handling and multibyte
16388 copy routines. Before the DN field creation was handled in an ad hoc
16389 way in req, ca, and x509 which was rather broken and didn't support
16390 BMPStrings or UTF8Strings. Since some software doesn't implement
16391 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16392 using the dirstring_type option. See the new comment in the default
16393 openssl.cnf for more info.
16397 * Make crypto/rand/md_rand.c more robust:
16398 - Assure unique random numbers after fork().
16399 - Make sure that concurrent threads access the global counter and
16400 md serializably so that we never lose entropy in them
16401 or use exactly the same state in multiple threads.
16402 Access to the large state is not always serializable because
16403 the additional locking could be a performance killer, and
16404 md should be large enough anyway.
16408 * New file apps/app_rand.c with commonly needed functionality
16409 for handling the random seed file.
16411 Use the random seed file in some applications that previously did not:
16413 dsaparam -genkey (which also ignored its '-rand' option),
16416 x509 (when signing).
16417 Except on systems with /dev/urandom, it is crucial to have a random
16418 seed file at least for key creation, DSA signing, and for DH exchanges;
16419 for RSA signatures we could do without one.
16421 gendh and gendsa (unlike genrsa) used to read only the first byte
16422 of each file listed in the '-rand' option. The function as previously
16423 found in genrsa is now in app_rand.c and is used by all programs
16424 that support '-rand'.
16428 * In RAND_write_file, use mode 0600 for creating files;
16429 don't just chmod when it may be too late.
16433 * Report an error from X509_STORE_load_locations
16434 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16438 * New function ASN1_mbstring_copy() this copies a string in either
16439 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16440 into an ASN1_STRING type. A mask of permissible types is passed
16441 and it chooses the "minimal" type to use or an error if not type
16446 * Add function equivalents to the various macros in asn1.h. The old
16447 macros are retained with an `M_` prefix. Code inside the library can
16448 use the `M_` macros. External code (including the openssl utility)
16449 should *NOT* in order to be "shared library friendly".
16453 * Add various functions that can check a certificate's extensions
16454 to see if it usable for various purposes such as SSL client,
16455 server or S/MIME and CAs of these types. This is currently
16456 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16457 verification. Also added a -purpose flag to x509 utility to
16458 print out all the purposes.
16462 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16467 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16468 for, obtain and decode and extension and obtain its critical flag.
16469 This allows all the necessary extension code to be handled in a
16470 single function call.
16474 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16475 platforms. See crypto/rc4/rc4_enc.c for further details.
16479 * New -noout option to asn1parse. This causes no output to be produced
16480 its main use is when combined with -strparse and -out to extract data
16481 from a file (which may not be in ASN.1 format).
16485 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16486 when producing the local key id.
16488 *Richard Levitte <levitte@stacken.kth.se>*
16490 * New option -dhparam in s_server. This allows a DH parameter file to be
16491 stated explicitly. If it is not stated then it tries the first server
16492 certificate file. The previous behaviour hard coded the filename
16497 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16498 a public key to be input or output. For example:
16499 openssl rsa -in key.pem -pubout -out pubkey.pem
16500 Also added necessary DSA public key functions to handle this.
16504 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16505 in the message. This was handled by allowing
16506 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16508 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16510 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16511 to the end of the strings whereas this didn't. This would cause problems
16512 if strings read with d2i_ASN1_bytes() were later modified.
16514 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16516 * Fix for base64 decode bug. When a base64 bio reads only one line of
16517 data and it contains EOF it will end up returning an error. This is
16518 caused by input 46 bytes long. The cause is due to the way base64
16519 BIOs find the start of base64 encoded data. They do this by trying a
16520 trial decode on each line until they find one that works. When they
16521 do a flag is set and it starts again knowing it can pass all the
16522 data directly through the decoder. Unfortunately it doesn't reset
16523 the context it uses. This means that if EOF is reached an attempt
16524 is made to pass two EOFs through the context and this causes the
16525 resulting error. This can also cause other problems as well. As is
16526 usual with these problems it takes *ages* to find and the fix is
16527 trivial: move one line.
16529 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16531 * Ugly workaround to get s_client and s_server working under Windows. The
16532 old code wouldn't work because it needed to select() on sockets and the
16533 tty (for keypresses and to see if data could be written). Win32 only
16534 supports select() on sockets so we select() with a 1s timeout on the
16535 sockets and then see if any characters are waiting to be read, if none
16536 are present then we retry, we also assume we can always write data to
16537 the tty. This isn't nice because the code then blocks until we've
16538 received a complete line of data and it is effectively polling the
16539 keyboard at 1s intervals: however it's quite a bit better than not
16540 working at all :-) A dedicated Windows application might handle this
16541 with an event loop for example.
16545 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16546 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16547 will be called when RSA_sign() and RSA_verify() are used. This is useful
16548 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16549 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16550 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16551 This necessitated the support of an extra signature type NID_md5_sha1
16552 for SSL signatures and modifications to the SSL library to use it instead
16553 of calling RSA_public_decrypt() and RSA_private_encrypt().
16557 * Add new -verify -CAfile and -CApath options to the crl program, these
16558 will lookup a CRL issuers certificate and verify the signature in a
16559 similar way to the verify program. Tidy up the crl program so it
16560 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16561 less strict. It will now permit CRL extensions even if it is not
16562 a V2 CRL: this will allow it to tolerate some broken CRLs.
16566 * Initialize all non-automatic variables each time one of the openssl
16567 sub-programs is started (this is necessary as they may be started
16568 multiple times from the "OpenSSL>" prompt).
16570 *Lennart Bang, Bodo Moeller*
16572 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16573 removing all other RSA functionality (this is what NO_RSA does). This
16574 is so (for example) those in the US can disable those operations covered
16575 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16580 * Non-copying interface to BIO pairs.
16581 (still largely untested)
16585 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16586 ASCII string. This was handled independently in various places before.
16590 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16591 UTF8 strings a character at a time.
16595 * Use client_version from client hello to select the protocol
16596 (s23_srvr.c) and for RSA client key exchange verification
16597 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16601 * Add various utility functions to handle SPKACs, these were previously
16602 handled by poking round in the structure internals. Added new function
16603 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16604 print, verify and generate SPKACs. Based on an original idea from
16605 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16609 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16613 * Allow the config file extension section to be overwritten on the
16614 command line. Based on an original idea from Massimiliano Pala
16615 <madwolf@comune.modena.it>. The new option is called -extensions
16616 and can be applied to ca, req and x509. Also -reqexts to override
16617 the request extensions in req and -crlexts to override the crl extensions
16622 * Add new feature to the SPKAC handling in ca. Now you can include
16623 the same field multiple times by preceding it by "XXXX." for example:
16626 this is the same syntax as used in the req config file.
