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_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
470 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
471 DH_generate_parameters_ex, 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)>.
486 Additionally functions that read and write DH objects such as d2i_DHparams,
487 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
488 functions have also been deprecated. Applications should instead use the
489 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
491 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
492 EVP_PKEY_assign_DH(), EVP_PKEY_get0_DH, EVP_PKEY_get1_DH, EVP_PKEY_set1_DH
493 are also deprecated. Applications should instead either read or write an
494 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs. Or load an
495 EVP_PKEY directly from DH data using EVP_PKEY_fromdata().
497 *Paul Dale and Matt Caswell*
499 * All of the low level DSA functions have been deprecated including:
501 DSA_do_sign, DSA_do_verify, DSA_OpenSSL, DSA_set_default_method,
502 DSA_get_default_method, DSA_set_method, DSA_get_method,
503 DSA_new_method, DSA_size, DSA_security_bits, DSA_sign_setup, DSA_sign,
504 DSA_verify, DSA_get_ex_new_index, DSA_set_ex_data, DSA_get_ex_data,
505 DSA_generate_parameters_ex, DSA_generate_key, DSA_meth_new, DSA_get0_engine,
506 DSA_meth_free, DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name,
507 DSA_meth_get_flags, DSA_meth_set_flags, DSA_meth_get0_app_data,
508 DSA_meth_set0_app_data, DSA_meth_get_sign, DSA_meth_set_sign,
509 DSA_meth_get_sign_setup, DSA_meth_set_sign_setup, DSA_meth_get_verify,
510 DSA_meth_set_verify, DSA_meth_get_mod_exp, DSA_meth_set_mod_exp,
511 DSA_meth_get_bn_mod_exp, DSA_meth_set_bn_mod_exp, DSA_meth_get_init,
512 DSA_meth_set_init, DSA_meth_get_finish, DSA_meth_set_finish,
513 DSA_meth_get_paramgen, DSA_meth_set_paramgen, DSA_meth_get_keygen and
516 Use of these low level functions has been informally discouraged for a long
517 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
518 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
522 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
523 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
524 This means that applications don't have to look at the curve NID and
525 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
526 However, they still can, that EVP_PKEY_set_alias_type() call acts as
527 a no-op when the EVP_PKEY is already of the given type.
529 Parameter and key generation is also reworked to make it possible
530 to generate EVP_PKEY_SM2 parameters and keys without having to go
531 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
532 However, code that does the latter will still work as before.
536 * Deprecated low level ECDH and ECDSA functions. These include:
538 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
539 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
542 Use of these low level functions has been informally discouraged for a long
543 time. Instead applications should use the EVP_PKEY_derive(3),
544 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
548 * Deprecated the EC_KEY_METHOD functions. These include:
550 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
551 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
552 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
553 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
554 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
555 EC_KEY_METHOD_get_verify.
557 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
561 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
562 and EVP_PKEY_decrypt() instead.
563 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
564 and EVP_PKEY_encrypt() instead.
568 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
569 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
570 a new formulation to include all the things it can be used for,
571 as well as words of caution.
575 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
576 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
580 * All of the low level HMAC functions have been deprecated including:
582 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
583 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
586 Use of these low level functions has been informally discouraged for a long
587 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
588 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
589 and L<EVP_MAC_final(3)>.
593 * Over two thousand fixes were made to the documentation, including:
594 - Common options (such as -rand/-writerand, TLS version control, etc)
595 were refactored and point to newly-enhanced descriptions in openssl.pod.
596 - Added style conformance for all options (with help from Richard Levitte),
597 documented all reported missing options, added a CI build to check
598 that all options are documented and that no unimplemented options
600 - Documented some internals, such as all use of environment variables.
601 - Addressed all internal broken L<> references.
605 * All of the low level CMAC functions have been deprecated including:
607 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
608 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
610 Use of these low level functions has been informally discouraged for a long
611 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
612 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
613 and L<EVP_MAC_final(3)>.
617 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
618 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
621 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
622 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
623 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
624 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
625 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
626 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
627 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
628 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
629 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
630 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
632 Use of these low level functions has been informally discouraged
633 for a long time. Applications should use the EVP_DigestInit_ex(3),
634 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
638 * Corrected the documentation of the return values from the `EVP_DigestSign*`
639 set of functions. The documentation mentioned negative values for some
640 errors, but this was never the case, so the mention of negative values
643 Code that followed the documentation and thereby check with something
644 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
648 * All of the low level cipher functions have been deprecated including:
650 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
651 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
652 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
653 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
654 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
655 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
656 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
657 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
658 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
659 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
660 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
661 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
662 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
663 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
664 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
665 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
666 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
667 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
668 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
669 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
670 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
671 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
672 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
673 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
674 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
675 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
676 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
677 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
678 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
680 Use of these low level functions has been informally discouraged for
681 a long time. Applications should use the high level EVP APIs, e.g.
682 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
683 equivalently named decrypt functions instead.
685 *Matt Caswell and Paul Dale*
687 * Removed include/openssl/opensslconf.h.in and replaced it with
688 include/openssl/configuration.h.in, which differs in not including
689 <openssl/macros.h>. A short header include/openssl/opensslconf.h
690 was added to include both.
692 This allows internal hacks where one might need to modify the set
693 of configured macros, for example this if deprecated symbols are
694 still supposed to be available internally:
696 #include <openssl/configuration.h>
698 #undef OPENSSL_NO_DEPRECATED
699 #define OPENSSL_SUPPRESS_DEPRECATED
701 #include <openssl/macros.h>
703 This should not be used by applications that use the exported
704 symbols, as that will lead to linking errors.
708 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
709 used in exponentiation with 512-bit moduli. No EC algorithms are
710 affected. Analysis suggests that attacks against 2-prime RSA1024,
711 3-prime RSA1536, and DSA1024 as a result of this defect would be very
712 difficult to perform and are not believed likely. Attacks against DH512
713 are considered just feasible. However, for an attack the target would
714 have to re-use the DH512 private key, which is not recommended anyway.
715 Also applications directly using the low level API BN_mod_exp may be
716 affected if they use BN_FLG_CONSTTIME.
721 * Most memory-debug features have been deprecated, and the functionality
722 replaced with no-ops.
726 * Added documentation for the STACK API.
730 * Introduced a new method type and API, OSSL_ENCODER, to
731 represent generic encoders. An implementation is expected to
732 be able to encode an object associated with a given name (such
733 as an algorithm name for an asymmetric key) into forms given by
734 implementation properties.
736 Encoders are primarily used from inside libcrypto, through
737 calls to functions like EVP_PKEY_print_private(),
738 PEM_write_bio_PrivateKey() and similar.
740 Encoders are specified in such a way that they can be made to
741 directly handle the provider side portion of an object, if this
742 provider side part comes from the same provider as the encoder
743 itself, but can also be made to handle objects in parametrized
744 form (as an OSSL_PARAM array of data). This allows a provider to
745 offer generic encoders as a service for any other provider.
749 * Added a .pragma directive to the syntax of configuration files, to
750 allow varying behavior in a supported and predictable manner.
751 Currently added pragma:
755 This allows dollar signs to be a keyword character unless it's
756 followed by a opening brace or parenthesis. This is useful for
757 platforms where dollar signs are commonly used in names, such as
758 volume names and system directory names on VMS.
762 * Added functionality to create an EVP_PKEY from user data. This
763 is effectively the same as creating a RSA, DH or DSA object and
764 then assigning them to an EVP_PKEY, but directly using algorithm
765 agnostic EVP functions. A benefit is that this should be future
766 proof for public key algorithms to come.
770 * Change the interpretation of the '--api' configuration option to
771 mean that this is a desired API compatibility level with no
772 further meaning. The previous interpretation, that this would
773 also mean to remove all deprecated symbols up to and including
774 the given version, no requires that 'no-deprecated' is also used
775 in the configuration.
777 When building applications, the desired API compatibility level
778 can be set with the OPENSSL_API_COMPAT macro like before. For
779 API compatibility version below 3.0, the old style numerical
780 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
781 For version 3.0 and on, the value is expected to be the decimal
782 value calculated from the major and minor version like this:
784 MAJOR * 10000 + MINOR * 100
788 -DOPENSSL_API_COMPAT=30000 For 3.0
789 -DOPENSSL_API_COMPAT=30200 For 3.2
791 To hide declarations that are deprecated up to and including the
792 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
793 given when building the application as well.
797 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
798 access to certificate and CRL stores via URIs and OSSL_STORE
801 This adds the following functions:
803 - X509_LOOKUP_store()
804 - X509_STORE_load_file()
805 - X509_STORE_load_path()
806 - X509_STORE_load_store()
807 - SSL_add_store_cert_subjects_to_stack()
808 - SSL_CTX_set_default_verify_store()
809 - SSL_CTX_load_verify_file()
810 - SSL_CTX_load_verify_dir()
811 - SSL_CTX_load_verify_store()
815 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
816 The presence of this system service is determined at run-time.
820 * Added functionality to create an EVP_PKEY context based on data
821 for methods from providers. This takes an algorithm name and a
822 property query string and simply stores them, with the intent
823 that any operation that uses this context will use those strings
824 to fetch the needed methods implicitly, thereby making the port
825 of application written for pre-3.0 OpenSSL easier.
829 * The undocumented function NCONF_WIN32() has been deprecated; for
830 conversion details see the HISTORY section of doc/man5/config.pod
834 * Introduced the new functions EVP_DigestSignInit_ex() and
835 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
836 EVP_DigestVerifyUpdate() have been converted to functions. See the man
837 pages for further details.
841 * Over two thousand fixes were made to the documentation, including:
842 adding missing command flags, better style conformance, documentation
845 *Rich Salz, Richard Levitte*
847 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
848 X25519, X448, Ed25519 and Ed448.
852 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
857 * Deprecated the public definition of ERR_STATE as well as the function
858 ERR_get_state(). This is done in preparation of making ERR_STATE an
863 * Added ERR functionality to give callers access to the stored function
864 names that have replaced the older function code based functions.
866 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
867 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
868 ERR_peek_error_all() and ERR_peek_last_error_all().
870 These functions have become deprecated: ERR_get_error_line(),
871 ERR_get_error_line_data(), ERR_peek_error_line_data(),
872 ERR_peek_last_error_line_data() and ERR_func_error_string().
874 Users are recommended to use ERR_get_error_all(), or to pick information
875 with ERR_peek functions and finish off with getting the error code by using
880 * Extended testing to be verbose for failing tests only. The make variables
881 VERBOSE_FAILURE or VF can be used to enable this:
883 $ make VF=1 test # Unix
884 $ mms /macro=(VF=1) test ! OpenVMS
885 $ nmake VF=1 test # Windows
889 * Added several checks to X509_verify_cert() according to requirements in
890 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
891 (which may be done by using the CLI option `-x509_strict`):
892 * The basicConstraints of CA certificates must be marked critical.
893 * CA certificates must explicitly include the keyUsage extension.
894 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
895 * The issuer name of any certificate must not be empty.
896 * The subject name of CA certs, certs with keyUsage crlSign,
897 and certs without subjectAlternativeName must not be empty.
898 * If a subjectAlternativeName extension is given it must not be empty.
899 * The signatureAlgorithm field and the cert signature must be consistent.
900 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
901 must not be marked critical.
902 * The authorityKeyIdentifier must be given for X.509v3 certs
903 unless they are self-signed.
904 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
908 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
909 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
913 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
914 used even when parsing explicit parameters, when loading a encoded key
915 or calling `EC_GROUP_new_from_ecpkparameters()`/
916 `EC_GROUP_new_from_ecparameters()`.
917 This prevents bypass of security hardening and performance gains,
918 especially for curves with specialized EC_METHODs.
919 By default, if a key encoded with explicit parameters is loaded and later
920 encoded, the output is still encoded with explicit parameters, even if
921 internally a "named" EC_GROUP is used for computation.
925 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
926 this change, EC_GROUP_set_generator would accept order and/or cofactor as
927 NULL. After this change, only the cofactor parameter can be NULL. It also
928 does some minimal sanity checks on the passed order.
933 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
934 An attack is simple, if the first CMS_recipientInfo is valid but the
935 second CMS_recipientInfo is chosen ciphertext. If the second
936 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
937 encryption key will be replaced by garbage, and the message cannot be
938 decoded, but if the RSA decryption fails, the correct encryption key is
939 used and the recipient will not notice the attack.
940 As a work around for this potential attack the length of the decrypted
941 key must be equal to the cipher default key length, in case the
942 certifiate is not given and all recipientInfo are tried out.
943 The old behaviour can be re-enabled in the CMS code by setting the
944 CMS_DEBUG_DECRYPT flag.
948 * Early start up entropy quality from the DEVRANDOM seed source has been
949 improved for older Linux systems. The RAND subsystem will wait for
950 /dev/random to be producing output before seeding from /dev/urandom.
951 The seeded state is stored for future library initialisations using
952 a system global shared memory segment. The shared memory identifier
953 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
954 the desired value. The default identifier is 114.
958 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
959 when primes for RSA keys are computed.
960 Since we previously always generated primes == 2 (mod 3) for RSA keys,
961 the 2-prime and 3-prime RSA modules were easy to distinguish, since
962 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
963 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
964 This avoids possible fingerprinting of newly generated RSA modules.
968 * Correct the extended master secret constant on EBCDIC systems. Without this
969 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
970 negotiate EMS will fail. Unfortunately this also means that TLS connections
971 between EBCDIC systems with this fix, and EBCDIC systems without this
972 fix will fail if they negotiate EMS.
976 * Changed the library initialisation so that the config file is now loaded
977 by default. This was already the case for libssl. It now occurs for both
978 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
979 OPENSSL_init_crypto() to suppress automatic loading of a config file.
983 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
984 where the former acts as a replacement for ERR_put_error(), and the
985 latter replaces the combination ERR_put_error()+ERR_add_error_data().
986 ERR_raise_data() adds more flexibility by taking a format string and
987 an arbitrary number of arguments following it, to be processed with
992 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
993 to check if a named provider is loaded and available. When called, it
994 will also activate all fallback providers if such are still present.
998 * Enforce a minimum DH modulus size of 512 bits.
1002 * Changed DH parameters to generate the order q subgroup instead of 2q.
1003 Previously generated DH parameters are still accepted by DH_check
1004 but DH_generate_key works around that by clearing bit 0 of the
1005 private key for those. This avoids leaking bit 0 of the private key.
1009 * Significantly reduce secure memory usage by the randomness pools.
1013 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1018 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1019 algorithms. An implementation of a key exchange algorithm can be obtained
1020 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1021 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1022 the older EVP_PKEY_derive_init() function. See the man pages for the new
1023 functions for further details.
1027 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1031 * Removed the function names from error messages and deprecated the
1034 * Removed NextStep support and the macro OPENSSL_UNISTD
1038 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1039 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1040 Also removed "export var as function" capability; we do not export
1041 variables, only functions.
1045 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1046 an error and 1 indicating success. In previous versions of OpenSSL this
1047 was a void type. If a key was set longer than the maximum possible this
1052 * Support SM2 signing and verification schemes with X509 certificate.
1056 * Use SHA256 as the default digest for TS query in the ts app.
1060 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1061 This checks that the salt length is at least 128 bits, the derived key
1062 length is at least 112 bits, and that the iteration count is at least 1000.
1063 For backwards compatibility these checks are disabled by default in the
1064 default provider, but are enabled by default in the fips provider.
1065 To enable or disable these checks use the control
1066 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1070 * Default cipher lists/suites are now available via a function, the
1071 #defines are deprecated.
1075 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1076 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1077 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1081 * Join the directories crypto/x509 and crypto/x509v3
1085 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1086 This changes the size when using the genpkey app when no size is given. It
1087 fixes an omission in earlier changes that changed all RSA, DSA and DH
1088 generation apps to use 2048 bits by default.
1092 * Added command 'openssl kdf' that uses the EVP_KDF API.
1096 * Added command 'openssl mac' that uses the EVP_MAC API.
1100 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1101 as default directories. Also added the command 'openssl info'
1102 for scripting purposes.
1106 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1107 deprecated. These undocumented functions were never integrated into the EVP
1108 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1109 Bi-directional IGE mode. These modes were never formally standardised and
1110 usage of these functions is believed to be very small. In particular
1111 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1112 is ever used. The security implications are believed to be minimal, but
1113 this issue was never fixed for backwards compatibility reasons. New code
1114 should not use these modes.
1118 * Add prediction resistance to the DRBG reseeding process.
1122 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1123 mandated by IEEE Std 1619-2018.
1127 * Added newline escaping functionality to a filename when using openssl dgst.
1128 This output format is to replicate the output format found in the `*sum`
1129 checksum programs. This aims to preserve backward compatibility.
1131 *Matt Eaton, Richard Levitte, and Paul Dale*
1133 * Removed the heartbeat message in DTLS feature, as it has very
1134 little usage and doesn't seem to fulfill a valuable purpose.
1135 The configuration option is now deprecated.
1139 * Changed the output of 'openssl {digestname} < file' to display the
1140 digest name in its output.
1144 * Added a new generic trace API which provides support for enabling
1145 instrumentation through trace output. This feature is mainly intended
1146 as an aid for developers and is disabled by default. To utilize it,
1147 OpenSSL needs to be configured with the `enable-trace` option.
1149 If the tracing API is enabled, the application can activate trace output
1150 by registering BIOs as trace channels for a number of tracing and debugging
1153 The 'openssl' application has been expanded to enable any of the types
1154 available via environment variables defined by the user, and serves as
1155 one possible example on how to use this functionality.
1157 *Richard Levitte & Matthias St. Pierre*
1159 * Added build tests for C++. These are generated files that only do one
1160 thing, to include one public OpenSSL head file each. This tests that
1161 the public header files can be usefully included in a C++ application.
1163 This test isn't enabled by default. It can be enabled with the option
1164 'enable-buildtest-c++'.
1168 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1172 * Add KMAC to EVP_MAC.
1176 * Added property based algorithm implementation selection framework to
1181 * Added SCA hardening for modular field inversion in EC_GROUP through
1182 a new dedicated field_inv() pointer in EC_METHOD.
1183 This also addresses a leakage affecting conversions from projective
1184 to affine coordinates.
1186 *Billy Bob Brumley, Nicola Tuveri*
1188 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1189 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1190 those algorithms that were already supported through the EVP_PKEY API
1191 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1192 and scrypt are now wrappers that call EVP_KDF.
1196 * Build devcrypto engine as a dynamic engine.
1198 *Eneas U de Queiroz*
1200 * Add keyed BLAKE2 to EVP_MAC.
1204 * Fix a bug in the computation of the endpoint-pair shared secret used
1205 by DTLS over SCTP. This breaks interoperability with older versions
1206 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1207 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1208 interoperability with such broken implementations. However, enabling
1209 this switch breaks interoperability with correct implementations.
1211 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1212 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1216 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1220 * Change the license to the Apache License v2.0.
1224 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1226 - Major releases (indicated by incrementing the MAJOR release number)
1227 may introduce incompatible API/ABI changes.
1228 - Minor releases (indicated by incrementing the MINOR release number)
1229 may introduce new features but retain API/ABI compatibility.
1230 - Patch releases (indicated by incrementing the PATCH number)
1231 are intended for bug fixes and other improvements of existing
1232 features only (like improving performance or adding documentation)
1233 and retain API/ABI compatibility.
1237 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1241 * Remove the 'dist' target and add a tarball building script. The
1242 'dist' target has fallen out of use, and it shouldn't be
1243 necessary to configure just to create a source distribution.
1247 * Recreate the OS390-Unix config target. It no longer relies on a
1248 special script like it did for OpenSSL pre-1.1.0.
1252 * Instead of having the source directories listed in Configure, add
1253 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1258 * Add GMAC to EVP_MAC.
1262 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1266 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1267 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1268 to facilitate the continued use of MACs through raw private keys in
1269 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1273 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1274 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1278 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1279 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1280 are retained for backwards compatibility.
1284 * AES-XTS mode now enforces that its two keys are different to mitigate
1285 the attacked described in "Efficient Instantiations of Tweakable
1286 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1287 Details of this attack can be obtained from:
1288 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1292 * Rename the object files, i.e. give them other names than in previous
1293 versions. Their names now include the name of the final product, as
1294 well as its type mnemonic (bin, lib, shlib).
1298 * Added new option for 'openssl list', '-objects', which will display the
1299 list of built in objects, i.e. OIDs with names.
1303 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1304 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1309 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1310 improves application performance by removing data copies and providing
1311 applications with zero-copy system calls such as sendfile and splice.
1315 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1316 option is set, openssl cleanses (zeroize) plaintext bytes from
1317 internal buffers after delivering them to the application. Note,
1318 the application is still responsible for cleansing other copies
1319 (e.g.: data received by SSL_read(3)).
1323 * `PKCS12_parse` now maintains the order of the parsed certificates
1324 when outputting them via `*ca` (rather than reversing it).
1328 * Deprecated pthread fork support methods. These were unused so no
1329 replacement is required.
1331 - OPENSSL_fork_prepare()
1332 - OPENSSL_fork_parent()
1333 - OPENSSL_fork_child()
1340 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1342 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1343 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1344 If an attacker can control both items being compared then this could lead
1345 to a possible denial of service attack. OpenSSL itself uses the
1346 GENERAL_NAME_cmp function for two purposes:
1347 1) Comparing CRL distribution point names between an available CRL and a
1348 CRL distribution point embedded in an X509 certificate
1349 2) When verifying that a timestamp response token signer matches the
1350 timestamp authority name (exposed via the API functions
1351 TS_RESP_verify_response and TS_RESP_verify_token)
1356 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1358 * Certificates with explicit curve parameters are now disallowed in
1359 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1363 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1364 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1365 conversely, silently ignore DTLS protocol version bounds when configuring
1366 TLS-based contexts. The commands can be repeated to set bounds of both
1367 types. The same applies with the corresponding "min_protocol" and
1368 "max_protocol" command-line switches, in case some application uses both TLS
1371 SSL_CTX instances that are created for a fixed protocol version (e.g.
1372 TLSv1_server_method()) also silently ignore version bounds. Previously
1373 attempts to apply bounds to these protocol versions would result in an
1374 error. Now only the "version-flexible" SSL_CTX instances are subject to
1375 limits in configuration files in command-line options.
1379 * Handshake now fails if Extended Master Secret extension is dropped
1384 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1386 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1388 * Fixed segmentation fault in SSL_check_chain()
1389 Server or client applications that call the SSL_check_chain() function
1390 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1391 dereference as a result of incorrect handling of the
1392 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1393 or unrecognised signature algorithm is received from the peer. This could
1394 be exploited by a malicious peer in a Denial of Service attack.
1399 * Added AES consttime code for no-asm configurations
1400 an optional constant time support for AES was added
1401 when building openssl for no-asm.
1402 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1403 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1404 At this time this feature is by default disabled.
1405 It will be enabled by default in 3.0.
1409 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1411 * Revert the change of EOF detection while reading in libssl to avoid
1412 regressions in applications depending on the current way of reporting
1413 the EOF. As the existing method is not fully accurate the change to
1414 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1415 branch and will be present in the 3.0 release.
1419 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1420 when primes for RSA keys are computed.
1421 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1422 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1423 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1424 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1425 This avoids possible fingerprinting of newly generated RSA modules.
1429 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1431 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1432 while reading in libssl then we would report an error back to the
1433 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1434 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1435 therefore give a hint as to what went wrong.
1439 * Check that ed25519 and ed448 are allowed by the security level. Previously
1440 signature algorithms not using an MD were not being checked that they were
1441 allowed by the security level.
1445 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1446 was not quite right. The behaviour was not consistent between resumption
1447 and normal handshakes, and also not quite consistent with historical
1448 behaviour. The behaviour in various scenarios has been clarified and
1449 it has been updated to make it match historical behaviour as closely as
1454 * *[VMS only]* The header files that the VMS compilers include automatically,
1455 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1456 that the C++ compiler doesn't understand. This is a shortcoming in the
1457 compiler, but can be worked around with `__cplusplus` guards.
1459 C++ applications that use OpenSSL libraries must be compiled using the
1460 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1461 functions. Otherwise, only functions with symbols of less than 31
1462 characters can be used, as the linker will not be able to successfully
1463 resolve symbols with longer names.
1467 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1468 The presence of this system service is determined at run-time.
1472 * Added newline escaping functionality to a filename when using openssl dgst.
1473 This output format is to replicate the output format found in the `*sum`
1474 checksum programs. This aims to preserve backward compatibility.
1476 *Matt Eaton, Richard Levitte, and Paul Dale*
1478 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1483 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1485 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1486 number generator (RNG). This was intended to include protection in the
1487 event of a fork() system call in order to ensure that the parent and child
1488 processes did not share the same RNG state. However this protection was not
1489 being used in the default case.
1491 A partial mitigation for this issue is that the output from a high
1492 precision timer is mixed into the RNG state so the likelihood of a parent
1493 and child process sharing state is significantly reduced.
1495 If an application already calls OPENSSL_init_crypto() explicitly using
1496 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1499 *Matthias St. Pierre*
1501 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1502 used even when parsing explicit parameters, when loading a encoded key
1503 or calling `EC_GROUP_new_from_ecpkparameters()`/
1504 `EC_GROUP_new_from_ecparameters()`.
1505 This prevents bypass of security hardening and performance gains,
1506 especially for curves with specialized EC_METHODs.
1507 By default, if a key encoded with explicit parameters is loaded and later
1508 encoded, the output is still encoded with explicit parameters, even if
1509 internally a "named" EC_GROUP is used for computation.
1513 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1514 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1515 NULL. After this change, only the cofactor parameter can be NULL. It also
1516 does some minimal sanity checks on the passed order.
1521 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1522 An attack is simple, if the first CMS_recipientInfo is valid but the
1523 second CMS_recipientInfo is chosen ciphertext. If the second
1524 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1525 encryption key will be replaced by garbage, and the message cannot be
1526 decoded, but if the RSA decryption fails, the correct encryption key is
1527 used and the recipient will not notice the attack.
1528 As a work around for this potential attack the length of the decrypted
1529 key must be equal to the cipher default key length, in case the
1530 certifiate is not given and all recipientInfo are tried out.
1531 The old behaviour can be re-enabled in the CMS code by setting the
1532 CMS_DEBUG_DECRYPT flag.
1537 * Early start up entropy quality from the DEVRANDOM seed source has been
1538 improved for older Linux systems. The RAND subsystem will wait for
1539 /dev/random to be producing output before seeding from /dev/urandom.
1540 The seeded state is stored for future library initialisations using
1541 a system global shared memory segment. The shared memory identifier
1542 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1543 the desired value. The default identifier is 114.
1547 * Correct the extended master secret constant on EBCDIC systems. Without this
1548 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1549 negotiate EMS will fail. Unfortunately this also means that TLS connections
1550 between EBCDIC systems with this fix, and EBCDIC systems without this
1551 fix will fail if they negotiate EMS.
1555 * Use Windows installation paths in the mingw builds
1557 Mingw isn't a POSIX environment per se, which means that Windows
1558 paths should be used for installation.
1563 * Changed DH_check to accept parameters with order q and 2q subgroups.
1564 With order 2q subgroups the bit 0 of the private key is not secret
1565 but DH_generate_key works around that by clearing bit 0 of the
1566 private key for those. This avoids leaking bit 0 of the private key.
1570 * Significantly reduce secure memory usage by the randomness pools.
1574 * Revert the DEVRANDOM_WAIT feature for Linux systems
1576 The DEVRANDOM_WAIT feature added a select() call to wait for the
1577 /dev/random device to become readable before reading from the
1578 /dev/urandom device.
1580 It turned out that this change had negative side effects on
1581 performance which were not acceptable. After some discussion it
1582 was decided to revert this feature and leave it up to the OS
1583 resp. the platform maintainer to ensure a proper initialization
1584 during early boot time.
1586 *Matthias St. Pierre*
1588 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1590 * Add build tests for C++. These are generated files that only do one
1591 thing, to include one public OpenSSL head file each. This tests that
1592 the public header files can be usefully included in a C++ application.
1594 This test isn't enabled by default. It can be enabled with the option
1595 'enable-buildtest-c++'.
1599 * Enable SHA3 pre-hashing for ECDSA and DSA.
1603 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1604 This changes the size when using the genpkey app when no size is given. It
1605 fixes an omission in earlier changes that changed all RSA, DSA and DH
1606 generation apps to use 2048 bits by default.
1610 * Reorganize the manual pages to consistently have RETURN VALUES,
1611 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1612 util/fix-doc-nits accordingly.
1614 *Paul Yang, Joshua Lock*
1616 * Add the missing accessor EVP_PKEY_get0_engine()
1620 * Have apps like 's_client' and 's_server' output the signature scheme
1621 along with other cipher suite parameters when debugging.
1625 * Make OPENSSL_config() error agnostic again.
1629 * Do the error handling in RSA decryption constant time.
1633 * Prevent over long nonces in ChaCha20-Poly1305.
1635 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1636 for every encryption operation. RFC 7539 specifies that the nonce value
1637 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1638 and front pads the nonce with 0 bytes if it is less than 12
1639 bytes. However it also incorrectly allows a nonce to be set of up to 16
1640 bytes. In this case only the last 12 bytes are significant and any
1641 additional leading bytes are ignored.
1643 It is a requirement of using this cipher that nonce values are
1644 unique. Messages encrypted using a reused nonce value are susceptible to
1645 serious confidentiality and integrity attacks. If an application changes
1646 the default nonce length to be longer than 12 bytes and then makes a
1647 change to the leading bytes of the nonce expecting the new value to be a
1648 new unique nonce then such an application could inadvertently encrypt
1649 messages with a reused nonce.
1651 Additionally the ignored bytes in a long nonce are not covered by the
1652 integrity guarantee of this cipher. Any application that relies on the
1653 integrity of these ignored leading bytes of a long nonce may be further
1654 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1655 is safe because no such use sets such a long nonce value. However user
1656 applications that use this cipher directly and set a non-default nonce
1657 length to be longer than 12 bytes may be vulnerable.
1659 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1665 * Add DEVRANDOM_WAIT feature for Linux systems
1667 On older Linux systems where the getrandom() system call is not available,
1668 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1669 Contrary to getrandom(), the /dev/urandom device will not block during
1670 early boot when the kernel CSPRNG has not been seeded yet.
1672 To mitigate this known weakness, use select() to wait for /dev/random to
1673 become readable before reading from /dev/urandom.
1675 * Ensure that SM2 only uses SM3 as digest algorithm
1679 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1681 * Change the info callback signals for the start and end of a post-handshake
1682 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1683 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1684 confused by this and assume that a TLSv1.2 renegotiation has started. This
1685 can break KeyUpdate handling. Instead we no longer signal the start and end
1686 of a post handshake message exchange (although the messages themselves are
1687 still signalled). This could break some applications that were expecting
1688 the old signals. However without this KeyUpdate is not usable for many
1693 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1695 * Timing vulnerability in DSA signature generation
1697 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1698 timing side channel attack. An attacker could use variations in the signing
1699 algorithm to recover the private key.
1701 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1706 * Timing vulnerability in ECDSA signature generation
1708 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1709 timing side channel attack. An attacker could use variations in the signing
1710 algorithm to recover the private key.
1712 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1717 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1718 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1719 of two gigabytes and the error handling improved.
1721 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1722 categorized as a normal bug, not a security issue, because the DRBG reseeds
1723 automatically and is fully functional even without additional randomness
1724 provided by the application.
1726 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1728 * Add a new ClientHello callback. Provides a callback interface that gives
1729 the application the ability to adjust the nascent SSL object at the
1730 earliest stage of ClientHello processing, immediately after extensions have
1731 been collected but before they have been processed. In particular, this
1732 callback can adjust the supported TLS versions in response to the contents
1737 * Add SM2 base algorithm support.
1741 * s390x assembly pack: add (improved) hardware-support for the following
1742 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1743 aes-cfb/cfb8, aes-ecb.
1747 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1748 parameter is no longer accepted, as it leads to a corrupt table. NULL
1749 pem_str is reserved for alias entries only.
1753 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1754 step for prime curves. The new implementation is based on formulae from
1755 differential addition-and-doubling in homogeneous projective coordinates
1756 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1757 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1758 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1759 to work in projective coordinates.
1761 *Billy Bob Brumley, Nicola Tuveri*
1763 * Change generating and checking of primes so that the error rate of not
1764 being prime depends on the intended use based on the size of the input.
1765 For larger primes this will result in more rounds of Miller-Rabin.
1766 The maximal error rate for primes with more than 1080 bits is lowered
1769 *Kurt Roeckx, Annie Yousar*
1771 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1775 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1776 moving between systems, and to avoid confusion when a Windows build is
1777 done with mingw vs with MSVC. For POSIX installs, there's still a
1778 symlink or copy named 'tsget' to avoid that confusion as well.
1782 * Revert blinding in ECDSA sign and instead make problematic addition
1783 length-invariant. Switch even to fixed-length Montgomery multiplication.
1787 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1788 step for binary curves. The new implementation is based on formulae from
1789 differential addition-and-doubling in mixed Lopez-Dahab projective
1790 coordinates, modified to independently blind the operands.
1792 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1794 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1795 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1796 EC_METHODs to implement their own specialized "ladder step", to take
1797 advantage of more favorable coordinate systems or more efficient
1798 differential addition-and-doubling algorithms.
1800 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1802 * Modified the random device based seed sources to keep the relevant
1803 file descriptors open rather than reopening them on each access.
1804 This allows such sources to operate in a chroot() jail without
1805 the associated device nodes being available. This behaviour can be
1806 controlled using RAND_keep_random_devices_open().
1810 * Numerous side-channel attack mitigations have been applied. This may have
1811 performance impacts for some algorithms for the benefit of improved
1812 security. Specific changes are noted in this change log by their respective
1817 * AIX shared library support overhaul. Switch to AIX "natural" way of
1818 handling shared libraries, which means collecting shared objects of
1819 different versions and bitnesses in one common archive. This allows to
1820 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1821 doesn't affect the way 3rd party applications are linked, only how
1822 multi-version installation is managed.
1826 * Make ec_group_do_inverse_ord() more robust and available to other
1827 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1828 mitigations are applied to the fallback BN_mod_inverse().
1829 When using this function rather than BN_mod_inverse() directly, new
1830 EC cryptosystem implementations are then safer-by-default.
1834 * Add coordinate blinding for EC_POINT and implement projective
1835 coordinate blinding for generic prime curves as a countermeasure to
1836 chosen point SCA attacks.
1838 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1840 * Add blinding to ECDSA and DSA signatures to protect against side channel
1841 attacks discovered by Keegan Ryan (NCC Group).
1845 * Enforce checking in the pkeyutl command line app to ensure that the input
1846 length does not exceed the maximum supported digest length when performing
1847 a sign, verify or verifyrecover operation.
1851 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1852 I/O in combination with something like select() or poll() will hang. This
1853 can be turned off again using SSL_CTX_clear_mode().
1854 Many applications do not properly handle non-application data records, and
1855 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1856 around the problems in those applications, but can also break some.
1857 It's recommended to read the manpages about SSL_read(), SSL_write(),
1858 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1859 SSL_CTX_set_read_ahead() again.
1863 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1864 now allow empty (zero character) pass phrases.
1868 * Apply blinding to binary field modular inversion and remove patent
1869 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1873 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1874 binary and prime elliptic curves.
1878 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1879 constant time fixed point multiplication.
1883 * Revise elliptic curve scalar multiplication with timing attack
1884 defenses: ec_wNAF_mul redirects to a constant time implementation
1885 when computing fixed point and variable point multiplication (which
1886 in OpenSSL are mostly used with secret scalars in keygen, sign,
1887 ECDH derive operations).
1888 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1891 * Updated CONTRIBUTING
1895 * Updated DRBG / RAND to request nonce and additional low entropy
1896 randomness from the system.
1898 *Matthias St. Pierre*
1900 * Updated 'openssl rehash' to use OpenSSL consistent default.
1904 * Moved the load of the ssl_conf module to libcrypto, which helps
1905 loading engines that libssl uses before libssl is initialised.
1909 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1913 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1915 *Ingo Schwarze, Rich Salz*
1917 * Added output of accepting IP address and port for 'openssl s_server'
1921 * Added a new API for TLSv1.3 ciphersuites:
1922 SSL_CTX_set_ciphersuites()
1923 SSL_set_ciphersuites()
1927 * Memory allocation failures consistently add an error to the error
1932 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1933 in libcrypto when run as setuid/setgid.
1937 * Load any config file by default when libssl is used.
1941 * Added new public header file <openssl/rand_drbg.h> and documentation
1942 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1944 *Matthias St. Pierre*
1946 * QNX support removed (cannot find contributors to get their approval
1947 for the license change).
1951 * TLSv1.3 replay protection for early data has been implemented. See the
1952 SSL_read_early_data() man page for further details.
1956 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1957 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1958 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1959 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1960 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1961 configuration has been separated out. See the ciphers man page or the
1962 SSL_CTX_set_ciphersuites() man page for more information.
1966 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1967 in responder mode now supports the new "-multi" option, which
1968 spawns the specified number of child processes to handle OCSP
1969 requests. The "-timeout" option now also limits the OCSP
1970 responder's patience to wait to receive the full client request
1971 on a newly accepted connection. Child processes are respawned
1972 as needed, and the CA index file is automatically reloaded
1973 when changed. This makes it possible to run the "ocsp" responder
1974 as a long-running service, making the OpenSSL CA somewhat more
1975 feature-complete. In this mode, most diagnostic messages logged
1976 after entering the event loop are logged via syslog(3) rather than
1981 * Added support for X448 and Ed448. Heavily based on original work by
1986 * Extend OSSL_STORE with capabilities to search and to narrow the set of
1987 objects loaded. This adds the functions OSSL_STORE_expect() and
1988 OSSL_STORE_find() as well as needed tools to construct searches and
1989 get the search data out of them.
1993 * Support for TLSv1.3 added. Note that users upgrading from an earlier
1994 version of OpenSSL should review their configuration settings to ensure
1995 that they are still appropriate for TLSv1.3. For further information see:
1996 <https://wiki.openssl.org/index.php/TLS1.3>
2000 * Grand redesign of the OpenSSL random generator
2002 The default RAND method now utilizes an AES-CTR DRBG according to
2003 NIST standard SP 800-90Ar1. The new random generator is essentially
2004 a port of the default random generator from the OpenSSL FIPS 2.0
2005 object module. It is a hybrid deterministic random bit generator
2006 using an AES-CTR bit stream and which seeds and reseeds itself
2007 automatically using trusted system entropy sources.
2009 Some of its new features are:
2010 - Support for multiple DRBG instances with seed chaining.
2011 - The default RAND method makes use of a DRBG.
2012 - There is a public and private DRBG instance.
2013 - The DRBG instances are fork-safe.
2014 - Keep all global DRBG instances on the secure heap if it is enabled.
2015 - The public and private DRBG instance are per thread for lock free
2018 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2020 * Changed Configure so it only says what it does and doesn't dump
2021 so much data. Instead, ./configdata.pm should be used as a script
2022 to display all sorts of configuration data.
2026 * Added processing of "make variables" to Configure.
2030 * Added SHA512/224 and SHA512/256 algorithm support.
2034 * The last traces of Netware support, first removed in 1.1.0, have
2039 * Get rid of Makefile.shared, and in the process, make the processing
2040 of certain files (rc.obj, or the .def/.map/.opt files produced from
2041 the ordinal files) more visible and hopefully easier to trace and
2042 debug (or make silent).
2046 * Make it possible to have environment variable assignments as
2047 arguments to config / Configure.
2051 * Add multi-prime RSA (RFC 8017) support.
2055 * Add SM3 implemented according to GB/T 32905-2016
2056 *Jack Lloyd <jack.lloyd@ribose.com>,*
2057 *Ronald Tse <ronald.tse@ribose.com>,*
2058 *Erick Borsboom <erick.borsboom@ribose.com>*
2060 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2061 as documented in RFC6066.
2062 Based on a patch from Tomasz Moń
2064 *Filipe Raimundo da Silva*
2066 * Add SM4 implemented according to GB/T 32907-2016.
2067 *Jack Lloyd <jack.lloyd@ribose.com>,*
2068 *Ronald Tse <ronald.tse@ribose.com>,*
2069 *Erick Borsboom <erick.borsboom@ribose.com>*
2071 * Reimplement -newreq-nodes and ERR_error_string_n; the
2072 original author does not agree with the license change.
2076 * Add ARIA AEAD TLS support.
2080 * Some macro definitions to support VS6 have been removed. Visual
2081 Studio 6 has not worked since 1.1.0
2085 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2086 without clearing the errors.
2090 * Add "atfork" functions. If building on a system that without
2091 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2092 requirements. The RAND facility now uses/requires this.