16630 * Allow certificate extensions to be added to certificate requests. These
16631 are specified in a 'req_extensions' option of the req section of the
16632 config file. They can be printed out with the -text option to req but
16633 are otherwise ignored at present.
16637 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16638 data read consists of only the final block it would not decrypted because
16639 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16640 A misplaced 'break' also meant the decrypted final block might not be
16641 copied until the next read.
16645 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16646 a few extra parameters to the DH structure: these will be useful if
16647 for example we want the value of 'q' or implement X9.42 DH.
16651 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16652 provides hooks that allow the default DSA functions or functions on a
16653 "per key" basis to be replaced. This allows hardware acceleration and
16654 hardware key storage to be handled without major modification to the
16655 library. Also added low level modexp hooks and CRYPTO_EX structure and
16656 associated functions.
16660 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16661 as "read only": it can't be written to and the buffer it points to will
16662 not be freed. Reading from a read only BIO is much more efficient than
16663 a normal memory BIO. This was added because there are several times when
16664 an area of memory needs to be read from a BIO. The previous method was
16665 to create a memory BIO and write the data to it, this results in two
16666 copies of the data and an O(n^2) reading algorithm. There is a new
16667 function BIO_new_mem_buf() which creates a read only memory BIO from
16668 an area of memory. Also modified the PKCS#7 routines to use read only
16673 * Bugfix: ssl23_get_client_hello did not work properly when called in
16674 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16675 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16676 but a retry condition occurred while trying to read the rest.
16680 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16681 NID_pkcs7_encrypted by default: this was wrong since this should almost
16682 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16683 the encrypted data type: this is a more sensible place to put it and it
16684 allows the PKCS#12 code to be tidied up that duplicated this
16689 * Changed obj_dat.pl script so it takes its input and output files on
16690 the command line. This should avoid shell escape redirection problems
16695 * Initial support for certificate extension requests, these are included
16696 in things like Xenroll certificate requests. Included functions to allow
16697 extensions to be obtained and added.
16701 * -crlf option to s_client and s_server for sending newlines as
16702 CRLF (as required by many protocols).
16706 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16708 * Install libRSAglue.a when OpenSSL is built with RSAref.
16710 *Ralf S. Engelschall*
16712 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16714 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16716 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16721 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16722 DH parameters/keys (q is lost during that conversion, but the resulting
16723 DH parameters contain its length).
16725 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16726 much faster than DH_generate_parameters (which creates parameters
16727 where `p = 2*q + 1`), and also the smaller q makes DH computations
16728 much more efficient (160-bit exponentiation instead of 1024-bit
16729 exponentiation); so this provides a convenient way to support DHE
16730 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16731 utter importance to use
16732 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16734 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16735 when such DH parameters are used, because otherwise small subgroup
16736 attacks may become possible!
16740 * Avoid memory leak in i2d_DHparams.
16744 * Allow the -k option to be used more than once in the enc program:
16745 this allows the same encrypted message to be read by multiple recipients.
16749 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16750 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16751 it will always use the numerical form of the OID, even if it has a short
16756 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16757 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16758 otherwise bn_mod_exp was called. In the case of hardware keys for example
16759 no private key components need be present and it might store extra data
16760 in the RSA structure, which cannot be accessed from bn_mod_exp.
16761 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16762 private key operations.
16766 * Added support for SPARC Linux.
16770 * pem_password_cb function type incompatibly changed from
16771 typedef int pem_password_cb(char *buf, int size, int rwflag);
16773 ....(char *buf, int size, int rwflag, void *userdata);
16774 so that applications can pass data to their callbacks:
16775 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16776 additional void * argument, which is just handed through whenever
16777 the password callback is called.
16779 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16781 New function SSL_CTX_set_default_passwd_cb_userdata.
16783 Compatibility note: As many C implementations push function arguments
16784 onto the stack in reverse order, the new library version is likely to
16785 interoperate with programs that have been compiled with the old
16786 pem_password_cb definition (PEM_whatever takes some data that
16787 happens to be on the stack as its last argument, and the callback
16788 just ignores this garbage); but there is no guarantee whatsoever that
16791 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16792 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16793 problems not only on Windows, but also on some Unix platforms.
16794 To avoid problematic command lines, these definitions are now in an
16795 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16796 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16800 * MIPS III/IV assembler module is reimplemented.
16804 * More DES library cleanups: remove references to srand/rand and
16805 delete an unused file.
16809 * Add support for the free Netwide assembler (NASM) under Win32,
16810 since not many people have MASM (ml) and it can be hard to obtain.
16811 This is currently experimental but it seems to work OK and pass all
16812 the tests. Check out INSTALL.W32 for info.
16816 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16817 without temporary keys kept an extra copy of the server key,
16818 and connections with temporary keys did not free everything in case
16823 * New function RSA_check_key and new openssl rsa option -check
16824 for verifying the consistency of RSA keys.
16826 *Ulf Moeller, Bodo Moeller*
16828 * Various changes to make Win32 compile work:
16829 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16830 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16831 comparison" warnings.
16832 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16836 * Add a debugging option to PKCS#5 v2 key generation function: when
16837 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16838 derived keys are printed to stderr.
16842 * Copy the flags in ASN1_STRING_dup().
16844 *Roman E. Pavlov <pre@mo.msk.ru>*
16846 * The x509 application mishandled signing requests containing DSA
16847 keys when the signing key was also DSA and the parameters didn't match.
16849 It was supposed to omit the parameters when they matched the signing key:
16850 the verifying software was then supposed to automatically use the CA's
16851 parameters if they were absent from the end user certificate.
16853 Omitting parameters is no longer recommended. The test was also
16854 the wrong way round! This was probably due to unusual behaviour in
16855 EVP_cmp_parameters() which returns 1 if the parameters match.
16856 This meant that parameters were omitted when they *didn't* match and
16857 the certificate was useless. Certificates signed with 'ca' didn't have
16860 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16862 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16863 The interface is as follows:
16864 Applications can use
16865 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16866 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16867 "off" is now the default.
16868 The library internally uses
16869 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16870 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16871 to disable memory-checking temporarily.
16873 Some inconsistent states that previously were possible (and were
16874 even the default) are now avoided.
16876 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16877 with each memory chunk allocated; this is occasionally more helpful
16878 than just having a counter.
16880 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16882 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16887 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16888 which largely parallels "options", but is for changing API behaviour,
16889 whereas "options" are about protocol behaviour.
16890 Initial "mode" flags are:
16892 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16893 a single record has been written.
16894 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16895 retries use the same buffer location.
16896 (But all of the contents must be
16901 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16904 * Fix problems with no-hmac etc.
16906 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16908 * New functions RSA_get_default_method(), RSA_set_method() and
16909 RSA_get_method(). These allows replacement of RSA_METHODs without having
16910 to mess around with the internals of an RSA structure.
16914 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16915 Also really enable memory leak checks in openssl.c and in some
16918 *Chad C. Mulligan, Bodo Moeller*
16920 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16921 up the length of negative integers. This has now been simplified to just
16922 store the length when it is first determined and use it later, rather
16923 than trying to keep track of where data is copied and updating it to
16925 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16927 * Add a new function PKCS7_signatureVerify. This allows the verification
16928 of a PKCS#7 signature but with the signing certificate passed to the
16929 function itself. This contrasts with PKCS7_dataVerify which assumes the
16930 certificate is present in the PKCS#7 structure. This isn't always the
16931 case: certificates can be omitted from a PKCS#7 structure and be
16932 distributed by "out of band" means (such as a certificate database).
16936 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16937 function prototypes in pem.h, also change util/mkdef.pl to add the
16938 necessary function names.
16942 * mk1mf.pl (used by Windows builds) did not properly read the
16943 options set by Configure in the top level Makefile, and Configure
16944 was not even able to write more than one option correctly.
16945 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16949 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16950 file to be loaded from a BIO or FILE pointer. The BIO version will
16951 for example allow memory BIOs to contain config info.
16955 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16956 Whoever hopes to achieve shared-library compatibility across versions
16957 must use this, not the compile-time macro.
16958 (Exercise 0.9.4: Which is the minimum library version required by
16960 Note: All this applies only to multi-threaded programs, others don't
16965 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16966 through a BIO pair triggered the default case, i.e.
16967 SSLerr(...,SSL_R_UNKNOWN_STATE).
16971 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16972 can use the SSL library even if none of the specific BIOs is
16977 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
16978 for the encoded length.
16980 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
16982 * Add initial documentation of the X509V3 functions.
16986 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
16987 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
16988 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
16989 secure PKCS#8 private key format with a high iteration count.
16993 * Fix determination of Perl interpreter: A perl or perl5
16994 *directory* in $PATH was also accepted as the interpreter.
16996 *Ralf S. Engelschall*
16998 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
16999 wrong with it but it was very old and did things like calling
17000 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17001 unusual formatting.
17005 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17006 to use the new extension code.
17010 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17011 with macros. This should make it easier to change their form, add extra
17012 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17017 * Add to configuration table a new entry that can specify an alternative
17018 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17019 according to Mark Crispin <MRC@Panda.COM>.
17023 * DES CBC did not update the IV. Weird.
17027 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17028 Changing the behaviour of the former might break existing programs --
17029 where IV updating is needed, des_ncbc_encrypt can be used.
17032 * When bntest is run from "make test" it drives bc to check its
17033 calculations, as well as internally checking them. If an internal check
17034 fails, it needs to cause bc to give a non-zero result or make test carries
17035 on without noticing the failure. Fixed.
17039 * DES library cleanups.
17043 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17044 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17045 ciphers. NOTE: although the key derivation function has been verified
17046 against some published test vectors it has not been extensively tested
17047 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17052 * Instead of "mkdir -p", which is not fully portable, use new
17053 Perl script "util/mkdir-p.pl".
17057 * Rewrite the way password based encryption (PBE) is handled. It used to
17058 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17059 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17060 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17061 the 'parameter' field of the AlgorithmIdentifier is passed to the
17062 underlying key generation function so it must do its own ASN1 parsing.
17063 This has also changed the EVP_PBE_CipherInit() function which now has a
17064 'parameter' argument instead of literal salt and iteration count values
17065 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17069 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17070 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17071 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17072 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17073 value was just used as a "magic string" and not used directly its
17074 value doesn't matter.
17078 * Introduce some semblance of const correctness to BN. Shame C doesn't
17083 * "linux-sparc64" configuration (ultrapenguin).
17085 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17086 "linux-sparc" configuration.
17088 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17090 * config now generates no-xxx options for missing ciphers.
17094 * Support the EBCDIC character set (work in progress).
17095 File ebcdic.c not yet included because it has a different license.
17097 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17099 * Support BS2000/OSD-POSIX.
17101 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17103 * Make callbacks for key generation use `void *` instead of `char *`.
17107 * Make S/MIME samples compile (not yet tested).
17111 * Additional typesafe stacks.
17115 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17119 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17121 * New configuration variant "sco5-gcc".
17123 * Updated some demos.
17125 *Sean O Riordain, Wade Scholine*
17127 * Add missing BIO_free at exit of pkcs12 application.
17131 * Fix memory leak in conf.c.
17135 * Updates for Win32 to assembler version of MD5.
17139 * Set #! path to perl in apps/der_chop to where we found it
17140 instead of using a fixed path.
17144 * SHA library changes for irix64-mips4-cc.
17148 * Improvements for VMS support.
17152 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17154 * Bignum library bug fix. IRIX 6 passes "make test" now!
17155 This also avoids the problems with SC4.2 and unpatched SC5.
17157 *Andy Polyakov <appro@fy.chalmers.se>*
17159 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17160 These are required because of the typesafe stack would otherwise break
17161 existing code. If old code used a structure member which used to be STACK
17162 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17163 sk_num or sk_value it would produce an error because the num, data members
17164 are not present in STACK_OF. Now it just produces a warning. sk_set
17165 replaces the old method of assigning a value to sk_value
17166 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17167 that does this will no longer work (and should use sk_set instead) but
17168 this could be regarded as a "questionable" behaviour anyway.
17172 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17173 correctly handle encrypted S/MIME data.
17177 * Change type of various DES function arguments from des_cblock
17178 (which means, in function argument declarations, pointer to char)
17179 to des_cblock * (meaning pointer to array with 8 char elements),
17180 which allows the compiler to do more typechecking; it was like
17181 that back in SSLeay, but with lots of ugly casts.
17183 Introduce new type const_des_cblock.
17187 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17188 problems: find RecipientInfo structure that matches recipient certificate
17189 and initialise the ASN1 structures properly based on passed cipher.
17193 * Belatedly make the BN tests actually check the results.
17197 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17198 to and from BNs: it was completely broken. New compilation option
17199 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17200 key elements as negative integers.
17204 * Reorganize and speed up MD5.
17206 *Andy Polyakov <appro@fy.chalmers.se>*
17210 *Richard Levitte <richard@levitte.org>*
17212 * New option -out to asn1parse to allow the parsed structure to be
17213 output to a file. This is most useful when combined with the -strparse
17214 option to examine the output of things like OCTET STRINGS.