2100 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2101 not possible to disable entirely. However, it's still possible to
2102 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2105 To disable, configure with 'no-ui-console'. 'no-ui' is still
2106 possible to use as an alias. Check at compile time with the
2107 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2108 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2112 * Add a STORE module, which implements a uniform and URI based reader of
2113 stores that can contain keys, certificates, CRLs and numerous other
2114 objects. The main API is loosely based on a few stdio functions,
2115 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2116 OSSL_STORE_error and OSSL_STORE_close.
2117 The implementation uses backends called "loaders" to implement arbitrary
2118 URI schemes. There is one built in "loader" for the 'file' scheme.
2122 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2123 then adjusted to work on FreeBSD 8.4 as well.
2124 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2125 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2129 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2130 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2131 error code calls like this:
2133 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2135 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2136 that can be encoded in C. For the foreseeable future, this will only
2139 *Richard Levitte and Tim Hudson*
2141 * Removed BSD cryptodev engine.
2145 * Add a build target 'build_all_generated', to build all generated files
2146 and only that. This can be used to prepare everything that requires
2147 things like perl for a system that lacks perl and then move everything
2148 to that system and do the rest of the build there.
2152 * In the UI interface, make it possible to duplicate the user data. This
2153 can be used by engines that need to retain the data for a longer time
2154 than just the call where this user data is passed.
2158 * Ignore the '-named_curve auto' value for compatibility of applications
2161 *Tomas Mraz <tmraz@fedoraproject.org>*
2163 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2164 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2165 alerts across multiple records (some of which could be empty). In practice
2166 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2167 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2168 support this at all. Supporting it adds significant complexity to the
2169 record layer, and its removal is unlikely to cause interoperability
2174 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2175 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2176 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2181 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2182 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2184 *Richard Levitte, Andy Polyakov*
2186 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2191 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2192 platform rather than 'mingw'.
2196 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2197 success if they are asked to add an object which already exists
2198 in the store. This change cascades to other functions which load
2199 certificates and CRLs.
2203 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2204 facilitate stack unwinding even from assembly subroutines.
2208 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2209 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2213 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2214 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2215 which is the minimum version we support.
2219 * Certificate time validation (X509_cmp_time) enforces stricter
2220 compliance with RFC 5280. Fractional seconds and timezone offsets
2221 are no longer allowed.
2225 * Add support for ARIA
2229 * s_client will now send the Server Name Indication (SNI) extension by
2230 default unless the new "-noservername" option is used. The server name is
2231 based on the host provided to the "-connect" option unless overridden by
2232 using "-servername".
2236 * Add support for SipHash
2240 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2241 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2242 prevent issues where no progress is being made and the peer continually
2243 sends unrecognised record types, using up resources processing them.
2247 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2248 using the algorithm defined in
2249 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2253 * Heartbeat support has been removed; the ABI is changed for now.
2255 *Richard Levitte, Rich Salz*
2257 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2261 * The RSA "null" method, which was partially supported to avoid patent
2262 issues, has been replaced to always returns NULL.
2269 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2271 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2272 used even when parsing explicit parameters, when loading a encoded key
2273 or calling `EC_GROUP_new_from_ecpkparameters()`/
2274 `EC_GROUP_new_from_ecparameters()`.
2275 This prevents bypass of security hardening and performance gains,
2276 especially for curves with specialized EC_METHODs.
2277 By default, if a key encoded with explicit parameters is loaded and later
2278 encoded, the output is still encoded with explicit parameters, even if
2279 internally a "named" EC_GROUP is used for computation.
2283 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2284 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2285 NULL. After this change, only the cofactor parameter can be NULL. It also
2286 does some minimal sanity checks on the passed order.
2291 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2292 An attack is simple, if the first CMS_recipientInfo is valid but the
2293 second CMS_recipientInfo is chosen ciphertext. If the second
2294 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2295 encryption key will be replaced by garbage, and the message cannot be
2296 decoded, but if the RSA decryption fails, the correct encryption key is
2297 used and the recipient will not notice the attack.
2298 As a work around for this potential attack the length of the decrypted
2299 key must be equal to the cipher default key length, in case the
2300 certifiate is not given and all recipientInfo are tried out.
2301 The old behaviour can be re-enabled in the CMS code by setting the
2302 CMS_DEBUG_DECRYPT flag.
2307 * Use Windows installation paths in the mingw builds
2309 Mingw isn't a POSIX environment per se, which means that Windows
2310 paths should be used for installation.
2315 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2317 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2318 This changes the size when using the genpkey app when no size is given. It
2319 fixes an omission in earlier changes that changed all RSA, DSA and DH
2320 generation apps to use 2048 bits by default.
2324 * Prevent over long nonces in ChaCha20-Poly1305.
2326 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2327 for every encryption operation. RFC 7539 specifies that the nonce value
2328 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2329 and front pads the nonce with 0 bytes if it is less than 12
2330 bytes. However it also incorrectly allows a nonce to be set of up to 16
2331 bytes. In this case only the last 12 bytes are significant and any
2332 additional leading bytes are ignored.
2334 It is a requirement of using this cipher that nonce values are
2335 unique. Messages encrypted using a reused nonce value are susceptible to
2336 serious confidentiality and integrity attacks. If an application changes
2337 the default nonce length to be longer than 12 bytes and then makes a
2338 change to the leading bytes of the nonce expecting the new value to be a
2339 new unique nonce then such an application could inadvertently encrypt
2340 messages with a reused nonce.
2342 Additionally the ignored bytes in a long nonce are not covered by the
2343 integrity guarantee of this cipher. Any application that relies on the
2344 integrity of these ignored leading bytes of a long nonce may be further
2345 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2346 is safe because no such use sets such a long nonce value. However user
2347 applications that use this cipher directly and set a non-default nonce
2348 length to be longer than 12 bytes may be vulnerable.
2350 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2356 * Added SCA hardening for modular field inversion in EC_GROUP through
2357 a new dedicated field_inv() pointer in EC_METHOD.
2358 This also addresses a leakage affecting conversions from projective
2359 to affine coordinates.
2361 *Billy Bob Brumley, Nicola Tuveri*
2363 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2364 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2368 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2372 * Remove the 'dist' target and add a tarball building script. The
2373 'dist' target has fallen out of use, and it shouldn't be
2374 necessary to configure just to create a source distribution.
2378 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2380 * Timing vulnerability in DSA signature generation
2382 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2383 timing side channel attack. An attacker could use variations in the signing
2384 algorithm to recover the private key.
2386 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2391 * Timing vulnerability in ECDSA signature generation
2393 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2394 timing side channel attack. An attacker could use variations in the signing
2395 algorithm to recover the private key.
2397 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2402 * Add coordinate blinding for EC_POINT and implement projective
2403 coordinate blinding for generic prime curves as a countermeasure to
2404 chosen point SCA attacks.
2406 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2408 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2410 * Client DoS due to large DH parameter
2412 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2413 malicious server can send a very large prime value to the client. This will
2414 cause the client to spend an unreasonably long period of time generating a
2415 key for this prime resulting in a hang until the client has finished. This
2416 could be exploited in a Denial Of Service attack.
2418 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2423 * Cache timing vulnerability in RSA Key Generation
2425 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2426 a cache timing side channel attack. An attacker with sufficient access to
2427 mount cache timing attacks during the RSA key generation process could
2428 recover the private key.
2430 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2431 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2436 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2437 parameter is no longer accepted, as it leads to a corrupt table. NULL
2438 pem_str is reserved for alias entries only.
2442 * Revert blinding in ECDSA sign and instead make problematic addition
2443 length-invariant. Switch even to fixed-length Montgomery multiplication.
2447 * Change generating and checking of primes so that the error rate of not
2448 being prime depends on the intended use based on the size of the input.
2449 For larger primes this will result in more rounds of Miller-Rabin.
2450 The maximal error rate for primes with more than 1080 bits is lowered
2453 *Kurt Roeckx, Annie Yousar*
2455 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2459 * Add blinding to ECDSA and DSA signatures to protect against side channel
2460 attacks discovered by Keegan Ryan (NCC Group).
2464 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2465 now allow empty (zero character) pass phrases.
2469 * Certificate time validation (X509_cmp_time) enforces stricter
2470 compliance with RFC 5280. Fractional seconds and timezone offsets
2471 are no longer allowed.
2475 * Fixed a text canonicalisation bug in CMS
2477 Where a CMS detached signature is used with text content the text goes
2478 through a canonicalisation process first prior to signing or verifying a
2479 signature. This process strips trailing space at the end of lines, converts
2480 line terminators to CRLF and removes additional trailing line terminators
2481 at the end of a file. A bug in the canonicalisation process meant that
2482 some characters, such as form-feed, were incorrectly treated as whitespace
2483 and removed. This is contrary to the specification (RFC5485). This fix
2484 could mean that detached text data signed with an earlier version of
2485 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2486 signed with a fixed OpenSSL may fail to verify with an earlier version of
2487 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2488 and use the "-binary" flag (for the "cms" command line application) or set
2489 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2493 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2495 * Constructed ASN.1 types with a recursive definition could exceed the stack
2497 Constructed ASN.1 types with a recursive definition (such as can be found
2498 in PKCS7) could eventually exceed the stack given malicious input with
2499 excessive recursion. This could result in a Denial Of Service attack. There
2500 are no such structures used within SSL/TLS that come from untrusted sources
2501 so this is considered safe.
2503 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2509 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2511 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2512 effectively reduced to only comparing the least significant bit of each
2513 byte. This allows an attacker to forge messages that would be considered as
2514 authenticated in an amount of tries lower than that guaranteed by the
2515 security claims of the scheme. The module can only be compiled by the
2516 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2518 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2524 * Add a build target 'build_all_generated', to build all generated files
2525 and only that. This can be used to prepare everything that requires
2526 things like perl for a system that lacks perl and then move everything
2527 to that system and do the rest of the build there.
2531 * Backport SSL_OP_NO_RENGOTIATION
2533 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2534 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2535 changes this is no longer possible in 1.1.0. Therefore the new
2536 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2537 1.1.0 to provide equivalent functionality.
2539 Note that if an application built against 1.1.0h headers (or above) is run
2540 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2541 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2545 * Removed the OS390-Unix config target. It relied on a script that doesn't
2550 * rsaz_1024_mul_avx2 overflow bug on x86_64
2552 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2553 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2554 Analysis suggests that attacks against RSA and DSA as a result of this
2555 defect would be very difficult to perform and are not believed likely.
2556 Attacks against DH1024 are considered just feasible, because most of the
2557 work necessary to deduce information about a private key may be performed
2558 offline. The amount of resources required for such an attack would be
2559 significant. However, for an attack on TLS to be meaningful, the server
2560 would have to share the DH1024 private key among multiple clients, which is
2561 no longer an option since CVE-2016-0701.
2563 This only affects processors that support the AVX2 but not ADX extensions
2564 like Intel Haswell (4th generation).
2566 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2567 was originally found via the OSS-Fuzz project.
2572 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2574 * bn_sqrx8x_internal carry bug on x86_64
2576 There is a carry propagating bug in the x86_64 Montgomery squaring
2577 procedure. No EC algorithms are affected. Analysis suggests that attacks
2578 against RSA and DSA as a result of this defect would be very difficult to
2579 perform and are not believed likely. Attacks against DH are considered just
2580 feasible (although very difficult) because most of the work necessary to
2581 deduce information about a private key may be performed offline. The amount
2582 of resources required for such an attack would be very significant and
2583 likely only accessible to a limited number of attackers. An attacker would
2584 additionally need online access to an unpatched system using the target
2585 private key in a scenario with persistent DH parameters and a private
2586 key that is shared between multiple clients.
2588 This only affects processors that support the BMI1, BMI2 and ADX extensions
2589 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2591 This issue was reported to OpenSSL by the OSS-Fuzz project.
2596 * Malformed X.509 IPAddressFamily could cause OOB read
2598 If an X.509 certificate has a malformed IPAddressFamily extension,
2599 OpenSSL could do a one-byte buffer overread. The most likely result
2600 would be an erroneous display of the certificate in text format.
2602 This issue was reported to OpenSSL by the OSS-Fuzz project.
2607 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2609 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2610 platform rather than 'mingw'.
2614 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2615 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2616 which is the minimum version we support.
2620 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2622 * Encrypt-Then-Mac renegotiation crash
2624 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2625 negotiated where it was not in the original handshake (or vice-versa) then
2626 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2627 and servers are affected.
2629 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2634 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2636 * Truncated packet could crash via OOB read
2638 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2639 cipher is being used, then a truncated packet can cause that host to
2640 perform an out-of-bounds read, usually resulting in a crash.
2642 This issue was reported to OpenSSL by Robert Święcki of Google.
2647 * Bad (EC)DHE parameters cause a client crash
2649 If a malicious server supplies bad parameters for a DHE or ECDHE key
2650 exchange then this can result in the client attempting to dereference a
2651 NULL pointer leading to a client crash. This could be exploited in a Denial
2654 This issue was reported to OpenSSL by Guido Vranken.
2659 * BN_mod_exp may produce incorrect results on x86_64
2661 There is a carry propagating bug in the x86_64 Montgomery squaring
2662 procedure. No EC algorithms are affected. Analysis suggests that attacks
2663 against RSA and DSA as a result of this defect would be very difficult to
2664 perform and are not believed likely. Attacks against DH are considered just
2665 feasible (although very difficult) because most of the work necessary to
2666 deduce information about a private key may be performed offline. The amount
2667 of resources required for such an attack would be very significant and
2668 likely only accessible to a limited number of attackers. An attacker would
2669 additionally need online access to an unpatched system using the target
2670 private key in a scenario with persistent DH parameters and a private
2671 key that is shared between multiple clients. For example this can occur by
2672 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2673 similar to CVE-2015-3193 but must be treated as a separate problem.
2675 This issue was reported to OpenSSL by the OSS-Fuzz project.
2680 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2682 * ChaCha20/Poly1305 heap-buffer-overflow
2684 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2685 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2686 crash. This issue is not considered to be exploitable beyond a DoS.
2688 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2693 * CMS Null dereference
2695 Applications parsing invalid CMS structures can crash with a NULL pointer
2696 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2697 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2698 structure callback if an attempt is made to free certain invalid encodings.
2699 Only CHOICE structures using a callback which do not handle NULL value are
2702 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2707 * Montgomery multiplication may produce incorrect results
2709 There is a carry propagating bug in the Broadwell-specific Montgomery
2710 multiplication procedure that handles input lengths divisible by, but
2711 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2712 and DH private keys are impossible. This is because the subroutine in
2713 question is not used in operations with the private key itself and an input
2714 of the attacker's direct choice. Otherwise the bug can manifest itself as
2715 transient authentication and key negotiation failures or reproducible
2716 erroneous outcome of public-key operations with specially crafted input.
2717 Among EC algorithms only Brainpool P-512 curves are affected and one
2718 presumably can attack ECDH key negotiation. Impact was not analyzed in
2719 detail, because pre-requisites for attack are considered unlikely. Namely
2720 multiple clients have to choose the curve in question and the server has to
2721 share the private key among them, neither of which is default behaviour.
2722 Even then only clients that chose the curve will be affected.
2724 This issue was publicly reported as transient failures and was not
2725 initially recognized as a security issue. Thanks to Richard Morgan for
2726 providing reproducible case.
2731 * Removed automatic addition of RPATH in shared libraries and executables,
2732 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2736 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2738 * Fix Use After Free for large message sizes
2740 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2741 message larger than approx 16k is received then the underlying buffer to
2742 store the incoming message is reallocated and moved. Unfortunately a
2743 dangling pointer to the old location is left which results in an attempt to
2744 write to the previously freed location. This is likely to result in a
2745 crash, however it could potentially lead to execution of arbitrary code.
2747 This issue only affects OpenSSL 1.1.0a.
2749 This issue was reported to OpenSSL by Robert Święcki.
2754 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2756 * OCSP Status Request extension unbounded memory growth
2758 A malicious client can send an excessively large OCSP Status Request
2759 extension. If that client continually requests renegotiation, sending a
2760 large OCSP Status Request extension each time, then there will be unbounded
2761 memory growth on the server. This will eventually lead to a Denial Of
2762 Service attack through memory exhaustion. Servers with a default
2763 configuration are vulnerable even if they do not support OCSP. Builds using
2764 the "no-ocsp" build time option are not affected.
2766 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2771 * SSL_peek() hang on empty record
2773 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2774 sends an empty record. This could be exploited by a malicious peer in a
2775 Denial Of Service attack.
2777 This issue was reported to OpenSSL by Alex Gaynor.
2782 * Excessive allocation of memory in tls_get_message_header() and
2783 dtls1_preprocess_fragment()
2785 A (D)TLS message includes 3 bytes for its length in the header for the
2786 message. This would allow for messages up to 16Mb in length. Messages of
2787 this length are excessive and OpenSSL includes a check to ensure that a
2788 peer is sending reasonably sized messages in order to avoid too much memory
2789 being consumed to service a connection. A flaw in the logic of version
2790 1.1.0 means that memory for the message is allocated too early, prior to
2791 the excessive message length check. Due to way memory is allocated in
2792 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2793 to service a connection. This could lead to a Denial of Service through
2794 memory exhaustion. However, the excessive message length check still takes
2795 place, and this would cause the connection to immediately fail. Assuming
2796 that the application calls SSL_free() on the failed connection in a timely
2797 manner then the 21Mb of allocated memory will then be immediately freed
2798 again. Therefore the excessive memory allocation will be transitory in
2799 nature. This then means that there is only a security impact if:
2801 1) The application does not call SSL_free() in a timely manner in the event
2802 that the connection fails
2804 2) The application is working in a constrained environment where there is
2805 very little free memory
2807 3) The attacker initiates multiple connection attempts such that there are
2808 multiple connections in a state where memory has been allocated for the
2809 connection; SSL_free() has not yet been called; and there is insufficient
2810 memory to service the multiple requests.
2812 Except in the instance of (1) above any Denial Of Service is likely to be
2813 transitory because as soon as the connection fails the memory is
2814 subsequently freed again in the SSL_free() call. However there is an
2815 increased risk during this period of application crashes due to the lack of
2816 memory - which would then mean a more serious Denial of Service.
2818 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2819 (CVE-2016-6307 and CVE-2016-6308)
2823 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2824 had to be removed. Primary reason is that vendor assembler can't
2825 assemble our modules with -KPIC flag. As result it, assembly
2826 support, was not even available as option. But its lack means
2827 lack of side-channel resistant code, which is incompatible with
2828 security by todays standards. Fortunately gcc is readily available
2829 prepackaged option, which we firmly point at...
2833 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2835 * Windows command-line tool supports UTF-8 opt-in option for arguments
2836 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2837 (to any value) allows Windows user to access PKCS#12 file generated
2838 with Windows CryptoAPI and protected with non-ASCII password, as well
2839 as files generated under UTF-8 locale on Linux also protected with
2844 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2845 have been disabled by default and removed from DEFAULT, just like RC4.
2846 See the RC4 item below to re-enable both.
2850 * The method for finding the storage location for the Windows RAND seed file
2851 has changed. First we check %RANDFILE%. If that is not set then we check
2852 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2853 all else fails we fall back to C:\.
2857 * The EVP_EncryptUpdate() function has had its return type changed from void
2858 to int. A return of 0 indicates and error while a return of 1 indicates
2863 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2864 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2865 off the constant time implementation for RSA, DSA and DH have been made
2866 no-ops and deprecated.
2870 * Windows RAND implementation was simplified to only get entropy by
2871 calling CryptGenRandom(). Various other RAND-related tickets
2874 *Joseph Wylie Yandle, Rich Salz*
2876 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2877 and `OPENSSL_LH_`, respectively. The old names are available
2878 with API compatibility. They new names are now completely documented.
2882 * Unify TYPE_up_ref(obj) methods signature.
2883 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2884 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2885 int (instead of void) like all others TYPE_up_ref() methods.
2886 So now these methods also check the return value of CRYPTO_atomic_add(),
2887 and the validity of object reference counter.
2889 *fdasilvayy@gmail.com*
2891 * With Windows Visual Studio builds, the .pdb files are installed
2892 alongside the installed libraries and executables. For a static
2893 library installation, ossl_static.pdb is the associate compiler
2894 generated .pdb file to be used when linking programs.
2898 * Remove openssl.spec. Packaging files belong with the packagers.
2902 * Automatic Darwin/OSX configuration has had a refresh, it will now
2903 recognise x86_64 architectures automatically. You can still decide
2904 to build for a different bitness with the environment variable
2905 KERNEL_BITS (can be 32 or 64), for example:
2907 KERNEL_BITS=32 ./config
2911 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2912 256 bit AES and HMAC with SHA256.
2916 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2920 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2924 * To enable users to have their own config files and build file templates,
2925 Configure looks in the directory indicated by the environment variable
2926 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2927 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2928 name and is used as is.
2932 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2933 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2934 X509_CERT_FILE_CTX was removed.
2938 * "shared" builds are now the default. To create only static libraries use
2939 the "no-shared" Configure option.
2943 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2944 All of these option have not worked for some while and are fundamental
2949 * Make various cleanup routines no-ops and mark them as deprecated. Most
2950 global cleanup functions are no longer required because they are handled
2951 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2952 Explicitly de-initing can cause problems (e.g. where a library that uses
2953 OpenSSL de-inits, but an application is still using it). The affected
2954 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2955 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2956 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2957 COMP_zlib_cleanup().
2961 * --strict-warnings no longer enables runtime debugging options
2962 such as REF_DEBUG. Instead, debug options are automatically
2963 enabled with '--debug' builds.
2965 *Andy Polyakov, Emilia Käsper*
2967 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2968 have been moved out of the public header files. New functions for managing
2969 these have been added.
2973 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2974 objects have been moved out of the public header files. New
2975 functions for managing these have been added.
2979 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2980 have been moved out of the public header files. New functions for managing
2981 these have been added.
2985 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
2986 moved out of the public header files. New functions for managing these
2991 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
2995 * Removed the mk1mf build scripts.
2999 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3000 it is always safe to #include a header now.
3004 * Removed the aged BC-32 config and all its supporting scripts
3008 * Removed support for Ultrix, Netware, and OS/2.
3012 * Add support for HKDF.
3014 *Alessandro Ghedini*
3016 * Add support for blake2b and blake2s
3020 * Added support for "pipelining". Ciphers that have the
3021 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3022 encryptions/decryptions simultaneously. There are currently no built-in
3023 ciphers with this property but the expectation is that engines will be able
3024 to offer it to significantly improve throughput. Support has been extended
3025 into libssl so that multiple records for a single connection can be
3026 processed in one go (for >=TLS 1.1).
3030 * Added the AFALG engine. This is an async capable engine which is able to
3031 offload work to the Linux kernel. In this initial version it only supports
3032 AES128-CBC. The kernel must be version 4.1.0 or greater.
3036 * OpenSSL now uses a new threading API. It is no longer necessary to
3037 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3038 are two supported threading models: pthreads and windows threads. It is
3039 also possible to configure OpenSSL at compile time for "no-threads". The
3040 old threading API should no longer be used. The functions have been
3041 replaced with "no-op" compatibility macros.
3043 *Alessandro Ghedini, Matt Caswell*
3045 * Modify behavior of ALPN to invoke callback after SNI/servername
3046 callback, such that updates to the SSL_CTX affect ALPN.
3050 * Add SSL_CIPHER queries for authentication and key-exchange.
3054 * Changes to the DEFAULT cipherlist:
3055 - Prefer (EC)DHE handshakes over plain RSA.
3056 - Prefer AEAD ciphers over legacy ciphers.
3057 - Prefer ECDSA over RSA when both certificates are available.
3058 - Prefer TLSv1.2 ciphers/PRF.
3059 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3064 * Change the ECC default curve list to be this, in order: x25519,
3065 secp256r1, secp521r1, secp384r1.
3069 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3070 disabled by default. They can be re-enabled using the
3071 enable-weak-ssl-ciphers option to Configure.
3075 * If the server has ALPN configured, but supports no protocols that the
3076 client advertises, send a fatal "no_application_protocol" alert.
3077 This behaviour is SHALL in RFC 7301, though it isn't universally
3078 implemented by other servers.
3082 * Add X25519 support.
3083 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3084 for public and private key encoding using the format documented in
3085 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3086 key generation and key derivation.
3088 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3093 * Deprecate SRP_VBASE_get_by_user.
3094 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3095 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3096 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3097 seed, even if the seed is configured.
3099 Users should use SRP_VBASE_get1_by_user instead. Note that in
3100 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3101 also that even though configuring the SRP seed attempts to hide
3102 invalid usernames by continuing the handshake with fake
3103 credentials, this behaviour is not constant time and no strong
3104 guarantees are made that the handshake is indistinguishable from
3105 that of a valid user.
3109 * Configuration change; it's now possible to build dynamic engines
3110 without having to build shared libraries and vice versa. This
3111 only applies to the engines in engines/, those in crypto/engine/
3112 will always be built into libcrypto (i.e. "static").
3114 Building dynamic engines is enabled by default; to disable, use
3115 the configuration option "disable-dynamic-engine".
3117 The only requirements for building dynamic engines are the
3118 presence of the DSO module and building with position independent
3119 code, so they will also automatically be disabled if configuring
3120 with "disable-dso" or "disable-pic".
3122 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3123 are also taken away from openssl/opensslconf.h, as they are
3128 * Configuration change; if there is a known flag to compile
3129 position independent code, it will always be applied on the
3130 libcrypto and libssl object files, and never on the application
3131 object files. This means other libraries that use routines from
3132 libcrypto / libssl can be made into shared libraries regardless
3133 of how OpenSSL was configured.
3135 If this isn't desirable, the configuration options "disable-pic"
3136 or "no-pic" can be used to disable the use of PIC. This will
3137 also disable building shared libraries and dynamic engines.
3141 * Removed JPAKE code. It was experimental and has no wide use.
3145 * The INSTALL_PREFIX Makefile variable has been renamed to
3146 DESTDIR. That makes for less confusion on what this variable
3147 is for. Also, the configuration option --install_prefix is
3152 * Heartbeat for TLS has been removed and is disabled by default
3153 for DTLS; configure with enable-heartbeats. Code that uses the
3154 old #define's might need to be updated.
3156 *Emilia Käsper, Rich Salz*
3158 * Rename REF_CHECK to REF_DEBUG.
3162 * New "unified" build system
3164 The "unified" build system is aimed to be a common system for all
3165 platforms we support. With it comes new support for VMS.
3167 This system builds supports building in a different directory tree
3168 than the source tree. It produces one Makefile (for unix family
3169 or lookalikes), or one descrip.mms (for VMS).
3171 The source of information to make the Makefile / descrip.mms is
3172 small files called 'build.info', holding the necessary
3173 information for each directory with source to compile, and a
3174 template in Configurations, like unix-Makefile.tmpl or
3177 With this change, the library names were also renamed on Windows
3178 and on VMS. They now have names that are closer to the standard
3179 on Unix, and include the major version number, and in certain
3180 cases, the architecture they are built for. See "Notes on shared
3181 libraries" in INSTALL.
3183 We rely heavily on the perl module Text::Template.
3187 * Added support for auto-initialisation and de-initialisation of the library.
3188 OpenSSL no longer requires explicit init or deinit routines to be called,
3189 except in certain circumstances. See the OPENSSL_init_crypto() and
3190 OPENSSL_init_ssl() man pages for further information.
3194 * The arguments to the DTLSv1_listen function have changed. Specifically the
3195 "peer" argument is now expected to be a BIO_ADDR object.
3197 * Rewrite of BIO networking library. The BIO library lacked consistent
3198 support of IPv6, and adding it required some more extensive
3199 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3200 which hold all types of addresses and chains of address information.
3201 It also introduces a new API, with functions like BIO_socket,
3202 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3203 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3204 have been adapted accordingly.
3208 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3213 * CRIME protection: disable compression by default, even if OpenSSL is
3214 compiled with zlib enabled. Applications can still enable compression
3215 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3216 using the SSL_CONF library to configure compression.
3220 * The signature of the session callback configured with
3221 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3222 was explicitly marked as `const unsigned char*` instead of
3227 * Always DPURIFY. Remove the use of uninitialized memory in the
3228 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3232 * Removed many obsolete configuration items, including
3233 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3234 MD2_CHAR, MD2_INT, MD2_LONG
3236 IDEA_SHORT, IDEA_LONG
3237 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3239 *Rich Salz, with advice from Andy Polyakov*
3241 * Many BN internals have been moved to an internal header file.
3243 *Rich Salz with help from Andy Polyakov*
3245 * Configuration and writing out the results from it has changed.
3246 Files such as Makefile include/openssl/opensslconf.h and are now
3247 produced through general templates, such as Makefile.in and
3248 crypto/opensslconf.h.in and some help from the perl module
3251 Also, the center of configuration information is no longer
3252 Makefile. Instead, Configure produces a perl module in
3253 configdata.pm which holds most of the config data (in the hash
3254 table %config), the target data that comes from the target
3255 configuration in one of the `Configurations/*.conf` files (in
3260 * To clarify their intended purposes, the Configure options
3261 --prefix and --openssldir change their semantics, and become more
3262 straightforward and less interdependent.
3264 --prefix shall be used exclusively to give the location INSTALLTOP
3265 where programs, scripts, libraries, include files and manuals are
3266 going to be installed. The default is now /usr/local.
3268 --openssldir shall be used exclusively to give the default
3269 location OPENSSLDIR where certificates, private keys, CRLs are
3270 managed. This is also where the default openssl.cnf gets
3272 If the directory given with this option is a relative path, the
3273 values of both the --prefix value and the --openssldir value will
3274 be combined to become OPENSSLDIR.
3275 The default for --openssldir is INSTALLTOP/ssl.
3277 Anyone who uses --openssldir to specify where OpenSSL is to be
3278 installed MUST change to use --prefix instead.
3282 * The GOST engine was out of date and therefore it has been removed. An up
3283 to date GOST engine is now being maintained in an external repository.
3284 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3285 support for GOST ciphersuites (these are only activated if a GOST engine
3290 * EGD is no longer supported by default; use enable-egd when
3293 *Ben Kaduk and Rich Salz*
3295 * The distribution now has Makefile.in files, which are used to
3296 create Makefile's when Configure is run. *Configure must be run
3297 before trying to build now.*
3301 * The return value for SSL_CIPHER_description() for error conditions
3306 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3308 Obtaining and performing DNSSEC validation of TLSA records is
3309 the application's responsibility. The application provides
3310 the TLSA records of its choice to OpenSSL, and these are then
3311 used to authenticate the peer.
3313 The TLSA records need not even come from DNS. They can, for
3314 example, be used to implement local end-entity certificate or
3315 trust-anchor "pinning", where the "pin" data takes the form
3316 of TLSA records, which can augment or replace verification
3317 based on the usual WebPKI public certification authorities.
3321 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3322 continues to support deprecated interfaces in default builds.
3323 However, applications are strongly advised to compile their
3324 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3325 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3326 or the 1.1.0 releases.
3328 In environments in which all applications have been ported to
3329 not use any deprecated interfaces OpenSSL's Configure script
3330 should be used with the --api=1.1.0 option to entirely remove
3331 support for the deprecated features from the library and
3332 unconditionally disable them in the installed headers.
3333 Essentially the same effect can be achieved with the "no-deprecated"
3334 argument to Configure, except that this will always restrict
3335 the build to just the latest API, rather than a fixed API
3338 As applications are ported to future revisions of the API,
3339 they should update their compile-time OPENSSL_API_COMPAT define
3340 accordingly, but in most cases should be able to continue to
3341 compile with later releases.
3343 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3344 0x10000000L and 0x00908000L, respectively. However those
3345 versions did not support the OPENSSL_API_COMPAT feature, and
3346 so applications are not typically tested for explicit support
3347 of just the undeprecated features of either release.
3351 * Add support for setting the minimum and maximum supported protocol.
3352 It can bet set via the SSL_set_min_proto_version() and
3353 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3354 MaxProtocol. It's recommended to use the new APIs to disable
3355 protocols instead of disabling individual protocols using
3356 SSL_set_options() or SSL_CONF's Protocol. This change also
3357 removes support for disabling TLS 1.2 in the OpenSSL TLS
3358 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3362 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3366 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3367 and integrates ECDSA and ECDH functionality into EC. Implementations can
3368 now redirect key generation and no longer need to convert to or from
3371 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3372 include the ec.h header file instead.
3376 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3377 ciphers who are no longer supported and drops support the ephemeral RSA key
3378 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3382 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3383 opaque. For HMAC_CTX, the following constructors and destructors
3386 HMAC_CTX *HMAC_CTX_new(void);
3387 void HMAC_CTX_free(HMAC_CTX *ctx);
3389 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3390 destroy such methods has been added. See EVP_MD_meth_new(3) and
3391 EVP_CIPHER_meth_new(3) for documentation.
3394 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3395 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3396 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3397 an already created structure.
3398 2) For consistency with the majority of our object creators and
3399 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3400 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3401 for deprecated builds.
3405 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3406 cryptographic operations to be performed asynchronously as long as an
3407 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3408 further details. Libssl has also had this capability integrated with the
3409 introduction of the new mode SSL_MODE_ASYNC and associated error
3410 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3411 pages. This work was developed in partnership with Intel Corp.
3415 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3416 always enabled now. If you want to disable the support you should
3417 exclude it using the list of supported ciphers. This also means that the
3418 "-no_ecdhe" option has been removed from s_server.
3422 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3423 SSL_{CTX_}set1_curves() which can set a list.
3427 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3428 curve you want to support using SSL_{CTX_}set1_curves().
3432 * State machine rewrite. The state machine code has been significantly
3433 refactored in order to remove much duplication of code and solve issues
3434 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3435 further details). This change does have some associated API changes.
3436 Notably the SSL_state() function has been removed and replaced by
3437 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3438 SSL_set_state() has been removed altogether. The previous handshake states
3439 defined in ssl.h and ssl3.h have also been removed.
3443 * All instances of the string "ssleay" in the public API were replaced
3444 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3445 Some error codes related to internal RSA_eay API's were renamed.
3449 * The demo files in crypto/threads were moved to demo/threads.
3453 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3456 *Matt Caswell, Rich Salz*
3458 * New ASN.1 embed macro.
3460 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3461 structure is not allocated: it is part of the parent. That is instead of
3469 This reduces memory fragmentation and make it impossible to accidentally
3470 set a mandatory field to NULL.
3472 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3473 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3474 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3479 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3483 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3484 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3485 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3486 DES and RC4 ciphersuites.
3490 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3491 This changes the decoding behaviour for some invalid messages,
3492 though the change is mostly in the more lenient direction, and
3493 legacy behaviour is preserved as much as possible.
3497 * Fix no-stdio build.
3498 *David Woodhouse <David.Woodhouse@intel.com> and also*
3499 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3501 * New testing framework
3502 The testing framework has been largely rewritten and is now using
3503 perl and the perl modules Test::Harness and an extended variant of
3504 Test::More called OpenSSL::Test to do its work. All test scripts in
3505 test/ have been rewritten into test recipes, and all direct calls to
3506 executables in test/Makefile have become individual recipes using the
3507 simplified testing OpenSSL::Test::Simple.
3509 For documentation on our testing modules, do:
3511 perldoc test/testlib/OpenSSL/Test/Simple.pm
3512 perldoc test/testlib/OpenSSL/Test.pm
3516 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3517 are used; the latter aborts on memory leaks (usually checked on exit).
3518 Some undocumented "set malloc, etc., hooks" functions were removed
3519 and others were changed. All are now documented.
3523 * In DSA_generate_parameters_ex, if the provided seed is too short,
3526 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3528 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3529 from RFC4279, RFC4785, RFC5487, RFC5489.
3531 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3532 original RSA_PSK patch.
3536 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3537 era flag was never set throughout the codebase (only read). Also removed
3538 SSL3_FLAGS_POP_BUFFER which was only used if
3539 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3543 * Changed the default name options in the "ca", "crl", "req" and "x509"
3544 to be "oneline" instead of "compat".
3548 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3549 not aware of clients that still exhibit this bug, and the workaround
3550 hasn't been working properly for a while.
3554 * The return type of BIO_number_read() and BIO_number_written() as well as
3555 the corresponding num_read and num_write members in the BIO structure has
3556 changed from unsigned long to uint64_t. On platforms where an unsigned
3557 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3562 * Given the pervasive nature of TLS extensions it is inadvisable to run
3563 OpenSSL without support for them. It also means that maintaining
3564 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3565 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3569 * Removed support for the two export grade static DH ciphersuites
3570 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3571 were newly added (along with a number of other static DH ciphersuites) to
3572 1.0.2. However the two export ones have *never* worked since they were
3573 introduced. It seems strange in any case to be adding new export
3574 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3578 * Version negotiation has been rewritten. In particular SSLv23_method(),
3579 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3580 and turned into macros which simply call the new preferred function names
3581 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3582 should use the new names instead. Also as part of this change the ssl23.h
3583 header file has been removed.
3587 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3588 code and the associated standard is no longer considered fit-for-purpose.
3592 * RT2547 was closed. When generating a private key, try to make the
3593 output file readable only by the owner. This behavior change might
3594 be noticeable when interacting with other software.
3596 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3601 * Added HTTP GET support to the ocsp command.
3605 * Changed default digest for the dgst and enc commands from MD5 to
3610 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3614 * Added support for TLS extended master secret from
3615 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3616 initial patch which was a great help during development.
3620 * All libssl internal structures have been removed from the public header
3621 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3622 now redundant). Users should not attempt to access internal structures
3623 directly. Instead they should use the provided API functions.
3627 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3628 Access to deprecated functions can be re-enabled by running config with
3629 "enable-deprecated". In addition applications wishing to use deprecated
3630 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3631 will, by default, disable some transitive includes that previously existed
3632 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3636 * Added support for OCB mode. OpenSSL has been granted a patent license
3637 compatible with the OpenSSL license for use of OCB. Details are available
3638 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3639 for OCB can be removed by calling config with no-ocb.
3643 * SSLv2 support has been removed. It still supports receiving a SSLv2
3644 compatible client hello.
3648 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3649 done while fixing the error code for the key-too-small case.
3651 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3653 * CA.sh has been removed; use CA.pl instead.
3657 * Removed old DES API.
3661 * Remove various unsupported platforms:
3667 Sinix/ReliantUNIX RM400
3672 16-bit platforms such as WIN16
3676 * Clean up OPENSSL_NO_xxx #define's
3677 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3678 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3679 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3680 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3681 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3682 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3683 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3684 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3685 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3686 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3690 * Cleaned up dead code
3691 Remove all but one '#ifdef undef' which is to be looked at.
3695 * Clean up calling of xxx_free routines.
3696 Just like free(), fix most of the xxx_free routines to accept
3697 NULL. Remove the non-null checks from callers. Save much code.
3701 * Add secure heap for storage of private keys (when possible).
3702 Add BIO_s_secmem(), CBIGNUM, etc.
3703 Contributed by Akamai Technologies under our Corporate CLA.
3707 * Experimental support for a new, fast, unbiased prime candidate generator,
3708 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3710 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3712 * New output format NSS in the sess_id command line tool. This allows
3713 exporting the session id and the master key in NSS keylog format.
3715 *Martin Kaiser <martin@kaiser.cx>*
3717 * Harmonize version and its documentation. -f flag is used to display
3720 *mancha <mancha1@zoho.com>*
3722 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3723 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3725 *mancha <mancha1@zoho.com>*
3727 * Fix some double frees. These are not thought to be exploitable.
3729 *mancha <mancha1@zoho.com>*
3731 * A missing bounds check in the handling of the TLS heartbeat extension
3732 can be used to reveal up to 64k of memory to a connected client or
3735 Thanks for Neel Mehta of Google Security for discovering this bug and to
3736 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3737 preparing the fix ([CVE-2014-0160])
3739 *Adam Langley, Bodo Moeller*
3741 * Fix for the attack described in the paper "Recovering OpenSSL
3742 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3743 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3744 <http://eprint.iacr.org/2014/140>
3746 Thanks to Yuval Yarom and Naomi Benger for discovering this
3747 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3749 *Yuval Yarom and Naomi Benger*
3751 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3752 this fixes a limitation in previous versions of OpenSSL.
3756 * Experimental encrypt-then-mac support.
3758 Experimental support for encrypt then mac from
3759 draft-gutmann-tls-encrypt-then-mac-02.txt
3761 To enable it set the appropriate extension number (0x42 for the test
3762 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3764 For non-compliant peers (i.e. just about everything) this should have no
3767 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3771 * Add EVP support for key wrapping algorithms, to avoid problems with
3772 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3773 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3774 algorithms and include tests cases.
3778 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3783 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3784 MGF1 digest and OAEP label.
3788 * Make openssl verify return errors.