17218 * Make SSL library a little more fool-proof by not requiring any longer
17219 that `SSL_set_{accept,connect}_state` be called before
17220 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17221 in many applications because usually everything *appeared* to work as
17222 intended anyway -- now it really works as intended).
17226 * Move openssl.cnf out of lib/.
17230 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17231 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17232 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17234 *Ralf S. Engelschall*
17236 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17237 handle PKCS#7 enveloped data properly.
17239 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17241 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17242 copying pointers. The cert_st handling is changed by this in
17243 various ways (and thus what used to be known as ctx->default_cert
17244 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17245 any longer when s->cert does not give us what we need).
17246 ssl_cert_instantiate becomes obsolete by this change.
17247 As soon as we've got the new code right (possibly it already is?),
17248 we have solved a couple of bugs of the earlier code where s->cert
17249 was used as if it could not have been shared with other SSL structures.
17251 Note that using the SSL API in certain dirty ways now will result
17252 in different behaviour than observed with earlier library versions:
17253 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17254 does not influence s as it used to.
17256 In order to clean up things more thoroughly, inside SSL_SESSION
17257 we don't use CERT any longer, but a new structure SESS_CERT
17258 that holds per-session data (if available); currently, this is
17259 the peer's certificate chain and, for clients, the server's certificate
17260 and temporary key. CERT holds only those values that can have
17261 meaningful defaults in an SSL_CTX.
17265 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17266 from the internal representation. Various PKCS#7 fixes: remove some
17267 evil casts and set the enc_dig_alg field properly based on the signing
17272 * Allow PKCS#12 password to be set from the command line or the
17273 environment. Let 'ca' get its config file name from the environment
17274 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17279 * Allow certificate policies extension to use an IA5STRING for the
17280 organization field. This is contrary to the PKIX definition but
17281 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17286 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17287 without disallowing inline assembler and the like for non-pedantic builds.
17291 * Support Borland C++ builder.
17293 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17299 * SHA-1 cleanups and performance enhancements.
17301 *Andy Polyakov <appro@fy.chalmers.se>*
17303 * Sparc v8plus assembler for the bignum library.
17305 *Andy Polyakov <appro@fy.chalmers.se>*
17307 * Accept any -xxx and +xxx compiler options in Configure.
17311 * Update HPUX configuration.
17315 * Add missing `sk_<type>_unshift()` function to safestack.h
17317 *Ralf S. Engelschall*
17319 * New function SSL_CTX_use_certificate_chain_file that sets the
17320 "extra_cert"s in addition to the certificate. (This makes sense
17321 only for "PEM" format files, as chains as a whole are not
17326 * Support verify_depth from the SSL API.
17327 x509_vfy.c had what can be considered an off-by-one-error:
17328 Its depth (which was not part of the external interface)
17329 was actually counting the number of certificates in a chain;
17330 now it really counts the depth.
17334 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17335 instead of X509err, which often resulted in confusing error
17336 messages since the error codes are not globally unique
17337 (e.g. an alleged error in ssl3_accept when a certificate
17338 didn't match the private key).
17340 * New function SSL_CTX_set_session_id_context that allows to set a default
17341 value (so that you don't need SSL_set_session_id_context for each
17342 connection using the SSL_CTX).
17346 * OAEP decoding bug fix.
17350 * Support INSTALL_PREFIX for package builders, as proposed by
17355 * New Configure options "threads" and "no-threads". For systems
17356 where the proper compiler options are known (currently Solaris
17357 and Linux), "threads" is the default.
17361 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17365 * Install various scripts to $(OPENSSLDIR)/misc, not to
17366 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17367 such as /usr/local/bin.
17371 * "make linux-shared" to build shared libraries.
17373 *Niels Poppe <niels@netbox.org>*
17375 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17379 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17380 extension adding in x509 utility.
17384 * Remove NOPROTO sections and error code comments.
17388 * Partial rewrite of the DEF file generator to now parse the ANSI
17393 * New Configure options --prefix=DIR and --openssldir=DIR.
17397 * Complete rewrite of the error code script(s). It is all now handled
17398 by one script at the top level which handles error code gathering,
17399 header rewriting and C source file generation. It should be much better
17400 than the old method: it now uses a modified version of Ulf's parser to
17401 read the ANSI prototypes in all header files (thus the old K&R definitions
17402 aren't needed for error creation any more) and do a better job of
17403 translating function codes into names. The old 'ASN1 error code embedded
17404 in a comment' is no longer necessary and it doesn't use .err files which
17405 have now been deleted. Also the error code call doesn't have to appear all
17406 on one line (which resulted in some large lines...).
17410 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17414 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17415 0 (which usually indicates a closed connection), but continue reading.
17419 * Fix some race conditions.
17423 * Add support for CRL distribution points extension. Add Certificate
17424 Policies and CRL distribution points documentation.
17428 * Move the autogenerated header file parts to crypto/opensslconf.h.
17432 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17433 8 of keying material. Merlin has also confirmed interop with this fix
17434 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17436 *Merlin Hughes <merlin@baltimore.ie>*
17438 * Fix lots of warnings.
17440 *Richard Levitte <levitte@stacken.kth.se>*
17442 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17443 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17445 *Richard Levitte <levitte@stacken.kth.se>*
17447 * Fix problems with sizeof(long) == 8.
17449 *Andy Polyakov <appro@fy.chalmers.se>*
17451 * Change functions to ANSI C.
17455 * Fix typos in error codes.
17457 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17459 * Remove defunct assembler files from Configure.
17463 * SPARC v8 assembler BIGNUM implementation.
17465 *Andy Polyakov <appro@fy.chalmers.se>*
17467 * Support for Certificate Policies extension: both print and set.
17468 Various additions to support the r2i method this uses.
17472 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17473 return a const string when you are expecting an allocated buffer.
17477 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17478 types DirectoryString and DisplayText.
17482 * Add code to allow r2i extensions to access the configuration database,
17483 add an LHASH database driver and add several ctx helper functions.
17487 * Fix an evil bug in bn_expand2() which caused various BN functions to
17488 fail when they extended the size of a BIGNUM.
17492 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17493 support typesafe stack.
17497 * Fix typo in SSL_[gs]et_options().
17499 *Nils Frostberg <nils@medcom.se>*
17501 * Delete various functions and files that belonged to the (now obsolete)
17502 old X509V3 handling code.
17506 * New Configure option "rsaref".
17510 * Don't auto-generate pem.h.
17514 * Introduce type-safe ASN.1 SETs.
17518 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17520 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17522 * Introduce type-safe STACKs. This will almost certainly break lots of code
17523 that links with OpenSSL (well at least cause lots of warnings), but fear
17524 not: the conversion is trivial, and it eliminates loads of evil casts. A
17525 few STACKed things have been converted already. Feel free to convert more.
17526 In the fullness of time, I'll do away with the STACK type altogether.
17530 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17531 specified in `<certfile>` by updating the entry in the index.txt file.
17532 This way one no longer has to edit the index.txt file manually for
17533 revoking a certificate. The -revoke option does the gory details now.
17535 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17537 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17538 `-text` option at all and this way the `-noout -text` combination was
17539 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17541 *Ralf S. Engelschall*
17543 * Make sure a corresponding plain text error message exists for the
17544 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17545 verify callback function determined that a certificate was revoked.
17547 *Ralf S. Engelschall*
17549 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17550 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17551 all available ciphers including rc5, which was forgotten until now.
17552 In order to let the testing shell script know which algorithms
17553 are available, a new (up to now undocumented) command
17554 `openssl list-cipher-commands` is used.
17558 * Bugfix: s_client occasionally would sleep in select() when
17559 it should have checked SSL_pending() first.
17563 * New functions DSA_do_sign and DSA_do_verify to provide access to
17564 the raw DSA values prior to ASN.1 encoding.
17568 * Tweaks to Configure
17570 *Niels Poppe <niels@netbox.org>*
17572 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17577 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17581 * New config option to avoid instructions that are illegal on the 80386.
17582 The default code is faster, but requires at least a 486.
17586 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17587 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17588 same as SSL2_VERSION anyway.
17592 * New "-showcerts" option for s_client.
17596 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17597 application. Various cleanups and fixes.
17601 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17602 modify error routines to work internally. Add error codes and PBE init
17603 to library startup routines.
17607 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17608 packing functions to asn1 and evp. Changed function names and error
17609 codes along the way.
17613 * PKCS12 integration: and so it begins... First of several patches to
17614 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17615 objects to objects.h
17619 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17620 and display support for Thawte strong extranet extension.
17624 * Add LinuxPPC support.
17626 *Jeff Dubrule <igor@pobox.org>*
17628 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17629 bn_div_words in alpha.s.
17631 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17633 * Make sure the RSA OAEP test is skipped under -DRSAref because
17634 OAEP isn't supported when OpenSSL is built with RSAref.
17636 *Ulf Moeller <ulf@fitug.de>*
17638 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17639 so they no longer are missing under -DNOPROTO.
17641 *Soren S. Jorvang <soren@t.dk>*
17643 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17645 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17646 doesn't work when the session is reused. Coming soon!
17650 * Fix a security hole, that allows sessions to be reused in the wrong
17651 context thus bypassing client cert protection! All software that uses
17652 client certs and session caches in multiple contexts NEEDS PATCHING to
17653 allow session reuse! A fuller solution is in the works.
17655 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17657 * Some more source tree cleanups (removed obsolete files
17658 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17659 permission on "config" script to be executable) and a fix for the INSTALL
17662 *Ulf Moeller <ulf@fitug.de>*
17664 * Remove some legacy and erroneous uses of malloc, free instead of
17667 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17669 * Make rsa_oaep_test return non-zero on error.
17671 *Ulf Moeller <ulf@fitug.de>*
17673 * Add support for native Solaris shared libraries. Configure
17674 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17675 if someone would make that last step automatic.
17677 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17679 * ctx_size was not built with the right compiler during "make links". Fixed.
17683 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17684 except NULL ciphers". This means the default cipher list will no longer
17685 enable NULL ciphers. They need to be specifically enabled e.g. with
17686 the string "DEFAULT:eNULL".
17690 * Fix to RSA private encryption routines: if p < q then it would
17691 occasionally produce an invalid result. This will only happen with
17692 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17696 * Be less restrictive and allow also `perl util/perlpath.pl
17697 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17698 because this way one can also use an interpreter named `perl5` (which is
17699 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17700 installed as `perl`).
17702 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17704 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17706 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17708 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17709 advapi32.lib to Win32 build and change the pem test comparison
17710 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17711 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17712 and crypto/des/ede_cbcm_enc.c.
17716 * DES quad checksum was broken on big-endian architectures. Fixed.
17720 * Comment out two functions in bio.h that aren't implemented. Fix up the
17721 Win32 test batch file so it (might) work again. The Win32 test batch file
17722 is horrible: I feel ill....
17726 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17727 in e_os.h. Audit of header files to check ANSI and non ANSI
17728 sections: 10 functions were absent from non ANSI section and not exported
17729 from Windows DLLs. Fixed up libeay.num for new functions.
17733 * Make `openssl version` output lines consistent.
17735 *Ralf S. Engelschall*
17737 * Fix Win32 symbol export lists for BIO functions: Added
17738 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17739 to ms/libeay{16,32}.def.
17741 *Ralf S. Engelschall*
17743 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17744 fine under Unix and passes some trivial tests I've now added. But the
17745 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17746 added to make sure no one expects that this stuff really works in the
17747 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17748 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17751 *Ralf S. Engelschall*
17753 * Fix the generation of two part addresses in perl.
17755 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17757 * Add config entry for Linux on MIPS.
17759 *John Tobey <jtobey@channel1.com>*
17761 * Make links whenever Configure is run, unless we are on Windoze.
17765 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17766 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17771 * Add a useful kludge to allow package maintainers to specify compiler and
17772 other platforms details on the command line without having to patch the
17773 Configure script every time: One now can use
17774 `perl Configure <id>:<details>`,
17775 i.e. platform ids are allowed to have details appended
17776 to them (separated by colons). This is treated as there would be a static
17777 pre-configured entry in Configure's %table under key `<id>` with value
17778 `<details>` and `perl Configure <id>` is called. So, when you want to
17779 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17780 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17781 now, which overrides the FreeBSD-elf entry on-the-fly.
17783 *Ralf S. Engelschall*
17785 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17789 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17790 on the `perl Configure ...` command line. This way one can compile
17791 OpenSSL libraries with Position Independent Code (PIC) which is needed
17792 for linking it into DSOs.
17794 *Ralf S. Engelschall*
17796 * Remarkably, export ciphers were totally broken and no-one had noticed!
17801 * Cleaned up the LICENSE document: The official contact for any license
17802 questions now is the OpenSSL core team under openssl-core@openssl.org.
17803 And add a paragraph about the dual-license situation to make sure people
17804 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17805 to the OpenSSL toolkit.