3790 *Chris Palmer <palmer@google.com> and Ben Laurie*
3792 * New function ASN1_TIME_diff to calculate the difference between two
3793 ASN1_TIME structures or one structure and the current time.
3797 * Update fips_test_suite to support multiple command line options. New
3798 test to induce all self test errors in sequence and check expected
3803 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3804 sign or verify all in one operation.
3808 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3809 test programs and fips_test_suite. Includes functionality to parse
3810 the minimal script output of fipsalgest.pl directly.
3814 * Add authorisation parameter to FIPS_module_mode_set().
3818 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3822 * Use separate DRBG fields for internal and external flags. New function
3823 FIPS_drbg_health_check() to perform on demand health checking. Add
3824 generation tests to fips_test_suite with reduced health check interval to
3825 demonstrate periodic health checking. Add "nodh" option to
3826 fips_test_suite to skip very slow DH test.
3830 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3835 * More extensive health check for DRBG checking many more failure modes.
3836 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3837 combination: call this in fips_test_suite.
3841 * Add support for canonical generation of DSA parameter 'g'. See
3844 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3845 POST to handle HMAC cases.
3849 * Add functions FIPS_module_version() and FIPS_module_version_text()
3850 to return numerical and string versions of the FIPS module number.
3854 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3855 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3856 outside the validated module in the FIPS capable OpenSSL.
3860 * Minor change to DRBG entropy callback semantics. In some cases
3861 there is no multiple of the block length between min_len and
3862 max_len. Allow the callback to return more than max_len bytes
3863 of entropy but discard any extra: it is the callback's responsibility
3864 to ensure that the extra data discarded does not impact the
3865 requested amount of entropy.
3869 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3870 information in FIPS186-3, SP800-57 and SP800-131A.
3874 * CCM support via EVP. Interface is very similar to GCM case except we
3875 must supply all data in one chunk (i.e. no update, final) and the
3876 message length must be supplied if AAD is used. Add algorithm test
3881 * Initial version of POST overhaul. Add POST callback to allow the status
3882 of POST to be monitored and/or failures induced. Modify fips_test_suite
3883 to use callback. Always run all selftests even if one fails.
3887 * XTS support including algorithm test driver in the fips_gcmtest program.
3888 Note: this does increase the maximum key length from 32 to 64 bytes but
3889 there should be no binary compatibility issues as existing applications
3890 will never use XTS mode.
3894 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3895 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3896 performs algorithm blocking for unapproved PRNG types. Also do not
3897 set PRNG type in FIPS_mode_set(): leave this to the application.
3898 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3899 the standard OpenSSL PRNG: set additional data to a date time vector.
3903 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3904 This shouldn't present any incompatibility problems because applications
3905 shouldn't be using these directly and any that are will need to rethink
3906 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3910 * Extensive self tests and health checking required by SP800-90 DRBG.
3911 Remove strength parameter from FIPS_drbg_instantiate and always
3912 instantiate at maximum supported strength.
3916 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3920 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3924 * New function DH_compute_key_padded() to compute a DH key and pad with
3925 leading zeroes if needed: this complies with SP800-56A et al.
3929 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3930 anything, incomplete, subject to change and largely untested at present.
3934 * Modify fipscanisteronly build option to only build the necessary object
3935 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3939 * Add experimental option FIPSSYMS to give all symbols in
3940 fipscanister.o and FIPS or fips prefix. This will avoid
3941 conflicts with future versions of OpenSSL. Add perl script
3942 util/fipsas.pl to preprocess assembly language source files
3943 and rename any affected symbols.
3947 * Add selftest checks and algorithm block of non-fips algorithms in
3948 FIPS mode. Remove DES2 from selftests.
3952 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3953 return internal method without any ENGINE dependencies. Add new
3954 tiny fips sign and verify functions.
3958 * New build option no-ec2m to disable characteristic 2 code.
3962 * New build option "fipscanisteronly". This only builds fipscanister.o
3963 and (currently) associated fips utilities. Uses the file Makefile.fips
3964 instead of Makefile.org as the prototype.
3968 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3969 Update fips_gcmtest to use IV generator.
3973 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3974 setting output buffer to NULL. The `*Final` function must be
3975 called although it will not retrieve any additional data. The tag
3976 can be set or retrieved with a ctrl. The IV length is by default 12
3977 bytes (96 bits) but can be set to an alternative value. If the IV
3978 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3983 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3984 underlying do_cipher function handles all cipher semantics itself
3985 including padding and finalisation. This is useful if (for example)
3986 an ENGINE cipher handles block padding itself. The behaviour of
3987 do_cipher is subtly changed if this flag is set: the return value
3988 is the number of characters written to the output buffer (zero is
3989 no longer an error code) or a negative error code. Also if the
3990 input buffer is NULL and length 0 finalisation should be performed.
3994 * If a candidate issuer certificate is already part of the constructed
3995 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
3999 * Improve forward-security support: add functions
4001 void SSL_CTX_set_not_resumable_session_callback(
4002 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4003 void SSL_set_not_resumable_session_callback(
4004 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4006 for use by SSL/TLS servers; the callback function will be called whenever a
4007 new session is created, and gets to decide whether the session may be
4008 cached to make it resumable (return 0) or not (return 1). (As by the
4009 SSL/TLS protocol specifications, the session_id sent by the server will be
4010 empty to indicate that the session is not resumable; also, the server will
4011 not generate RFC 4507 (RFC 5077) session tickets.)
4013 A simple reasonable callback implementation is to return is_forward_secure.
4014 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4015 by the SSL/TLS server library, indicating whether it can provide forward
4018 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4020 * New -verify_name option in command line utilities to set verification
4025 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4026 Add CMAC pkey methods.
4030 * Experimental renegotiation in s_server -www mode. If the client
4031 browses /reneg connection is renegotiated. If /renegcert it is
4032 renegotiated requesting a certificate.
4036 * Add an "external" session cache for debugging purposes to s_server. This
4037 should help trace issues which normally are only apparent in deployed
4038 multi-process servers.
4042 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4043 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4044 BIO_set_cipher() and some obscure PEM functions were changed so they
4045 can now return an error. The RAND changes required a change to the
4046 RAND_METHOD structure.
4050 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4051 a gcc attribute to warn if the result of a function is ignored. This
4052 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4053 whose return value is often ignored.
4057 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4058 These allow SCTs (signed certificate timestamps) to be requested and
4059 validated when establishing a connection.
4061 *Rob Percival <robpercival@google.com>*
4066 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4068 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4069 used even when parsing explicit parameters, when loading a encoded key
4070 or calling `EC_GROUP_new_from_ecpkparameters()`/
4071 `EC_GROUP_new_from_ecparameters()`.
4072 This prevents bypass of security hardening and performance gains,
4073 especially for curves with specialized EC_METHODs.
4074 By default, if a key encoded with explicit parameters is loaded and later
4075 encoded, the output is still encoded with explicit parameters, even if
4076 internally a "named" EC_GROUP is used for computation.
4080 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4081 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4082 NULL. After this change, only the cofactor parameter can be NULL. It also
4083 does some minimal sanity checks on the passed order.
4088 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4089 An attack is simple, if the first CMS_recipientInfo is valid but the
4090 second CMS_recipientInfo is chosen ciphertext. If the second
4091 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4092 encryption key will be replaced by garbage, and the message cannot be
4093 decoded, but if the RSA decryption fails, the correct encryption key is
4094 used and the recipient will not notice the attack.
4095 As a work around for this potential attack the length of the decrypted
4096 key must be equal to the cipher default key length, in case the
4097 certifiate is not given and all recipientInfo are tried out.
4098 The old behaviour can be re-enabled in the CMS code by setting the
4099 CMS_DEBUG_DECRYPT flag.
4104 * Document issue with installation paths in diverse Windows builds
4106 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4107 binaries and run-time config file.
4112 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4114 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4115 This changes the size when using the genpkey app when no size is given. It
4116 fixes an omission in earlier changes that changed all RSA, DSA and DH
4117 generation apps to use 2048 bits by default.
4121 * Add FIPS support for Android Arm 64-bit
4123 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4124 Module in Version 2.0.10. For some reason, the corresponding target
4125 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4126 built with FIPS support on Android Arm 64-bit. This omission has been
4129 *Matthias St. Pierre*
4131 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4133 * 0-byte record padding oracle
4135 If an application encounters a fatal protocol error and then calls
4136 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4137 then OpenSSL can respond differently to the calling application if a 0 byte
4138 record is received with invalid padding compared to if a 0 byte record is
4139 received with an invalid MAC. If the application then behaves differently
4140 based on that in a way that is detectable to the remote peer, then this
4141 amounts to a padding oracle that could be used to decrypt data.
4143 In order for this to be exploitable "non-stitched" ciphersuites must be in
4144 use. Stitched ciphersuites are optimised implementations of certain
4145 commonly used ciphersuites. Also the application must call SSL_shutdown()
4146 twice even if a protocol error has occurred (applications should not do
4147 this but some do anyway).
4149 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4150 Aviram, with additional investigation by Steven Collison and Andrew
4151 Hourselt. It was reported to OpenSSL on 10th December 2018.
4156 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4160 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4162 * Microarchitecture timing vulnerability in ECC scalar multiplication
4164 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4165 shown to be vulnerable to a microarchitecture timing side channel attack.
4166 An attacker with sufficient access to mount local timing attacks during
4167 ECDSA signature generation could recover the private key.
4169 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4170 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4176 * Timing vulnerability in DSA signature generation
4178 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4179 timing side channel attack. An attacker could use variations in the signing
4180 algorithm to recover the private key.
4182 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4187 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4188 Module, accidentally introduced while backporting security fixes from the
4189 development branch and hindering the use of ECC in FIPS mode.
4193 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4195 * Client DoS due to large DH parameter
4197 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4198 malicious server can send a very large prime value to the client. This will
4199 cause the client to spend an unreasonably long period of time generating a
4200 key for this prime resulting in a hang until the client has finished. This
4201 could be exploited in a Denial Of Service attack.
4203 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4208 * Cache timing vulnerability in RSA Key Generation
4210 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4211 a cache timing side channel attack. An attacker with sufficient access to
4212 mount cache timing attacks during the RSA key generation process could
4213 recover the private key.
4215 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4216 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4221 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4222 parameter is no longer accepted, as it leads to a corrupt table. NULL
4223 pem_str is reserved for alias entries only.
4227 * Revert blinding in ECDSA sign and instead make problematic addition
4228 length-invariant. Switch even to fixed-length Montgomery multiplication.
4232 * Change generating and checking of primes so that the error rate of not
4233 being prime depends on the intended use based on the size of the input.
4234 For larger primes this will result in more rounds of Miller-Rabin.
4235 The maximal error rate for primes with more than 1080 bits is lowered
4238 *Kurt Roeckx, Annie Yousar*
4240 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4244 * Add blinding to ECDSA and DSA signatures to protect against side channel
4245 attacks discovered by Keegan Ryan (NCC Group).
4249 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4250 now allow empty (zero character) pass phrases.
4254 * Certificate time validation (X509_cmp_time) enforces stricter
4255 compliance with RFC 5280. Fractional seconds and timezone offsets
4256 are no longer allowed.
4260 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4262 * Constructed ASN.1 types with a recursive definition could exceed the stack
4264 Constructed ASN.1 types with a recursive definition (such as can be found
4265 in PKCS7) could eventually exceed the stack given malicious input with
4266 excessive recursion. This could result in a Denial Of Service attack. There
4267 are no such structures used within SSL/TLS that come from untrusted sources
4268 so this is considered safe.
4270 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4276 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4278 * Read/write after SSL object in error state
4280 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4281 mechanism. The intent was that if a fatal error occurred during a handshake
4282 then OpenSSL would move into the error state and would immediately fail if
4283 you attempted to continue the handshake. This works as designed for the
4284 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4285 SSL_connect()), however due to a bug it does not work correctly if
4286 SSL_read() or SSL_write() is called directly. In that scenario, if the
4287 handshake fails then a fatal error will be returned in the initial function
4288 call. If SSL_read()/SSL_write() is subsequently called by the application
4289 for the same SSL object then it will succeed and the data is passed without
4290 being decrypted/encrypted directly from the SSL/TLS record layer.
4292 In order to exploit this issue an application bug would have to be present
4293 that resulted in a call to SSL_read()/SSL_write() being issued after having
4294 already received a fatal error.
4296 This issue was reported to OpenSSL by David Benjamin (Google).
4301 * rsaz_1024_mul_avx2 overflow bug on x86_64
4303 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4304 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4305 Analysis suggests that attacks against RSA and DSA as a result of this
4306 defect would be very difficult to perform and are not believed likely.
4307 Attacks against DH1024 are considered just feasible, because most of the
4308 work necessary to deduce information about a private key may be performed
4309 offline. The amount of resources required for such an attack would be
4310 significant. However, for an attack on TLS to be meaningful, the server
4311 would have to share the DH1024 private key among multiple clients, which is
4312 no longer an option since CVE-2016-0701.
4314 This only affects processors that support the AVX2 but not ADX extensions
4315 like Intel Haswell (4th generation).
4317 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4318 was originally found via the OSS-Fuzz project.
4323 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4325 * bn_sqrx8x_internal carry bug on x86_64
4327 There is a carry propagating bug in the x86_64 Montgomery squaring
4328 procedure. No EC algorithms are affected. Analysis suggests that attacks
4329 against RSA and DSA as a result of this defect would be very difficult to
4330 perform and are not believed likely. Attacks against DH are considered just
4331 feasible (although very difficult) because most of the work necessary to
4332 deduce information about a private key may be performed offline. The amount
4333 of resources required for such an attack would be very significant and
4334 likely only accessible to a limited number of attackers. An attacker would
4335 additionally need online access to an unpatched system using the target
4336 private key in a scenario with persistent DH parameters and a private
4337 key that is shared between multiple clients.
4339 This only affects processors that support the BMI1, BMI2 and ADX extensions
4340 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4342 This issue was reported to OpenSSL by the OSS-Fuzz project.
4347 * Malformed X.509 IPAddressFamily could cause OOB read
4349 If an X.509 certificate has a malformed IPAddressFamily extension,
4350 OpenSSL could do a one-byte buffer overread. The most likely result
4351 would be an erroneous display of the certificate in text format.
4353 This issue was reported to OpenSSL by the OSS-Fuzz project.
4358 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4360 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4361 platform rather than 'mingw'.
4365 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4367 * Truncated packet could crash via OOB read
4369 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4370 cipher is being used, then a truncated packet can cause that host to
4371 perform an out-of-bounds read, usually resulting in a crash.
4373 This issue was reported to OpenSSL by Robert Święcki of Google.
4378 * BN_mod_exp may produce incorrect results on x86_64
4380 There is a carry propagating bug in the x86_64 Montgomery squaring
4381 procedure. No EC algorithms are affected. Analysis suggests that attacks
4382 against RSA and DSA as a result of this defect would be very difficult to
4383 perform and are not believed likely. Attacks against DH are considered just
4384 feasible (although very difficult) because most of the work necessary to
4385 deduce information about a private key may be performed offline. The amount
4386 of resources required for such an attack would be very significant and
4387 likely only accessible to a limited number of attackers. An attacker would
4388 additionally need online access to an unpatched system using the target
4389 private key in a scenario with persistent DH parameters and a private
4390 key that is shared between multiple clients. For example this can occur by
4391 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4392 similar to CVE-2015-3193 but must be treated as a separate problem.
4394 This issue was reported to OpenSSL by the OSS-Fuzz project.
4399 * Montgomery multiplication may produce incorrect results
4401 There is a carry propagating bug in the Broadwell-specific Montgomery
4402 multiplication procedure that handles input lengths divisible by, but
4403 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4404 and DH private keys are impossible. This is because the subroutine in
4405 question is not used in operations with the private key itself and an input
4406 of the attacker's direct choice. Otherwise the bug can manifest itself as
4407 transient authentication and key negotiation failures or reproducible
4408 erroneous outcome of public-key operations with specially crafted input.
4409 Among EC algorithms only Brainpool P-512 curves are affected and one
4410 presumably can attack ECDH key negotiation. Impact was not analyzed in
4411 detail, because pre-requisites for attack are considered unlikely. Namely
4412 multiple clients have to choose the curve in question and the server has to
4413 share the private key among them, neither of which is default behaviour.
4414 Even then only clients that chose the curve will be affected.
4416 This issue was publicly reported as transient failures and was not
4417 initially recognized as a security issue. Thanks to Richard Morgan for
4418 providing reproducible case.
4423 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4424 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4425 prevent issues where no progress is being made and the peer continually
4426 sends unrecognised record types, using up resources processing them.
4430 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4432 * Missing CRL sanity check
4434 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4435 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4436 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4438 This issue only affects the OpenSSL 1.0.2i
4443 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4445 * OCSP Status Request extension unbounded memory growth
4447 A malicious client can send an excessively large OCSP Status Request
4448 extension. If that client continually requests renegotiation, sending a
4449 large OCSP Status Request extension each time, then there will be unbounded
4450 memory growth on the server. This will eventually lead to a Denial Of
4451 Service attack through memory exhaustion. Servers with a default
4452 configuration are vulnerable even if they do not support OCSP. Builds using
4453 the "no-ocsp" build time option are not affected.
4455 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4460 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4463 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4469 * OOB write in MDC2_Update()
4471 An overflow can occur in MDC2_Update() either if called directly or
4472 through the EVP_DigestUpdate() function using MDC2. If an attacker
4473 is able to supply very large amounts of input data after a previous
4474 call to EVP_EncryptUpdate() with a partial block then a length check
4475 can overflow resulting in a heap corruption.
4477 The amount of data needed is comparable to SIZE_MAX which is impractical
4480 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4485 * Malformed SHA512 ticket DoS
4487 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4488 DoS attack where a malformed ticket will result in an OOB read which will
4491 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4492 a custom server callback and ticket lookup mechanism.
4494 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4499 * OOB write in BN_bn2dec()
4501 The function BN_bn2dec() does not check the return value of BN_div_word().
4502 This can cause an OOB write if an application uses this function with an
4503 overly large BIGNUM. This could be a problem if an overly large certificate
4504 or CRL is printed out from an untrusted source. TLS is not affected because
4505 record limits will reject an oversized certificate before it is parsed.
4507 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4512 * OOB read in TS_OBJ_print_bio()
4514 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4515 the total length the OID text representation would use and not the amount
4516 of data written. This will result in OOB reads when large OIDs are
4519 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4524 * Pointer arithmetic undefined behaviour
4526 Avoid some undefined pointer arithmetic
4528 A common idiom in the codebase is to check limits in the following manner:
4531 Where "p" points to some malloc'd data of SIZE bytes and
4534 "len" here could be from some externally supplied data (e.g. from a TLS
4537 The rules of C pointer arithmetic are such that "p + len" is only well
4538 defined where len <= SIZE. Therefore the above idiom is actually
4539 undefined behaviour.
4541 For example this could cause problems if some malloc implementation
4542 provides an address for "p" such that "p + len" actually overflows for
4543 values of len that are too big and therefore p + len < limit.
4545 This issue was reported to OpenSSL by Guido Vranken
4550 * Constant time flag not preserved in DSA signing
4552 Operations in the DSA signing algorithm should run in constant time in
4553 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4554 implementation means that a non-constant time codepath is followed for
4555 certain operations. This has been demonstrated through a cache-timing
4556 attack to be sufficient for an attacker to recover the private DSA key.
4558 This issue was reported by César Pereida (Aalto University), Billy Brumley
4559 (Tampere University of Technology), and Yuval Yarom (The University of
4560 Adelaide and NICTA).
4565 * DTLS buffered message DoS
4567 In a DTLS connection where handshake messages are delivered out-of-order
4568 those messages that OpenSSL is not yet ready to process will be buffered
4569 for later use. Under certain circumstances, a flaw in the logic means that
4570 those messages do not get removed from the buffer even though the handshake
4571 has been completed. An attacker could force up to approx. 15 messages to
4572 remain in the buffer when they are no longer required. These messages will
4573 be cleared when the DTLS connection is closed. The default maximum size for
4574 a message is 100k. Therefore the attacker could force an additional 1500k
4575 to be consumed per connection. By opening many simulataneous connections an
4576 attacker could cause a DoS attack through memory exhaustion.
4578 This issue was reported to OpenSSL by Quan Luo.
4583 * DTLS replay protection DoS
4585 A flaw in the DTLS replay attack protection mechanism means that records
4586 that arrive for future epochs update the replay protection "window" before
4587 the MAC for the record has been validated. This could be exploited by an
4588 attacker by sending a record for the next epoch (which does not have to
4589 decrypt or have a valid MAC), with a very large sequence number. This means
4590 that all subsequent legitimate packets are dropped causing a denial of
4591 service for a specific DTLS connection.
4593 This issue was reported to OpenSSL by the OCAP audit team.
4598 * Certificate message OOB reads
4600 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4601 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4602 theoretical DoS risk but this has not been observed in practice on common
4605 The messages affected are client certificate, client certificate request
4606 and server certificate. As a result the attack can only be performed
4607 against a client or a server which enables client authentication.
4609 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4614 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4616 * Prevent padding oracle in AES-NI CBC MAC check
4618 A MITM attacker can use a padding oracle attack to decrypt traffic
4619 when the connection uses an AES CBC cipher and the server support
4622 This issue was introduced as part of the fix for Lucky 13 padding
4623 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4624 constant time by making sure that always the same bytes are read and
4625 compared against either the MAC or padding bytes. But it no longer
4626 checked that there was enough data to have both the MAC and padding
4629 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4634 * Fix EVP_EncodeUpdate overflow
4636 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4637 Base64 encoding of binary data. If an attacker is able to supply very large
4638 amounts of input data then a length check can overflow resulting in a heap
4641 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4642 the `PEM_write_bio*` family of functions. These are mainly used within the
4643 OpenSSL command line applications, so any application which processes data
4644 from an untrusted source and outputs it as a PEM file should be considered
4645 vulnerable to this issue. User applications that call these APIs directly
4646 with large amounts of untrusted data may also be vulnerable.
4648 This issue was reported by Guido Vranken.
4653 * Fix EVP_EncryptUpdate overflow
4655 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4656 is able to supply very large amounts of input data after a previous call to
4657 EVP_EncryptUpdate() with a partial block then a length check can overflow
4658 resulting in a heap corruption. Following an analysis of all OpenSSL
4659 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4660 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4661 the first called function after an EVP_EncryptInit(), and therefore that
4662 specific call must be safe. The second form is where the length passed to
4663 EVP_EncryptUpdate() can be seen from the code to be some small value and
4664 therefore there is no possibility of an overflow. Since all instances are
4665 one of these two forms, it is believed that there can be no overflows in
4666 internal code due to this problem. It should be noted that
4667 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4668 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4669 of these calls have also been analysed too and it is believed there are no
4670 instances in internal usage where an overflow could occur.
4672 This issue was reported by Guido Vranken.
4677 * Prevent ASN.1 BIO excessive memory allocation
4679 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4680 a short invalid encoding can cause allocation of large amounts of memory
4681 potentially consuming excessive resources or exhausting memory.
4683 Any application parsing untrusted data through d2i BIO functions is
4684 affected. The memory based functions such as d2i_X509() are *not* affected.
4685 Since the memory based functions are used by the TLS library, TLS
4686 applications are not affected.
4688 This issue was reported by Brian Carpenter.
4695 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4696 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4697 in arbitrary stack data being returned in the buffer.
4699 This issue was reported by Guido Vranken.
4704 * Modify behavior of ALPN to invoke callback after SNI/servername
4705 callback, such that updates to the SSL_CTX affect ALPN.
4709 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4714 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4715 methods are enabled and ssl2 is disabled the methods return NULL.
4719 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4721 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4722 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4723 provide any "EXPORT" or "LOW" strength ciphers.
4727 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4728 is by default disabled at build-time. Builds that are not configured with
4729 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4730 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4731 will need to explicitly call either of:
4733 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4735 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4737 as appropriate. Even if either of those is used, or the application
4738 explicitly uses the version-specific SSLv2_method() or its client and
4739 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4740 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4741 ciphers, and SSLv2 56-bit DES are no longer available.
4746 * Fix a double-free in DSA code
4748 A double free bug was discovered when OpenSSL parses malformed DSA private
4749 keys and could lead to a DoS attack or memory corruption for applications
4750 that receive DSA private keys from untrusted sources. This scenario is
4753 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4759 * Disable SRP fake user seed to address a server memory leak.
4761 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4763 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4764 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4765 was changed to ignore the "fake user" SRP seed, even if the seed
4768 Users should use SRP_VBASE_get1_by_user instead. Note that in
4769 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4770 also that even though configuring the SRP seed attempts to hide
4771 invalid usernames by continuing the handshake with fake
4772 credentials, this behaviour is not constant time and no strong
4773 guarantees are made that the handshake is indistinguishable from
4774 that of a valid user.
4779 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4781 In the BN_hex2bn function the number of hex digits is calculated using an
4782 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4783 large values of `i` this can result in `bn_expand` not allocating any
4784 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4785 field as NULL leading to a subsequent NULL ptr deref. For very large values
4786 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4787 In this case memory is allocated to the internal BIGNUM data field, but it
4788 is insufficiently sized leading to heap corruption. A similar issue exists
4789 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4790 is ever called by user applications with very large untrusted hex/dec data.
4791 This is anticipated to be a rare occurrence.
4793 All OpenSSL internal usage of these functions use data that is not expected
4794 to be untrusted, e.g. config file data or application command line
4795 arguments. If user developed applications generate config file data based
4796 on untrusted data then it is possible that this could also lead to security
4797 consequences. This is also anticipated to be rare.
4799 This issue was reported to OpenSSL by Guido Vranken.
4804 * Fix memory issues in `BIO_*printf` functions
4806 The internal `fmtstr` function used in processing a "%s" format string in
4807 the `BIO_*printf` functions could overflow while calculating the length of a
4808 string and cause an OOB read when printing very long strings.
4810 Additionally the internal `doapr_outch` function can attempt to write to an
4811 OOB memory location (at an offset from the NULL pointer) in the event of a
4812 memory allocation failure. In 1.0.2 and below this could be caused where
4813 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4814 could be in processing a very long "%s" format string. Memory leaks can
4817 The first issue may mask the second issue dependent on compiler behaviour.
4818 These problems could enable attacks where large amounts of untrusted data
4819 is passed to the `BIO_*printf` functions. If applications use these functions
4820 in this way then they could be vulnerable. OpenSSL itself uses these
4821 functions when printing out human-readable dumps of ASN.1 data. Therefore
4822 applications that print this data could be vulnerable if the data is from
4823 untrusted sources. OpenSSL command line applications could also be
4824 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4825 as command line arguments.
4827 Libssl is not considered directly vulnerable. Additionally certificates etc
4828 received via remote connections via libssl are also unlikely to be able to
4829 trigger these issues because of message size limits enforced within libssl.
4831 This issue was reported to OpenSSL Guido Vranken.
4836 * Side channel attack on modular exponentiation
4838 A side-channel attack was found which makes use of cache-bank conflicts on
4839 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4840 of RSA keys. The ability to exploit this issue is limited as it relies on
4841 an attacker who has control of code in a thread running on the same
4842 hyper-threaded core as the victim thread which is performing decryptions.
4844 This issue was reported to OpenSSL by Yuval Yarom, The University of
4845 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4846 Nadia Heninger, University of Pennsylvania with more information at
4847 <http://cachebleed.info>.
4852 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4853 if no keysize is specified with default_bits. This fixes an
4854 omission in an earlier change that changed all RSA/DSA key generation
4855 apps to use 2048 bits by default.
4859 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4861 * DH small subgroups
4863 Historically OpenSSL only ever generated DH parameters based on "safe"
4864 primes. More recently (in version 1.0.2) support was provided for
4865 generating X9.42 style parameter files such as those required for RFC 5114
4866 support. The primes used in such files may not be "safe". Where an
4867 application is using DH configured with parameters based on primes that are
4868 not "safe" then an attacker could use this fact to find a peer's private
4869 DH exponent. This attack requires that the attacker complete multiple
4870 handshakes in which the peer uses the same private DH exponent. For example
4871 this could be used to discover a TLS server's private DH exponent if it's
4872 reusing the private DH exponent or it's using a static DH ciphersuite.
4874 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4875 TLS. It is not on by default. If the option is not set then the server
4876 reuses the same private DH exponent for the life of the server process and
4877 would be vulnerable to this attack. It is believed that many popular
4878 applications do set this option and would therefore not be at risk.
4880 The fix for this issue adds an additional check where a "q" parameter is
4881 available (as is the case in X9.42 based parameters). This detects the
4882 only known attack, and is the only possible defense for static DH
4883 ciphersuites. This could have some performance impact.
4885 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4886 default and cannot be disabled. This could have some performance impact.
4888 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4893 * SSLv2 doesn't block disabled ciphers
4895 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4896 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4897 been disabled, provided that the SSLv2 protocol was not also disabled via
4900 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4901 and Sebastian Schinzel.
4906 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4908 * BN_mod_exp may produce incorrect results on x86_64
4910 There is a carry propagating bug in the x86_64 Montgomery squaring
4911 procedure. No EC algorithms are affected. Analysis suggests that attacks
4912 against RSA and DSA as a result of this defect would be very difficult to
4913 perform and are not believed likely. Attacks against DH are considered just
4914 feasible (although very difficult) because most of the work necessary to
4915 deduce information about a private key may be performed offline. The amount
4916 of resources required for such an attack would be very significant and
4917 likely only accessible to a limited number of attackers. An attacker would
4918 additionally need online access to an unpatched system using the target
4919 private key in a scenario with persistent DH parameters and a private
4920 key that is shared between multiple clients. For example this can occur by
4921 default in OpenSSL DHE based SSL/TLS ciphersuites.
4923 This issue was reported to OpenSSL by Hanno Böck.
4928 * Certificate verify crash with missing PSS parameter
4930 The signature verification routines will crash with a NULL pointer
4931 dereference if presented with an ASN.1 signature using the RSA PSS
4932 algorithm and absent mask generation function parameter. Since these
4933 routines are used to verify certificate signature algorithms this can be
4934 used to crash any certificate verification operation and exploited in a
4935 DoS attack. Any application which performs certificate verification is
4936 vulnerable including OpenSSL clients and servers which enable client
4939 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4944 * X509_ATTRIBUTE memory leak
4946 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4947 memory. This structure is used by the PKCS#7 and CMS routines so any
4948 application which reads PKCS#7 or CMS data from untrusted sources is
4949 affected. SSL/TLS is not affected.
4951 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4957 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4958 This changes the decoding behaviour for some invalid messages,
4959 though the change is mostly in the more lenient direction, and
4960 legacy behaviour is preserved as much as possible.
4964 * In DSA_generate_parameters_ex, if the provided seed is too short,
4967 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4969 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4971 * Alternate chains certificate forgery
4973 During certificate verification, OpenSSL will attempt to find an
4974 alternative certificate chain if the first attempt to build such a chain
4975 fails. An error in the implementation of this logic can mean that an
4976 attacker could cause certain checks on untrusted certificates to be
4977 bypassed, such as the CA flag, enabling them to use a valid leaf
4978 certificate to act as a CA and "issue" an invalid certificate.
4980 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4985 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
4987 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
4988 incompatibility in the handling of HMAC. The previous ABI has now been
4993 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
4995 * Malformed ECParameters causes infinite loop
4997 When processing an ECParameters structure OpenSSL enters an infinite loop
4998 if the curve specified is over a specially malformed binary polynomial
5001 This can be used to perform denial of service against any
5002 system which processes public keys, certificate requests or
5003 certificates. This includes TLS clients and TLS servers with
5004 client authentication enabled.
5006 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5011 * Exploitable out-of-bounds read in X509_cmp_time
5013 X509_cmp_time does not properly check the length of the ASN1_TIME
5014 string and can read a few bytes out of bounds. In addition,
5015 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5018 An attacker can use this to craft malformed certificates and CRLs of
5019 various sizes and potentially cause a segmentation fault, resulting in
5020 a DoS on applications that verify certificates or CRLs. TLS clients
5021 that verify CRLs are affected. TLS clients and servers with client
5022 authentication enabled may be affected if they use custom verification
5025 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5026 independently by Hanno Böck.
5031 * PKCS7 crash with missing EnvelopedContent
5033 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5034 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5035 with missing content and trigger a NULL pointer dereference on parsing.
5037 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5038 structures from untrusted sources are affected. OpenSSL clients and
5039 servers are not affected.
5041 This issue was reported to OpenSSL by Michal Zalewski (Google).
5046 * CMS verify infinite loop with unknown hash function
5048 When verifying a signedData message the CMS code can enter an infinite loop
5049 if presented with an unknown hash function OID. This can be used to perform
5050 denial of service against any system which verifies signedData messages using
5052 This issue was reported to OpenSSL by Johannes Bauer.
5057 * Race condition handling NewSessionTicket
5059 If a NewSessionTicket is received by a multi-threaded client when attempting to
5060 reuse a previous ticket then a race condition can occur potentially leading to
5061 a double free of the ticket data.
5066 * Only support 256-bit or stronger elliptic curves with the
5067 'ecdh_auto' setting (server) or by default (client). Of supported
5068 curves, prefer P-256 (both).
5072 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5074 * ClientHello sigalgs DoS fix
5076 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5077 invalid signature algorithms extension a NULL pointer dereference will
5078 occur. This can be exploited in a DoS attack against the server.
5080 This issue was was reported to OpenSSL by David Ramos of Stanford
5084 *Stephen Henson and Matt Caswell*
5086 * Multiblock corrupted pointer fix
5088 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5089 feature only applies on 64 bit x86 architecture platforms that support AES
5090 NI instructions. A defect in the implementation of "multiblock" can cause
5091 OpenSSL's internal write buffer to become incorrectly set to NULL when
5092 using non-blocking IO. Typically, when the user application is using a
5093 socket BIO for writing, this will only result in a failed connection.
5094 However if some other BIO is used then it is likely that a segmentation
5095 fault will be triggered, thus enabling a potential DoS attack.
5097 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5102 * Segmentation fault in DTLSv1_listen fix
5104 The DTLSv1_listen function is intended to be stateless and processes the
5105 initial ClientHello from many peers. It is common for user code to loop
5106 over the call to DTLSv1_listen until a valid ClientHello is received with
5107 an associated cookie. A defect in the implementation of DTLSv1_listen means
5108 that state is preserved in the SSL object from one invocation to the next
5109 that can lead to a segmentation fault. Errors processing the initial
5110 ClientHello can trigger this scenario. An example of such an error could be
5111 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5114 This issue was reported to OpenSSL by Per Allansson.
5119 * Segmentation fault in ASN1_TYPE_cmp fix
5121 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5122 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5123 certificate signature algorithm consistency this can be used to crash any
5124 certificate verification operation and exploited in a DoS attack. Any
5125 application which performs certificate verification is vulnerable including
5126 OpenSSL clients and servers which enable client authentication.
5131 * Segmentation fault for invalid PSS parameters fix
5133 The signature verification routines will crash with a NULL pointer
5134 dereference if presented with an ASN.1 signature using the RSA PSS
5135 algorithm and invalid parameters. Since these routines are used to verify
5136 certificate signature algorithms this can be used to crash any
5137 certificate verification operation and exploited in a DoS attack. Any
5138 application which performs certificate verification is vulnerable including
5139 OpenSSL clients and servers which enable client authentication.
5141 This issue was was reported to OpenSSL by Brian Carpenter.
5146 * ASN.1 structure reuse memory corruption fix
5148 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5149 memory corruption via an invalid write. Such reuse is and has been
5150 strongly discouraged and is believed to be rare.
5152 Applications that parse structures containing CHOICE or ANY DEFINED BY
5153 components may be affected. Certificate parsing (d2i_X509 and related
5154 functions) are however not affected. OpenSSL clients and servers are
5160 * PKCS7 NULL pointer dereferences fix
5162 The PKCS#7 parsing code does not handle missing outer ContentInfo
5163 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5164 missing content and trigger a NULL pointer dereference on parsing.
5166 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5167 otherwise parse PKCS#7 structures from untrusted sources are
5168 affected. OpenSSL clients and servers are not affected.
5170 This issue was reported to OpenSSL by Michal Zalewski (Google).
5175 * DoS via reachable assert in SSLv2 servers fix
5177 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5178 servers that both support SSLv2 and enable export cipher suites by sending
5179 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5181 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5182 (OpenSSL development team).
5187 * Empty CKE with client auth and DHE fix
5189 If client auth is used then a server can seg fault in the event of a DHE
5190 ciphersuite being selected and a zero length ClientKeyExchange message
5191 being sent by the client. This could be exploited in a DoS attack.
5196 * Handshake with unseeded PRNG fix
5198 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5199 with an unseeded PRNG. The conditions are:
5200 - The client is on a platform where the PRNG has not been seeded
5201 automatically, and the user has not seeded manually
5202 - A protocol specific client method version has been used (i.e. not
5203 SSL_client_methodv23)
5204 - A ciphersuite is used that does not require additional random data from
5205 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5207 If the handshake succeeds then the client random that has been used will
5208 have been generated from a PRNG with insufficient entropy and therefore the
5209 output may be predictable.
5211 For example using the following command with an unseeded openssl will
5212 succeed on an unpatched platform:
5214 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5219 * Use After Free following d2i_ECPrivatekey error fix
5221 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5222 could cause a use after free condition. This, in turn, could cause a double
5223 free in several private key parsing functions (such as d2i_PrivateKey
5224 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5225 for applications that receive EC private keys from untrusted
5226 sources. This scenario is considered rare.
5228 This issue was discovered by the BoringSSL project and fixed in their
5234 * X509_to_X509_REQ NULL pointer deref fix
5236 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5237 the certificate key is invalid. This function is rarely used in practice.
5239 This issue was discovered by Brian Carpenter.
5244 * Removed the export ciphers from the DEFAULT ciphers
5248 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5250 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5251 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5252 So far those who have to target multiple platforms would compromise
5253 and argue that binary targeting say ARMv5 would still execute on
5254 ARMv8. "Universal" build resolves this compromise by providing
5255 near-optimal performance even on newer platforms.
5259 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5260 (other platforms pending).
5262 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5264 * Add support for the SignedCertificateTimestampList certificate and
5265 OCSP response extensions from RFC6962.
5269 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5270 for corner cases. (Certain input points at infinity could lead to
5271 bogus results, with non-infinity inputs mapped to infinity too.)
5275 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5276 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5277 common cases are optimized and there still is room for further
5278 improvements. Vector Permutation AES for Altivec is also added.
5282 * Add support for little-endian ppc64 Linux target.
5284 *Marcelo Cerri (IBM)*
5286 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5287 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5288 are optimized and there still is room for further improvements.
5289 Both 32- and 64-bit modes are supported.
5291 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5293 * Improved ARMv7 NEON support.
5297 * Support for SPARC Architecture 2011 crypto extensions, first
5298 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5299 SHA256/512, MD5, GHASH and modular exponentiation.
5301 *Andy Polyakov, David Miller*
5303 * Accelerated modular exponentiation for Intel processors, a.k.a.
5306 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5308 * Support for new and upcoming Intel processors, including AVX2,
5309 BMI and SHA ISA extensions. This includes additional "stitched"
5310 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5313 This work was sponsored by Intel Corp.
5317 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5318 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5319 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5323 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5324 this fixes a limitation in previous versions of OpenSSL.
5328 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5329 MGF1 digest and OAEP label.
5333 * Add EVP support for key wrapping algorithms, to avoid problems with
5334 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5335 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5336 algorithms and include tests cases.
5340 * Add functions to allocate and set the fields of an ECDSA_METHOD
5343 *Douglas E. Engert, Steve Henson*
5345 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5346 difference in days and seconds between two tm or ASN1_TIME structures.
5350 * Add -rev test option to s_server to just reverse order of characters
5351 received by client and send back to server. Also prints an abbreviated
5352 summary of the connection parameters.
5356 * New option -brief for s_client and s_server to print out a brief summary
5357 of connection parameters.
5361 * Add callbacks for arbitrary TLS extensions.
5363 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5365 * New option -crl_download in several openssl utilities to download CRLs
5366 from CRLDP extension in certificates.
5370 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5374 * New function X509_CRL_diff to generate a delta CRL from the difference
5375 of two full CRLs. Add support to "crl" utility.
5379 * New functions to set lookup_crls function and to retrieve
5380 X509_STORE from X509_STORE_CTX.
5384 * Print out deprecated issuer and subject unique ID fields in
5389 * Extend OCSP I/O functions so they can be used for simple general purpose
5390 HTTP as well as OCSP. New wrapper function which can be used to download
5391 CRLs using the OCSP API.
5395 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5399 * `SSL_CONF*` functions. These provide a common framework for application
5400 configuration using configuration files or command lines.
5404 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5405 message callback and prints the results. Needs compile time option
5406 "enable-ssl-trace". New options to s_client and s_server to enable
5411 * New ctrl and macro to retrieve supported points extensions.
5412 Print out extension in s_server and s_client.
5416 * New functions to retrieve certificate signature and signature
5421 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5426 * New Suite B modes for TLS code. These use and enforce the requirements
5427 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5428 only use Suite B curves. The Suite B modes can be set by using the
5429 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5433 * New chain verification flags for Suite B levels of security. Check
5434 algorithms are acceptable when flags are set in X509_verify_cert.
5438 * Make tls1_check_chain return a set of flags indicating checks passed
5439 by a certificate chain. Add additional tests to handle client
5440 certificates: checks for matching certificate type and issuer name
5445 * If an attempt is made to use a signature algorithm not in the peer
5446 preference list abort the handshake. If client has no suitable
5447 signature algorithms in response to a certificate request do not
5448 use the certificate.
5452 * If server EC tmp key is not in client preference list abort handshake.
5456 * Add support for certificate stores in CERT structure. This makes it
5457 possible to have different stores per SSL structure or one store in
5458 the parent SSL_CTX. Include distinct stores for certificate chain
5459 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5460 to build and store a certificate chain in CERT structure: returning
5461 an error if the chain cannot be built: this will allow applications
5462 to test if a chain is correctly configured.
5464 Note: if the CERT based stores are not set then the parent SSL_CTX
5465 store is used to retain compatibility with existing behaviour.
5469 * New function ssl_set_client_disabled to set a ciphersuite disabled
5470 mask based on the current session, check mask when sending client
5471 hello and checking the requested ciphersuite.
5475 * New ctrls to retrieve and set certificate types in a certificate
5476 request message. Print out received values in s_client. If certificate
5477 types is not set with custom values set sensible values based on
5478 supported signature algorithms.
5482 * Support for distinct client and server supported signature algorithms.
5486 * Add certificate callback. If set this is called whenever a certificate
5487 is required by client or server. An application can decide which
5488 certificate chain to present based on arbitrary criteria: for example
5489 supported signature algorithms. Add very simple example to s_server.
5490 This fixes many of the problems and restrictions of the existing client
5491 certificate callback: for example you can now clear an existing
5492 certificate and specify the whole chain.
5496 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5497 the certificate can be used for (if anything). Set valid_flags field
5498 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5499 to have similar checks in it.
5501 Add new "cert_flags" field to CERT structure and include a "strict mode".
5502 This enforces some TLS certificate requirements (such as only permitting
5503 certificate signature algorithms contained in the supported algorithms
5504 extension) which some implementations ignore: this option should be used
5505 with caution as it could cause interoperability issues.
5509 * Update and tidy signature algorithm extension processing. Work out
5510 shared signature algorithms based on preferences and peer algorithms
5511 and print them out in s_client and s_server. Abort handshake if no
5512 shared signature algorithms.
5516 * Add new functions to allow customised supported signature algorithms
5517 for SSL and SSL_CTX structures. Add options to s_client and s_server
5522 * New function SSL_certs_clear() to delete all references to certificates
5523 from an SSL structure. Before this once a certificate had been added
5524 it couldn't be removed.
5528 * Integrate hostname, email address and IP address checking with certificate
5529 verification. New verify options supporting checking in openssl utility.
5533 * Fixes and wildcard matching support to hostname and email checking
5534 functions. Add manual page.
5536 *Florian Weimer (Red Hat Product Security Team)*
5538 * New functions to check a hostname email or IP address against a
5539 certificate. Add options x509 utility to print results of checks against
5544 * Fix OCSP checking.
5546 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5548 * Initial experimental support for explicitly trusted non-root CAs.
5549 OpenSSL still tries to build a complete chain to a root but if an
5550 intermediate CA has a trust setting included that is used. The first
5551 setting is used: whether to trust (e.g., -addtrust option to the x509
5556 * Add -trusted_first option which attempts to find certificates in the
5557 trusted store even if an untrusted chain is also supplied.
5561 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5562 platform support for Linux and Android.
5566 * Support for linux-x32, ILP32 environment in x86_64 framework.
5570 * Experimental multi-implementation support for FIPS capable OpenSSL.
5571 When in FIPS mode the approved implementations are used as normal,
5572 when not in FIPS mode the internal unapproved versions are used instead.
5573 This means that the FIPS capable OpenSSL isn't forced to use the
5574 (often lower performance) FIPS implementations outside FIPS mode.
5578 * Transparently support X9.42 DH parameters when calling
5579 PEM_read_bio_DHparameters. This means existing applications can handle
5580 the new parameter format automatically.
5584 * Initial experimental support for X9.42 DH parameter format: mainly
5585 to support use of 'q' parameter for RFC5114 parameters.
5589 * Add DH parameters from RFC5114 including test data to dhtest.
5593 * Support for automatic EC temporary key parameter selection. If enabled
5594 the most preferred EC parameters are automatically used instead of
5595 hardcoded fixed parameters. Now a server just has to call:
5596 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5597 support ECDH and use the most appropriate parameters.
5601 * Enhance and tidy EC curve and point format TLS extension code. Use
5602 static structures instead of allocation if default values are used.
5603 New ctrls to set curves we wish to support and to retrieve shared curves.
5604 Print out shared curves in s_server. New options to s_server and s_client
5605 to set list of supported curves.
5609 * New ctrls to retrieve supported signature algorithms and
5610 supported curve values as an array of NIDs. Extend openssl utility
5611 to print out received values.
5615 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5616 between NIDs and the more common NIST names such as "P-256". Enhance
5617 ecparam utility and ECC method to recognise the NIST names for curves.
5621 * Enhance SSL/TLS certificate chain handling to support different
5622 chains for each certificate instead of one chain in the parent SSL_CTX.
5626 * Support for fixed DH ciphersuite client authentication: where both
5627 server and client use DH certificates with common parameters.
5631 * Support for fixed DH ciphersuites: those requiring DH server
5636 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5638 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5639 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5640 X509_CINF_get_signature were reverted post internal team review.
5645 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5647 * OCSP Status Request extension unbounded memory growth
5649 A malicious client can send an excessively large OCSP Status Request
5650 extension. If that client continually requests renegotiation, sending a
5651 large OCSP Status Request extension each time, then there will be unbounded
5652 memory growth on the server. This will eventually lead to a Denial Of
5653 Service attack through memory exhaustion. Servers with a default
5654 configuration are vulnerable even if they do not support OCSP. Builds using
5655 the "no-ocsp" build time option are not affected.
5657 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5662 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5665 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5671 * OOB write in MDC2_Update()
5673 An overflow can occur in MDC2_Update() either if called directly or
5674 through the EVP_DigestUpdate() function using MDC2. If an attacker
5675 is able to supply very large amounts of input data after a previous
5676 call to EVP_EncryptUpdate() with a partial block then a length check
5677 can overflow resulting in a heap corruption.
5679 The amount of data needed is comparable to SIZE_MAX which is impractical
5682 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5687 * Malformed SHA512 ticket DoS
5689 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5690 DoS attack where a malformed ticket will result in an OOB read which will
5693 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5694 a custom server callback and ticket lookup mechanism.
5696 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5701 * OOB write in BN_bn2dec()
5703 The function BN_bn2dec() does not check the return value of BN_div_word().
5704 This can cause an OOB write if an application uses this function with an
5705 overly large BIGNUM. This could be a problem if an overly large certificate
5706 or CRL is printed out from an untrusted source. TLS is not affected because
5707 record limits will reject an oversized certificate before it is parsed.
5709 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5714 * OOB read in TS_OBJ_print_bio()
5716 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5717 the total length the OID text representation would use and not the amount
5718 of data written. This will result in OOB reads when large OIDs are
5721 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5726 * Pointer arithmetic undefined behaviour
5728 Avoid some undefined pointer arithmetic
5730 A common idiom in the codebase is to check limits in the following manner:
5733 Where "p" points to some malloc'd data of SIZE bytes and
5736 "len" here could be from some externally supplied data (e.g. from a TLS
5739 The rules of C pointer arithmetic are such that "p + len" is only well
5740 defined where len <= SIZE. Therefore the above idiom is actually
5741 undefined behaviour.
5743 For example this could cause problems if some malloc implementation
5744 provides an address for "p" such that "p + len" actually overflows for
5745 values of len that are too big and therefore p + len < limit.
5747 This issue was reported to OpenSSL by Guido Vranken
5752 * Constant time flag not preserved in DSA signing
5754 Operations in the DSA signing algorithm should run in constant time in
5755 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5756 implementation means that a non-constant time codepath is followed for
5757 certain operations. This has been demonstrated through a cache-timing
5758 attack to be sufficient for an attacker to recover the private DSA key.
5760 This issue was reported by César Pereida (Aalto University), Billy Brumley
5761 (Tampere University of Technology), and Yuval Yarom (The University of
5762 Adelaide and NICTA).
5767 * DTLS buffered message DoS
5769 In a DTLS connection where handshake messages are delivered out-of-order
5770 those messages that OpenSSL is not yet ready to process will be buffered
5771 for later use. Under certain circumstances, a flaw in the logic means that
5772 those messages do not get removed from the buffer even though the handshake
5773 has been completed. An attacker could force up to approx. 15 messages to
5774 remain in the buffer when they are no longer required. These messages will
5775 be cleared when the DTLS connection is closed. The default maximum size for
5776 a message is 100k. Therefore the attacker could force an additional 1500k
5777 to be consumed per connection. By opening many simulataneous connections an
5778 attacker could cause a DoS attack through memory exhaustion.
5780 This issue was reported to OpenSSL by Quan Luo.
5785 * DTLS replay protection DoS
5787 A flaw in the DTLS replay attack protection mechanism means that records
5788 that arrive for future epochs update the replay protection "window" before
5789 the MAC for the record has been validated. This could be exploited by an
5790 attacker by sending a record for the next epoch (which does not have to
5791 decrypt or have a valid MAC), with a very large sequence number. This means
5792 that all subsequent legitimate packets are dropped causing a denial of
5793 service for a specific DTLS connection.
5795 This issue was reported to OpenSSL by the OCAP audit team.
5800 * Certificate message OOB reads
5802 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5803 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5804 theoretical DoS risk but this has not been observed in practice on common
5807 The messages affected are client certificate, client certificate request
5808 and server certificate. As a result the attack can only be performed
5809 against a client or a server which enables client authentication.
5811 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5816 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5818 * Prevent padding oracle in AES-NI CBC MAC check
5820 A MITM attacker can use a padding oracle attack to decrypt traffic
5821 when the connection uses an AES CBC cipher and the server support
5824 This issue was introduced as part of the fix for Lucky 13 padding
5825 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5826 constant time by making sure that always the same bytes are read and
5827 compared against either the MAC or padding bytes. But it no longer
5828 checked that there was enough data to have both the MAC and padding
5831 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5836 * Fix EVP_EncodeUpdate overflow
5838 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5839 Base64 encoding of binary data. If an attacker is able to supply very large
5840 amounts of input data then a length check can overflow resulting in a heap
5843 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5844 the `PEM_write_bio*` family of functions. These are mainly used within the
5845 OpenSSL command line applications, so any application which processes data
5846 from an untrusted source and outputs it as a PEM file should be considered
5847 vulnerable to this issue. User applications that call these APIs directly
5848 with large amounts of untrusted data may also be vulnerable.
5850 This issue was reported by Guido Vranken.
5855 * Fix EVP_EncryptUpdate overflow
5857 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5858 is able to supply very large amounts of input data after a previous call to
5859 EVP_EncryptUpdate() with a partial block then a length check can overflow
5860 resulting in a heap corruption. Following an analysis of all OpenSSL
5861 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5862 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5863 the first called function after an EVP_EncryptInit(), and therefore that
5864 specific call must be safe. The second form is where the length passed to
5865 EVP_EncryptUpdate() can be seen from the code to be some small value and
5866 therefore there is no possibility of an overflow. Since all instances are
5867 one of these two forms, it is believed that there can be no overflows in
5868 internal code due to this problem. It should be noted that
5869 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5870 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5871 of these calls have also been analysed too and it is believed there are no
5872 instances in internal usage where an overflow could occur.
5874 This issue was reported by Guido Vranken.
5879 * Prevent ASN.1 BIO excessive memory allocation
5881 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5882 a short invalid encoding can casuse allocation of large amounts of memory
5883 potentially consuming excessive resources or exhausting memory.
5885 Any application parsing untrusted data through d2i BIO functions is
5886 affected. The memory based functions such as d2i_X509() are *not* affected.
5887 Since the memory based functions are used by the TLS library, TLS
5888 applications are not affected.
5890 This issue was reported by Brian Carpenter.
5897 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5898 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5899 in arbitrary stack data being returned in the buffer.
5901 This issue was reported by Guido Vranken.
5906 * Modify behavior of ALPN to invoke callback after SNI/servername
5907 callback, such that updates to the SSL_CTX affect ALPN.
5911 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5916 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5917 methods are enabled and ssl2 is disabled the methods return NULL.
5921 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5923 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5924 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5925 provide any "EXPORT" or "LOW" strength ciphers.
5929 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5930 is by default disabled at build-time. Builds that are not configured with
5931 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5932 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5933 will need to explicitly call either of:
5935 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5937 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5939 as appropriate. Even if either of those is used, or the application
5940 explicitly uses the version-specific SSLv2_method() or its client and
5941 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5942 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5943 ciphers, and SSLv2 56-bit DES are no longer available.
5948 * Fix a double-free in DSA code
5950 A double free bug was discovered when OpenSSL parses malformed DSA private
5951 keys and could lead to a DoS attack or memory corruption for applications
5952 that receive DSA private keys from untrusted sources. This scenario is
5955 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5961 * Disable SRP fake user seed to address a server memory leak.
5963 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5965 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5966 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5967 was changed to ignore the "fake user" SRP seed, even if the seed
5970 Users should use SRP_VBASE_get1_by_user instead. Note that in
5971 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5972 also that even though configuring the SRP seed attempts to hide
5973 invalid usernames by continuing the handshake with fake
5974 credentials, this behaviour is not constant time and no strong
5975 guarantees are made that the handshake is indistinguishable from
5976 that of a valid user.
5981 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5983 In the BN_hex2bn function the number of hex digits is calculated using an
5984 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5985 large values of `i` this can result in `bn_expand` not allocating any
5986 memory because `i * 4` is negative. This can leave the internal BIGNUM data
5987 field as NULL leading to a subsequent NULL ptr deref. For very large values
5988 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
5989 In this case memory is allocated to the internal BIGNUM data field, but it
5990 is insufficiently sized leading to heap corruption. A similar issue exists
5991 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
5992 is ever called by user applications with very large untrusted hex/dec data.
5993 This is anticipated to be a rare occurrence.
5995 All OpenSSL internal usage of these functions use data that is not expected
5996 to be untrusted, e.g. config file data or application command line
5997 arguments. If user developed applications generate config file data based
5998 on untrusted data then it is possible that this could also lead to security
5999 consequences. This is also anticipated to be rare.
6001 This issue was reported to OpenSSL by Guido Vranken.
6006 * Fix memory issues in `BIO_*printf` functions
6008 The internal `fmtstr` function used in processing a "%s" format string in
6009 the `BIO_*printf` functions could overflow while calculating the length of a
6010 string and cause an OOB read when printing very long strings.
6012 Additionally the internal `doapr_outch` function can attempt to write to an
6013 OOB memory location (at an offset from the NULL pointer) in the event of a
6014 memory allocation failure. In 1.0.2 and below this could be caused where
6015 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6016 could be in processing a very long "%s" format string. Memory leaks can
6019 The first issue may mask the second issue dependent on compiler behaviour.
6020 These problems could enable attacks where large amounts of untrusted data
6021 is passed to the `BIO_*printf` functions. If applications use these functions
6022 in this way then they could be vulnerable. OpenSSL itself uses these
6023 functions when printing out human-readable dumps of ASN.1 data. Therefore
6024 applications that print this data could be vulnerable if the data is from
6025 untrusted sources. OpenSSL command line applications could also be
6026 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6027 as command line arguments.
6029 Libssl is not considered directly vulnerable. Additionally certificates etc
6030 received via remote connections via libssl are also unlikely to be able to
6031 trigger these issues because of message size limits enforced within libssl.
6033 This issue was reported to OpenSSL Guido Vranken.
6038 * Side channel attack on modular exponentiation
6040 A side-channel attack was found which makes use of cache-bank conflicts on
6041 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6042 of RSA keys. The ability to exploit this issue is limited as it relies on
6043 an attacker who has control of code in a thread running on the same
6044 hyper-threaded core as the victim thread which is performing decryptions.
6046 This issue was reported to OpenSSL by Yuval Yarom, The University of
6047 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6048 Nadia Heninger, University of Pennsylvania with more information at
6049 <http://cachebleed.info>.
6054 * Change the req app to generate a 2048-bit RSA/DSA key by default,
6055 if no keysize is specified with default_bits. This fixes an
6056 omission in an earlier change that changed all RSA/DSA key generation
6057 apps to use 2048 bits by default.
6061 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6063 * Protection for DH small subgroup attacks
6065 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6066 switched on by default and cannot be disabled. This could have some
6071 * SSLv2 doesn't block disabled ciphers
6073 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6074 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6075 been disabled, provided that the SSLv2 protocol was not also disabled via
6078 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6079 and Sebastian Schinzel.
6084 * Reject DH handshakes with parameters shorter than 1024 bits.
6088 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6090 * Certificate verify crash with missing PSS parameter
6092 The signature verification routines will crash with a NULL pointer
6093 dereference if presented with an ASN.1 signature using the RSA PSS
6094 algorithm and absent mask generation function parameter. Since these
6095 routines are used to verify certificate signature algorithms this can be
6096 used to crash any certificate verification operation and exploited in a
6097 DoS attack. Any application which performs certificate verification is
6098 vulnerable including OpenSSL clients and servers which enable client
6101 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6106 * X509_ATTRIBUTE memory leak
6108 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6109 memory. This structure is used by the PKCS#7 and CMS routines so any
6110 application which reads PKCS#7 or CMS data from untrusted sources is
6111 affected. SSL/TLS is not affected.
6113 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6119 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6120 This changes the decoding behaviour for some invalid messages,
6121 though the change is mostly in the more lenient direction, and
6122 legacy behaviour is preserved as much as possible.
6126 * In DSA_generate_parameters_ex, if the provided seed is too short,
6127 use a random seed, as already documented.
6129 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6131 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6133 * Alternate chains certificate forgery
6135 During certificate verfification, OpenSSL will attempt to find an
6136 alternative certificate chain if the first attempt to build such a chain
6137 fails. An error in the implementation of this logic can mean that an
6138 attacker could cause certain checks on untrusted certificates to be
6139 bypassed, such as the CA flag, enabling them to use a valid leaf
6140 certificate to act as a CA and "issue" an invalid certificate.
6142 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6148 * Race condition handling PSK identify hint
6150 If PSK identity hints are received by a multi-threaded client then
6151 the values are wrongly updated in the parent SSL_CTX structure. This can
6152 result in a race condition potentially leading to a double free of the
6158 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6160 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6161 incompatibility in the handling of HMAC. The previous ABI has now been
6164 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6166 * Malformed ECParameters causes infinite loop
6168 When processing an ECParameters structure OpenSSL enters an infinite loop
6169 if the curve specified is over a specially malformed binary polynomial
6172 This can be used to perform denial of service against any
6173 system which processes public keys, certificate requests or
6174 certificates. This includes TLS clients and TLS servers with
6175 client authentication enabled.
6177 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6182 * Exploitable out-of-bounds read in X509_cmp_time
6184 X509_cmp_time does not properly check the length of the ASN1_TIME
6185 string and can read a few bytes out of bounds. In addition,
6186 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6189 An attacker can use this to craft malformed certificates and CRLs of
6190 various sizes and potentially cause a segmentation fault, resulting in
6191 a DoS on applications that verify certificates or CRLs. TLS clients
6192 that verify CRLs are affected. TLS clients and servers with client
6193 authentication enabled may be affected if they use custom verification
6196 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6197 independently by Hanno Böck.
6202 * PKCS7 crash with missing EnvelopedContent
6204 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6205 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6206 with missing content and trigger a NULL pointer dereference on parsing.
6208 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6209 structures from untrusted sources are affected. OpenSSL clients and
6210 servers are not affected.
6212 This issue was reported to OpenSSL by Michal Zalewski (Google).
6217 * CMS verify infinite loop with unknown hash function
6219 When verifying a signedData message the CMS code can enter an infinite loop
6220 if presented with an unknown hash function OID. This can be used to perform
6221 denial of service against any system which verifies signedData messages using
6223 This issue was reported to OpenSSL by Johannes Bauer.
6228 * Race condition handling NewSessionTicket
6230 If a NewSessionTicket is received by a multi-threaded client when attempting to
6231 reuse a previous ticket then a race condition can occur potentially leading to
6232 a double free of the ticket data.
6237 * Reject DH handshakes with parameters shorter than 768 bits.
6239 *Kurt Roeckx and Emilia Kasper*
6241 * dhparam: generate 2048-bit parameters by default.
6243 *Kurt Roeckx and Emilia Kasper*
6245 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6247 * Segmentation fault in ASN1_TYPE_cmp fix
6249 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6250 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6251 certificate signature algorithm consistency this can be used to crash any
6252 certificate verification operation and exploited in a DoS attack. Any
6253 application which performs certificate verification is vulnerable including
6254 OpenSSL clients and servers which enable client authentication.
6259 * ASN.1 structure reuse memory corruption fix
6261 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6262 memory corruption via an invalid write. Such reuse is and has been
6263 strongly discouraged and is believed to be rare.
6265 Applications that parse structures containing CHOICE or ANY DEFINED BY
6266 components may be affected. Certificate parsing (d2i_X509 and related
6267 functions) are however not affected. OpenSSL clients and servers are
6273 * PKCS7 NULL pointer dereferences fix
6275 The PKCS#7 parsing code does not handle missing outer ContentInfo
6276 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6277 missing content and trigger a NULL pointer dereference on parsing.
6279 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6280 otherwise parse PKCS#7 structures from untrusted sources are
6281 affected. OpenSSL clients and servers are not affected.
6283 This issue was reported to OpenSSL by Michal Zalewski (Google).
6288 * DoS via reachable assert in SSLv2 servers fix
6290 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6291 servers that both support SSLv2 and enable export cipher suites by sending
6292 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6294 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6295 (OpenSSL development team).
6300 * Use After Free following d2i_ECPrivatekey error fix
6302 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6303 could cause a use after free condition. This, in turn, could cause a double
6304 free in several private key parsing functions (such as d2i_PrivateKey
6305 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6306 for applications that receive EC private keys from untrusted
6307 sources. This scenario is considered rare.
6309 This issue was discovered by the BoringSSL project and fixed in their
6315 * X509_to_X509_REQ NULL pointer deref fix
6317 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6318 the certificate key is invalid. This function is rarely used in practice.
6320 This issue was discovered by Brian Carpenter.
6325 * Removed the export ciphers from the DEFAULT ciphers
6329 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6331 * Build fixes for the Windows and OpenVMS platforms
6333 *Matt Caswell and Richard Levitte*
6335 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6337 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6338 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6339 dereference. This could lead to a Denial Of Service attack. Thanks to
6340 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6345 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6346 dtls1_buffer_record function under certain conditions. In particular this
6347 could occur if an attacker sent repeated DTLS records with the same
6348 sequence number but for the next epoch. The memory leak could be exploited
6349 by an attacker in a Denial of Service attack through memory exhaustion.
6350 Thanks to Chris Mueller for reporting this issue.
6355 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6356 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6357 method would be set to NULL which could later result in a NULL pointer
6358 dereference. Thanks to Frank Schmirler for reporting this issue.
6363 * Abort handshake if server key exchange message is omitted for ephemeral
6366 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6367 reporting this issue.
6372 * Remove non-export ephemeral RSA code on client and server. This code
6373 violated the TLS standard by allowing the use of temporary RSA keys in
6374 non-export ciphersuites and could be used by a server to effectively
6375 downgrade the RSA key length used to a value smaller than the server
6376 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6377 INRIA or reporting this issue.
6382 * Fixed issue where DH client certificates are accepted without verification.
6383 An OpenSSL server will accept a DH certificate for client authentication
6384 without the certificate verify message. This effectively allows a client to
6385 authenticate without the use of a private key. This only affects servers
6386 which trust a client certificate authority which issues certificates
6387 containing DH keys: these are extremely rare and hardly ever encountered.
6388 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6394 * Ensure that the session ID context of an SSL is updated when its
6395 SSL_CTX is updated via SSL_set_SSL_CTX.
6397 The session ID context is typically set from the parent SSL_CTX,
6398 and can vary with the CTX.
6402 * Fix various certificate fingerprint issues.
6404 By using non-DER or invalid encodings outside the signed portion of a
6405 certificate the fingerprint can be changed without breaking the signature.
6406 Although no details of the signed portion of the certificate can be changed
6407 this can cause problems with some applications: e.g. those using the
6408 certificate fingerprint for blacklists.
6410 1. Reject signatures with non zero unused bits.
6412 If the BIT STRING containing the signature has non zero unused bits reject
6413 the signature. All current signature algorithms require zero unused bits.
6415 2. Check certificate algorithm consistency.
6417 Check the AlgorithmIdentifier inside TBS matches the one in the
6418 certificate signature. NB: this will result in signature failure
6419 errors for some broken certificates.
6421 Thanks to Konrad Kraszewski from Google for reporting this issue.
6423 3. Check DSA/ECDSA signatures use DER.
6425 Re-encode DSA/ECDSA signatures and compare with the original received
6426 signature. Return an error if there is a mismatch.
6428 This will reject various cases including garbage after signature
6429 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6430 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6431 (negative or with leading zeroes).
6433 Further analysis was conducted and fixes were developed by Stephen Henson
6434 of the OpenSSL core team.
6440 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6441 results on some platforms, including x86_64. This bug occurs at random
6442 with a very low probability, and is not known to be exploitable in any
6443 way, though its exact impact is difficult to determine. Thanks to Pieter
6444 Wuille (Blockstream) who reported this issue and also suggested an initial
6445 fix. Further analysis was conducted by the OpenSSL development team and
6446 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6447 the OpenSSL core team.
6452 * Do not resume sessions on the server if the negotiated protocol
6453 version does not match the session's version. Resuming with a different
6454 version, while not strictly forbidden by the RFC, is of questionable
6455 sanity and breaks all known clients.
6457 *David Benjamin, Emilia Käsper*
6459 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6460 early CCS messages during renegotiation. (Note that because
6461 renegotiation is encrypted, this early CCS was not exploitable.)
6465 * Tighten client-side session ticket handling during renegotiation:
6466 ensure that the client only accepts a session ticket if the server sends
6467 the extension anew in the ServerHello. Previously, a TLS client would
6468 reuse the old extension state and thus accept a session ticket if one was
6469 announced in the initial ServerHello.
6471 Similarly, ensure that the client requires a session ticket if one
6472 was advertised in the ServerHello. Previously, a TLS client would
6473 ignore a missing NewSessionTicket message.
6477 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6481 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6482 sends a carefully crafted handshake message, to cause OpenSSL to fail
6483 to free up to 64k of memory causing a memory leak. This could be
6484 exploited in a Denial Of Service attack. This issue affects OpenSSL
6485 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6486 whether SRTP is used or configured. Implementations of OpenSSL that
6487 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6489 The fix was developed by the OpenSSL team.
6494 * Session Ticket Memory Leak.
6496 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6497 integrity of that ticket is first verified. In the event of a session
6498 ticket integrity check failing, OpenSSL will fail to free memory
6499 causing a memory leak. By sending a large number of invalid session
6500 tickets an attacker could exploit this issue in a Denial Of Service
6506 * Build option no-ssl3 is incomplete.
6508 When OpenSSL is configured with "no-ssl3" as a build option, servers
6509 could accept and complete a SSL 3.0 handshake, and clients could be
6510 configured to send them.
6513 *Akamai and the OpenSSL team*
6515 * Add support for TLS_FALLBACK_SCSV.
6516 Client applications doing fallback retries should call
6517 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6520 *Adam Langley, Bodo Moeller*
6522 * Add additional DigestInfo checks.
6524 Re-encode DigestInto in DER and check against the original when
6525 verifying RSA signature: this will reject any improperly encoded
6526 DigestInfo structures.
6528 Note: this is a precautionary measure and no attacks are currently known.
6532 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6534 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6535 SRP code can be overrun an internal buffer. Add sanity check that
6536 g, A, B < N to SRP code.
6538 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6539 Group for discovering this issue.
6544 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6545 TLS 1.0 instead of higher protocol versions when the ClientHello message
6546 is badly fragmented. This allows a man-in-the-middle attacker to force a
6547 downgrade to TLS 1.0 even if both the server and the client support a
6548 higher protocol version, by modifying the client's TLS records.
6550 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6551 researching this issue.
6556 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6557 to a denial of service attack. A malicious server can crash the client
6558 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6559 ciphersuite and sending carefully crafted handshake messages.
6561 Thanks to Felix Gröbert (Google) for discovering and researching this
6567 * By sending carefully crafted DTLS packets an attacker could cause openssl
6568 to leak memory. This can be exploited through a Denial of Service attack.
6569 Thanks to Adam Langley for discovering and researching this issue.
6574 * An attacker can force openssl to consume large amounts of memory whilst
6575 processing DTLS handshake messages. This can be exploited through a
6576 Denial of Service attack.
6577 Thanks to Adam Langley for discovering and researching this issue.
6582 * An attacker can force an error condition which causes openssl to crash
6583 whilst processing DTLS packets due to memory being freed twice. This
6584 can be exploited through a Denial of Service attack.
6585 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6591 * If a multithreaded client connects to a malicious server using a resumed
6592 session and the server sends an ec point format extension it could write
6593 up to 255 bytes to freed memory.
6595 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6601 * A malicious server can crash an OpenSSL client with a null pointer
6602 dereference (read) by specifying an SRP ciphersuite even though it was not
6603 properly negotiated with the client. This can be exploited through a
6604 Denial of Service attack.
6606 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6607 discovering and researching this issue.
6612 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6613 X509_name_oneline, X509_name_print_ex et al. to leak some information
6614 from the stack. Applications may be affected if they echo pretty printing
6615 output to the attacker.
6617 Thanks to Ivan Fratric (Google) for discovering this issue.
6620 *Emilia Käsper, and Steve Henson*
6622 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6623 for corner cases. (Certain input points at infinity could lead to
6624 bogus results, with non-infinity inputs mapped to infinity too.)
6628 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6630 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6631 handshake can force the use of weak keying material in OpenSSL
6632 SSL/TLS clients and servers.
6634 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6635 researching this issue. ([CVE-2014-0224])
6637 *KIKUCHI Masashi, Steve Henson*
6639 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6640 OpenSSL DTLS client the code can be made to recurse eventually crashing
6643 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6646 *Imre Rad, Steve Henson*
6648 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6649 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6650 client or server. This is potentially exploitable to run arbitrary
6651 code on a vulnerable client or server.
6653 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6655 *Jüri Aedla, Steve Henson*
6657 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6658 are subject to a denial of service attack.
6660 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6661 this issue. ([CVE-2014-3470])
6663 *Felix Gröbert, Ivan Fratric, Steve Henson*
6665 * Harmonize version and its documentation. -f flag is used to display
6668 *mancha <mancha1@zoho.com>*
6670 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6671 in i2d_ECPrivateKey.
6673 *mancha <mancha1@zoho.com>*
6675 * Fix some double frees. These are not thought to be exploitable.
6677 *mancha <mancha1@zoho.com>*
6679 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6681 * A missing bounds check in the handling of the TLS heartbeat extension
6682 can be used to reveal up to 64k of memory to a connected client or
6685 Thanks for Neel Mehta of Google Security for discovering this bug and to
6686 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6687 preparing the fix ([CVE-2014-0160])
6689 *Adam Langley, Bodo Moeller*
6691 * Fix for the attack described in the paper "Recovering OpenSSL
6692 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6693 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6694 <http://eprint.iacr.org/2014/140>
6696 Thanks to Yuval Yarom and Naomi Benger for discovering this
6697 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6699 *Yuval Yarom and Naomi Benger*
6701 * TLS pad extension: draft-agl-tls-padding-03
6703 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6704 TLS client Hello record length value would otherwise be > 255 and
6705 less that 512 pad with a dummy extension containing zeroes so it
6706 is at least 512 bytes long.
6708 *Adam Langley, Steve Henson*
6710 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6712 * Fix for TLS record tampering bug. A carefully crafted invalid
6713 handshake could crash OpenSSL with a NULL pointer exception.
6714 Thanks to Anton Johansson for reporting this issues.
6717 * Keep original DTLS digest and encryption contexts in retransmission
6718 structures so we can use the previous session parameters if they need
6719 to be resent. ([CVE-2013-6450])
6723 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6724 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6725 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6726 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6727 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6728 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6730 *Rob Stradling, Adam Langley*
6732 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6734 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6735 supporting platforms or when small records were transferred.
6737 *Andy Polyakov, Steve Henson*
6739 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6741 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6743 This addresses the flaw in CBC record processing discovered by
6744 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6745 at: <http://www.isg.rhul.ac.uk/tls/>
6747 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6748 Security Group at Royal Holloway, University of London
6749 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6750 Emilia Käsper for the initial patch.
6753 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6755 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6756 ciphersuites which can be exploited in a denial of service attack.
6757 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6758 and detecting this bug and to Wolfgang Ettlinger
6759 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6764 * Return an error when checking OCSP signatures when key is NULL.
6765 This fixes a DoS attack. ([CVE-2013-0166])
6769 * Make openssl verify return errors.
6771 *Chris Palmer <palmer@google.com> and Ben Laurie*
6773 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6774 the right response is stapled. Also change SSL_get_certificate()
6775 so it returns the certificate actually sent.
6776 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6778 *Rob Stradling <rob.stradling@comodo.com>*
6780 * Fix possible deadlock when decoding public keys.
6784 * Don't use TLS 1.0 record version number in initial client hello
6789 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6791 * Sanity check record length before skipping explicit IV in TLS
6792 1.2, 1.1 and DTLS to fix DoS attack.
6794 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6795 fuzzing as a service testing platform.
6800 * Initialise tkeylen properly when encrypting CMS messages.
6801 Thanks to Solar Designer of Openwall for reporting this issue.
6805 * In FIPS mode don't try to use composite ciphers as they are not
6810 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6812 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6813 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6814 mean any application compiled against OpenSSL 1.0.0 headers setting
6815 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6816 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6817 0x10000000L Any application which was previously compiled against
6818 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6819 will need to be recompiled as a result. Letting be results in
6820 inability to disable specifically TLS 1.1 and in client context,
6821 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6825 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6826 disable just protocol X, but all protocols above X *if* there are
6827 protocols *below* X still enabled. In more practical terms it means
6828 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6829 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6830 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6835 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6837 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6838 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6839 in CRYPTO_realloc_clean.
6841 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6842 issue and to Adam Langley <agl@chromium.org> for fixing it.
6845 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6847 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6851 * Workarounds for some broken servers that "hang" if a client hello
6852 record length exceeds 255 bytes.
6854 1. Do not use record version number > TLS 1.0 in initial client
6855 hello: some (but not all) hanging servers will now work.
6856 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6857 the number of ciphers sent in the client hello. This should be
6858 set to an even number, such as 50, for example by passing:
6859 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6860 Most broken servers should now work.
6861 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6862 TLS 1.2 client support entirely.
6866 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6870 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6872 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6873 STRING form instead of a DigestInfo.
6877 * The format used for MDC2 RSA signatures is inconsistent between EVP
6878 and the RSA_sign/RSA_verify functions. This was made more apparent when
6879 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6880 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6881 the correct format in RSA_verify so both forms transparently work.
6885 * Some servers which support TLS 1.0 can choke if we initially indicate
6886 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6887 encrypted premaster secret. As a workaround use the maximum permitted
6888 client version in client hello, this should keep such servers happy
6889 and still work with previous versions of OpenSSL.
6893 * Add support for TLS/DTLS heartbeats.
6895 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6897 * Add support for SCTP.
6899 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6901 * Improved PRNG seeding for VOS.
6903 *Paul Green <Paul.Green@stratus.com>*
6905 * Extensive assembler packs updates, most notably:
6907 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6908 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6909 - x86_64: bit-sliced AES implementation;
6910 - ARM: NEON support, contemporary platforms optimizations;
6911 - s390x: z196 support;
6912 - `*`: GHASH and GF(2^m) multiplication implementations;
6916 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6917 (removal of unnecessary code)
6919 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6921 * Add TLS key material exporter from RFC 5705.
6925 * Add DTLS-SRTP negotiation from RFC 5764.
6929 * Add Next Protocol Negotiation,
6930 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6931 disabled with a no-npn flag to config or Configure. Code donated
6934 *Adam Langley <agl@google.com> and Ben Laurie*
6936 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6937 NIST-P256, NIST-P521, with constant-time single point multiplication on
6938 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6939 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6940 Code made available under Apache License version 2.0.
6942 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6943 line to include this in your build of OpenSSL, and run "make depend" (or
6944 "make update"). This enables the following EC_METHODs:
6946 EC_GFp_nistp224_method()
6947 EC_GFp_nistp256_method()
6948 EC_GFp_nistp521_method()
6950 EC_GROUP_new_by_curve_name() will automatically use these (while
6951 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6954 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6956 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6957 all platforms. Move ssize_t definition from e_os.h to the public
6958 header file e_os2.h as it now appears in public header file cms.h
6962 * New -sigopt option to the ca, req and x509 utilities. Additional
6963 signature parameters can be passed using this option and in
6968 * Add RSA PSS signing function. This will generate and set the
6969 appropriate AlgorithmIdentifiers for PSS based on those in the
6970 corresponding EVP_MD_CTX structure. No application support yet.
6974 * Support for companion algorithm specific ASN1 signing routines.
6975 New function ASN1_item_sign_ctx() signs a pre-initialised
6976 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6977 the appropriate parameters.
6981 * Add new algorithm specific ASN1 verification initialisation function
6982 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6983 handling will be the same no matter what EVP_PKEY_METHOD is used.
6984 Add a PSS handler to support verification of PSS signatures: checked
6985 against a number of sample certificates.
6989 * Add signature printing for PSS. Add PSS OIDs.
6991 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
6993 * Add algorithm specific signature printing. An individual ASN1 method
6994 can now print out signatures instead of the standard hex dump.
6996 More complex signatures (e.g. PSS) can print out more meaningful
6997 information. Include DSA version that prints out the signature
7002 * Password based recipient info support for CMS library: implementing
7007 * Split password based encryption into PBES2 and PBKDF2 functions. This
7008 neatly separates the code into cipher and PBE sections and is required
7009 for some algorithms that split PBES2 into separate pieces (such as
7010 password based CMS).
7014 * Session-handling fixes:
7015 - Fix handling of connections that are resuming with a session ID,
7016 but also support Session Tickets.
7017 - Fix a bug that suppressed issuing of a new ticket if the client
7018 presented a ticket with an expired session.
7019 - Try to set the ticket lifetime hint to something reasonable.
7020 - Make tickets shorter by excluding irrelevant information.
7021 - On the client side, don't ignore renewed tickets.
7023 *Adam Langley, Bodo Moeller (Google)*
7025 * Fix PSK session representation.
7029 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7031 This work was sponsored by Intel.
7035 * Add GCM support to TLS library. Some custom code is needed to split
7036 the IV between the fixed (from PRF) and explicit (from TLS record)
7037 portions. This adds all GCM ciphersuites supported by RFC5288 and
7038 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7039 add a special AESGCM string for GCM only.
7043 * Expand range of ctrls for AES GCM. Permit setting invocation
7044 field on decrypt and retrieval of invocation field only on encrypt.
7048 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7049 As required by RFC5289 these ciphersuites cannot be used if for
7050 versions of TLS earlier than 1.2.
7054 * For FIPS capable OpenSSL interpret a NULL default public key method
7055 as unset and return the appropriate default but do *not* set the default.
7056 This means we can return the appropriate method in applications that
7057 switch between FIPS and non-FIPS modes.
7061 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7062 ENGINE is used then we cannot handle that in the FIPS module so we
7063 keep original code iff non-FIPS operations are allowed.
7067 * Add -attime option to openssl utilities.
7069 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7071 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7075 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7076 FIPS EC methods unconditionally for now.
7080 * New build option no-ec2m to disable characteristic 2 code.
7084 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7085 all cases can be covered as some introduce binary incompatibilities.
7089 * Redirect RSA operations to FIPS module including keygen,
7090 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7094 * Add similar low level API blocking to ciphers.
7098 * Low level digest APIs are not approved in FIPS mode: any attempt
7099 to use these will cause a fatal error. Applications that *really* want
7100 to use them can use the `private_*` version instead.
7104 * Redirect cipher operations to FIPS module for FIPS builds.