17807 *Ralf S. Engelschall*
17809 * General source tree makefile cleanups: Made `making xxx in yyy...`
17810 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17811 Additionally cleaned up the `make links` target: Remove unnecessary
17812 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17813 to speed processing and no longer clutter the display with confusing
17814 stuff. Instead only the actually done links are displayed.
17816 *Ralf S. Engelschall*
17818 * Permit null encryption ciphersuites, used for authentication only. It used
17819 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17820 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17825 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17826 signed attributes when verifying signatures (this would break them),
17827 the detached data encoding was wrong and public keys obtained using
17828 X509_get_pubkey() weren't freed.
17832 * Add text documentation for the BUFFER functions. Also added a work around
17833 to a Win95 console bug. This was triggered by the password read stuff: the
17834 last character typed gets carried over to the next fread(). If you were
17835 generating a new cert request using 'req' for example then the last
17836 character of the passphrase would be CR which would then enter the first
17841 * Added the new 'Includes OpenSSL Cryptography Software' button as
17842 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17843 button and can be used by applications based on OpenSSL to show the
17844 relationship to the OpenSSL project.
17846 *Ralf S. Engelschall*
17848 * Remove confusing variables in function signatures in files
17849 ssl/ssl_lib.c and ssl/ssl.h.
17851 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17853 * Don't install bss_file.c under PREFIX/include/
17855 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17857 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17858 functions that return function pointers and has support for NT specific
17859 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17860 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17861 unsigned to signed types: this was killing the Win32 compile.
17865 * Add new certificate file to stack functions,
17866 SSL_add_dir_cert_subjects_to_stack() and
17867 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17868 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17869 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17870 This means that Apache-SSL and similar packages don't have to mess around
17871 to add as many CAs as they want to the preferred list.
17875 * Experiment with doxygen documentation. Currently only partially applied to
17877 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17878 openssl.doxy as the configuration file.
17882 * Get rid of remaining C++-style comments which strict C compilers hate.
17884 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17886 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17887 compiled in by default: it has problems with large keys.
17891 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17892 DH private keys and/or callback functions which directly correspond to
17893 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17894 is needed for applications which have to configure certificates on a
17895 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17897 For the RSA certificate situation is makes no difference, but
17898 for the DSA certificate situation this fixes the "no shared cipher"
17899 problem where the OpenSSL cipher selection procedure failed because the
17900 temporary keys were not overtaken from the context and the API provided
17901 no way to reconfigure them.
17902 The new functions now let applications reconfigure the stuff and they
17903 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17904 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17905 non-public-API function ssl_cert_instantiate() is used as a helper
17906 function and also to reduce code redundancy inside ssl_rsa.c.
17908 *Ralf S. Engelschall*
17910 * Move s_server -dcert and -dkey options out of the undocumented feature
17911 area because they are useful for the DSA situation and should be
17912 recognized by the users.
17914 *Ralf S. Engelschall*
17916 * Fix the cipher decision scheme for export ciphers: the export bits are
17917 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17918 SSL_EXP_MASK. So, the original variable has to be used instead of the
17919 already masked variable.
17921 *Richard Levitte <levitte@stacken.kth.se>*
17923 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17925 *Richard Levitte <levitte@stacken.kth.se>*
17927 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17928 from `int` to `unsigned int` because it is a length and initialized by
17929 EVP_DigestFinal() which expects an `unsigned int *`.
17931 *Richard Levitte <levitte@stacken.kth.se>*
17933 * Don't hard-code path to Perl interpreter on shebang line of Configure
17934 script. Instead use the usual Shell->Perl transition trick.
17936 *Ralf S. Engelschall*
17938 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17939 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17940 -noout -modulus` as it's already the case for `openssl rsa -noout
17941 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17942 currently the public key is printed (a decision which was already done by
17943 `openssl dsa -modulus` in the past) which serves a similar purpose.
17944 Additionally the NO_RSA no longer completely removes the whole -modulus
17945 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17948 *Ralf S. Engelschall*
17950 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17951 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17953 *Arne Ansper <arne@ats.cyber.ee>*
17955 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17956 to be added. Now both 'req' and 'ca' can use new objects defined in the
17961 * Add cool BIO that does syslog (or event log on NT).
17963 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17965 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17966 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17967 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17968 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17972 * Add preliminary config info for new extension code.
17976 * Make RSA_NO_PADDING really use no padding.
17978 *Ulf Moeller <ulf@fitug.de>*
17980 * Generate errors when private/public key check is done.
17984 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
17985 for some CRL extensions and new objects added.
17989 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
17990 key usage extension and fuller support for authority key id.
17994 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
17995 padding method for RSA, which is recommended for new applications in PKCS
17996 #1 v2.0 (RFC 2437, October 1998).
17997 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
17998 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
17999 against Bleichbacher's attack on RSA.
18000 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18003 * Updates to the new SSL compression code
18005 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18007 * Fix so that the version number in the master secret, when passed
18008 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18009 (because the server will not accept higher), that the version number
18010 is 0x03,0x01, not 0x03,0x00
18012 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18014 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
18015 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
18016 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
18020 * Support for RAW extensions where an arbitrary extension can be
18021 created by including its DER encoding. See apps/openssl.cnf for
18026 * Make sure latest Perl versions don't interpret some generated C array
18027 code as Perl array code in the crypto/err/err_genc.pl script.
18029 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18031 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18032 not many people have the assembler. Various Win32 compilation fixes and
18033 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18034 build instructions.
18038 * Modify configure script 'Configure' to automatically create crypto/date.h
18039 file under Win32 and also build pem.h from pem.org. New script
18040 util/mkfiles.pl to create the MINFO file on environments that can't do a
18041 'make files': perl util/mkfiles.pl >MINFO should work.
18045 * Major rework of DES function declarations, in the pursuit of correctness
18046 and purity. As a result, many evil casts evaporated, and some weirdness,
18047 too. You may find this causes warnings in your code. Zapping your evil
18048 casts will probably fix them. Mostly.
18052 * Fix for a typo in asn1.h. Bug fix to object creation script
18053 obj_dat.pl. It considered a zero in an object definition to mean
18054 "end of object": none of the objects in objects.h have any zeros
18055 so it wasn't spotted.
18057 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18059 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18060 Masking (CBCM). In the absence of test vectors, the best I have been able
18061 to do is check that the decrypt undoes the encrypt, so far. Send me test
18062 vectors if you have them.
18066 * Correct calculation of key length for export ciphers (too much space was
18067 allocated for null ciphers). This has not been tested!
18071 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18072 message is now correct (it understands "crypto" and "ssl" on its
18073 command line). There is also now an "update" option. This will update
18074 the util/ssleay.num and util/libeay.num files with any new functions.