7108 * Redirect digest operations to FIPS module for FIPS builds.
7112 * Update build system to add "fips" flag which will link in fipscanister.o
7113 for static and shared library builds embedding a signature if needed.
7117 * Output TLS supported curves in preference order instead of numerical
7118 order. This is currently hardcoded for the highest order curves first.
7119 This should be configurable so applications can judge speed vs strength.
7123 * Add TLS v1.2 server support for client authentication.
7127 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7132 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7133 FIPS modules versions.
7137 * Add TLS v1.2 client side support for client authentication. Keep cache
7138 of handshake records longer as we don't know the hash algorithm to use
7139 until after the certificate request message is received.
7143 * Initial TLS v1.2 client support. Add a default signature algorithms
7144 extension including all the algorithms we support. Parse new signature
7145 format in client key exchange. Relax some ECC signing restrictions for
7146 TLS v1.2 as indicated in RFC5246.
7150 * Add server support for TLS v1.2 signature algorithms extension. Switch
7151 to new signature format when needed using client digest preference.
7152 All server ciphersuites should now work correctly in TLS v1.2. No client
7153 support yet and no support for client certificates.
7157 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7158 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7159 ciphersuites. At present only RSA key exchange ciphersuites work with
7160 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7161 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7162 and version checking.
7166 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7167 with this defined it will not be affected by any changes to ssl internal
7168 structures. Add several utility functions to allow openssl application
7169 to work with OPENSSL_NO_SSL_INTERN defined.
7173 * A long standing patch to add support for SRP from EdelWeb (Peter
7174 Sylvester and Christophe Renou) was integrated.
7175 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7176 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7179 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7183 * Permit abbreviated handshakes when renegotiating using the function
7184 SSL_renegotiate_abbreviated().
7186 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7188 * Add call to ENGINE_register_all_complete() to
7189 ENGINE_load_builtin_engines(), so some implementations get used
7190 automatically instead of needing explicit application support.
7194 * Add support for TLS key exporter as described in RFC5705.
7196 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7198 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7199 a few changes are required:
7201 Add SSL_OP_NO_TLSv1_1 flag.
7202 Add TLSv1_1 methods.
7203 Update version checking logic to handle version 1.1.
7204 Add explicit IV handling (ported from DTLS code).
7205 Add command line options to s_client/s_server.
7212 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7214 * X509_ATTRIBUTE memory leak
7216 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7217 memory. This structure is used by the PKCS#7 and CMS routines so any
7218 application which reads PKCS#7 or CMS data from untrusted sources is
7219 affected. SSL/TLS is not affected.
7221 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7227 * Race condition handling PSK identify hint
7229 If PSK identity hints are received by a multi-threaded client then
7230 the values are wrongly updated in the parent SSL_CTX structure. This can
7231 result in a race condition potentially leading to a double free of the
7237 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7239 * Malformed ECParameters causes infinite loop
7241 When processing an ECParameters structure OpenSSL enters an infinite loop
7242 if the curve specified is over a specially malformed binary polynomial
7245 This can be used to perform denial of service against any
7246 system which processes public keys, certificate requests or
7247 certificates. This includes TLS clients and TLS servers with
7248 client authentication enabled.
7250 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7255 * Exploitable out-of-bounds read in X509_cmp_time
7257 X509_cmp_time does not properly check the length of the ASN1_TIME
7258 string and can read a few bytes out of bounds. In addition,
7259 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7262 An attacker can use this to craft malformed certificates and CRLs of
7263 various sizes and potentially cause a segmentation fault, resulting in
7264 a DoS on applications that verify certificates or CRLs. TLS clients
7265 that verify CRLs are affected. TLS clients and servers with client
7266 authentication enabled may be affected if they use custom verification
7269 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7270 independently by Hanno Böck.
7275 * PKCS7 crash with missing EnvelopedContent
7277 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7278 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7279 with missing content and trigger a NULL pointer dereference on parsing.
7281 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7282 structures from untrusted sources are affected. OpenSSL clients and
7283 servers are not affected.
7285 This issue was reported to OpenSSL by Michal Zalewski (Google).
7290 * CMS verify infinite loop with unknown hash function
7292 When verifying a signedData message the CMS code can enter an infinite loop
7293 if presented with an unknown hash function OID. This can be used to perform
7294 denial of service against any system which verifies signedData messages using
7296 This issue was reported to OpenSSL by Johannes Bauer.
7301 * Race condition handling NewSessionTicket
7303 If a NewSessionTicket is received by a multi-threaded client when attempting to
7304 reuse a previous ticket then a race condition can occur potentially leading to
7305 a double free of the ticket data.
7310 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7312 * Segmentation fault in ASN1_TYPE_cmp fix
7314 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7315 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7316 certificate signature algorithm consistency this can be used to crash any
7317 certificate verification operation and exploited in a DoS attack. Any
7318 application which performs certificate verification is vulnerable including
7319 OpenSSL clients and servers which enable client authentication.
7324 * ASN.1 structure reuse memory corruption fix
7326 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7327 memory corruption via an invalid write. Such reuse is and has been
7328 strongly discouraged and is believed to be rare.
7330 Applications that parse structures containing CHOICE or ANY DEFINED BY
7331 components may be affected. Certificate parsing (d2i_X509 and related
7332 functions) are however not affected. OpenSSL clients and servers are
7338 * PKCS7 NULL pointer dereferences fix
7340 The PKCS#7 parsing code does not handle missing outer ContentInfo
7341 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7342 missing content and trigger a NULL pointer dereference on parsing.
7344 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7345 otherwise parse PKCS#7 structures from untrusted sources are
7346 affected. OpenSSL clients and servers are not affected.
7348 This issue was reported to OpenSSL by Michal Zalewski (Google).
7353 * DoS via reachable assert in SSLv2 servers fix
7355 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7356 servers that both support SSLv2 and enable export cipher suites by sending
7357 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7359 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7360 (OpenSSL development team).
7365 * Use After Free following d2i_ECPrivatekey error fix
7367 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7368 could cause a use after free condition. This, in turn, could cause a double
7369 free in several private key parsing functions (such as d2i_PrivateKey
7370 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7371 for applications that receive EC private keys from untrusted
7372 sources. This scenario is considered rare.
7374 This issue was discovered by the BoringSSL project and fixed in their
7380 * X509_to_X509_REQ NULL pointer deref fix
7382 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7383 the certificate key is invalid. This function is rarely used in practice.
7385 This issue was discovered by Brian Carpenter.
7390 * Removed the export ciphers from the DEFAULT ciphers
7394 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7396 * Build fixes for the Windows and OpenVMS platforms
7398 *Matt Caswell and Richard Levitte*
7400 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7402 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7403 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7404 dereference. This could lead to a Denial Of Service attack. Thanks to
7405 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7410 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7411 dtls1_buffer_record function under certain conditions. In particular this
7412 could occur if an attacker sent repeated DTLS records with the same
7413 sequence number but for the next epoch. The memory leak could be exploited
7414 by an attacker in a Denial of Service attack through memory exhaustion.
7415 Thanks to Chris Mueller for reporting this issue.
7420 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7421 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7422 method would be set to NULL which could later result in a NULL pointer
7423 dereference. Thanks to Frank Schmirler for reporting this issue.
7428 * Abort handshake if server key exchange message is omitted for ephemeral
7431 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7432 reporting this issue.
7437 * Remove non-export ephemeral RSA code on client and server. This code
7438 violated the TLS standard by allowing the use of temporary RSA keys in
7439 non-export ciphersuites and could be used by a server to effectively
7440 downgrade the RSA key length used to a value smaller than the server
7441 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7442 INRIA or reporting this issue.
7447 * Fixed issue where DH client certificates are accepted without verification.
7448 An OpenSSL server will accept a DH certificate for client authentication
7449 without the certificate verify message. This effectively allows a client to
7450 authenticate without the use of a private key. This only affects servers
7451 which trust a client certificate authority which issues certificates
7452 containing DH keys: these are extremely rare and hardly ever encountered.
7453 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7459 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7460 results on some platforms, including x86_64. This bug occurs at random
7461 with a very low probability, and is not known to be exploitable in any
7462 way, though its exact impact is difficult to determine. Thanks to Pieter
7463 Wuille (Blockstream) who reported this issue and also suggested an initial
7464 fix. Further analysis was conducted by the OpenSSL development team and
7465 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7466 the OpenSSL core team.
7471 * Fix various certificate fingerprint issues.
7473 By using non-DER or invalid encodings outside the signed portion of a
7474 certificate the fingerprint can be changed without breaking the signature.
7475 Although no details of the signed portion of the certificate can be changed
7476 this can cause problems with some applications: e.g. those using the
7477 certificate fingerprint for blacklists.
7479 1. Reject signatures with non zero unused bits.
7481 If the BIT STRING containing the signature has non zero unused bits reject
7482 the signature. All current signature algorithms require zero unused bits.
7484 2. Check certificate algorithm consistency.
7486 Check the AlgorithmIdentifier inside TBS matches the one in the
7487 certificate signature. NB: this will result in signature failure
7488 errors for some broken certificates.
7490 Thanks to Konrad Kraszewski from Google for reporting this issue.
7492 3. Check DSA/ECDSA signatures use DER.
7494 Reencode DSA/ECDSA signatures and compare with the original received
7495 signature. Return an error if there is a mismatch.
7497 This will reject various cases including garbage after signature
7498 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7499 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7500 (negative or with leading zeroes).
7502 Further analysis was conducted and fixes were developed by Stephen Henson
7503 of the OpenSSL core team.
7509 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7511 * Session Ticket Memory Leak.
7513 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7514 integrity of that ticket is first verified. In the event of a session
7515 ticket integrity check failing, OpenSSL will fail to free memory
7516 causing a memory leak. By sending a large number of invalid session
7517 tickets an attacker could exploit this issue in a Denial Of Service
7523 * Build option no-ssl3 is incomplete.
7525 When OpenSSL is configured with "no-ssl3" as a build option, servers
7526 could accept and complete a SSL 3.0 handshake, and clients could be
7527 configured to send them.
7530 *Akamai and the OpenSSL team*
7532 * Add support for TLS_FALLBACK_SCSV.
7533 Client applications doing fallback retries should call
7534 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7537 *Adam Langley, Bodo Moeller*
7539 * Add additional DigestInfo checks.
7541 Reencode DigestInto in DER and check against the original when
7542 verifying RSA signature: this will reject any improperly encoded
7543 DigestInfo structures.
7545 Note: this is a precautionary measure and no attacks are currently known.
7549 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7551 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7552 to a denial of service attack. A malicious server can crash the client
7553 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7554 ciphersuite and sending carefully crafted handshake messages.
7556 Thanks to Felix Gröbert (Google) for discovering and researching this
7562 * By sending carefully crafted DTLS packets an attacker could cause openssl
7563 to leak memory. This can be exploited through a Denial of Service attack.
7564 Thanks to Adam Langley for discovering and researching this issue.
7569 * An attacker can force openssl to consume large amounts of memory whilst
7570 processing DTLS handshake messages. This can be exploited through a
7571 Denial of Service attack.
7572 Thanks to Adam Langley for discovering and researching this issue.
7577 * An attacker can force an error condition which causes openssl to crash
7578 whilst processing DTLS packets due to memory being freed twice. This
7579 can be exploited through a Denial of Service attack.
7580 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7586 * If a multithreaded client connects to a malicious server using a resumed
7587 session and the server sends an ec point format extension it could write
7588 up to 255 bytes to freed memory.
7590 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7596 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7597 X509_name_oneline, X509_name_print_ex et al. to leak some information
7598 from the stack. Applications may be affected if they echo pretty printing
7599 output to the attacker.
7601 Thanks to Ivan Fratric (Google) for discovering this issue.
7604 *Emilia Käsper, and Steve Henson*
7606 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7607 for corner cases. (Certain input points at infinity could lead to
7608 bogus results, with non-infinity inputs mapped to infinity too.)
7612 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7614 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7615 handshake can force the use of weak keying material in OpenSSL
7616 SSL/TLS clients and servers.
7618 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7619 researching this issue. ([CVE-2014-0224])
7621 *KIKUCHI Masashi, Steve Henson*
7623 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7624 OpenSSL DTLS client the code can be made to recurse eventually crashing
7627 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7630 *Imre Rad, Steve Henson*
7632 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7633 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7634 client or server. This is potentially exploitable to run arbitrary
7635 code on a vulnerable client or server.
7637 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7639 *Jüri Aedla, Steve Henson*
7641 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7642 are subject to a denial of service attack.
7644 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7645 this issue. ([CVE-2014-3470])
7647 *Felix Gröbert, Ivan Fratric, Steve Henson*
7649 * Harmonize version and its documentation. -f flag is used to display
7652 *mancha <mancha1@zoho.com>*
7654 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7655 in i2d_ECPrivateKey.
7657 *mancha <mancha1@zoho.com>*
7659 * Fix some double frees. These are not thought to be exploitable.
7661 *mancha <mancha1@zoho.com>*
7663 * Fix for the attack described in the paper "Recovering OpenSSL
7664 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7665 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7666 <http://eprint.iacr.org/2014/140>
7668 Thanks to Yuval Yarom and Naomi Benger for discovering this
7669 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7671 *Yuval Yarom and Naomi Benger*
7673 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7675 * Keep original DTLS digest and encryption contexts in retransmission
7676 structures so we can use the previous session parameters if they need
7677 to be resent. ([CVE-2013-6450])
7681 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7682 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7683 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7684 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7685 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7686 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7688 *Rob Stradling, Adam Langley*
7690 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7692 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7694 This addresses the flaw in CBC record processing discovered by
7695 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7696 at: <http://www.isg.rhul.ac.uk/tls/>
7698 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7699 Security Group at Royal Holloway, University of London
7700 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7701 Emilia Käsper for the initial patch.
7704 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7706 * Return an error when checking OCSP signatures when key is NULL.
7707 This fixes a DoS attack. ([CVE-2013-0166])
7711 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7712 the right response is stapled. Also change SSL_get_certificate()
7713 so it returns the certificate actually sent.
7714 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7715 (This is a backport)
7717 *Rob Stradling <rob.stradling@comodo.com>*
7719 * Fix possible deadlock when decoding public keys.
7723 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7725 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7728 * Sanity check record length before skipping explicit IV in DTLS
7731 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7732 fuzzing as a service testing platform.
7737 * Initialise tkeylen properly when encrypting CMS messages.
7738 Thanks to Solar Designer of Openwall for reporting this issue.
7742 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7744 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7745 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7746 in CRYPTO_realloc_clean.
7748 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7749 issue and to Adam Langley <agl@chromium.org> for fixing it.
7752 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7754 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7756 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7757 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7758 content decryption and always return the same error. Note: this attack
7759 needs on average 2^20 messages so it only affects automated senders. The
7760 old behaviour can be re-enabled in the CMS code by setting the
7761 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7762 an MMA defence is not necessary.
7763 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7764 this issue. ([CVE-2012-0884])
7768 * Fix CVE-2011-4619: make sure we really are receiving a
7769 client hello before rejecting multiple SGC restarts. Thanks to
7770 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7774 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7776 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7777 Thanks to Antonio Martin, Enterprise Secure Access Research and
7778 Development, Cisco Systems, Inc. for discovering this bug and
7779 preparing a fix. ([CVE-2012-0050])
7783 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7785 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7786 of the Vaudenay padding oracle attack on CBC mode encryption
7787 which enables an efficient plaintext recovery attack against
7788 the OpenSSL implementation of DTLS. Their attack exploits timing
7789 differences arising during decryption processing. A research
7790 paper describing this attack can be found at:
7791 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7792 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7793 Security Group at Royal Holloway, University of London
7794 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7795 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7796 for preparing the fix. ([CVE-2011-4108])
7798 *Robin Seggelmann, Michael Tuexen*
7800 * Clear bytes used for block padding of SSL 3.0 records.
7803 *Adam Langley (Google)*
7805 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7806 Kadianakis <desnacked@gmail.com> for discovering this issue and
7807 Adam Langley for preparing the fix. ([CVE-2011-4619])
7809 *Adam Langley (Google)*
7811 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7813 *Andrey Kulikov <amdeich@gmail.com>*
7815 * Prevent malformed RFC3779 data triggering an assertion failure.
7816 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7817 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7819 *Rob Austein <sra@hactrn.net>*
7821 * Improved PRNG seeding for VOS.
7823 *Paul Green <Paul.Green@stratus.com>*
7825 * Fix ssl_ciph.c set-up race.
7827 *Adam Langley (Google)*
7829 * Fix spurious failures in ecdsatest.c.
7831 *Emilia Käsper (Google)*
7833 * Fix the BIO_f_buffer() implementation (which was mixing different
7834 interpretations of the `..._len` fields).
7836 *Adam Langley (Google)*
7838 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7839 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7840 threads won't reuse the same blinding coefficients.
7842 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7843 lock to call BN_BLINDING_invert_ex, and avoids one use of
7844 BN_BLINDING_update for each BN_BLINDING structure (previously,
7845 the last update always remained unused).
7847 *Emilia Käsper (Google)*
7849 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7851 *Bob Buckholz (Google)*
7853 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7855 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7856 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7858 *Kaspar Brand <ossl@velox.ch>*
7860 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7861 for multi-threaded use of ECDH. ([CVE-2011-3210])
7863 *Adam Langley (Google)*
7865 * Fix x509_name_ex_d2i memory leak on bad inputs.
7869 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7870 signature public key algorithm by using OID xref utilities instead.
7871 Before this you could only use some ECC ciphersuites with SHA1 only.
7875 * Add protection against ECDSA timing attacks as mentioned in the paper
7876 by Billy Bob Brumley and Nicola Tuveri, see:
7877 <http://eprint.iacr.org/2011/232.pdf>
7879 *Billy Bob Brumley and Nicola Tuveri*
7881 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7883 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7885 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7887 * Fix bug in string printing code: if *any* escaping is enabled we must
7888 escape the escape character (backslash) or the resulting string is
7893 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7895 * Disable code workaround for ancient and obsolete Netscape browsers
7896 and servers: an attacker can use it in a ciphersuite downgrade attack.
7897 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7901 * Fixed J-PAKE implementation error, originally discovered by
7902 Sebastien Martini, further info and confirmation from Stefan
7903 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7907 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7909 * Fix extension code to avoid race conditions which can result in a buffer
7910 overrun vulnerability: resumed sessions must not be modified as they can
7911 be shared by multiple threads. CVE-2010-3864
7915 * Fix WIN32 build system to correctly link an ENGINE directory into
7920 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7922 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7925 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7927 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7929 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7930 context. The operation can be customised via the ctrl mechanism in
7931 case ENGINEs want to include additional functionality.
7935 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7939 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7940 output hashes compatible with older versions of OpenSSL.
7942 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7944 * Fix compression algorithm handling: if resuming a session use the
7945 compression algorithm of the resumed session instead of determining
7946 it from client hello again. Don't allow server to change algorithm.
7950 * Add load_crls() function to apps tidying load_certs() too. Add option
7951 to verify utility to allow additional CRLs to be included.
7955 * Update OCSP request code to permit adding custom headers to the request:
7956 some responders need this.
7960 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7963 *Julia Lawall <julia@diku.dk>*
7965 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7966 needlessly dereferenced structures, used obsolete functions and
7967 didn't handle all updated verify codes correctly.
7971 * Disable MD2 in the default configuration.
7975 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7976 indicate the initial BIO being pushed or popped. This makes it possible
7977 to determine whether the BIO is the one explicitly called or as a result
7978 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7979 it handles reference counts correctly and doesn't zero out the I/O bio
7980 when it is not being explicitly popped. WARNING: applications which
7981 included workarounds for the old buggy behaviour will need to be modified
7982 or they could free up already freed BIOs.
7986 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
7987 renaming to all platforms (within the 0.9.8 branch, this was
7988 done conditionally on Netware platforms to avoid a name clash).
7990 *Guenter <lists@gknw.net>*
7992 * Add ECDHE and PSK support to DTLS.
7994 *Michael Tuexen <tuexen@fh-muenster.de>*
7996 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8001 * Add "missing" function EVP_MD_flags() (without this the only way to
8002 retrieve a digest flags is by accessing the structure directly. Update
8003 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8004 or cipher is registered as in the "from" argument. Print out all
8005 registered digests in the dgst usage message instead of manually
8006 attempting to work them out.
8010 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8011 this allows the use of compression and extensions. Change default cipher
8012 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8013 by default unless an application cipher string requests it.
8017 * Alter match criteria in PKCS12_parse(). It used to try to use local
8018 key ids to find matching certificates and keys but some PKCS#12 files
8019 don't follow the (somewhat unwritten) rules and this strategy fails.
8020 Now just gather all certificates together and the first private key
8021 then look for the first certificate that matches the key.
8025 * Support use of registered digest and cipher names for dgst and cipher
8026 commands instead of having to add each one as a special case. So now
8033 openssl dgst -sha256 foo
8035 and this works for ENGINE based algorithms too.
8039 * Update Gost ENGINE to support parameter files.
8041 *Victor B. Wagner <vitus@cryptocom.ru>*
8043 * Support GeneralizedTime in ca utility.
8045 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8047 * Enhance the hash format used for certificate directory links. The new
8048 form uses the canonical encoding (meaning equivalent names will work
8049 even if they aren't identical) and uses SHA1 instead of MD5. This form
8050 is incompatible with the older format and as a result c_rehash should
8051 be used to rebuild symbolic links.
8055 * Make PKCS#8 the default write format for private keys, replacing the
8056 traditional format. This form is standardised, more secure and doesn't
8057 include an implicit MD5 dependency.
8061 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8062 committed to OpenSSL should pass this lot as a minimum.
8066 * Add session ticket override functionality for use by EAP-FAST.
8068 *Jouni Malinen <j@w1.fi>*
8070 * Modify HMAC functions to return a value. Since these can be implemented
8071 in an ENGINE errors can occur.
8075 * Type-checked OBJ_bsearch_ex.
8079 * Type-checked OBJ_bsearch. Also some constification necessitated
8080 by type-checking. Still to come: TXT_DB, bsearch(?),
8081 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8086 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8087 seconds to a tm structure directly, instead of going through OS
8088 specific date routines. This avoids any issues with OS routines such
8089 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8090 and X509_time_adj_ex() to cover the extended range. The existing
8091 X509_time_adj() is still usable and will no longer have any date issues.
8095 * Delta CRL support. New use deltas option which will attempt to locate
8096 and search any appropriate delta CRLs available.
8098 This work was sponsored by Google.
8102 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8103 code and add additional score elements. Validate alternate CRL paths
8104 as part of the CRL checking and indicate a new error "CRL path validation
8105 error" in this case. Applications wanting additional details can use
8106 the verify callback and check the new "parent" field. If this is not
8107 NULL CRL path validation is taking place. Existing applications won't
8108 see this because it requires extended CRL support which is off by
8111 This work was sponsored by Google.
8115 * Support for freshest CRL extension.
8117 This work was sponsored by Google.
8121 * Initial indirect CRL support. Currently only supported in the CRLs
8122 passed directly and not via lookup. Process certificate issuer
8123 CRL entry extension and lookup CRL entries by bother issuer name
8124 and serial number. Check and process CRL issuer entry in IDP extension.
8126 This work was sponsored by Google.
8130 * Add support for distinct certificate and CRL paths. The CRL issuer
8131 certificate is validated separately in this case. Only enabled if
8132 an extended CRL support flag is set: this flag will enable additional
8133 CRL functionality in future.
8135 This work was sponsored by Google.
8139 * Add support for policy mappings extension.
8141 This work was sponsored by Google.
8145 * Fixes to pathlength constraint, self issued certificate handling,
8146 policy processing to align with RFC3280 and PKITS tests.
8148 This work was sponsored by Google.
8152 * Support for name constraints certificate extension. DN, email, DNS
8153 and URI types are currently supported.
8155 This work was sponsored by Google.
8159 * To cater for systems that provide a pointer-based thread ID rather
8160 than numeric, deprecate the current numeric thread ID mechanism and
8161 replace it with a structure and associated callback type. This
8162 mechanism allows a numeric "hash" to be extracted from a thread ID in
8163 either case, and on platforms where pointers are larger than 'long',
8164 mixing is done to help ensure the numeric 'hash' is usable even if it
8165 can't be guaranteed unique. The default mechanism is to use "&errno"
8166 as a pointer-based thread ID to distinguish between threads.
8168 Applications that want to provide their own thread IDs should now use
8169 CRYPTO_THREADID_set_callback() to register a callback that will call
8170 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8172 Note that ERR_remove_state() is now deprecated, because it is tied
8173 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8174 to free the current thread's error state should be replaced by
8175 ERR_remove_thread_state(NULL).
8177 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8178 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8179 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8180 application was previously providing a numeric thread callback that
8181 was inappropriate for distinguishing threads, then uniqueness might
8182 have been obtained with &errno that happened immediately in the
8183 intermediate development versions of OpenSSL; this is no longer the
8184 case, the numeric thread callback will now override the automatic use
8187 *Geoff Thorpe, with help from Bodo Moeller*
8189 * Initial support for different CRL issuing certificates. This covers a
8190 simple case where the self issued certificates in the chain exist and
8191 the real CRL issuer is higher in the existing chain.
8193 This work was sponsored by Google.
8197 * Removed effectively defunct crypto/store from the build.
8201 * Revamp of STACK to provide stronger type-checking. Still to come:
8202 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8203 ASN1_STRING, CONF_VALUE.
8207 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8208 RAM on SSL connections. This option can save about 34k per idle SSL.
8212 * Revamp of LHASH to provide stronger type-checking. Still to come:
8213 STACK, TXT_DB, bsearch, qsort.
8217 * Initial support for Cryptographic Message Syntax (aka CMS) based
8218 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8219 support for data, signedData, compressedData, digestedData and
8220 encryptedData, envelopedData types included. Scripts to check against
8221 RFC4134 examples draft and interop and consistency checks of many
8222 content types and variants.
8226 * Add options to enc utility to support use of zlib compression BIO.
8230 * Extend mk1mf to support importing of options and assembly language
8231 files from Configure script, currently only included in VC-WIN32.
8232 The assembly language rules can now optionally generate the source
8233 files from the associated perl scripts.
8237 * Implement remaining functionality needed to support GOST ciphersuites.
8238 Interop testing has been performed using CryptoPro implementations.
8240 *Victor B. Wagner <vitus@cryptocom.ru>*
8242 * s390x assembler pack.
8246 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8251 * Implement Opaque PRF Input TLS extension as specified in
8252 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8253 official specification yet and no extension type assignment by
8254 IANA exists, this extension (for now) will have to be explicitly
8255 enabled when building OpenSSL by providing the extension number
8256 to use. For example, specify an option
8258 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8260 to the "config" or "Configure" script to enable the extension,
8261 assuming extension number 0x9527 (which is a completely arbitrary
8262 and unofficial assignment based on the MD5 hash of the Internet
8263 Draft). Note that by doing so, you potentially lose
8264 interoperability with other TLS implementations since these might
8265 be using the same extension number for other purposes.
8267 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8268 opaque PRF input value to use in the handshake. This will create
8269 an internal copy of the length-'len' string at 'src', and will
8270 return non-zero for success.
8272 To get more control and flexibility, provide a callback function
8275 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8276 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8280 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8283 Callback function 'cb' will be called in handshakes, and is
8284 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8285 Argument 'arg' is for application purposes (the value as given to
8286 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8287 be provided to the callback function). The callback function
8288 has to return non-zero to report success: usually 1 to use opaque
8289 PRF input just if possible, or 2 to enforce use of the opaque PRF
8290 input. In the latter case, the library will abort the handshake
8291 if opaque PRF input is not successfully negotiated.
8293 Arguments 'peerinput' and 'len' given to the callback function
8294 will always be NULL and 0 in the case of a client. A server will
8295 see the client's opaque PRF input through these variables if
8296 available (NULL and 0 otherwise). Note that if the server
8297 provides an opaque PRF input, the length must be the same as the
8298 length of the client's opaque PRF input.
8300 Note that the callback function will only be called when creating
8301 a new session (session resumption can resume whatever was
8302 previously negotiated), and will not be called in SSL 2.0
8303 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8304 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8305 for applications that need to enforce opaque PRF input.
8309 * Update ssl code to support digests other than SHA1+MD5 for handshake
8312 *Victor B. Wagner <vitus@cryptocom.ru>*
8314 * Add RFC4507 support to OpenSSL. This includes the corrections in
8315 RFC4507bis. The encrypted ticket format is an encrypted encoded
8316 SSL_SESSION structure, that way new session features are automatically
8319 If a client application caches session in an SSL_SESSION structure
8320 support is transparent because tickets are now stored in the encoded
8323 The SSL_CTX structure automatically generates keys for ticket
8324 protection in servers so again support should be possible
8325 with no application modification.
8327 If a client or server wishes to disable RFC4507 support then the option
8328 SSL_OP_NO_TICKET can be set.
8330 Add a TLS extension debugging callback to allow the contents of any client
8331 or server extensions to be examined.
8333 This work was sponsored by Google.
8337 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8338 OpenSSL should now compile cleanly on gcc 4.2
8340 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8342 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8343 support including streaming MAC support: this is required for GOST
8344 ciphersuite support.
8346 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8348 * Add option -stream to use PKCS#7 streaming in smime utility. New
8349 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8350 to output in BER and PEM format.
8354 * Experimental support for use of HMAC via EVP_PKEY interface. This
8355 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8356 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8357 ENGINE support for HMAC keys which are unextractable. New -mac and
8358 -macopt options to dgst utility.
8362 * New option -sigopt to dgst utility. Update dgst to use
8363 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8364 alternative signing parameters such as X9.31 or PSS in the dgst
8369 * Change ssl_cipher_apply_rule(), the internal function that does
8370 the work each time a ciphersuite string requests enabling
8371 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8372 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8373 the order of disabled ciphersuites such that those ciphersuites
8374 that most recently went from enabled to disabled not only stay
8375 in order with respect to each other, but also have higher priority
8376 than other disabled ciphersuites the next time ciphersuites are
8379 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8380 the same ciphersuites as with "HIGH" alone, but in a specific
8381 order where the PSK ciphersuites come first (since they are the
8382 most recently disabled ciphersuites when "HIGH" is parsed).
8384 Also, change ssl_create_cipher_list() (using this new
8385 functionality) such that between otherwise identical
8386 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8391 * Change ssl_create_cipher_list() so that it automatically
8392 arranges the ciphersuites in reasonable order before starting
8393 to process the rule string. Thus, the definition for "DEFAULT"
8394 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8395 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8396 This makes it much easier to arrive at a reasonable default order
8397 in applications for which anonymous ciphers are OK (meaning
8398 that you can't actually use DEFAULT).
8400 *Bodo Moeller; suggested by Victor Duchovni*
8402 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8403 processing) into multiple integers instead of setting
8404 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8405 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8406 (These masks as well as the individual bit definitions are hidden
8407 away into the non-exported interface ssl/ssl_locl.h, so this
8408 change to the definition of the SSL_CIPHER structure shouldn't
8409 affect applications.) This give us more bits for each of these
8410 categories, so there is no longer a need to coagulate AES128 and
8411 AES256 into a single algorithm bit, and to coagulate Camellia128
8412 and Camellia256 into a single algorithm bit, which has led to all
8415 Thus, among other things, the kludge introduced in 0.9.7m and
8416 0.9.8e for masking out AES256 independently of AES128 or masking
8417 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8419 With the change, we also introduce new ciphersuite aliases that
8420 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8425 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8426 Use the leftmost N bytes of the signature input if the input is
8427 larger than the prime q (with N being the size in bytes of q).
8431 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8432 it yet and it is largely untested.
8436 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8440 * Initial incomplete changes to avoid need for function casts in OpenSSL
8441 some compilers (gcc 4.2 and later) reject their use. Safestack is
8442 reimplemented. Update ASN1 to avoid use of legacy functions.
8446 * Win32/64 targets are linked with Winsock2.
8450 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8451 to external functions. This can be used to increase CRL handling
8452 efficiency especially when CRLs are very large by (for example) storing
8453 the CRL revoked certificates in a database.
8457 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8458 new CRLs added to a directory can be used. New command line option
8459 -verify_return_error to s_client and s_server. This causes real errors
8460 to be returned by the verify callback instead of carrying on no matter
8461 what. This reflects the way a "real world" verify callback would behave.
8465 * GOST engine, supporting several GOST algorithms and public key formats.
8466 Kindly donated by Cryptocom.
8470 * Partial support for Issuing Distribution Point CRL extension. CRLs
8471 partitioned by DP are handled but no indirect CRL or reason partitioning
8472 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8473 selected via a scoring technique which handles IDP and AKID in CRLs.
8477 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8478 will ultimately be used for all verify operations: this will remove the
8479 X509_STORE dependency on certificate verification and allow alternative
8480 lookup methods. X509_STORE based implementations of these two callbacks.
8484 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8485 Modify get_crl() to find a valid (unexpired) CRL if possible.
8489 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8490 this would be called X509_CRL_cmp() but that name is already used by
8491 a function that just compares CRL issuer names. Cache several CRL
8492 extensions in X509_CRL structure and cache CRLDP in X509.
8496 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8497 this maps equivalent X509_NAME structures into a consistent structure.
8498 Name comparison can then be performed rapidly using memcmp().
8502 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8507 * Allow digests to supply their own micalg string for S/MIME type using
8508 the ctrl EVP_MD_CTRL_MICALG.
8512 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8513 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8514 ctrl. It can then customise the structure before and/or after signing
8519 * New function OBJ_add_sigid() to allow application defined signature OIDs
8520 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8521 to free up any added signature OIDs.
8525 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8526 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8527 digest and cipher tables. New options added to openssl utility:
8528 list-message-digest-algorithms and list-cipher-algorithms.
8532 * Change the array representation of binary polynomials: the list
8533 of degrees of non-zero coefficients is now terminated with -1.
8534 Previously it was terminated with 0, which was also part of the
8535 value; thus, the array representation was not applicable to
8536 polynomials where t^0 has coefficient zero. This change makes
8537 the array representation useful in a more general context.
8541 * Various modifications and fixes to SSL/TLS cipher string
8542 handling. For ECC, the code now distinguishes between fixed ECDH
8543 with RSA certificates on the one hand and with ECDSA certificates
8544 on the other hand, since these are separate ciphersuites. The
8545 unused code for Fortezza ciphersuites has been removed.
8547 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8548 (not "ECDHE"). For consistency with the code for DH
8549 certificates, use of ECDH certificates is now considered ECDH
8550 authentication, not RSA or ECDSA authentication (the latter is
8551 merely the CA's signing algorithm and not actively used in the
8554 The temporary ciphersuite alias "ECCdraft" is no longer
8555 available, and ECC ciphersuites are no longer excluded from "ALL"
8556 and "DEFAULT". The following aliases now exist for RFC 4492
8557 ciphersuites, most of these by analogy with the DH case:
8559 kECDHr - ECDH cert, signed with RSA
8560 kECDHe - ECDH cert, signed with ECDSA
8561 kECDH - ECDH cert (signed with either RSA or ECDSA)
8562 kEECDH - ephemeral ECDH
8563 ECDH - ECDH cert or ephemeral ECDH
8569 AECDH - anonymous ECDH
8570 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8574 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8575 Use correct micalg parameters depending on digest(s) in signed message.
8579 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8580 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8584 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8585 an engine to register a method. Add ENGINE lookups for methods and
8586 functional reference processing.
8590 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8591 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8596 * New -resign option to smime utility. This adds one or more signers
8597 to an existing PKCS#7 signedData structure. Also -md option to use an
8598 alternative message digest algorithm for signing.
8602 * Tidy up PKCS#7 routines and add new functions to make it easier to
8603 create PKCS7 structures containing multiple signers. Update smime
8604 application to support multiple signers.
8608 * New -macalg option to pkcs12 utility to allow setting of an alternative
8613 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8614 Reorganize PBE internals to lookup from a static table using NIDs,
8615 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8616 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8617 PRF which will be automatically used with PBES2.
8621 * Replace the algorithm specific calls to generate keys in "req" with the
8626 * Update PKCS#7 enveloped data routines to use new API. This is now
8627 supported by any public key method supporting the encrypt operation. A
8628 ctrl is added to allow the public key algorithm to examine or modify
8629 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8634 * Add a ctrl to asn1 method to allow a public key algorithm to express
8635 a default digest type to use. In most cases this will be SHA1 but some
8636 algorithms (such as GOST) need to specify an alternative digest. The
8637 return value indicates how strong the preference is 1 means optional and
8638 2 is mandatory (that is it is the only supported type). Modify
8639 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8640 use the default md. Update openssl utilities to use the default digest
8641 type for signing if it is not explicitly indicated.
8645 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8646 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8647 signing method from the key type. This effectively removes the link
8648 between digests and public key types.
8652 * Add an OID cross reference table and utility functions. Its purpose is to
8653 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8654 rsaEncryption. This will allow some of the algorithm specific hackery
8655 needed to use the correct OID to be removed.
8659 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8660 structures for PKCS7_sign(). They are now set up by the relevant public
8665 * Add provisional EC pkey method with support for ECDSA and ECDH.
8669 * Add support for key derivation (agreement) in the API, DH method and
8674 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8675 public and private key formats. As a side effect these add additional
8676 command line functionality not previously available: DSA signatures can be
8677 generated and verified using pkeyutl and DH key support and generation in
8684 *Oliver Tappe <zooey@hirschkaefer.de>*
8686 * New make target "install_html_docs" installs HTML renditions of the
8689 *Oliver Tappe <zooey@hirschkaefer.de>*
8691 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8692 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8693 support key and parameter generation and add initial key generation
8694 functionality for RSA.
8698 * Add functions for main EVP_PKEY_method operations. The undocumented
8699 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8700 `EVP_PKEY_{encrypt,decrypt}_old`.
8704 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8705 key API, doesn't do much yet.
8709 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8710 public key algorithms. New option to openssl utility:
8711 "list-public-key-algorithms" to print out info.
8715 * Implement the Supported Elliptic Curves Extension for
8716 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8720 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8721 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8725 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8726 utilities such as rsa, dsa, dsaparam etc except they process any key
8731 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8732 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8733 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8738 * Initial support for pluggable public key ASN1.
8739 De-spaghettify the public key ASN1 handling. Move public and private
8740 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8741 algorithm specific handling to a single module within the relevant
8742 algorithm directory. Add functions to allow (near) opaque processing
8743 of public and private key structures.
8747 * Implement the Supported Point Formats Extension for
8748 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8752 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8753 for the psk identity [hint] and the psk callback functions to the
8754 SSL_SESSION, SSL and SSL_CTX structure.
8757 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8761 SSL_CTX_use_psk_identity_hint
8762 SSL_get_psk_identity_hint
8763 SSL_get_psk_identity
8764 SSL_use_psk_identity_hint
8766 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8768 * Add RFC 3161 compliant time stamp request creation, response generation
8769 and response verification functionality.
8771 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8773 * Add initial support for TLS extensions, specifically for the server_name
8774 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8775 have new members for a host name. The SSL data structure has an
8776 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8777 stored in that context to allow for session resumption, even after the
8778 SSL has been switched to a new SSL_CTX in reaction to a client's
8779 server_name extension.
8781 New functions (subject to change):
8783 SSL_get_servername()
8784 SSL_get_servername_type()
8787 New CTRL codes and macros (subject to change):
8789 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8790 - SSL_CTX_set_tlsext_servername_callback()
8791 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8792 - SSL_CTX_set_tlsext_servername_arg()
8793 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8795 openssl s_client has a new '-servername ...' option.
8797 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8798 '-key2 ...', '-servername_fatal' (subject to change). This allows
8799 testing the HostName extension for a specific single host name ('-cert'
8800 and '-key' remain fallbacks for handshakes without HostName
8801 negotiation). If the unrecognized_name alert has to be sent, this by
8802 default is a warning; it becomes fatal with the '-servername_fatal'
8805 *Peter Sylvester, Remy Allais, Christophe Renou*
8807 * Whirlpool hash implementation is added.
8811 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8812 bn(64,32). Because of instruction set limitations it doesn't have
8813 any negative impact on performance. This was done mostly in order
8814 to make it possible to share assembler modules, such as bn_mul_mont
8815 implementations, between 32- and 64-bit builds without hassle.
8819 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8820 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8825 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8826 dedicated Montgomery multiplication procedure, is introduced.
8827 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8828 "64-bit" performance on certain 32-bit targets.
8832 * New option SSL_OP_NO_COMP to disable use of compression selectively
8833 in SSL structures. New SSL ctrl to set maximum send fragment size.
8834 Save memory by setting the I/O buffer sizes dynamically instead of
8835 using the maximum available value.
8839 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8840 in addition to the text details.
8844 * Very, very preliminary EXPERIMENTAL support for printing of general
8845 ASN1 structures. This currently produces rather ugly output and doesn't
8846 handle several customised structures at all.
8850 * Integrated support for PVK file format and some related formats such
8851 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8852 these in the 'rsa' and 'dsa' utilities.
8856 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8860 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8861 place for the (very old) "NETSCAPE" format certificates which are now
8862 handled using new ASN1 code equivalents.
8866 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8867 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8868 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8872 * Modify CRL distribution points extension code to print out previously
8873 unsupported fields. Enhance extension setting code to allow setting of
8878 * Add print and set support for Issuing Distribution Point CRL extension.
8882 * Change 'Configure' script to enable Camellia by default.
8889 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8891 * When rejecting SSL/TLS records due to an incorrect version number, never
8892 update s->server with a new major version number. As of
8893 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8894 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8895 the previous behavior could result in a read attempt at NULL when
8896 receiving specific incorrect SSL/TLS records once record payload
8897 protection is active. ([CVE-2010-0740])
8899 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8901 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8902 could be crashed if the relevant tables were not present (e.g. chrooted).
8904 *Tomas Hoger <thoger@redhat.com>*
8906 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8908 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8910 *Martin Olsson, Neel Mehta*
8912 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8913 accommodate for stack sorting, always a write lock!).
8917 * On some versions of WIN32 Heap32Next is very slow. This can cause
8918 excessive delays in the RAND_poll(): over a minute. As a workaround
8919 include a time check in the inner Heap32Next loop too.
8923 * The code that handled flushing of data in SSL/TLS originally used the
8924 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8925 the problem outlined in PR#1949. The fix suggested there however can
8926 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8927 of Apache). So instead simplify the code to flush unconditionally.
8928 This should be fine since flushing with no data to flush is a no op.
8932 * Handle TLS versions 2.0 and later properly and correctly use the
8933 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8934 off ancient servers have a habit of sticking around for a while...
8938 * Modify compression code so it frees up structures without using the
8939 ex_data callbacks. This works around a problem where some applications
8940 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8941 restarting) then use compression (e.g. SSL with compression) later.
8942 This results in significant per-connection memory leaks and
8943 has caused some security issues including CVE-2008-1678 and
8948 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8949 change when encrypting or decrypting.
8953 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8954 connect and renegotiate with servers which do not support RI.
8955 Until RI is more widely deployed this option is enabled by default.
8959 * Add "missing" ssl ctrls to clear options and mode.
8963 * If client attempts to renegotiate and doesn't support RI respond with
8964 a no_renegotiation alert as required by RFC5746. Some renegotiating
8965 TLS clients will continue a connection gracefully when they receive
8966 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8967 waiting for a server hello which it will never receive. Now we treat a
8968 received no_renegotiation alert as a fatal error. This is because
8969 applications requesting a renegotiation might well expect it to succeed
8970 and would have no code in place to handle the server denying it so the
8971 only safe thing to do is to terminate the connection.
8975 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8976 peer supports secure renegotiation and 0 otherwise. Print out peer
8977 renegotiation support in s_client/s_server.
8981 * Replace the highly broken and deprecated SPKAC certification method with
8982 the updated NID creation version. This should correctly handle UTF8.
8986 * Implement RFC5746. Re-enable renegotiation but require the extension
8987 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
8988 turns out to be a bad idea. It has been replaced by
8989 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
8990 SSL_CTX_set_options(). This is really not recommended unless you
8991 know what you are doing.
8993 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
8995 * Fixes to stateless session resumption handling. Use initial_ctx when
8996 issuing and attempting to decrypt tickets in case it has changed during
8997 servername handling. Use a non-zero length session ID when attempting
8998 stateless session resumption: this makes it possible to determine if
8999 a resumption has occurred immediately after receiving server hello
9000 (several places in OpenSSL subtly assume this) instead of later in
9005 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9006 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9007 fixes for a few places where the return code is not checked
9010 *Julia Lawall <julia@diku.dk>*
9012 * Add --strict-warnings option to Configure script to include devteam
9013 warnings in other configurations.
9017 * Add support for --libdir option and LIBDIR variable in makefiles. This
9018 makes it possible to install openssl libraries in locations which
9019 have names other than "lib", for example "/usr/lib64" which some
9022 *Steve Henson, based on patch from Jeremy Utley*
9024 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9025 X690 8.9.12 and can produce some misleading textual output of OIDs.
9027 *Steve Henson, reported by Dan Kaminsky*
9029 * Delete MD2 from algorithm tables. This follows the recommendation in
9030 several standards that it is not used in new applications due to
9031 several cryptographic weaknesses. For binary compatibility reasons
9032 the MD2 API is still compiled in by default.
9036 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9041 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9042 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9045 *Guenter <lists@gknw.net>*
9047 * Fix the server certificate chain building code to use X509_verify_cert(),
9048 it used to have an ad-hoc builder which was unable to cope with anything
9049 other than a simple chain.
9051 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9053 * Don't check self signed certificate signatures in X509_verify_cert()
9054 by default (a flag can override this): it just wastes time without
9055 adding any security. As a useful side effect self signed root CAs
9056 with non-FIPS digests are now usable in FIPS mode.
9060 * In dtls1_process_out_of_seq_message() the check if the current message
9061 is already buffered was missing. For every new message was memory
9062 allocated, allowing an attacker to perform an denial of service attack
9063 with sending out of seq handshake messages until there is no memory
9064 left. Additionally every future message was buffered, even if the
9065 sequence number made no sense and would be part of another handshake.
9066 So only messages with sequence numbers less than 10 in advance will be
9067 buffered. ([CVE-2009-1378])
9069 *Robin Seggelmann, discovered by Daniel Mentz*
9071 * Records are buffered if they arrive with a future epoch to be
9072 processed after finishing the corresponding handshake. There is
9073 currently no limitation to this buffer allowing an attacker to perform
9074 a DOS attack with sending records with future epochs until there is no
9075 memory left. This patch adds the pqueue_size() function to determine
9076 the size of a buffer and limits the record buffer to 100 entries.
9079 *Robin Seggelmann, discovered by Daniel Mentz*
9081 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9082 parent structure is freed. ([CVE-2009-1379])
9086 * Handle non-blocking I/O properly in SSL_shutdown() call.
9088 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9090 * Add `2.5.4.*` OIDs
9092 *Ilya O. <vrghost@gmail.com>*
9094 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9096 * Disable renegotiation completely - this fixes a severe security
9097 problem ([CVE-2009-3555]) at the cost of breaking all
9098 renegotiation. Renegotiation can be re-enabled by setting
9099 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9100 run-time. This is really not recommended unless you know what
9105 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9107 * Don't set val to NULL when freeing up structures, it is freed up by
9108 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9109 zeroing past the valid field. ([CVE-2009-0789])
9111 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9113 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9114 checked correctly. This would allow some invalid signed attributes to
9115 appear to verify correctly. ([CVE-2009-0591])
9117 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9119 * Reject UniversalString and BMPString types with invalid lengths. This
9120 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9121 a legal length. ([CVE-2009-0590])
9125 * Set S/MIME signing as the default purpose rather than setting it
9126 unconditionally. This allows applications to override it at the store
9131 * Permit restricted recursion of ASN1 strings. This is needed in practice
9132 to handle some structures.
9136 * Improve efficiency of mem_gets: don't search whole buffer each time
9139 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9141 * New -hex option for openssl rand.
9145 * Print out UTF8String and NumericString when parsing ASN1.
9149 * Support NumericString type for name components.
9153 * Allow CC in the environment to override the automatically chosen
9154 compiler. Note that nothing is done to ensure flags work with the
9159 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9161 * Properly check EVP_VerifyFinal() and similar return values
9164 *Ben Laurie, Bodo Moeller, Google Security Team*
9166 * Enable TLS extensions by default.
9170 * Allow the CHIL engine to be loaded, whether the application is
9171 multithreaded or not. (This does not release the developer from the
9172 obligation to set up the dynamic locking callbacks.)
9174 *Sander Temme <sander@temme.net>*
9176 * Use correct exit code if there is an error in dgst command.
9178 *Steve Henson; problem pointed out by Roland Dirlewanger*
9180 * Tweak Configure so that you need to say "experimental-jpake" to enable
9181 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9185 * Add experimental JPAKE support, including demo authentication in
9186 s_client and s_server.
9190 * Set the comparison function in v3_addr_canonize().
9192 *Rob Austein <sra@hactrn.net>*
9194 * Add support for XMPP STARTTLS in s_client.
9196 *Philip Paeps <philip@freebsd.org>*
9198 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9199 to ensure that even with this option, only ciphersuites in the
9200 server's preference list will be accepted. (Note that the option
9201 applies only when resuming a session, so the earlier behavior was
9202 just about the algorithm choice for symmetric cryptography.)
9206 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9208 * Fix NULL pointer dereference if a DTLS server received
9209 ChangeCipherSpec as first record ([CVE-2009-1386]).
9213 * Fix a state transition in s3_srvr.c and d1_srvr.c
9214 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9218 * The fix in 0.9.8c that supposedly got rid of unsafe
9219 double-checked locking was incomplete for RSA blinding,
9220 addressing just one layer of what turns out to have been
9221 doubly unsafe triple-checked locking.
9223 So now fix this for real by retiring the MONT_HELPER macro
9224 in crypto/rsa/rsa_eay.c.
9226 *Bodo Moeller; problem pointed out by Marius Schilder*
9228 * Various precautionary measures:
9230 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9232 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9233 (NB: This would require knowledge of the secret session ticket key
9234 to exploit, in which case you'd be SOL either way.)
9236 - Change bn_nist.c so that it will properly handle input BIGNUMs
9237 outside the expected range.
9239 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9242 *Neel Mehta, Bodo Moeller*
9244 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9245 the load fails. Useful for distros.
9247 *Ben Laurie and the FreeBSD team*
9249 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9253 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9257 * Expand ENGINE to support engine supplied SSL client certificate functions.
9259 This work was sponsored by Logica.
9263 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9264 keystores. Support for SSL/TLS client authentication too.
9265 Not compiled unless enable-capieng specified to Configure.
9267 This work was sponsored by Logica.
9271 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9272 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9273 attribute creation routines such as certificate requests and PKCS#12
9278 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9280 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9281 handshake which could lead to a client crash as found using the
9282 Codenomicon TLS test suite ([CVE-2008-1672])
9284 *Steve Henson, Mark Cox*
9286 * Fix double free in TLS server name extensions which could lead to
9287 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9291 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9293 Clear the error queue to ensure that error entries left from
9294 older function calls do not interfere with the correct operation.
9296 *Lutz Jaenicke, Erik de Castro Lopo*
9298 * Remove root CA certificates of commercial CAs:
9300 The OpenSSL project does not recommend any specific CA and does not
9301 have any policy with respect to including or excluding any CA.
9302 Therefore it does not make any sense to ship an arbitrary selection
9303 of root CA certificates with the OpenSSL software.
9307 * RSA OAEP patches to fix two separate invalid memory reads.
9308 The first one involves inputs when 'lzero' is greater than
9309 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9310 before the beginning of from). The second one involves inputs where
9311 the 'db' section contains nothing but zeroes (there is a one-byte
9312 invalid read after the end of 'db').
9314 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9316 * Partial backport from 0.9.9-dev:
9318 Introduce bn_mul_mont (dedicated Montgomery multiplication
9319 procedure) as a candidate for BIGNUM assembler implementation.
9320 While 0.9.9-dev uses assembler for various architectures, only
9321 x86_64 is available by default here in the 0.9.8 branch, and
9322 32-bit x86 is available through a compile-time setting.
9324 To try the 32-bit x86 assembler implementation, use Configure
9325 option "enable-montasm" (which exists only for this backport).
9327 As "enable-montasm" for 32-bit x86 disclaims code stability
9328 anyway, in this constellation we activate additional code
9329 backported from 0.9.9-dev for further performance improvements,
9330 namely BN_from_montgomery_word. (To enable this otherwise,
9331 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9333 *Andy Polyakov (backport partially by Bodo Moeller)*
9335 * Add TLS session ticket callback. This allows an application to set
9336 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9337 values. This is useful for key rollover for example where several key
9338 sets may exist with different names.
9342 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9343 This was broken until now in 0.9.8 releases, such that the only way
9344 a registered ENGINE could be used (assuming it initialises
9345 successfully on the host) was to explicitly set it as the default
9346 for the relevant algorithms. This is in contradiction with 0.9.7
9347 behaviour and the documentation. With this fix, when an ENGINE is
9348 registered into a given algorithm's table of implementations, the
9349 'uptodate' flag is reset so that auto-discovery will be used next
9350 time a new context for that algorithm attempts to select an
9353 *Ian Lister (tweaked by Geoff Thorpe)*
9355 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9356 implementation in the following ways:
9358 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9361 Lack of BER streaming support means one pass streaming processing is
9362 only supported if data is detached: setting the streaming flag is
9363 ignored for embedded content.
9365 CMS support is disabled by default and must be explicitly enabled
9366 with the enable-cms configuration option.
9370 * Update the GMP engine glue to do direct copies between BIGNUM and
9371 mpz_t when openssl and GMP use the same limb size. Otherwise the
9372 existing "conversion via a text string export" trick is still used.
9374 *Paul Sheer <paulsheer@gmail.com>*
9376 * Zlib compression BIO. This is a filter BIO which compressed and
9377 uncompresses any data passed through it.
9381 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9382 RFC3394 compatible AES key wrapping.
9386 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9387 sets string data without copying. X509_ALGOR_set0() and
9388 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9389 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9390 from an X509_ATTRIBUTE structure optionally checking it occurs only
9391 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9396 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9397 to get the expected BN_FLG_CONSTTIME behavior.
9399 *Bodo Moeller (Google)*
9403 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9404 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9405 - added some more tests to do_tests.pl
9406 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9407 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9408 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9409 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9410 - various changes to netware.pl to enable gcc-cross builds on Win32
9412 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9413 - various changes to fix missing prototype warnings
9414 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9415 - added AES, WHIRLPOOL and CPUID assembler code to build files
9416 - added missing AES assembler make rules to mk1mf.pl
9417 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9419 *Guenter Knauf <eflash@gmx.net>*
9421 * Implement certificate status request TLS extension defined in RFC3546.
9422 A client can set the appropriate parameters and receive the encoded
9423 OCSP response via a callback. A server can query the supplied parameters
9424 and set the encoded OCSP response in the callback. Add simplified examples
9425 to s_client and s_server.
9429 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9432 + Binary incompatibility of ssl_ctx_st structure
9433 + DTLS interoperation with non-compliant servers
9434 + Don't call get_session_cb() without proposed session
9435 + Fix ia64 assembler code
9437 *Andy Polyakov, Steve Henson*
9439 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9441 * DTLS Handshake overhaul. There were longstanding issues with
9442 OpenSSL DTLS implementation, which were making it impossible for
9443 RFC 4347 compliant client to communicate with OpenSSL server.
9444 Unfortunately just fixing these incompatibilities would "cut off"
9445 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9446 server keeps tolerating non RFC compliant syntax. The opposite is
9447 not true, 0.9.8f client can not communicate with earlier server.
9448 This update even addresses CVE-2007-4995.
9452 * Changes to avoid need for function casts in OpenSSL: some compilers
9453 (gcc 4.2 and later) reject their use.
9454 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9457 * Add RFC4507 support to OpenSSL. This includes the corrections in
9458 RFC4507bis. The encrypted ticket format is an encrypted encoded
9459 SSL_SESSION structure, that way new session features are automatically
9462 If a client application caches session in an SSL_SESSION structure
9463 support is transparent because tickets are now stored in the encoded
9466 The SSL_CTX structure automatically generates keys for ticket
9467 protection in servers so again support should be possible
9468 with no application modification.
9470 If a client or server wishes to disable RFC4507 support then the option
9471 SSL_OP_NO_TICKET can be set.
9473 Add a TLS extension debugging callback to allow the contents of any client
9474 or server extensions to be examined.
9476 This work was sponsored by Google.
9480 * Add initial support for TLS extensions, specifically for the server_name
9481 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9482 have new members for a host name. The SSL data structure has an
9483 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9484 stored in that context to allow for session resumption, even after the
9485 SSL has been switched to a new SSL_CTX in reaction to a client's
9486 server_name extension.
9488 New functions (subject to change):
9490 SSL_get_servername()
9491 SSL_get_servername_type()
9494 New CTRL codes and macros (subject to change):
9496 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9497 - SSL_CTX_set_tlsext_servername_callback()
9498 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9499 - SSL_CTX_set_tlsext_servername_arg()
9500 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9502 openssl s_client has a new '-servername ...' option.
9504 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9505 '-key2 ...', '-servername_fatal' (subject to change). This allows
9506 testing the HostName extension for a specific single host name ('-cert'
9507 and '-key' remain fallbacks for handshakes without HostName
9508 negotiation). If the unrecognized_name alert has to be sent, this by
9509 default is a warning; it becomes fatal with the '-servername_fatal'
9512 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9514 * Add AES and SSE2 assembly language support to VC++ build.
9518 * Mitigate attack on final subtraction in Montgomery reduction.
9522 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9523 (which previously caused an internal error).
9527 * Squeeze another 10% out of IGE mode when in != out.
9531 * AES IGE mode speedup.
9533 *Dean Gaudet (Google)*
9535 * Add the Korean symmetric 128-bit cipher SEED (see
9536 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9537 add SEED ciphersuites from RFC 4162:
9539 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9540 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9541 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9542 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9544 To minimize changes between patchlevels in the OpenSSL 0.9.8
9545 series, SEED remains excluded from compilation unless OpenSSL
9546 is configured with 'enable-seed'.
9548 *KISA, Bodo Moeller*
9550 * Mitigate branch prediction attacks, which can be practical if a
9551 single processor is shared, allowing a spy process to extract
9552 information. For detailed background information, see
9553 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9554 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9555 and Necessary Software Countermeasures"). The core of the change
9556 are new versions BN_div_no_branch() and
9557 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9558 respectively, which are slower, but avoid the security-relevant
9559 conditional branches. These are automatically called by BN_div()
9560 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9561 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9562 remove a conditional branch.
9564 BN_FLG_CONSTTIME is the new name for the previous
9565 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9566 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9567 in the exponent causes BN_mod_exp_mont() to use the alternative
9568 implementation in BN_mod_exp_mont_consttime().) The old name
9569 remains as a deprecated alias.
9571 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9572 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9573 constant-time implementations for more than just exponentiation.
9574 Here too the old name is kept as a deprecated alias.
9576 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9577 the BN_BLINDING structure gets an independent copy of the
9578 modulus. This means that the previous `BIGNUM *m` argument to
9579 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9580 essentially becomes `const BIGNUM *m`, although we can't actually
9581 change this in the header file before 0.9.9. It allows
9582 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9583 enable BN_FLG_CONSTTIME.
9585 *Matthew D Wood (Intel Corp)*
9587 * In the SSL/TLS server implementation, be strict about session ID
9588 context matching (which matters if an application uses a single
9589 external cache for different purposes). Previously,
9590 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9591 set. This did ensure strict client verification, but meant that,
9592 with applications using a single external cache for quite
9593 different requirements, clients could circumvent ciphersuite
9594 restrictions for a given session ID context by starting a session
9595 in a different context.
9599 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9600 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9601 authentication-only ciphersuites.
9605 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9606 not complete and could lead to a possible single byte overflow
9607 ([CVE-2007-5135]) [Ben Laurie]
9609 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9611 * Since AES128 and AES256 (and similarly Camellia128 and
9612 Camellia256) share a single mask bit in the logic of
9613 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9614 kludge to work properly if AES128 is available and AES256 isn't
9615 (or if Camellia128 is available and Camellia256 isn't).
9619 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9620 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9621 When a point or a seed is encoded in a BIT STRING, we need to
9622 prevent the removal of trailing zero bits to get the proper DER
9623 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9624 of a NamedBitList, for which trailing 0 bits need to be removed.)
9628 * Have SSL/TLS server implementation tolerate "mismatched" record
9629 protocol version while receiving ClientHello even if the
9630 ClientHello is fragmented. (The server can't insist on the
9631 particular protocol version it has chosen before the ServerHello
9632 message has informed the client about his choice.)
9636 * Add RFC 3779 support.
9638 *Rob Austein for ARIN, Ben Laurie*
9640 * Load error codes if they are not already present instead of using a
9641 static variable. This allows them to be cleanly unloaded and reloaded.
9642 Improve header file function name parsing.
9646 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9647 or CAPABILITY handshake as required by RFCs.
9651 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9653 * Introduce limits to prevent malicious keys being able to
9654 cause a denial of service. ([CVE-2006-2940])
9656 *Steve Henson, Bodo Moeller*
9658 * Fix ASN.1 parsing of certain invalid structures that can result
9659 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9661 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9662 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9664 * Fix SSL client code which could crash if connecting to a
9665 malicious SSLv2 server. ([CVE-2006-4343])
9667 *Tavis Ormandy and Will Drewry, Google Security Team*
9669 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9670 match only those. Before that, "AES256-SHA" would be interpreted
9671 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9672 the same strength classification in 0.9.7h) as we currently only
9673 have a single AES bit in the ciphersuite description bitmap.
9674 That change, however, also applied to ciphersuite strings such as
9675 "RC4-MD5" that intentionally matched multiple ciphersuites --
9676 namely, SSL 2.0 ciphersuites in addition to the more common ones
9677 from SSL 3.0/TLS 1.0.
9679 So we change the selection algorithm again: Naming an explicit
9680 ciphersuite selects this one ciphersuite, and any other similar
9681 ciphersuite (same bitmap) from *other* protocol versions.
9682 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9683 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9685 Since SSL 2.0 does not have any ciphersuites for which the
9686 128/256 bit distinction would be relevant, this works for now.
9687 The proper fix will be to use different bits for AES128 and
9688 AES256, which would have avoided the problems from the beginning;
9689 however, bits are scarce, so we can only do this in a new release
9690 (not just a patchlevel) when we can change the SSL_CIPHER
9691 definition to split the single 'unsigned long mask' bitmap into
9692 multiple values to extend the available space.
9696 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9698 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9699 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9701 * Add AES IGE and biIGE modes.
9705 * Change the Unix randomness entropy gathering to use poll() when
9706 possible instead of select(), since the latter has some
9707 undesirable limitations.
9709 *Darryl Miles via Richard Levitte and Bodo Moeller*
9711 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9712 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9713 cannot be implicitly activated as part of, e.g., the "AES" alias.
9714 However, please upgrade to OpenSSL 0.9.9[-dev] for
9715 non-experimental use of the ECC ciphersuites to get TLS extension
9716 support, which is required for curve and point format negotiation
9717 to avoid potential handshake problems.
9721 * Disable rogue ciphersuites:
9723 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9724 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9725 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9727 The latter two were purportedly from
9728 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9731 Also deactivate the remaining ciphersuites from
9732 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9733 unofficial, and the ID has long expired.
9737 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9738 dual-core machines) and other potential thread-safety issues.
9742 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9743 versions), which is now available for royalty-free use
9744 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9745 Also, add Camellia TLS ciphersuites from RFC 4132.
9747 To minimize changes between patchlevels in the OpenSSL 0.9.8
9748 series, Camellia remains excluded from compilation unless OpenSSL
9749 is configured with 'enable-camellia'.
9753 * Disable the padding bug check when compression is in use. The padding
9754 bug check assumes the first packet is of even length, this is not
9755 necessarily true if compression is enabled and can result in false
9756 positives causing handshake failure. The actual bug test is ancient
9757 code so it is hoped that implementations will either have fixed it by
9758 now or any which still have the bug do not support compression.
9762 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9764 * When applying a cipher rule check to see if string match is an explicit
9765 cipher suite and only match that one cipher suite if it is.
9769 * Link in manifests for VC++ if needed.
9771 *Austin Ziegler <halostatue@gmail.com>*
9773 * Update support for ECC-based TLS ciphersuites according to
9774 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9775 TLS extensions, which are supported starting with the 0.9.9
9776 branch, not in the OpenSSL 0.9.8 branch).
9780 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9781 opaque EVP_CIPHER_CTX handling.
9785 * Fixes and enhancements to zlib compression code. We now only use
9786 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9787 to conform with the standards mentioned here:
9788 <http://www.zlib.net/DLL_FAQ.txt>
9789 Static zlib linking now works on Windows and the new --with-zlib-include
9790 --with-zlib-lib options to Configure can be used to supply the location
9791 of the headers and library. Gracefully handle case where zlib library
9796 * Several fixes and enhancements to the OID generation code. The old code
9797 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9798 handle numbers larger than ULONG_MAX, truncated printing and had a
9799 non standard OBJ_obj2txt() behaviour.
9803 * Add support for building of engines under engine/ as shared libraries
9804 under VC++ build system.
9808 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9809 Hopefully, we will not see any false combination of paths any more.
9813 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9815 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9816 (part of SSL_OP_ALL). This option used to disable the
9817 countermeasure against man-in-the-middle protocol-version
9818 rollback in the SSL 2.0 server implementation, which is a bad
9819 idea. ([CVE-2005-2969])
9821 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9822 for Information Security, National Institute of Advanced Industrial
9823 Science and Technology [AIST], Japan)*
9825 * Add two function to clear and return the verify parameter flags.
9829 * Keep cipherlists sorted in the source instead of sorting them at
9830 runtime, thus removing the need for a lock.
9834 * Avoid some small subgroup attacks in Diffie-Hellman.
9836 *Nick Mathewson and Ben Laurie*
9838 * Add functions for well-known primes.
9842 * Extended Windows CE support.
9844 *Satoshi Nakamura and Andy Polyakov*
9846 * Initialize SSL_METHOD structures at compile time instead of during
9847 runtime, thus removing the need for a lock.
9851 * Make PKCS7_decrypt() work even if no certificate is supplied by
9852 attempting to decrypt each encrypted key in turn. Add support to
9857 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9859 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9862 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9866 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9867 key into the same file any more.
9871 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9875 * Add -utf8 command line and config file option to 'ca'.
9877 *Stefan <stf@udoma.org*
9879 * Removed the macro des_crypt(), as it seems to conflict with some
9880 libraries. Use DES_crypt().
9884 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9885 involves renaming the source and generated shared-libs for
9886 both. The engines will accept the corrected or legacy ids
9887 ('ncipher' and '4758_cca' respectively) when binding. NB,
9888 this only applies when building 'shared'.
9890 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9892 * Add attribute functions to EVP_PKEY structure. Modify
9893 PKCS12_create() to recognize a CSP name attribute and
9894 use it. Make -CSP option work again in pkcs12 utility.
9898 * Add new functionality to the bn blinding code:
9899 - automatic re-creation of the BN_BLINDING parameters after
9900 a fixed number of uses (currently 32)
9901 - add new function for parameter creation
9902 - introduce flags to control the update behaviour of the
9903 BN_BLINDING parameters
9904 - hide BN_BLINDING structure
9905 Add a second BN_BLINDING slot to the RSA structure to improve
9906 performance when a single RSA object is shared among several
9911 * Add support for DTLS.
9913 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9915 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9916 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9920 * Remove buggy and incomplete DH cert support from
9921 ssl/ssl_rsa.c and ssl/s3_both.c
9925 * Use SHA-1 instead of MD5 as the default digest algorithm for
9926 the apps/openssl applications.
9930 * Compile clean with "-Wall -Wmissing-prototypes
9931 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9932 DEBUG_SAFESTACK must also be set.
9936 * Change ./Configure so that certain algorithms can be disabled by default.
9937 The new counterpiece to "no-xxx" is "enable-xxx".
9939 The patented RC5 and MDC2 algorithms will now be disabled unless
9940 "enable-rc5" and "enable-mdc2", respectively, are specified.
9942 (IDEA remains enabled despite being patented. This is because IDEA
9943 is frequently required for interoperability, and there is no license
9944 fee for non-commercial use. As before, "no-idea" can be used to
9945 avoid this algorithm.)
9949 * Add processing of proxy certificates (see RFC 3820). This work was
9950 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9951 EGEE (Enabling Grids for E-science in Europe).
9955 * RC4 performance overhaul on modern architectures/implementations, such
9956 as Intel P4, IA-64 and AMD64.
9960 * New utility extract-section.pl. This can be used specify an alternative
9961 section number in a pod file instead of having to treat each file as
9962 a separate case in Makefile. This can be done by adding two lines to the
9965 =for comment openssl_section:XXX
9967 The blank line is mandatory.
9971 * New arguments -certform, -keyform and -pass for s_client and s_server
9972 to allow alternative format key and certificate files and passphrase
9977 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9978 update associated structures and add various utility functions.
9980 Add new policy related verify parameters, include policy checking in
9981 standard verify code. Enhance 'smime' application with extra parameters
9982 to support policy checking and print out.
9986 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
9987 Nehemiah processors. These extensions support AES encryption in hardware
9988 as well as RNG (though RNG support is currently disabled).
9990 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
9992 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
9996 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
9998 *Andy Polyakov and a number of other people*
10000 * Improved PowerPC platform support. Most notably BIGNUM assembler
10001 implementation contributed by IBM.
10003 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10005 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10006 exponent rather than 'unsigned long'. There is a corresponding change to
10007 the new 'rsa_keygen' element of the RSA_METHOD structure.
10009 *Jelte Jansen, Geoff Thorpe*
10011 * Functionality for creating the initial serial number file is now
10012 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10014 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10015 number file to 1, which is bound to cause problems. To avoid
10016 the problems while respecting compatibility between different 0.9.7
10017 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10018 CA.pl for serial number initialization. With the new release 0.9.8,
10019 we can fix the problem directly in the 'ca' utility.)
10023 * Reduced header interdependencies by declaring more opaque objects in
10024 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10025 give fewer recursive includes, which could break lazy source code - so
10026 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10027 developers should define this symbol when building and using openssl to
10028 ensure they track the recommended behaviour, interfaces, [etc], but
10029 backwards-compatible behaviour prevails when this isn't defined.
10033 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10037 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10038 This will generate a random key of the appropriate length based on the
10039 cipher context. The EVP_CIPHER can provide its own random key generation
10040 routine to support keys of a specific form. This is used in the des and
10041 3des routines to generate a key of the correct parity. Update S/MIME
10042 code to use new functions and hence generate correct parity DES keys.
10043 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10044 valid (weak or incorrect parity).
10048 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10049 as looking them up. This is useful when the verified structure may contain
10050 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10051 present unless the new PKCS7_NO_CRL flag is asserted.
10055 * Extend ASN1 oid configuration module. It now additionally accepts the
10058 shortName = some long name, 1.2.3.4
10062 * Reimplemented the BN_CTX implementation. There is now no more static
10063 limitation on the number of variables it can handle nor the depth of the
10064 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10065 information can now expand as required, and rather than having a single
10066 static array of bignums, BN_CTX now uses a linked-list of such arrays
10067 allowing it to expand on demand whilst maintaining the usefulness of
10068 BN_CTX's "bundling".
10072 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10073 to allow all RSA operations to function using a single BN_CTX.
10077 * Preliminary support for certificate policy evaluation and checking. This
10078 is initially intended to pass the tests outlined in "Conformance Testing
10079 of Relying Party Client Certificate Path Processing Logic" v1.07.
10083 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10084 remained unused and not that useful. A variety of other little bignum
10085 tweaks and fixes have also been made continuing on from the audit (see
10090 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10091 associated ASN1, EVP and SSL functions and old ASN1 macros.
10095 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10096 and this should never fail. So the return value from the use of
10097 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10098 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10102 * BN_CTX_get() should return zero-valued bignums, providing the same
10103 initialised value as BN_new().
10105 *Geoff Thorpe, suggested by Ulf Möller*
10107 * Support for inhibitAnyPolicy certificate extension.
10111 * An audit of the BIGNUM code is underway, for which debugging code is
10112 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10113 is considered valid when processing BIGNUMs, and causes execution to
10114 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10115 further steps are taken to deliberately pollute unused data in BIGNUM
10116 structures to try and expose faulty code further on. For now, openssl will
10117 (in its default mode of operation) continue to tolerate the inconsistent
10118 forms that it has tolerated in the past, but authors and packagers should
10119 consider trying openssl and their own applications when compiled with
10120 these debugging symbols defined. It will help highlight potential bugs in
10121 their own code, and will improve the test coverage for OpenSSL itself. At
10122 some point, these tighter rules will become openssl's default to improve
10123 maintainability, though the assert()s and other overheads will remain only
10124 in debugging configurations. See bn.h for more details.
10126 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10128 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10129 that can only be obtained through BN_CTX_new() (which implicitly
10130 initialises it). The presence of this function only made it possible
10131 to overwrite an existing structure (and cause memory leaks).
10135 * Because of the callback-based approach for implementing LHASH as a
10136 template type, lh_insert() adds opaque objects to hash-tables and
10137 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10138 to clean up those corresponding objects before destroying the hash table
10139 (and losing the object pointers). So some over-zealous constifications in
10140 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10141 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10142 prototyped to have "const" restrictions on the object pointers they are
10143 given (and so aren't required to cast them away any more).
10147 * The tmdiff.h API was so ugly and minimal that our own timing utility
10148 (speed) prefers to use its own implementation. The two implementations
10149 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10150 its object type properly exposed (MS_TM) instead of casting to/from
10151 `char *`. This may still change yet if someone realises MS_TM and
10153 aren't necessarily the greatest nomenclatures - but this is what was used
10154 internally to the implementation so I've used that for now.
10158 * Ensure that deprecated functions do not get compiled when
10159 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10160 the self-tests were still using deprecated key-generation functions so
10161 these have been updated also.
10165 * Reorganise PKCS#7 code to separate the digest location functionality
10166 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10167 New function PKCS7_set_digest() to set the digest type for PKCS#7
10168 digestedData type. Add additional code to correctly generate the
10169 digestedData type and add support for this type in PKCS7 initialization
10174 * New function PKCS7_set0_type_other() this initializes a PKCS7
10175 structure of type "other".
10179 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10180 sure the loop does correctly stop and breaking ("division by zero")
10181 modulus operations are not performed. The (pre-generated) prime
10182 table crypto/bn/bn_prime.h was already correct, but it could not be
10183 re-generated on some platforms because of the "division by zero"
10184 situation in the script.
10186 *Ralf S. Engelschall*
10188 * Update support for ECC-based TLS ciphersuites according to
10189 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10190 SHA-1 now is only used for "small" curves (where the
10191 representation of a field element takes up to 24 bytes); for
10192 larger curves, the field element resulting from ECDH is directly
10193 used as premaster secret.
10195 *Douglas Stebila (Sun Microsystems Laboratories)*
10197 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10198 curve secp160r1 to the tests.
10200 *Douglas Stebila (Sun Microsystems Laboratories)*
10202 * Add the possibility to load symbols globally with DSO.
10204 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10206 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10207 control of the error stack.
10211 * Add support for STORE in ENGINE.
10215 * Add the STORE type. The intention is to provide a common interface
10216 to certificate and key stores, be they simple file-based stores, or
10217 HSM-type store, or LDAP stores, or...
10218 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10222 * Add a generic structure called OPENSSL_ITEM. This can be used to
10223 pass a list of arguments to any function as well as provide a way
10224 for a function to pass data back to the caller.
10228 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10229 works like BUF_strdup() but can be used to duplicate a portion of
10230 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10235 * Add the function sk_find_ex() which works like sk_find(), but will
10236 return an index to an element even if an exact match couldn't be
10237 found. The index is guaranteed to point at the element where the
10238 searched-for key would be inserted to preserve sorting order.
10242 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10243 takes an extra flags argument for optional functionality. Currently,
10244 the following flags are defined:
10246 OBJ_BSEARCH_VALUE_ON_NOMATCH
10247 This one gets OBJ_bsearch_ex() to return a pointer to the first
10248 element where the comparing function returns a negative or zero
10251 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10252 This one gets OBJ_bsearch_ex() to return a pointer to the first
10253 element where the comparing function returns zero. This is useful
10254 if there are more than one element where the comparing function
10259 * Make it possible to create self-signed certificates with 'openssl ca'
10260 in such a way that the self-signed certificate becomes part of the
10261 CA database and uses the same mechanisms for serial number generation
10262 as all other certificate signing. The new flag '-selfsign' enables
10263 this functionality. Adapt CA.sh and CA.pl.in.
10267 * Add functionality to check the public key of a certificate request
10268 against a given private. This is useful to check that a certificate
10269 request can be signed by that key (self-signing).
10273 * Make it possible to have multiple active certificates with the same
10274 subject in the CA index file. This is done only if the keyword
10275 'unique_subject' is set to 'no' in the main CA section (default
10276 if 'CA_default') of the configuration file. The value is saved
10277 with the database itself in a separate index attribute file,
10278 named like the index file with '.attr' appended to the name.
10282 * Generate multi-valued AVAs using '+' notation in config files for
10287 * Support for nameConstraints certificate extension.
10291 * Support for policyConstraints certificate extension.
10295 * Support for policyMappings certificate extension.
10299 * Make sure the default DSA_METHOD implementation only uses its
10300 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10301 and change its own handlers to be NULL so as to remove unnecessary
10302 indirection. This lets alternative implementations fallback to the
10303 default implementation more easily.
10307 * Support for directoryName in GeneralName related extensions
10312 * Make it possible to link applications using Makefile.shared.
10313 Make that possible even when linking against static libraries!
10317 * Support for single pass processing for S/MIME signing. This now
10318 means that S/MIME signing can be done from a pipe, in addition
10319 cleartext signing (multipart/signed type) is effectively streaming
10320 and the signed data does not need to be all held in memory.
10322 This is done with a new flag PKCS7_STREAM. When this flag is set
10323 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10324 is done after the data is output (and digests calculated) in
10325 SMIME_write_PKCS7().
10329 * Add full support for -rpath/-R, both in shared libraries and
10330 applications, at least on the platforms where it's known how
10335 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10336 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10337 will now compute a table of multiples of the generator that
10338 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10339 faster (notably in the case of a single point multiplication,
10340 scalar * generator).
10342 *Nils Larsch, Bodo Moeller*
10344 * IPv6 support for certificate extensions. The various extensions
10345 which use the IP:a.b.c.d can now take IPv6 addresses using the
10346 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10351 * Added an ENGINE that implements RSA by performing private key
10352 exponentiations with the GMP library. The conversions to and from
10353 GMP's mpz_t format aren't optimised nor are any montgomery forms
10354 cached, and on x86 it appears OpenSSL's own performance has caught up.
10355 However there are likely to be other architectures where GMP could
10356 provide a boost. This ENGINE is not built in by default, but it can be
10357 specified at Configure time and should be accompanied by the necessary
10358 linker additions, eg;
10359 ./config -DOPENSSL_USE_GMP -lgmp
10363 * "openssl engine" will not display ENGINE/DSO load failure errors when
10364 testing availability of engines with "-t" - the old behaviour is
10365 produced by increasing the feature's verbosity with "-tt".
10369 * ECDSA routines: under certain error conditions uninitialized BN objects
10370 could be freed. Solution: make sure initialization is performed early
10371 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10376 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10377 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10378 software implementations. For DSA and DH, parameter generation can
10379 also be overridden by providing the appropriate method callbacks.
10383 * Change the "progress" mechanism used in key-generation and
10384 primality testing to functions that take a new BN_GENCB pointer in
10385 place of callback/argument pairs. The new API functions have `_ex`
10386 postfixes and the older functions are reimplemented as wrappers for
10387 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10388 declarations of the old functions to help (graceful) attempts to
10389 migrate to the new functions. Also, the new key-generation API
10390 functions operate on a caller-supplied key-structure and return
10391 success/failure rather than returning a key or NULL - this is to
10392 help make "keygen" another member function of RSA_METHOD etc.
10394 Example for using the new callback interface:
10396 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10397 void *my_arg = ...;
10400 BN_GENCB_set(&my_cb, my_callback, my_arg);
10402 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10403 /* For the meaning of a, b in calls to my_callback(), see the
10404 * documentation of the function that calls the callback.
10405 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10406 * my_callback should return 1 if it wants BN_is_prime_ex()
10407 * to continue, or 0 to stop.
10412 * Change the ZLIB compression method to be stateful, and make it
10413 available to TLS with the number defined in
10414 draft-ietf-tls-compression-04.txt.
10418 * Add the ASN.1 structures and functions for CertificatePair, which
10419 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10421 CertificatePair ::= SEQUENCE {
10422 forward [0] Certificate OPTIONAL,
10423 reverse [1] Certificate OPTIONAL,
10424 -- at least one of the pair shall be present -- }
10426 Also implement the PEM functions to read and write certificate
10427 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10429 This needed to be defined, mostly for the sake of the LDAP
10430 attribute crossCertificatePair, but may prove useful elsewhere as
10435 * Make it possible to inhibit symlinking of shared libraries in
10436 Makefile.shared, for Cygwin's sake.
10440 * Extend the BIGNUM API by creating a function
10441 void BN_set_negative(BIGNUM *a, int neg);
10442 and a macro that behave like
10443 int BN_is_negative(const BIGNUM *a);
10445 to avoid the need to access 'a->neg' directly in applications.