18076 perl util/mkdef.pl crypto ssl update
18077 it will update them.
18081 * Overhauled the Perl interface:
18082 - ported BN stuff to OpenSSL's different BN library
18083 - made the perl/ source tree CVS-aware
18084 - renamed the package from SSLeay to OpenSSL (the files still contain
18085 their history because I've copied them in the repository)
18086 - removed obsolete files (the test scripts will be replaced
18087 by better Test::Harness variants in the future)
18089 *Ralf S. Engelschall*
18091 * First cut for a very conservative source tree cleanup:
18092 1. merge various obsolete readme texts into doc/ssleay.txt
18093 where we collect the old documents and readme texts.
18094 2. remove the first part of files where I'm already sure that we no
18095 longer need them because of three reasons: either they are just temporary
18096 files which were left by Eric or they are preserved original files where
18097 I've verified that the diff is also available in the CVS via "cvs diff
18098 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18099 the crypto/md/ stuff).
18101 *Ralf S. Engelschall*
18103 * More extension code. Incomplete support for subject and issuer alt
18104 name, issuer and authority key id. Change the i2v function parameters
18105 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18106 what that's for :-) Fix to ASN1 macro which messed up
18107 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18111 * Preliminary support for ENUMERATED type. This is largely copied from the
18116 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18118 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18120 * Make sure `make rehash` target really finds the `openssl` program.
18122 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18124 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18125 like to hear about it if this slows down other processors.
18129 * Add CygWin32 platform information to Configure script.
18131 *Alan Batie <batie@aahz.jf.intel.com>*
18133 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18135 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18137 * New program nseq to manipulate netscape certificate sequences
18141 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18146 * Fixes to BN code. Previously the default was to define BN_RECURSION
18147 but the BN code had some problems that would cause failures when
18148 doing certificate verification and some other functions.
18150 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18152 * Add ASN1 and PEM code to support netscape certificate sequences.
18156 * Add ASN1 and PEM code to support netscape certificate sequences.
18160 * Add several PKIX and private extended key usage OIDs.
18164 * Modify the 'ca' program to handle the new extension code. Modify
18165 openssl.cnf for new extension format, add comments.
18169 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18170 and add a sample to openssl.cnf so req -x509 now adds appropriate
18175 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18176 error code, add initial support to X509_print() and x509 application.
18180 * Takes a deep breath and start adding X509 V3 extension support code. Add
18181 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18182 stuff is currently isolated and isn't even compiled yet.
18186 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18187 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18188 Removed the versions check from X509 routines when loading extensions:
18189 this allows certain broken certificates that don't set the version
18190 properly to be processed.
18194 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18195 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18196 can still be regenerated with "make depend".
18200 * Spelling mistake in C version of CAST-128.
18202 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18204 * Changes to the error generation code. The perl script err-code.pl
18205 now reads in the old error codes and retains the old numbers, only
18206 adding new ones if necessary. It also only changes the .err files if new
18207 codes are added. The makefiles have been modified to only insert errors
18208 when needed (to avoid needlessly modifying header files). This is done
18209 by only inserting errors if the .err file is newer than the auto generated
18210 C file. To rebuild all the error codes from scratch (the old behaviour)
18211 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18212 or delete all the .err files.
18216 * CAST-128 was incorrectly implemented for short keys. The C version has
18217 been fixed, but is untested. The assembler versions are also fixed, but
18218 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18219 to regenerate it if needed.
18220 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18221 Hagino <itojun@kame.net>*
18223 * File was opened incorrectly in randfile.c.
18225 *Ulf Möller <ulf@fitug.de>*
18227 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18228 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18229 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18230 al: it's just almost always a UTCTime. Note this patch adds new error
18231 codes so do a "make errors" if there are problems.
18235 * Correct Linux 1 recognition in config.
18237 *Ulf Möller <ulf@fitug.de>*
18239 * Remove pointless MD5 hash when using DSA keys in ca.
18241 *Anonymous <nobody@replay.com>*
18243 * Generate an error if given an empty string as a cert directory. Also
18244 generate an error if handed NULL (previously returned 0 to indicate an
18245 error, but didn't set one).
18247 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18249 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18253 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18254 parameters. This was causing a warning which killed off the Win32 compile.
18258 * Remove C++ style comments from crypto/bn/bn_local.h.
18260 *Neil Costigan <neil.costigan@celocom.com>*
18262 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18263 based on a text string, looking up short and long names and finally
18264 "dot" format. The "dot" format stuff didn't work. Added new function
18265 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18266 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18267 OID is not part of the table.
18271 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18272 X509_LOOKUP_by_alias().
18276 * Sort openssl functions by name.
18280 * Get the gendsa program working (hopefully) and add it to app list. Remove
18281 encryption from sample DSA keys (in case anyone is interested the password
18286 * Make *all* `*_free` functions accept a NULL pointer.
18288 *Frans Heymans <fheymans@isaserver.be>*
18290 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18293 *Anonymous <nobody@replay.com>*
18295 * s_server should send the CAfile as acceptable CAs, not its own cert.
18297 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18299 * Don't blow it for numeric -newkey arguments to apps/req.
18301 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18303 * Temp key "for export" tests were wrong in s3_srvr.c.
18305 *Anonymous <nobody@replay.com>*
18307 * Add prototype for temp key callback functions
18308 SSL_CTX_set_tmp_{rsa,dh}_callback().
18312 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18313 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18317 * X509_name_add_entry() freed the wrong thing after an error.
18319 *Arne Ansper <arne@ats.cyber.ee>*
18321 * rsa_eay.c would attempt to free a NULL context.
18323 *Arne Ansper <arne@ats.cyber.ee>*
18325 * BIO_s_socket() had a broken should_retry() on Windoze.
18327 *Arne Ansper <arne@ats.cyber.ee>*
18329 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18331 *Arne Ansper <arne@ats.cyber.ee>*
18333 * Make sure the already existing X509_STORE->depth variable is initialized
18334 in X509_STORE_new(), but document the fact that this variable is still
18335 unused in the certificate verification process.
18337 *Ralf S. Engelschall*
18339 * Fix the various library and apps files to free up pkeys obtained from
18340 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18344 * Fix reference counting in X509_PUBKEY_get(). This makes
18345 demos/maurice/example2.c work, amongst others, probably.
18347 *Steve Henson and Ben Laurie*
18349 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18350 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18351 are no longer created. This way we have a single and consistent command
18352 line interface `openssl <command>`, similar to `cvs <command>`.
18354 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18356 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18357 BIT STRING wrapper always have zero unused bits.