10449 * Implement fast modular reduction for pseudo-Mersenne primes
10450 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10451 EC_GROUP_new_curve_GFp() will now automatically use this
10454 *Nils Larsch <nla@trustcenter.de>*
10456 * Add new lock type (CRYPTO_LOCK_BN).
10460 * Change the ENGINE framework to automatically load engines
10461 dynamically from specific directories unless they could be
10462 found to already be built in or loaded. Move all the
10463 current engines except for the cryptodev one to a new
10464 directory engines/.
10465 The engines in engines/ are built as shared libraries if
10466 the "shared" options was given to ./Configure or ./config.
10467 Otherwise, they are inserted in libcrypto.a.
10468 /usr/local/ssl/engines is the default directory for dynamic
10469 engines, but that can be overridden at configure time through
10470 the usual use of --prefix and/or --openssldir, and at run
10471 time with the environment variable OPENSSL_ENGINES.
10473 *Geoff Thorpe and Richard Levitte*
10475 * Add Makefile.shared, a helper makefile to build shared
10476 libraries. Adapt Makefile.org.
10480 * Add version info to Win32 DLLs.
10482 *Peter 'Luna' Runestig" <peter@runestig.com>*
10484 * Add new 'medium level' PKCS#12 API. Certificates and keys
10485 can be added using this API to created arbitrary PKCS#12
10486 files while avoiding the low level API.
10488 New options to PKCS12_create(), key or cert can be NULL and
10489 will then be omitted from the output file. The encryption
10490 algorithm NIDs can be set to -1 for no encryption, the mac
10491 iteration count can be set to 0 to omit the mac.
10493 Enhance pkcs12 utility by making the -nokeys and -nocerts
10494 options work when creating a PKCS#12 file. New option -nomac
10495 to omit the mac, NONE can be set for an encryption algorithm.
10496 New code is modified to use the enhanced PKCS12_create()
10497 instead of the low level API.
10501 * Extend ASN1 encoder to support indefinite length constructed
10502 encoding. This can output sequences tags and octet strings in
10503 this form. Modify pk7_asn1.c to support indefinite length
10504 encoding. This is experimental and needs additional code to
10505 be useful, such as an ASN1 bio and some enhanced streaming
10508 Extend template encode functionality so that tagging is passed
10509 down to the template encoder.
10513 * Let 'openssl req' fail if an argument to '-newkey' is not
10514 recognized instead of using RSA as a default.
10518 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10519 As these are not official, they are not included in "ALL";
10520 the "ECCdraft" ciphersuite group alias can be used to select them.
10522 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10524 * Add ECDH engine support.
10526 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10528 * Add ECDH in new directory crypto/ecdh/.
10530 *Douglas Stebila (Sun Microsystems Laboratories)*
10532 * Let BN_rand_range() abort with an error after 100 iterations
10533 without success (which indicates a broken PRNG).
10537 * Change BN_mod_sqrt() so that it verifies that the input value
10538 is really the square of the return value. (Previously,
10539 BN_mod_sqrt would show GIGO behaviour.)
10543 * Add named elliptic curves over binary fields from X9.62, SECG,
10544 and WAP/WTLS; add OIDs that were still missing.
10546 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10548 * Extend the EC library for elliptic curves over binary fields
10549 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10552 EC_GF2m_simple_method
10556 EC_GROUP_new_curve_GF2m
10557 EC_GROUP_set_curve_GF2m
10558 EC_GROUP_get_curve_GF2m
10559 EC_POINT_set_affine_coordinates_GF2m
10560 EC_POINT_get_affine_coordinates_GF2m
10561 EC_POINT_set_compressed_coordinates_GF2m
10563 Point compression for binary fields is disabled by default for
10564 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10567 As binary polynomials are represented as BIGNUMs, various members
10568 of the EC_GROUP and EC_POINT data structures can be shared
10569 between the implementations for prime fields and binary fields;
10570 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10571 are essentially identical to their `..._GFp` counterparts.
10572 (For simplicity, the `..._GFp` prefix has been dropped from
10573 various internal method names.)
10575 An internal 'field_div' method (similar to 'field_mul' and
10576 'field_sqr') has been added; this is used only for binary fields.
10578 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10580 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10581 through methods ('mul', 'precompute_mult').
10583 The generic implementations (now internally called 'ec_wNAF_mul'
10584 and 'ec_wNAF_precomputed_mult') remain the default if these
10585 methods are undefined.
10587 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10589 * New function EC_GROUP_get_degree, which is defined through
10590 EC_METHOD. For curves over prime fields, this returns the bit
10591 length of the modulus.
10593 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10595 * New functions EC_GROUP_dup, EC_POINT_dup.
10596 (These simply call ..._new and ..._copy).
10598 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10600 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10601 Polynomials are represented as BIGNUMs (where the sign bit is not
10602 used) in the following functions [macros]:
10605 BN_GF2m_sub [= BN_GF2m_add]
10606 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10607 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10608 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10610 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10611 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10612 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10613 BN_GF2m_cmp [= BN_ucmp]
10615 (Note that only the 'mod' functions are actually for fields GF(2^m).
10616 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10618 For some functions, an the irreducible polynomial defining a
10619 field can be given as an 'unsigned int[]' with strictly
10620 decreasing elements giving the indices of those bits that are set;
10621 i.e., p[] represents the polynomial
10622 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10624 p[0] > p[1] > ... > p[k] = 0.
10625 This applies to the following functions:
10628 BN_GF2m_mod_mul_arr
10629 BN_GF2m_mod_sqr_arr
10630 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10631 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10632 BN_GF2m_mod_exp_arr
10633 BN_GF2m_mod_sqrt_arr
10634 BN_GF2m_mod_solve_quad_arr
10638 Conversion can be performed by the following functions:
10643 bntest.c has additional tests for binary polynomial arithmetic.
10645 Two implementations for BN_GF2m_mod_div() are available.
10646 The default algorithm simply uses BN_GF2m_mod_inv() and
10647 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10648 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10649 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10651 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10653 * Add new error code 'ERR_R_DISABLED' that can be used when some
10654 functionality is disabled at compile-time.
10656 *Douglas Stebila <douglas.stebila@sun.com>*
10658 * Change default behaviour of 'openssl asn1parse' so that more
10659 information is visible when viewing, e.g., a certificate:
10661 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10662 mode the content of non-printable OCTET STRINGs is output in a
10663 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10664 avoid the appearance of a printable string.
10666 *Nils Larsch <nla@trustcenter.de>*
10668 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10670 EC_GROUP_set_asn1_flag()
10671 EC_GROUP_get_asn1_flag()
10672 EC_GROUP_set_point_conversion_form()
10673 EC_GROUP_get_point_conversion_form()
10674 These control ASN1 encoding details:
10675 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10676 has been set to OPENSSL_EC_NAMED_CURVE.
10677 - Points are encoded in uncompressed form by default; options for
10678 asn1_for are as for point2oct, namely
10679 POINT_CONVERSION_COMPRESSED
10680 POINT_CONVERSION_UNCOMPRESSED
10681 POINT_CONVERSION_HYBRID
10683 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10685 EC_GROUP_set_seed()
10686 EC_GROUP_get0_seed()
10687 EC_GROUP_get_seed_len()
10688 This is used only for ASN1 purposes (so far).
10690 *Nils Larsch <nla@trustcenter.de>*
10692 * Add 'field_type' member to EC_METHOD, which holds the NID
10693 of the appropriate field type OID. The new function
10694 EC_METHOD_get_field_type() returns this value.
10696 *Nils Larsch <nla@trustcenter.de>*
10699 EC_POINT_point2bn()
10700 EC_POINT_bn2point()
10701 EC_POINT_point2hex()
10702 EC_POINT_hex2point()
10703 providing useful interfaces to EC_POINT_point2oct() and
10704 EC_POINT_oct2point().
10706 *Nils Larsch <nla@trustcenter.de>*
10708 * Change internals of the EC library so that the functions
10709 EC_GROUP_set_generator()
10710 EC_GROUP_get_generator()
10711 EC_GROUP_get_order()
10712 EC_GROUP_get_cofactor()
10713 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10714 to methods, which would lead to unnecessary code duplication when
10715 adding different types of curves.
10717 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10719 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10720 arithmetic, and such that modified wNAFs are generated
10721 (which avoid length expansion in many cases).
10725 * Add a function EC_GROUP_check_discriminant() (defined via
10726 EC_METHOD) that verifies that the curve discriminant is non-zero.
10728 Add a function EC_GROUP_check() that makes some sanity tests
10729 on a EC_GROUP, its generator and order. This includes
10730 EC_GROUP_check_discriminant().
10732 *Nils Larsch <nla@trustcenter.de>*
10734 * Add ECDSA in new directory crypto/ecdsa/.
10736 Add applications 'openssl ecparam' and 'openssl ecdsa'
10737 (these are based on 'openssl dsaparam' and 'openssl dsa').
10739 ECDSA support is also included in various other files across the
10740 library. Most notably,
10741 - 'openssl req' now has a '-newkey ecdsa:file' option;
10742 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10743 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10744 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10745 them suitable for ECDSA where domain parameters must be
10746 extracted before the specific public key;
10747 - ECDSA engine support has been added.
10749 *Nils Larsch <nla@trustcenter.de>*
10751 * Include some named elliptic curves, and add OIDs from X9.62,
10752 SECG, and WAP/WTLS. Each curve can be obtained from the new
10754 EC_GROUP_new_by_curve_name(),
10755 and the list of available named curves can be obtained with
10756 EC_get_builtin_curves().
10757 Also add a 'curve_name' member to EC_GROUP objects, which can be
10759 EC_GROUP_set_curve_name()
10760 EC_GROUP_get_curve_name()
10762 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10764 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10765 was actually never needed) and in BN_mul(). The removal in BN_mul()
10766 required a small change in bn_mul_part_recursive() and the addition
10767 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10768 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10769 bn_sub_words() and bn_add_words() except they take arrays with
10774 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10776 * Cleanse PEM buffers before freeing them since they may contain
10779 *Benjamin Bennett <ben@psc.edu>*
10781 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10782 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10783 authentication-only ciphersuites.
10787 * Since AES128 and AES256 share a single mask bit in the logic of
10788 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10789 kludge to work properly if AES128 is available and AES256 isn't.
10793 * Expand security boundary to match 1.1.1 module.
10797 * Remove redundant features: hash file source, editing of test vectors
10798 modify fipsld to use external fips_premain.c signature.
10802 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10803 run algorithm test programs.
10807 * Make algorithm test programs more tolerant of whitespace.
10811 * Have SSL/TLS server implementation tolerate "mismatched" record
10812 protocol version while receiving ClientHello even if the
10813 ClientHello is fragmented. (The server can't insist on the
10814 particular protocol version it has chosen before the ServerHello
10815 message has informed the client about his choice.)
10819 * Load error codes if they are not already present instead of using a
10820 static variable. This allows them to be cleanly unloaded and reloaded.
10824 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10826 * Introduce limits to prevent malicious keys being able to
10827 cause a denial of service. ([CVE-2006-2940])
10829 *Steve Henson, Bodo Moeller*
10831 * Fix ASN.1 parsing of certain invalid structures that can result
10832 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10834 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10835 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10837 * Fix SSL client code which could crash if connecting to a
10838 malicious SSLv2 server. ([CVE-2006-4343])
10840 *Tavis Ormandy and Will Drewry, Google Security Team*
10842 * Change ciphersuite string processing so that an explicit
10843 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10844 will no longer include "AES128-SHA"), and any other similar
10845 ciphersuite (same bitmap) from *other* protocol versions (so that
10846 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10847 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10848 changes from 0.9.8b and 0.9.8d.
10852 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10854 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10855 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10857 * Change the Unix randomness entropy gathering to use poll() when
10858 possible instead of select(), since the latter has some
10859 undesirable limitations.
10861 *Darryl Miles via Richard Levitte and Bodo Moeller*
10863 * Disable rogue ciphersuites:
10865 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10866 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10867 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10869 The latter two were purportedly from
10870 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10873 Also deactivate the remaining ciphersuites from
10874 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10875 unofficial, and the ID has long expired.
10879 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10880 dual-core machines) and other potential thread-safety issues.
10884 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10886 * Adapt fipsld and the build system to link against the validated FIPS
10887 module in FIPS mode.
10891 * Fixes for VC++ 2005 build under Windows.
10895 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10896 from a Windows bash shell such as MSYS. It is autodetected from the
10897 "config" script when run from a VC++ environment. Modify standard VC++
10898 build to use fipscanister.o from the GNU make build.
10902 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10904 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10905 The value now differs depending on if you build for FIPS or not.
10906 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10907 safely run with a non-FIPSed libcrypto, as it may crash because of
10908 the difference induced by this change.
10912 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10914 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10915 (part of SSL_OP_ALL). This option used to disable the
10916 countermeasure against man-in-the-middle protocol-version
10917 rollback in the SSL 2.0 server implementation, which is a bad
10918 idea. ([CVE-2005-2969])
10920 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10921 for Information Security, National Institute of Advanced Industrial
10922 Science and Technology [AIST, Japan)]*
10924 * Minimal support for X9.31 signatures and PSS padding modes. This is
10925 mainly for FIPS compliance and not fully integrated at this stage.
10929 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10930 the exponentiation using a fixed-length exponent. (Otherwise,
10931 the information leaked through timing could expose the secret key
10932 after many signatures; cf. Bleichenbacher's attack on DSA with
10937 * Make a new fixed-window mod_exp implementation the default for
10938 RSA, DSA, and DH private-key operations so that the sequence of
10939 squares and multiplies and the memory access pattern are
10940 independent of the particular secret key. This will mitigate
10941 cache-timing and potential related attacks.
10943 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10944 and this is automatically used by BN_mod_exp_mont() if the new flag
10945 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10946 will use this BN flag for private exponents unless the flag
10947 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10948 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10950 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10952 * Change the client implementation for SSLv23_method() and
10953 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10954 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10955 (Previously, the SSL 2.0 backwards compatible Client Hello
10956 message format would be used even with SSL_OP_NO_SSLv2.)
10960 * Add support for smime-type MIME parameter in S/MIME messages which some
10965 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10966 a threadsafe manner. Modify rsa code to use new function and add calls
10967 to dsa and dh code (which had race conditions before).
10971 * Include the fixed error library code in the C error file definitions
10972 instead of fixing them up at runtime. This keeps the error code
10973 structures constant.
10977 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10979 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10982 * Fixes for newer kerberos headers. NB: the casts are needed because
10983 the 'length' field is signed on one version and unsigned on another
10984 with no (?) obvious way to tell the difference, without these VC++
10985 complains. Also the "definition" of FAR (blank) is no longer included
10986 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
10987 some needed definitions.
10991 * Undo Cygwin change.
10995 * Added support for proxy certificates according to RFC 3820.
10996 Because they may be a security thread to unaware applications,
10997 they must be explicitly allowed in run-time. See
10998 docs/HOWTO/proxy_certificates.txt for further information.
11002 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11004 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11005 server and client random values. Previously
11006 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11007 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11009 This change has negligible security impact because:
11011 1. Server and client random values still have 24 bytes of pseudo random
11014 2. Server and client random values are sent in the clear in the initial
11017 3. The master secret is derived using the premaster secret (48 bytes in
11018 size for static RSA ciphersuites) as well as client server and random
11021 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11024 *Stephen Henson, reported by UK NISCC*
11026 * Use Windows randomness collection on Cygwin.
11030 * Fix hang in EGD/PRNGD query when communication socket is closed
11031 prematurely by EGD/PRNGD.
11033 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11035 * Prompt for pass phrases when appropriate for PKCS12 input format.
11039 * Back-port of selected performance improvements from development
11040 branch, as well as improved support for PowerPC platforms.
11044 * Add lots of checks for memory allocation failure, error codes to indicate
11045 failure and freeing up memory if a failure occurs.
11047 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11049 * Add new -passin argument to dgst.
11053 * Perform some character comparisons of different types in X509_NAME_cmp:
11054 this is needed for some certificates that re-encode DNs into UTF8Strings
11055 (in violation of RFC3280) and can't or won't issue name rollover
11060 * Make an explicit check during certificate validation to see that
11061 the CA setting in each certificate on the chain is correct. As a
11062 side effect always do the following basic checks on extensions,
11063 not just when there's an associated purpose to the check:
11065 - if there is an unhandled critical extension (unless the user
11066 has chosen to ignore this fault)
11067 - if the path length has been exceeded (if one is set at all)
11068 - that certain extensions fit the associated purpose (if one has
11073 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11075 * Avoid a race condition when CRLs are checked in a multi threaded
11076 environment. This would happen due to the reordering of the revoked
11077 entries during signature checking and serial number lookup. Now the
11078 encoding is cached and the serial number sort performed under a lock.
11079 Add new STACK function sk_is_sorted().
11083 * Add Delta CRL to the extension code.
11087 * Various fixes to s3_pkt.c so alerts are sent properly.
11089 *David Holmes <d.holmes@f5.com>*
11091 * Reduce the chances of duplicate issuer name and serial numbers (in
11092 violation of RFC3280) using the OpenSSL certificate creation utilities.
11093 This is done by creating a random 64 bit value for the initial serial
11094 number when a serial number file is created or when a self signed
11095 certificate is created using 'openssl req -x509'. The initial serial
11096 number file is created using 'openssl x509 -next_serial' in CA.pl
11097 rather than being initialized to 1.
11101 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11103 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11104 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11106 *Joe Orton, Steve Henson*
11108 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11111 *Joe Orton, Steve Henson*
11113 * Make it possible to have multiple active certificates with the same
11114 subject in the CA index file. This is done only if the keyword
11115 'unique_subject' is set to 'no' in the main CA section (default
11116 if 'CA_default') of the configuration file. The value is saved
11117 with the database itself in a separate index attribute file,
11118 named like the index file with '.attr' appended to the name.
11122 * X509 verify fixes. Disable broken certificate workarounds when
11123 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11124 keyUsage extension present. Don't accept CRLs with unhandled critical
11125 extensions: since verify currently doesn't process CRL extensions this
11126 rejects a CRL with *any* critical extensions. Add new verify error codes
11131 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11132 A clarification of RFC2560 will require the use of OCTET STRINGs and
11133 some implementations cannot handle the current raw format. Since OpenSSL
11134 copies and compares OCSP nonces as opaque blobs without any attempt at
11135 parsing them this should not create any compatibility issues.
11139 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11140 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11141 this HMAC (and other) operations are several times slower than OpenSSL
11146 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11148 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11150 * Use the correct content when signing type "other".
11154 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11156 * Fix various bugs revealed by running the NISCC test suite:
11158 Stop out of bounds reads in the ASN1 code when presented with
11159 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11161 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11163 If verify callback ignores invalid public key errors don't try to check
11164 certificate signature with the NULL public key.
11168 * New -ignore_err option in ocsp application to stop the server
11169 exiting on the first error in a request.
11173 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11174 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11179 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11180 extra data after the compression methods not only for TLS 1.0
11181 but also for SSL 3.0 (as required by the specification).
11183 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11185 * Change X509_certificate_type() to mark the key as exported/exportable
11186 when it's 512 *bits* long, not 512 bytes.
11190 * Change AES_cbc_encrypt() so it outputs exact multiple of
11191 blocks during encryption.
11195 * Various fixes to base64 BIO and non blocking I/O. On write
11196 flushes were not handled properly if the BIO retried. On read
11197 data was not being buffered properly and had various logic bugs.
11198 This also affects blocking I/O when the data being decoded is a
11203 * Various S/MIME bugfixes and compatibility changes:
11204 output correct application/pkcs7 MIME type if
11205 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11206 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11207 of files as .eml work). Correctly handle very long lines in MIME
11212 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11214 * Countermeasure against the Klima-Pokorny-Rosa extension of
11215 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11216 a protocol version number mismatch like a decryption error
11217 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11221 * Turn on RSA blinding by default in the default implementation
11222 to avoid a timing attack. Applications that don't want it can call
11223 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11224 They would be ill-advised to do so in most cases.
11226 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11228 * Change RSA blinding code so that it works when the PRNG is not
11229 seeded (in this case, the secret RSA exponent is abused as
11230 an unpredictable seed -- if it is not unpredictable, there
11231 is no point in blinding anyway). Make RSA blinding thread-safe
11232 by remembering the creator's thread ID in rsa->blinding and
11233 having all other threads use local one-time blinding factors
11234 (this requires more computation than sharing rsa->blinding, but
11235 avoids excessive locking; and if an RSA object is not shared
11236 between threads, blinding will still be very fast).
11240 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11241 ENGINE as defaults for all supported algorithms irrespective of
11242 the 'flags' parameter. 'flags' is now honoured, so applications
11243 should make sure they are passing it correctly.
11247 * Target "mingw" now allows native Windows code to be generated in
11248 the Cygwin environment as well as with the MinGW compiler.
11252 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11254 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11255 via timing by performing a MAC computation even if incorrect
11256 block cipher padding has been found. This is a countermeasure
11257 against active attacks where the attacker has to distinguish
11258 between bad padding and a MAC verification error. ([CVE-2003-0078])
11260 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11261 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11262 Martin Vuagnoux (EPFL, Ilion)*
11264 * Make the no-err option work as intended. The intention with no-err
11265 is not to have the whole error stack handling routines removed from
11266 libcrypto, it's only intended to remove all the function name and
11267 reason texts, thereby removing some of the footprint that may not
11268 be interesting if those errors aren't displayed anyway.
11270 NOTE: it's still possible for any application or module to have its
11271 own set of error texts inserted. The routines are there, just not
11272 used by default when no-err is given.
11276 * Add support for FreeBSD on IA64.
11278 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11280 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11281 Kerberos function mit_des_cbc_cksum(). Before this change,
11282 the value returned by DES_cbc_cksum() was like the one from
11283 mit_des_cbc_cksum(), except the bytes were swapped.
11285 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11287 * Allow an application to disable the automatic SSL chain building.
11288 Before this a rather primitive chain build was always performed in
11289 ssl3_output_cert_chain(): an application had no way to send the
11290 correct chain if the automatic operation produced an incorrect result.
11292 Now the chain builder is disabled if either:
11294 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11296 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11298 The reasoning behind this is that an application would not want the
11299 auto chain building to take place if extra chain certificates are
11300 present and it might also want a means of sending no additional
11301 certificates (for example the chain has two certificates and the
11306 * Add the possibility to build without the ENGINE framework.
11308 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11310 * Under Win32 gmtime() can return NULL: check return value in
11311 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11315 * DSA routines: under certain error conditions uninitialized BN objects
11316 could be freed. Solution: make sure initialization is performed early
11317 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11318 Nils Larsch <nla@trustcenter.de> via PR#459)
11322 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11323 checked on reconnect on the client side, therefore session resumption
11324 could still fail with a "ssl session id is different" error. This
11325 behaviour is masked when SSL_OP_ALL is used due to
11326 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11327 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11328 followup to PR #377.
11332 * IA-32 assembler support enhancements: unified ELF targets, support
11333 for SCO/Caldera platforms, fix for Cygwin shared build.
11337 * Add support for FreeBSD on sparc64. As a consequence, support for
11338 FreeBSD on non-x86 processors is separate from x86 processors on
11339 the config script, much like the NetBSD support.
11341 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11343 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11345 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11348 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11349 code (06) was taken as the first octet of the session ID and the last
11350 octet was ignored consequently. As a result SSLv2 client side session
11351 caching could not have worked due to the session ID mismatch between
11353 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11358 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11359 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11364 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11365 seems that in spite of existing for more than a year, many application
11366 author have done nothing to provide the necessary callbacks, which
11367 means that this particular engine will not work properly anywhere.
11368 This is a very unfortunate situation which forces us, in the name
11369 of usability, to give the hw_ncipher.c a static lock, which is part
11371 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11372 appear in 0.9.8 or later. We EXPECT application authors to have
11373 dealt properly with this when 0.9.8 is released (unless we actually
11374 make such changes in the libcrypto locking code that changes will
11375 have to be made anyway).
11379 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11380 octets have been read, EOF or an error occurs. Without this change
11381 some truncated ASN1 structures will not produce an error.
11385 * Disable Heimdal support, since it hasn't been fully implemented.
11386 Still give the possibility to force the use of Heimdal, but with
11387 warnings and a request that patches get sent to openssl-dev.
11391 * Add the VC-CE target, introduce the WINCE sysname, and add
11392 INSTALL.WCE and appropriate conditionals to make it build.
11394 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11396 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11397 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11398 edit numbers of the version.
11400 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11402 * Introduce safe string copy and catenation functions
11403 (BUF_strlcpy() and BUF_strlcat()).
11405 *Ben Laurie (CHATS) and Richard Levitte*
11407 * Avoid using fixed-size buffers for one-line DNs.
11409 *Ben Laurie (CHATS)*
11411 * Add BUF_MEM_grow_clean() to avoid information leakage when
11412 resizing buffers containing secrets, and use where appropriate.
11414 *Ben Laurie (CHATS)*
11416 * Avoid using fixed size buffers for configuration file location.
11418 *Ben Laurie (CHATS)*
11420 * Avoid filename truncation for various CA files.
11422 *Ben Laurie (CHATS)*
11424 * Use sizeof in preference to magic numbers.
11426 *Ben Laurie (CHATS)*
11428 * Avoid filename truncation in cert requests.
11430 *Ben Laurie (CHATS)*
11432 * Add assertions to check for (supposedly impossible) buffer
11435 *Ben Laurie (CHATS)*
11437 * Don't cache truncated DNS entries in the local cache (this could
11438 potentially lead to a spoofing attack).
11440 *Ben Laurie (CHATS)*
11442 * Fix various buffers to be large enough for hex/decimal
11443 representations in a platform independent manner.
11445 *Ben Laurie (CHATS)*
11447 * Add CRYPTO_realloc_clean() to avoid information leakage when
11448 resizing buffers containing secrets, and use where appropriate.
11450 *Ben Laurie (CHATS)*
11452 * Add BIO_indent() to avoid much slightly worrying code to do
11455 *Ben Laurie (CHATS)*
11457 * Convert sprintf()/BIO_puts() to BIO_printf().
11459 *Ben Laurie (CHATS)*
11461 * buffer_gets() could terminate with the buffer only half
11464 *Ben Laurie (CHATS)*
11466 * Add assertions to prevent user-supplied crypto functions from
11467 overflowing internal buffers by having large block sizes, etc.
11469 *Ben Laurie (CHATS)*
11471 * New OPENSSL_assert() macro (similar to assert(), but enabled
11474 *Ben Laurie (CHATS)*
11476 * Eliminate unused copy of key in RC4.
11478 *Ben Laurie (CHATS)*
11480 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11482 *Ben Laurie (CHATS)*
11484 * Fix off-by-one error in EGD path.
11486 *Ben Laurie (CHATS)*
11488 * If RANDFILE path is too long, ignore instead of truncating.
11490 *Ben Laurie (CHATS)*
11492 * Eliminate unused and incorrectly sized X.509 structure
11495 *Ben Laurie (CHATS)*
11497 * Eliminate unused and dangerous function knumber().
11499 *Ben Laurie (CHATS)*
11501 * Eliminate unused and dangerous structure, KSSL_ERR.
11503 *Ben Laurie (CHATS)*
11505 * Protect against overlong session ID context length in an encoded
11506 session object. Since these are local, this does not appear to be
11509 *Ben Laurie (CHATS)*
11511 * Change from security patch (see 0.9.6e below) that did not affect
11512 the 0.9.6 release series:
11514 Remote buffer overflow in SSL3 protocol - an attacker could
11515 supply an oversized master key in Kerberos-enabled versions.
11518 *Ben Laurie (CHATS)*
11520 * Change the SSL kerb5 codes to match RFC 2712.
11524 * Make -nameopt work fully for req and add -reqopt switch.
11526 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11528 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11530 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11532 * Make sure tests can be performed even if the corresponding algorithms
11533 have been removed entirely. This was also the last step to make
11534 OpenSSL compilable with DJGPP under all reasonable conditions.
11536 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11538 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11539 to allow version independent disabling of normally unselected ciphers,
11540 which may be activated as a side-effect of selecting a single cipher.
11542 (E.g., cipher list string "RSA" enables ciphersuites that are left
11543 out of "ALL" because they do not provide symmetric encryption.
11544 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11546 *Lutz Jaenicke, Bodo Moeller*
11548 * Add appropriate support for separate platform-dependent build
11549 directories. The recommended way to make a platform-dependent
11550 build directory is the following (tested on Linux), maybe with
11553 # Place yourself outside of the OpenSSL source tree. In
11554 # this example, the environment variable OPENSSL_SOURCE
11555 # is assumed to contain the absolute OpenSSL source directory.
11556 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11557 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11558 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11559 mkdir -p `dirname $F`
11560 ln -s $OPENSSL_SOURCE/$F $F
11563 To be absolutely sure not to disturb the source tree, a "make clean"
11564 is a good thing. If it isn't successful, don't worry about it,
11565 it probably means the source directory is very clean.
11569 * Make sure any ENGINE control commands make local copies of string
11570 pointers passed to them whenever necessary. Otherwise it is possible
11571 the caller may have overwritten (or deallocated) the original string
11572 data when a later ENGINE operation tries to use the stored values.
11574 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11576 * Improve diagnostics in file reading and command-line digests.
11578 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11580 * Add AES modes CFB and OFB to the object database. Correct an
11581 error in AES-CFB decryption.
11585 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11586 allows existing EVP_CIPHER_CTX structures to be reused after
11587 calling `EVP_*Final()`. This behaviour is used by encryption
11588 BIOs and some applications. This has the side effect that
11589 applications must explicitly clean up cipher contexts with
11590 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11594 * Check the values of dna and dnb in bn_mul_recursive before calling
11595 bn_mul_comba (a non zero value means the a or b arrays do not contain
11596 n2 elements) and fallback to bn_mul_normal if either is not zero.
11600 * Fix escaping of non-ASCII characters when using the -subj option
11601 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11605 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11606 form for "surname", serialNumber has no short form.
11607 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11608 therefore remove "mail" short name for "internet 7".
11609 The OID for unique identifiers in X509 certificates is
11610 x500UniqueIdentifier, not uniqueIdentifier.
11611 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11615 * Add an "init" command to the ENGINE config module and auto initialize
11616 ENGINEs. Without any "init" command the ENGINE will be initialized
11617 after all ctrl commands have been executed on it. If init=1 the
11618 ENGINE is initialized at that point (ctrls before that point are run
11619 on the uninitialized ENGINE and after on the initialized one). If
11620 init=0 then the ENGINE will not be initialized at all.
11624 * Fix the 'app_verify_callback' interface so that the user-defined
11625 argument is actually passed to the callback: In the
11626 SSL_CTX_set_cert_verify_callback() prototype, the callback
11627 declaration has been changed from
11630 int (*cb)(X509_STORE_CTX *,void *);
11631 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11632 i=s->ctx->app_verify_callback(&ctx)
11633 has been changed into
11634 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11636 To update applications using SSL_CTX_set_cert_verify_callback(),
11637 a dummy argument can be added to their callback functions.
11639 *D. K. Smetters <smetters@parc.xerox.com>*
11641 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11643 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11645 * Add and OPENSSL_LOAD_CONF define which will cause
11646 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11647 This allows older applications to transparently support certain
11648 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11649 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11650 load the config file and OPENSSL_add_all_algorithms_conf() which will
11651 always load it have also been added.
11655 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11656 Adjust NIDs and EVP layer.
11658 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11660 * Config modules support in openssl utility.
11662 Most commands now load modules from the config file,
11663 though in a few (such as version) this isn't done
11664 because it couldn't be used for anything.
11666 In the case of ca and req the config file used is
11667 the same as the utility itself: that is the -config
11668 command line option can be used to specify an
11673 * Move default behaviour from OPENSSL_config(). If appname is NULL
11674 use "openssl_conf" if filename is NULL use default openssl config file.
11678 * Add an argument to OPENSSL_config() to allow the use of an alternative
11679 config section name. Add a new flag to tolerate a missing config file
11680 and move code to CONF_modules_load_file().
11684 * Support for crypto accelerator cards from Accelerated Encryption
11685 Processing, www.aep.ie. (Use engine 'aep')
11686 The support was copied from 0.9.6c [engine] and adapted/corrected
11687 to work with the new engine framework.
11689 *AEP Inc. and Richard Levitte*
11691 * Support for SureWare crypto accelerator cards from Baltimore
11692 Technologies. (Use engine 'sureware')
11693 The support was copied from 0.9.6c [engine] and adapted
11694 to work with the new engine framework.
11698 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11699 make the newer ENGINE framework commands for the CHIL engine work.
11701 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11703 * Make it possible to produce shared libraries on ReliantUNIX.
11705 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11707 * Add the configuration target debug-linux-ppro.
11708 Make 'openssl rsa' use the general key loading routines
11709 implemented in apps.c, and make those routines able to
11710 handle the key format FORMAT_NETSCAPE and the variant
11713 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11715 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11717 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11719 * Add -keyform to rsautl, and document -engine.
11721 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11723 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11724 BIO_R_NO_SUCH_FILE error code rather than the generic
11725 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11729 * Add new functions
11730 ERR_peek_last_error
11731 ERR_peek_last_error_line
11732 ERR_peek_last_error_line_data.
11733 These are similar to
11735 ERR_peek_error_line
11736 ERR_peek_error_line_data,
11737 but report on the latest error recorded rather than the first one
11738 still in the error queue.
11740 *Ben Laurie, Bodo Moeller*
11742 * default_algorithms option in ENGINE config module. This allows things
11744 default_algorithms = ALL
11745 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11749 * Preliminary ENGINE config module.
11753 * New experimental application configuration code.
11757 * Change the AES code to follow the same name structure as all other
11758 symmetric ciphers, and behave the same way. Move everything to
11759 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11761 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11763 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11765 *Ben Laurie and Theo de Raadt*
11767 * Add option to output public keys in req command.
11769 *Massimiliano Pala madwolf@openca.org*
11771 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11772 (up to about 10% better than before for P-192 and P-224).
11776 * New functions/macros
11778 SSL_CTX_set_msg_callback(ctx, cb)
11779 SSL_CTX_set_msg_callback_arg(ctx, arg)
11780 SSL_set_msg_callback(ssl, cb)
11781 SSL_set_msg_callback_arg(ssl, arg)
11783 to request calling a callback function
11785 void cb(int write_p, int version, int content_type,
11786 const void *buf, size_t len, SSL *ssl, void *arg)
11788 whenever a protocol message has been completely received
11789 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11790 protocol version according to which the SSL library interprets
11791 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11792 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11793 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11794 specification (change_cipher_spec(20), alert(21), handshake(22)).
11795 'buf' and 'len' point to the actual message, 'ssl' to the
11796 SSL object, and 'arg' is the application-defined value set by
11797 SSL[_CTX]_set_msg_callback_arg().
11799 'openssl s_client' and 'openssl s_server' have new '-msg' options
11800 to enable a callback that displays all protocol messages.
11804 * Change the shared library support so shared libraries are built as
11805 soon as the corresponding static library is finished, and thereby get
11806 openssl and the test programs linked against the shared library.
11807 This still only happens when the keyword "shard" has been given to
11808 the configuration scripts.
11810 NOTE: shared library support is still an experimental thing, and
11811 backward binary compatibility is still not guaranteed.
11813 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11815 * Add support for Subject Information Access extension.
11817 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11819 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11820 additional bytes when new memory had to be allocated, not just
11821 when reusing an existing buffer.
11825 * New command line and configuration option 'utf8' for the req command.
11826 This allows field values to be specified as UTF8 strings.
11830 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11831 runs for the former and machine-readable output for the latter.
11835 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11836 of the e-mail address in the DN (i.e., it will go into a certificate
11837 extension only). The new configuration file option 'email_in_dn = no'
11838 has the same effect.
11840 *Massimiliano Pala madwolf@openca.org*
11842 * Change all functions with names starting with `des_` to be starting
11843 with `DES_` instead. Add wrappers that are compatible with libdes,
11844 but are named `_ossl_old_des_*`. Finally, add macros that map the
11845 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11846 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11847 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11850 Since we provide two compatibility mappings, the user needs to
11851 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11852 compatibility is desired. The default (i.e., when that macro
11853 isn't defined) is OpenSSL 0.9.6c compatibility.
11855 There are also macros that enable and disable the support of old
11856 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11857 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11858 are defined, the default will apply: to support the old des routines.
11860 In either case, one must include openssl/des.h to get the correct
11861 definitions. Do not try to just include openssl/des_old.h, that
11864 NOTE: This is a major break of an old API into a new one. Software
11865 authors are encouraged to switch to the `DES_` style functions. Some
11866 time in the future, des_old.h and the libdes compatibility functions
11867 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11868 default), and then completely removed.
11872 * Test for certificates which contain unsupported critical extensions.
11873 If such a certificate is found during a verify operation it is
11874 rejected by default: this behaviour can be overridden by either
11875 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11876 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11877 X509_supported_extension() has also been added which returns 1 if a
11878 particular extension is supported.
11882 * Modify the behaviour of EVP cipher functions in similar way to digests
11883 to retain compatibility with existing code.
11887 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11888 compatibility with existing code. In particular the 'ctx' parameter does
11889 not have to be to be initialized before the call to EVP_DigestInit() and
11890 it is tidied up after a call to EVP_DigestFinal(). New function
11891 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11892 EVP_MD_CTX_copy() changed to not require the destination to be
11893 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11894 requires the destination to be valid.
11896 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11897 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11901 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11902 so that complete 'Handshake' protocol structures are kept in memory
11903 instead of overwriting 'msg_type' and 'length' with 'body' data.
11907 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11909 *Massimo Santin via Richard Levitte*
11911 * Major restructuring to the underlying ENGINE code. This includes
11912 reduction of linker bloat, separation of pure "ENGINE" manipulation
11913 (initialisation, etc) from functionality dealing with implementations
11914 of specific crypto interfaces. This change also introduces integrated
11915 support for symmetric ciphers and digest implementations - so ENGINEs
11916 can now accelerate these by providing EVP_CIPHER and EVP_MD
11917 implementations of their own. This is detailed in
11918 [crypto/engine/README.md](crypto/engine/README.md)
11919 as it couldn't be adequately described here. However, there are a few
11920 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11921 were changed in the original introduction of ENGINE code have now
11922 reverted back - the hooking from this code to ENGINE is now a good
11923 deal more passive and at run-time, operations deal directly with
11924 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11925 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11926 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11927 they were not being used by the framework as there is no concept of a
11928 BIGNUM_METHOD and they could not be generalised to the new
11929 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11930 ENGINE_cpy() has been removed as it cannot be consistently defined in
11935 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11939 * Change mkdef.pl to sort symbols that get the same entry number,
11940 and make sure the automatically generated functions `ERR_load_*`
11941 become part of libeay.num as well.
11945 * New function SSL_renegotiate_pending(). This returns true once
11946 renegotiation has been requested (either SSL_renegotiate() call
11947 or HelloRequest/ClientHello received from the peer) and becomes
11948 false once a handshake has been completed.
11949 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11950 sends a HelloRequest, but does not ensure that a handshake takes
11951 place. SSL_renegotiate_pending() is useful for checking if the
11952 client has followed the request.)
11956 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11957 By default, clients may request session resumption even during
11958 renegotiation (if session ID contexts permit); with this option,
11959 session resumption is possible only in the first handshake.
11961 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11962 more bits available for options that should not be part of
11963 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11967 * Add some demos for certificate and certificate request creation.
11971 * Make maximum certificate chain size accepted from the peer application
11972 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11973 "Douglas E. Engert" <deengert@anl.gov>.
11977 * Add support for shared libraries for Unixware-7
11978 (Boyd Lynn Gerber <gerberb@zenez.com>).
11982 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11983 be done prior to destruction. Use this to unload error strings from
11984 ENGINEs that load their own error strings. NB: This adds two new API
11985 functions to "get" and "set" this destroy handler in an ENGINE.
11989 * Alter all existing ENGINE implementations (except "openssl" and
11990 "openbsd") to dynamically instantiate their own error strings. This
11991 makes them more flexible to be built both as statically-linked ENGINEs
11992 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
11993 Also, add stub code to each that makes building them as self-contained
11994 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
11998 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
11999 implementations into applications that are completely implemented in
12000 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12001 commands that can be used to configure what shared-library to load and
12002 to control aspects of the way it is handled. Also, made an update to
12003 the [README-Engine.md](README-Engine.md) file
12004 that brings its information up-to-date and
12005 provides some information and instructions on the "dynamic" ENGINE
12006 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12010 * Make it possible to unload ranges of ERR strings with a new
12011 "ERR_unload_strings" function.