18361 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18365 * Make the top-level INSTALL documentation easier to understand.
18369 * Makefiles updated to exit if an error occurs in a sub-directory
18370 make (including if user presses ^C) [Paul Sutton]
18372 * Make Montgomery context stuff explicit in RSA data structure.
18376 * Fix build order of pem and err to allow for generated pem.h.
18380 * Fix renumbering bug in X509_NAME_delete_entry().
18384 * Enhanced the err-ins.pl script so it makes the error library number
18385 global and can add a library name. This is needed for external ASN1 and
18386 other error libraries.
18390 * Fixed sk_insert which never worked properly.
18394 * Fix ASN1 macros so they can handle indefinite length constructed
18395 EXPLICIT tags. Some non standard certificates use these: they can now
18400 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18401 into a single doc/ssleay.txt bundle. This way the information is still
18402 preserved but no longer messes up this directory. Now it's new room for
18403 the new set of documentation files.
18405 *Ralf S. Engelschall*
18407 * SETs were incorrectly DER encoded. This was a major pain, because they
18408 shared code with SEQUENCEs, which aren't coded the same. This means that
18409 almost everything to do with SETs or SEQUENCEs has either changed name or
18410 number of arguments.
18412 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18414 * Fix test data to work with the above.
18418 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18419 was already fixed by Eric for 0.9.1 it seems.
18421 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18423 * Autodetect FreeBSD3.
18427 * Fix various bugs in Configure. This affects the following platforms:
18431 unixware-2.0-pentium
18436 * Eliminate generated files from CVS. Reorder tests to regenerate files
18437 before they are needed.
18441 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18445 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18447 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18448 changed SSLeay to OpenSSL in version strings.
18450 *Ralf S. Engelschall*
18452 * Some fixups to the top-level documents.
18456 * Fixed the nasty bug where rsaref.h was not found under compile-time
18457 because the symlink to include/ was missing.
18459 *Ralf S. Engelschall*
18461 * Incorporated the popular no-RSA/DSA-only patches
18462 which allow to compile a RSA-free SSLeay.
18464 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18466 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18467 when "ssleay" is still not found.
18469 *Ralf S. Engelschall*
18471 * Added more platforms to Configure: Cray T3E, HPUX 11,
18473 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18475 * Updated the README file.
18477 *Ralf S. Engelschall*
18479 * Added various .cvsignore files in the CVS repository subdirs
18480 to make a "cvs update" really silent.
18482 *Ralf S. Engelschall*
18484 * Recompiled the error-definition header files and added
18485 missing symbols to the Win32 linker tables.
18487 *Ralf S. Engelschall*
18489 * Cleaned up the top-level documents;
18490 o new files: CHANGES and LICENSE
18491 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18492 o merged COPYRIGHT into LICENSE
18493 o removed obsolete TODO file
18494 o renamed MICROSOFT to INSTALL.W32
18496 *Ralf S. Engelschall*
18498 * Removed dummy files from the 0.9.1b source tree:
18499 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18500 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18501 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18502 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18503 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18505 *Ralf S. Engelschall*
18507 * Added various platform portability fixes.
18511 * The Genesis of the OpenSSL rpject:
18512 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18513 Young and Tim J. Hudson created while they were working for C2Net until
18516 *The OpenSSL Project*
18518 ### Changes between 0.9.0b and 0.9.1b [not released]
18520 * Updated a few CA certificates under certs/
18524 * Changed some BIGNUM api stuff.
18528 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18529 DGUX x86, Linux Alpha, etc.
18533 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18534 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18539 * Add -strparse option to asn1pars program which parses nested
18542 *Dr Stephen Henson <shenson@bigfoot.com>*
18544 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18548 * DSA fix for "ca" program.
18552 * Added "-genkey" option to "dsaparam" program.
18556 * Added RIPE MD160 (rmd160) message digest.
18560 * Added -a (all) option to "ssleay version" command.
18564 * Added PLATFORM define which is the id given to Configure.
18568 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18572 * Extended the ASN.1 parser routines.
18576 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18580 * Added a BN_CTX to the BN library.
18584 * Fixed the weak key values in DES library
18588 * Changed API in EVP library for cipher aliases.
18592 * Added support for RC2/64bit cipher.
18596 * Converted the lhash library to the crypto/mem.c functions.
18600 * Added more recognized ASN.1 object ids.
18604 * Added more RSA padding checks for SSL/TLS.
18608 * Added BIO proxy/filter functionality.
18612 * Added extra_certs to SSL_CTX which can be used
18613 send extra CA certificates to the client in the CA cert chain sending
18614 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18618 * Now Fortezza is denied in the authentication phase because
18619 this is key exchange mechanism is not supported by SSLeay at all.
18623 * Additional PKCS1 checks.
18627 * Support the string "TLSv1" for all TLS v1 ciphers.
18631 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18632 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18636 * Fixed a few memory leaks.
18640 * Fixed various code and comment typos.
18644 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18645 bytes sent in the client random.
18647 *Edward Bishop <ebishop@spyglass.com>*
18651 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18652 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18653 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18654 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18655 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18656 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18657 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18658 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18659 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18660 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18661 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18662 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18663 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18664 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18665 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18666 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18667 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18668 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18669 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18670 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18671 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18672 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18673 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18674 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18675 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18676 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18677 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18678 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18679 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18680 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18681 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18682 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18683 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18684 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18685 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18686 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18687 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18688 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18689 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18690 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18691 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18692 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18693 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18694 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18695 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18696 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18697 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18698 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18699 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18700 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18701 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18702 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18703 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18704 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18705 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18706 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18707 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18708 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18709 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18710 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18711 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18712 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18713 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18714 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18715 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18716 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18717 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18718 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18719 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18720 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18721 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18722 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18723 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18724 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18725 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18726 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18727 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18728 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18729 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18730 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18731 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18732 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18733 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18734 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18735 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18736 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18737 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18738 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18739 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18740 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18741 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18742 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18743 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18744 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18745 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18746 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18747 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18748 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18749 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18750 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18751 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18752 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18753 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18754 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18755 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18756 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18757 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18758 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18759 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18760 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18761 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18762 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18763 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18764 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18765 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18766 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18767 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18768 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18769 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18770 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18771 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18772 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18773 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18774 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18775 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18776 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18777 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18778 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18779 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18780 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18781 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18782 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18783 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18784 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18785 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18786 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18787 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18788 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18789 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18790 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18791 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18792 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18793 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18794 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18795 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18796 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18797 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18798 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18799 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18800 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18801 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18802 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18803 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18804 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18805 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18806 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18807 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18808 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18809 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18810 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18811 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655