12015 * Add a copy() function to EVP_MD.
12019 * Make EVP_MD routines take a context pointer instead of just the
12020 md_data void pointer.
12024 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12025 that the digest can only process a single chunk of data
12026 (typically because it is provided by a piece of
12027 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12028 is only going to provide a single chunk of data, and hence the
12029 framework needn't accumulate the data for oneshot drivers.
12033 * As with "ERR", make it possible to replace the underlying "ex_data"
12034 functions. This change also alters the storage and management of global
12035 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12036 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12037 index counters. The API functions that use this state have been changed
12038 to take a "class_index" rather than pointers to the class's local STACK
12039 and counter, and there is now an API function to dynamically create new
12040 classes. This centralisation allows us to (a) plug a lot of the
12041 thread-safety problems that existed, and (b) makes it possible to clean
12042 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12043 such data would previously have always leaked in application code and
12044 workarounds were in place to make the memory debugging turn a blind eye
12045 to it. Application code that doesn't use this new function will still
12046 leak as before, but their memory debugging output will announce it now
12047 rather than letting it slide.
12049 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12050 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12051 has a return value to indicate success or failure.
12055 * Make it possible to replace the underlying "ERR" functions such that the
12056 global state (2 LHASH tables and 2 locks) is only used by the "default"
12057 implementation. This change also adds two functions to "get" and "set"
12058 the implementation prior to it being automatically set the first time
12059 any other ERR function takes place. Ie. an application can call "get",
12060 pass the return value to a module it has just loaded, and that module
12061 can call its own "set" function using that value. This means the
12062 module's "ERR" operations will use (and modify) the error state in the
12063 application and not in its own statically linked copy of OpenSSL code.
12067 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12068 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12069 the operation, and provides a more encapsulated way for external code
12070 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12071 to use these functions rather than manually incrementing the counts.
12073 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12077 * Add EVP test program.
12081 * Add symmetric cipher support to ENGINE. Expect the API to change!
12085 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12086 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12087 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12088 These allow a CRL to be built without having to access X509_CRL fields
12089 directly. Modify 'ca' application to use new functions.
12093 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12094 bug workarounds. Rollback attack detection is a security feature.
12095 The problem will only arise on OpenSSL servers when TLSv1 is not
12096 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12097 Software authors not wanting to support TLSv1 will have special reasons
12098 for their choice and can explicitly enable this option.
12100 *Bodo Moeller, Lutz Jaenicke*
12102 * Rationalise EVP so it can be extended: don't include a union of
12103 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12104 (similar to those existing for EVP_CIPHER_CTX).
12109 EVP_MD_CTX_init(&md); /* new function call */
12110 EVP_DigestInit(&md, EVP_sha1());
12111 EVP_DigestUpdate(&md, in, len);
12112 EVP_DigestFinal(&md, out, NULL);
12113 EVP_MD_CTX_cleanup(&md); /* new function call */
12117 * Make DES key schedule conform to the usual scheme, as well as
12118 correcting its structure. This means that calls to DES functions
12119 now have to pass a pointer to a des_key_schedule instead of a
12120 plain des_key_schedule (which was actually always a pointer
12123 des_key_schedule ks;
12125 des_set_key_checked(..., &ks);
12126 des_ncbc_encrypt(..., &ks, ...);
12128 (Note that a later change renames 'des_...' into 'DES_...'.)
12132 * Initial reduction of linker bloat: the use of some functions, such as
12133 PEM causes large amounts of unused functions to be linked in due to
12134 poor organisation. For example pem_all.c contains every PEM function
12135 which has a knock on effect of linking in large amounts of (unused)
12136 ASN1 code. Grouping together similar functions and splitting unrelated
12137 functions prevents this.
12141 * Cleanup of EVP macros.
12145 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12146 correct `_ecb suffix`.
12150 * Add initial OCSP responder support to ocsp application. The
12151 revocation information is handled using the text based index
12152 use by the ca application. The responder can either handle
12153 requests generated internally, supplied in files (for example
12154 via a CGI script) or using an internal minimal server.
12158 * Add configuration choices to get zlib compression for TLS.
12162 * Changes to Kerberos SSL for RFC 2712 compliance:
12163 1. Implemented real KerberosWrapper, instead of just using
12164 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12165 2. Implemented optional authenticator field of KerberosWrapper.
12167 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12168 and authenticator structs; see crypto/krb5/.
12170 Generalized Kerberos calls to support multiple Kerberos libraries.
12171 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12172 via Richard Levitte*
12174 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12175 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12176 values for each of the key sizes rather than having just
12177 parameters (and 'speed' generating keys each time).
12181 * Speed up EVP routines.
12184 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12185 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12186 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12187 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12189 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12190 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12191 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12194 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12196 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12200 * Added the OS2-EMX target.
12202 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12204 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12205 to support NCONF routines in extension code. New function CONF_set_nconf()
12206 to allow functions which take an NCONF to also handle the old LHASH
12207 structure: this means that the old CONF compatible routines can be
12208 retained (in particular wrt extensions) without having to duplicate the
12209 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12213 * Enhance the general user interface with mechanisms for inner control
12214 and with possibilities to have yes/no kind of prompts.
12218 * Change all calls to low level digest routines in the library and
12219 applications to use EVP. Add missing calls to HMAC_cleanup() and
12220 don't assume HMAC_CTX can be copied using memcpy().
12222 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12224 * Add the possibility to control engines through control names but with
12225 arbitrary arguments instead of just a string.
12226 Change the key loaders to take a UI_METHOD instead of a callback
12227 function pointer. NOTE: this breaks binary compatibility with earlier
12228 versions of OpenSSL [engine].
12229 Adapt the nCipher code for these new conditions and add a card insertion
12234 * Enhance the general user interface with mechanisms to better support
12235 dialog box interfaces, application-defined prompts, the possibility
12236 to use defaults (for example default passwords from somewhere else)
12237 and interrupts/cancellations.
12241 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12242 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12246 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12247 tidy up some unnecessarily weird code in 'sk_new()').
12249 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12251 * Change the key loading routines for ENGINEs to use the same kind
12252 callback (pem_password_cb) as all other routines that need this
12257 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12258 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12259 than this minimum value is recommended.
12263 * New random seeder for OpenVMS, using the system process statistics
12264 that are easily reachable.
12268 * Windows apparently can't transparently handle global
12269 variables defined in DLLs. Initialisations such as:
12271 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12273 won't compile. This is used by the any applications that need to
12274 declare their own ASN1 modules. This was fixed by adding the option
12275 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12276 needed for static libraries under Win32.
12280 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12281 setting of purpose and trust fields. New X509_STORE trust and
12282 purpose functions and tidy up setting in other SSL functions.
12286 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12287 structure. These are inherited by X509_STORE_CTX when it is
12288 initialised. This allows various defaults to be set in the
12289 X509_STORE structure (such as flags for CRL checking and custom
12290 purpose or trust settings) for functions which only use X509_STORE_CTX
12291 internally such as S/MIME.
12293 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12294 trust settings if they are not set in X509_STORE. This allows X509_STORE
12295 purposes and trust (in S/MIME for example) to override any set by default.
12297 Add command line options for CRL checking to smime, s_client and s_server
12302 * Initial CRL based revocation checking. If the CRL checking flag(s)
12303 are set then the CRL is looked up in the X509_STORE structure and
12304 its validity and signature checked, then if the certificate is found
12305 in the CRL the verify fails with a revoked error.
12307 Various new CRL related callbacks added to X509_STORE_CTX structure.
12309 Command line options added to 'verify' application to support this.
12311 This needs some additional work, such as being able to handle multiple
12312 CRLs with different times, extension based lookup (rather than just
12313 by subject name) and ultimately more complete V2 CRL extension
12318 * Add a general user interface API (crypto/ui/). This is designed
12319 to replace things like des_read_password and friends (backward
12320 compatibility functions using this new API are provided).
12321 The purpose is to remove prompting functions from the DES code
12322 section as well as provide for prompting through dialog boxes in
12323 a window system and the like.
12327 * Add "ex_data" support to ENGINE so implementations can add state at a
12328 per-structure level rather than having to store it globally.
12332 * Make it possible for ENGINE structures to be copied when retrieved by
12333 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12334 This causes the "original" ENGINE structure to act like a template,
12335 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12336 operational state can be localised to each ENGINE structure, despite the
12337 fact they all share the same "methods". New ENGINE structures returned in
12338 this case have no functional references and the return value is the single
12339 structural reference. This matches the single structural reference returned
12340 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12345 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12346 needs to match any other type at all we need to manually clear the
12351 * Changes to the "openssl engine" utility to include;
12352 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12353 about an ENGINE's available control commands.
12354 - executing control commands from command line arguments using the
12355 '-pre' and '-post' switches. '-post' is only used if '-t' is
12356 specified and the ENGINE is successfully initialised. The syntax for
12357 the individual commands are colon-separated, for example;
12358 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12362 * New dynamic control command support for ENGINEs. ENGINEs can now
12363 declare their own commands (numbers), names (strings), descriptions,
12364 and input types for run-time discovery by calling applications. A
12365 subset of these commands are implicitly classed as "executable"
12366 depending on their input type, and only these can be invoked through
12367 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12368 can be based on user input, config files, etc). The distinction is
12369 that "executable" commands cannot return anything other than a boolean
12370 result and can only support numeric or string input, whereas some
12371 discoverable commands may only be for direct use through
12372 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12373 pointers, or other custom uses. The "executable" commands are to
12374 support parameterisations of ENGINE behaviour that can be
12375 unambiguously defined by ENGINEs and used consistently across any
12376 OpenSSL-based application. Commands have been added to all the
12377 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12378 control over shared-library paths without source code alterations.
12382 * Changed all ENGINE implementations to dynamically allocate their
12383 ENGINEs rather than declaring them statically. Apart from this being
12384 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12385 this also allows the implementations to compile without using the
12386 internal engine_int.h header.
12390 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12391 'const' value. Any code that should be able to modify a RAND_METHOD
12392 should already have non-const pointers to it (ie. they should only
12393 modify their own ones).
12397 * Made a variety of little tweaks to the ENGINE code.
12398 - "atalla" and "ubsec" string definitions were moved from header files
12399 to C code. "nuron" string definitions were placed in variables
12400 rather than hard-coded - allowing parameterisation of these values
12401 later on via ctrl() commands.
12402 - Removed unused "#if 0"'d code.
12403 - Fixed engine list iteration code so it uses ENGINE_free() to release
12404 structural references.
12405 - Constified the RAND_METHOD element of ENGINE structures.
12406 - Constified various get/set functions as appropriate and added
12407 missing functions (including a catch-all ENGINE_cpy that duplicates
12408 all ENGINE values onto a new ENGINE except reference counts/state).
12409 - Removed NULL parameter checks in get/set functions. Setting a method
12410 or function to NULL is a way of cancelling out a previously set
12411 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12412 and doesn't justify the extra error symbols and code.
12413 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12414 flags from engine_int.h to engine.h.
12415 - Changed prototypes for ENGINE handler functions (init(), finish(),
12416 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12420 * Implement binary inversion algorithm for BN_mod_inverse in addition
12421 to the algorithm using long division. The binary algorithm can be
12422 used only if the modulus is odd. On 32-bit systems, it is faster
12423 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12424 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12425 up to 450 bits. In 64-bit environments, the binary algorithm
12426 appears to be advantageous for much longer moduli; here we use it
12427 for moduli up to 2048 bits.
12431 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12432 could not support the combine flag in choice fields.
12436 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12437 extensions from a certificate request to the certificate.
12441 * Allow multiple 'certopt' and 'nameopt' options to be separated
12442 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12443 file: this allows the display of the certificate about to be
12444 signed to be customised, to allow certain fields to be included
12445 or excluded and extension details. The old system didn't display
12446 multicharacter strings properly, omitted fields not in the policy
12447 and couldn't display additional details such as extensions.
12451 * Function EC_POINTs_mul for multiple scalar multiplication
12452 of an arbitrary number of elliptic curve points
12453 \sum scalars[i]*points[i],
12454 optionally including the generator defined for the EC_GROUP:
12455 scalar*generator + \sum scalars[i]*points[i].
12457 EC_POINT_mul is a simple wrapper function for the typical case
12458 that the point list has just one item (besides the optional
12463 * First EC_METHODs for curves over GF(p):
12465 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12466 operations and provides various method functions that can also
12467 operate with faster implementations of modular arithmetic.
12469 EC_GFp_mont_method() reuses most functions that are part of
12470 EC_GFp_simple_method, but uses Montgomery arithmetic.
12472 *Bodo Moeller; point addition and point doubling
12473 implementation directly derived from source code provided by
12474 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12476 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12477 crypto/ec/ec_lib.c):
12479 Curves are EC_GROUP objects (with an optional group generator)
12480 based on EC_METHODs that are built into the library.
12482 Points are EC_POINT objects based on EC_GROUP objects.
12484 Most of the framework would be able to handle curves over arbitrary
12485 finite fields, but as there are no obvious types for fields other
12486 than GF(p), some functions are limited to that for now.
12490 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12491 that the file contains a complete HTTP response.
12495 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12496 change the def and num file printf format specifier from "%-40sXXX"
12497 to "%-39s XXX". The latter will always guarantee a space after the
12498 field while the former will cause them to run together if the field
12499 is 40 of more characters long.
12503 * Constify the cipher and digest 'method' functions and structures
12504 and modify related functions to take constant EVP_MD and EVP_CIPHER
12509 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12510 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12514 * Modify `EVP_Digest*()` routines so they now return values. Although the
12515 internal software routines can never fail additional hardware versions
12520 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12522 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12523 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12526 ERR_R_NESTED_ASN1_ERROR
12528 ERR_R_MISSING_ASN1_EOS
12529 were 4 .. 9, conflicting with
12530 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12532 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12533 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12535 Add new error code 'ERR_R_INTERNAL_ERROR'.
12539 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12544 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12545 sets the subject name for a new request or supersedes the
12546 subject name in a given request. Formats that can be parsed are
12547 'CN=Some Name, OU=myOU, C=IT'
12549 'CN=Some Name/OU=myOU/C=IT'.
12551 Add options '-batch' and '-verbose' to 'openssl req'.
12553 *Massimiliano Pala <madwolf@hackmasters.net>*
12555 * Introduce the possibility to access global variables through
12556 functions on platform were that's the best way to handle exporting
12557 global variables in shared libraries. To enable this functionality,
12558 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12559 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12560 is normally done by Configure or something similar).
12562 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12563 in the source file (foo.c) like this:
12565 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12566 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12568 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12569 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12571 OPENSSL_DECLARE_GLOBAL(int,foo);
12572 #define foo OPENSSL_GLOBAL_REF(foo)
12573 OPENSSL_DECLARE_GLOBAL(double,bar);
12574 #define bar OPENSSL_GLOBAL_REF(bar)
12576 The #defines are very important, and therefore so is including the
12577 header file everywhere where the defined globals are used.
12579 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12580 of ASN.1 items, but that structure is a bit different.
12582 The largest change is in util/mkdef.pl which has been enhanced with
12583 better and easier to understand logic to choose which symbols should
12584 go into the Windows .def files as well as a number of fixes and code
12585 cleanup (among others, algorithm keywords are now sorted
12586 lexicographically to avoid constant rewrites).
12590 * In BN_div() keep a copy of the sign of 'num' before writing the
12591 result to 'rm' because if rm==num the value will be overwritten
12592 and produce the wrong result if 'num' is negative: this caused
12593 problems with BN_mod() and BN_nnmod().
12597 * Function OCSP_request_verify(). This checks the signature on an
12598 OCSP request and verifies the signer certificate. The signer
12599 certificate is just checked for a generic purpose and OCSP request
12604 * Add OCSP_check_validity() function to check the validity of OCSP
12605 responses. OCSP responses are prepared in real time and may only
12606 be a few seconds old. Simply checking that the current time lies
12607 between thisUpdate and nextUpdate max reject otherwise valid responses
12608 caused by either OCSP responder or client clock inaccuracy. Instead
12609 we allow thisUpdate and nextUpdate to fall within a certain period of
12610 the current time. The age of the response can also optionally be
12611 checked. Two new options -validity_period and -status_age added to
12616 * If signature or public key algorithm is unrecognized print out its
12617 OID rather that just UNKNOWN.
12621 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12622 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12623 ID to be generated from the issuer certificate alone which can then be
12624 passed to OCSP_id_issuer_cmp().
12628 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12629 ASN1 modules to export functions returning ASN1_ITEM pointers
12630 instead of the ASN1_ITEM structures themselves. This adds several
12631 new macros which allow the underlying ASN1 function/structure to
12632 be accessed transparently. As a result code should not use ASN1_ITEM
12633 references directly (such as &X509_it) but instead use the relevant
12634 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12635 use of the new ASN1 code on platforms where exporting structures
12636 is problematical (for example in shared libraries) but exporting
12637 functions returning pointers to structures is not.
12641 * Add support for overriding the generation of SSL/TLS session IDs.
12642 These callbacks can be registered either in an SSL_CTX or per SSL.
12643 The purpose of this is to allow applications to control, if they wish,
12644 the arbitrary values chosen for use as session IDs, particularly as it
12645 can be useful for session caching in multiple-server environments. A
12646 command-line switch for testing this (and any client code that wishes
12647 to use such a feature) has been added to "s_server".
12649 *Geoff Thorpe, Lutz Jaenicke*
12651 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12652 of the form `#if defined(...) || defined(...) || ...` and
12653 `#if !defined(...) && !defined(...) && ...`. This also avoids
12654 the growing number of special cases it was previously handling.
12658 * Make all configuration macros available for application by making
12659 sure they are available in opensslconf.h, by giving them names starting
12660 with `OPENSSL_` to avoid conflicts with other packages and by making
12661 sure e_os2.h will cover all platform-specific cases together with
12663 Additionally, it is now possible to define configuration/platform-
12664 specific names (called "system identities"). In the C code, these
12665 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12666 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12667 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12672 * New option -set_serial to 'req' and 'x509' this allows the serial
12673 number to use to be specified on the command line. Previously self
12674 signed certificates were hard coded with serial number 0 and the
12675 CA options of 'x509' had to use a serial number in a file which was
12680 * New options to 'ca' utility to support V2 CRL entry extensions.
12681 Currently CRL reason, invalidity date and hold instruction are
12682 supported. Add new CRL extensions to V3 code and some new objects.
12686 * New function EVP_CIPHER_CTX_set_padding() this is used to
12687 disable standard block padding (aka PKCS#5 padding) in the EVP
12688 API, which was previously mandatory. This means that the data is
12689 not padded in any way and so the total length much be a multiple
12690 of the block size, otherwise an error occurs.
12694 * Initial (incomplete) OCSP SSL support.
12698 * New function OCSP_parse_url(). This splits up a URL into its host,
12699 port and path components: primarily to parse OCSP URLs. New -url
12700 option to ocsp utility.
12704 * New nonce behavior. The return value of OCSP_check_nonce() now
12705 reflects the various checks performed. Applications can decide
12706 whether to tolerate certain situations such as an absent nonce
12707 in a response when one was present in a request: the ocsp application
12708 just prints out a warning. New function OCSP_add1_basic_nonce()
12709 this is to allow responders to include a nonce in a response even if
12710 the request is nonce-less.
12714 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12715 skipped when using openssl x509 multiple times on a single input file,
12716 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12720 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12721 set string type: to handle setting ASN1_TIME structures. Fix ca
12722 utility to correctly initialize revocation date of CRLs.
12726 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12727 the clients preferred ciphersuites and rather use its own preferences.
12728 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12729 Internet Explorer by ensuring unchanged hash method during stepup.
12730 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12734 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12735 to aes and add a new 'exist' option to print out symbols that don't
12740 * Additional options to ocsp utility to allow flags to be set and
12741 additional certificates supplied.
12745 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12746 OCSP client a number of certificate to only verify the response
12751 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12752 handle the new API. Currently only ECB, CBC modes supported. Add new
12755 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12756 Encryption Standard (AES) Ciphersuites for Transport Layer
12757 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12758 not enabled by default and were not part of the "ALL" ciphersuite
12759 alias because they were not yet official; they could be
12760 explicitly requested by specifying the "AESdraft" ciphersuite
12761 group alias. In the final release of OpenSSL 0.9.7, the group
12762 alias is called "AES" and is part of "ALL".)
12764 *Ben Laurie, Steve Henson, Bodo Moeller*
12766 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12767 request to response.
12771 * Functions for OCSP responders. OCSP_request_onereq_count(),
12772 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12773 extract information from a certificate request. OCSP_response_create()
12774 creates a response and optionally adds a basic response structure.
12775 OCSP_basic_add1_status() adds a complete single response to a basic
12776 response and returns the OCSP_SINGLERESP structure just added (to allow
12777 extensions to be included for example). OCSP_basic_add1_cert() adds a
12778 certificate to a basic response and OCSP_basic_sign() signs a basic
12779 response with various flags. New helper functions ASN1_TIME_check()
12780 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12781 (converts ASN1_TIME to GeneralizedTime).
12785 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12786 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12787 structure from a certificate. X509_pubkey_digest() digests the public_key
12788 contents: this is used in various key identifiers.
12792 * Make sk_sort() tolerate a NULL argument.
12794 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12796 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12797 passed by the function are trusted implicitly. If any of them signed the
12798 response then it is assumed to be valid and is not verified.
12802 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12803 to data. This was previously part of the PKCS7 ASN1 code. This
12804 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12805 *Steve Henson, reported by Kenneth R. Robinette
12806 <support@securenetterm.com>*
12808 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12809 routines: without these tracing memory leaks is very painful.
12810 Fix leaks in PKCS12 and PKCS7 routines.
12814 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12815 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12816 effectively meant GeneralizedTime would never be used. Now it
12817 is initialised to -1 but X509_time_adj() now has to check the value
12818 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12819 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12820 *Steve Henson, reported by Kenneth R. Robinette
12821 <support@securenetterm.com>*
12823 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12824 result in a zero length in the ASN1_INTEGER structure which was
12825 not consistent with the structure when d2i_ASN1_INTEGER() was used
12826 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12827 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12828 where it did not print out a minus for negative ASN1_INTEGER.
12832 * Add summary printout to ocsp utility. The various functions which
12833 convert status values to strings have been renamed to:
12834 OCSP_response_status_str(), OCSP_cert_status_str() and
12835 OCSP_crl_reason_str() and are no longer static. New options
12836 to verify nonce values and to disable verification. OCSP response
12837 printout format cleaned up.
12841 * Add additional OCSP certificate checks. These are those specified
12842 in RFC2560. This consists of two separate checks: the CA of the
12843 certificate being checked must either be the OCSP signer certificate
12844 or the issuer of the OCSP signer certificate. In the latter case the
12845 OCSP signer certificate must contain the OCSP signing extended key
12846 usage. This check is performed by attempting to match the OCSP
12847 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12848 in the OCSP_CERTID structures of the response.
12852 * Initial OCSP certificate verification added to OCSP_basic_verify()
12853 and related routines. This uses the standard OpenSSL certificate
12854 verify routines to perform initial checks (just CA validity) and
12855 to obtain the certificate chain. Then additional checks will be
12856 performed on the chain. Currently the root CA is checked to see
12857 if it is explicitly trusted for OCSP signing. This is used to set
12858 a root CA as a global signing root: that is any certificate that
12859 chains to that CA is an acceptable OCSP signing certificate.
12863 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12864 extensions from a separate configuration file.
12865 As when reading extensions from the main configuration file,
12866 the '-extensions ...' option may be used for specifying the
12869 *Massimiliano Pala <madwolf@comune.modena.it>*
12871 * New OCSP utility. Allows OCSP requests to be generated or
12872 read. The request can be sent to a responder and the output
12873 parsed, outputted or printed in text form. Not complete yet:
12874 still needs to check the OCSP response validity.
12878 * New subcommands for 'openssl ca':
12879 `openssl ca -status <serial>` prints the status of the cert with
12880 the given serial number (according to the index file).
12881 `openssl ca -updatedb` updates the expiry status of certificates
12884 *Massimiliano Pala <madwolf@comune.modena.it>*
12886 * New '-newreq-nodes' command option to CA.pl. This is like
12887 '-newreq', but calls 'openssl req' with the '-nodes' option
12888 so that the resulting key is not encrypted.
12890 *Damien Miller <djm@mindrot.org>*
12892 * New configuration for the GNU Hurd.
12894 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12896 * Initial code to implement OCSP basic response verify. This
12897 is currently incomplete. Currently just finds the signer's
12898 certificate and verifies the signature on the response.
12902 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12903 value of OPENSSLDIR. This is available via the new '-d' option
12904 to 'openssl version', and is also included in 'openssl version -a'.
12908 * Allowing defining memory allocation callbacks that will be given
12909 file name and line number information in additional arguments
12910 (a `const char*` and an int). The basic functionality remains, as
12911 well as the original possibility to just replace malloc(),
12912 realloc() and free() by functions that do not know about these
12913 additional arguments. To register and find out the current
12914 settings for extended allocation functions, the following
12915 functions are provided:
12917 CRYPTO_set_mem_ex_functions
12918 CRYPTO_set_locked_mem_ex_functions
12919 CRYPTO_get_mem_ex_functions
12920 CRYPTO_get_locked_mem_ex_functions
12922 These work the same way as CRYPTO_set_mem_functions and friends.
12923 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12924 extended allocation function is enabled.
12925 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12926 a conventional allocation function is enabled.
12928 *Richard Levitte, Bodo Moeller*
12930 * Finish off removing the remaining LHASH function pointer casts.
12931 There should no longer be any prototype-casting required when using
12932 the LHASH abstraction, and any casts that remain are "bugs". See
12933 the callback types and macros at the head of lhash.h for details
12934 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12938 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12939 If /dev/[u]random devices are not available or do not return enough
12940 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12942 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12943 /etc/entropy will be queried once each in this sequence, querying stops
12944 when enough entropy was collected without querying more sockets.
12948 * Change the Unix RAND_poll() variant to be able to poll several
12949 random devices, as specified by DEVRANDOM, until a sufficient amount
12950 of data has been collected. We spend at most 10 ms on each file
12951 (select timeout) and read in non-blocking mode. DEVRANDOM now
12952 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12953 (previously it was just the string "/dev/urandom"), so on typical
12954 platforms the 10 ms delay will never occur.
12955 Also separate out the Unix variant to its own file, rand_unix.c.
12956 For VMS, there's a currently-empty rand_vms.c.
12960 * Move OCSP client related routines to ocsp_cl.c. These
12961 provide utility functions which an application needing
12962 to issue a request to an OCSP responder and analyse the
12963 response will typically need: as opposed to those which an
12964 OCSP responder itself would need which will be added later.
12966 OCSP_request_sign() signs an OCSP request with an API similar
12967 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12968 response. OCSP_response_get1_basic() extracts basic response
12969 from response. OCSP_resp_find_status(): finds and extracts status
12970 information from an OCSP_CERTID structure (which will be created
12971 when the request structure is built). These are built from lower
12972 level functions which work on OCSP_SINGLERESP structures but
12973 won't normally be used unless the application wishes to examine
12974 extensions in the OCSP response for example.
12976 Replace nonce routines with a pair of functions.
12977 OCSP_request_add1_nonce() adds a nonce value and optionally
12978 generates a random value. OCSP_check_nonce() checks the
12979 validity of the nonce in an OCSP response.
12983 * Change function OCSP_request_add() to OCSP_request_add0_id().
12984 This doesn't copy the supplied OCSP_CERTID and avoids the
12985 need to free up the newly created id. Change return type
12986 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
12987 This can then be used to add extensions to the request.
12988 Deleted OCSP_request_new(), since most of its functionality
12989 is now in OCSP_REQUEST_new() (and the case insensitive name
12990 clash) apart from the ability to set the request name which
12991 will be added elsewhere.
12995 * Update OCSP API. Remove obsolete extensions argument from
12996 various functions. Extensions are now handled using the new
12997 OCSP extension code. New simple OCSP HTTP function which
12998 can be used to send requests and parse the response.
13002 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13003 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13004 uses the special reorder version of SET OF to sort the attributes
13005 and reorder them to match the encoded order. This resolves a long
13006 standing problem: a verify on a PKCS7 structure just after signing
13007 it used to fail because the attribute order did not match the
13008 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13009 it uses the received order. This is necessary to tolerate some broken
13010 software that does not order SET OF. This is handled by encoding
13011 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13012 to produce the required SET OF.
13016 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13017 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13018 files to get correct declarations of the ASN.1 item variables.
13022 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13023 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13024 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13025 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13026 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13027 ASN1_ITEM and no wrapper functions.
13031 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13032 replace the old function pointer based I/O routines. Change most of
13033 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13037 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13038 lines, recognize more "algorithms" that can be deselected, and make
13039 it complain about algorithm deselection that isn't recognised.
13043 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13044 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13045 to use new functions. Add NO_ASN1_OLD which can be set to remove
13046 some old style ASN1 functions: this can be used to determine if old
13047 code will still work when these eventually go away.
13051 * New extension functions for OCSP structures, these follow the
13052 same conventions as certificates and CRLs.
13056 * New function X509V3_add1_i2d(). This automatically encodes and
13057 adds an extension. Its behaviour can be customised with various
13058 flags to append, replace or delete. Various wrappers added for
13059 certificates and CRLs.
13063 * Fix to avoid calling the underlying ASN1 print routine when
13064 an extension cannot be parsed. Correct a typo in the
13065 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13069 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13070 entries for variables.
13074 * Add functionality to apps/openssl.c for detecting locking
13075 problems: As the program is single-threaded, all we have
13076 to do is register a locking callback using an array for
13077 storing which locks are currently held by the program.
13081 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13082 SSL_get_ex_data_X509_STORE_idx(), which is used in
13083 ssl_verify_cert_chain() and thus can be called at any time
13084 during TLS/SSL handshakes so that thread-safety is essential.
13085 Unfortunately, the ex_data design is not at all suited
13086 for multi-threaded use, so it probably should be abolished.
13090 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13092 *Broadcom, tweaked and integrated by Geoff Thorpe*
13094 * Move common extension printing code to new function
13095 X509V3_print_extensions(). Reorganise OCSP print routines and
13096 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13100 * New function X509_signature_print() to remove duplication in some
13105 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13106 set (this was treated exactly the same as SET OF previously). This
13107 is used to reorder the STACK representing the structure to match the
13108 encoding. This will be used to get round a problem where a PKCS7
13109 structure which was signed could not be verified because the STACK
13110 order did not reflect the encoded order.
13114 * Reimplement the OCSP ASN1 module using the new code.
13118 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13119 for its ASN1 operations. The old style function pointers still exist
13120 for now but they will eventually go away.
13124 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13125 completely replaces the old ASN1 functionality with a table driven
13126 encoder and decoder which interprets an ASN1_ITEM structure describing
13127 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13128 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13129 has also been converted to the new form.
13133 * Change BN_mod_exp_recp so that negative moduli are tolerated
13134 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13135 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13136 for negative moduli.
13140 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13141 of not touching the result's sign bit.
13145 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13150 * Changed the LHASH code to use prototypes for callbacks, and created
13151 macros to declare and implement thin (optionally static) functions
13152 that provide type-safety and avoid function pointer casting for the
13153 type-specific callbacks.
13157 * Added Kerberos Cipher Suites to be used with TLS, as written in
13159 *Veers Staats <staatsvr@asc.hpc.mil>,
13160 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13162 * Reformat the FAQ so the different questions and answers can be divided
13163 in sections depending on the subject.
13167 * Have the zlib compression code load ZLIB.DLL dynamically under
13172 * New function BN_mod_sqrt for computing square roots modulo a prime
13173 (using the probabilistic Tonelli-Shanks algorithm unless
13174 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13175 be handled deterministically).
13177 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13179 * Make BN_mod_inverse faster by explicitly handling small quotients
13180 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13181 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13185 * New function BN_kronecker.
13189 * Fix BN_gcd so that it works on negative inputs; the result is
13190 positive unless both parameters are zero.
13191 Previously something reasonably close to an infinite loop was
13192 possible because numbers could be growing instead of shrinking
13193 in the implementation of Euclid's algorithm.
13197 * Fix BN_is_word() and BN_is_one() macros to take into account the
13198 sign of the number in question.
13200 Fix BN_is_word(a,w) to work correctly for w == 0.
13202 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13203 because its test if the absolute value of 'a' equals 'w'.
13204 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13205 it exists mostly for use in the implementations of BN_is_zero(),
13206 BN_is_one(), and BN_is_word().
13210 * New function BN_swap.
13214 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13215 the exponentiation functions are more likely to produce reasonable
13216 results on negative inputs.
13220 * Change BN_mod_mul so that the result is always non-negative.
13221 Previously, it could be negative if one of the factors was negative;
13222 I don't think anyone really wanted that behaviour.
13226 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13227 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13228 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13229 and add new functions:
13238 BN_mod_lshift1_quick
13240 BN_mod_lshift_quick
13242 These functions always generate non-negative results.
13244 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13245 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13247 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13248 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13249 be reduced modulo `m`.
13251 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13254 The following entry accidentally appeared in the CHANGES file
13255 distributed with OpenSSL 0.9.7. The modifications described in
13256 it do *not* apply to OpenSSL 0.9.7.
13258 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13259 was actually never needed) and in BN_mul(). The removal in BN_mul()
13260 required a small change in bn_mul_part_recursive() and the addition
13261 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13262 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13263 bn_sub_words() and bn_add_words() except they take arrays with
13269 * In 'openssl passwd', verify passwords read from the terminal
13270 unless the '-salt' option is used (which usually means that
13271 verification would just waste user's time since the resulting
13272 hash is going to be compared with some given password hash)
13273 or the new '-noverify' option is used.
13275 This is an incompatible change, but it does not affect
13276 non-interactive use of 'openssl passwd' (passwords on the command
13277 line, '-stdin' option, '-in ...' option) and thus should not
13278 cause any problems.
13282 * Remove all references to RSAref, since there's no more need for it.
13286 * Make DSO load along a path given through an environment variable
13287 (SHLIB_PATH) with shl_load().
13291 * Constify the ENGINE code as a result of BIGNUM constification.
13292 Also constify the RSA code and most things related to it. In a
13293 few places, most notable in the depth of the ASN.1 code, ugly
13294 casts back to non-const were required (to be solved at a later
13299 * Make it so the openssl application has all engines loaded by default.
13303 * Constify the BIGNUM routines a little more.
13307 * Add the following functions:
13309 ENGINE_load_cswift()
13311 ENGINE_load_atalla()
13312 ENGINE_load_nuron()
13313 ENGINE_load_builtin_engines()
13315 That way, an application can itself choose if external engines that
13316 are built-in in OpenSSL shall ever be used or not. The benefit is
13317 that applications won't have to be linked with libdl or other dso
13318 libraries unless it's really needed.
13320 Changed 'openssl engine' to load all engines on demand.
13321 Changed the engine header files to avoid the duplication of some
13322 declarations (they differed!).
13326 * 'openssl engine' can now list capabilities.
13330 * Better error reporting in 'openssl engine'.
13334 * Never call load_dh_param(NULL) in s_server.
13338 * Add engine application. It can currently list engines by name and
13339 identity, and test if they are actually available.
13343 * Improve RPM specification file by forcing symbolic linking and making
13344 sure the installed documentation is also owned by root.root.
13346 *Damien Miller <djm@mindrot.org>*
13348 * Give the OpenSSL applications more possibilities to make use of
13349 keys (public as well as private) handled by engines.
13353 * Add OCSP code that comes from CertCo.
13357 * Add VMS support for the Rijndael code.
13361 * Added untested support for Nuron crypto accelerator.
13365 * Add support for external cryptographic devices. This code was
13366 previously distributed separately as the "engine" branch.
13368 *Geoff Thorpe, Richard Levitte*
13370 * Rework the filename-translation in the DSO code. It is now possible to
13371 have far greater control over how a "name" is turned into a filename
13372 depending on the operating environment and any oddities about the
13373 different shared library filenames on each system.
13377 * Support threads on FreeBSD-elf in Configure.
13381 * Fix for SHA1 assembly problem with MASM: it produces
13382 warnings about corrupt line number information when assembling
13383 with debugging information. This is caused by the overlapping
13386 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13389 NCONF_get_number() has no error checking at all. As a replacement,
13390 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13391 promoted strongly. The old NCONF_get_number is kept around for
13392 binary backward compatibility.
13393 Make it possible for methods to load from something other than a BIO,
13394 by providing a function pointer that is given a name instead of a BIO.
13395 For example, this could be used to load configuration data from an
13400 * Fix for non blocking accept BIOs. Added new I/O special reason
13401 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13402 with non blocking I/O was not possible because no retry code was
13403 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13408 * Added the beginnings of Rijndael support.
13412 * Fix for bug in DirectoryString mask setting. Add support for
13413 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13414 to allow certificate printing to more controllable, additional
13415 'certopt' option to 'x509' to allow new printing options to be
13420 * Clean old EAY MD5 hack from e_os.h.
13424 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13426 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13427 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13429 *Joe Orton, Steve Henson*
13431 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13433 * Fix additional bug revealed by the NISCC test suite:
13435 Stop bug triggering large recursion when presented with
13436 certain ASN.1 tags ([CVE-2003-0851])
13440 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13442 * Fix various bugs revealed by running the NISCC test suite:
13444 Stop out of bounds reads in the ASN1 code when presented with
13445 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13447 If verify callback ignores invalid public key errors don't try to check
13448 certificate signature with the NULL public key.
13452 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13453 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13458 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13459 extra data after the compression methods not only for TLS 1.0
13460 but also for SSL 3.0 (as required by the specification).
13462 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13464 * Change X509_certificate_type() to mark the key as exported/exportable
13465 when it's 512 *bits* long, not 512 bytes.
13469 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13471 * Countermeasure against the Klima-Pokorny-Rosa extension of
13472 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13473 a protocol version number mismatch like a decryption error
13474 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13478 * Turn on RSA blinding by default in the default implementation
13479 to avoid a timing attack. Applications that don't want it can call
13480 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13481 They would be ill-advised to do so in most cases.
13483 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13485 * Change RSA blinding code so that it works when the PRNG is not
13486 seeded (in this case, the secret RSA exponent is abused as
13487 an unpredictable seed -- if it is not unpredictable, there
13488 is no point in blinding anyway). Make RSA blinding thread-safe
13489 by remembering the creator's thread ID in rsa->blinding and
13490 having all other threads use local one-time blinding factors
13491 (this requires more computation than sharing rsa->blinding, but
13492 avoids excessive locking; and if an RSA object is not shared
13493 between threads, blinding will still be very fast).
13497 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13499 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13500 via timing by performing a MAC computation even if incorrect
13501 block cipher padding has been found. This is a countermeasure
13502 against active attacks where the attacker has to distinguish
13503 between bad padding and a MAC verification error. ([CVE-2003-0078])
13505 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13506 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13507 Martin Vuagnoux (EPFL, Ilion)*
13509 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13511 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13512 memory from its contents. This is done with a counter that will
13513 place alternating values in each byte. This can be used to solve
13514 two issues: 1) the removal of calls to memset() by highly optimizing
13515 compilers, and 2) cleansing with other values than 0, since those can
13516 be read through on certain media, for example a swap space on disk.
13520 * Bugfix: client side session caching did not work with external caching,
13521 because the session->cipher setting was not restored when reloading
13522 from the external cache. This problem was masked, when
13523 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13524 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13528 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13529 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13531 *Zeev Lieber <zeev-l@yahoo.com>*
13533 * Undo an undocumented change introduced in 0.9.6e which caused
13534 repeated calls to OpenSSL_add_all_ciphers() and
13535 OpenSSL_add_all_digests() to be ignored, even after calling
13540 * Change the default configuration reader to deal with last line not
13541 being properly terminated.
13545 * Change X509_NAME_cmp() so it applies the special rules on handling
13546 DN values that are of type PrintableString, as well as RDNs of type
13547 emailAddress where the value has the type ia5String.
13549 *stefank@valicert.com via Richard Levitte*
13551 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13552 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13553 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13554 the bitwise-OR of the two for use by the majority of applications
13555 wanting this behaviour, and update the docs. The documented
13556 behaviour and actual behaviour were inconsistent and had been
13557 changing anyway, so this is more a bug-fix than a behavioural
13560 *Geoff Thorpe, diagnosed by Nadav Har'El*
13562 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13563 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13567 * Fix initialization code race conditions in
13568 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13569 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13570 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13571 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13572 ssl2_get_cipher_by_char(),
13573 ssl3_get_cipher_by_char().
13575 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13577 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13578 the cached sessions are flushed, as the remove_cb() might use ex_data
13579 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13580 (see [openssl.org #212]).
13582 *Geoff Thorpe, Lutz Jaenicke*
13584 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13585 length, instead of the encoding length to d2i_ASN1_OBJECT.
13589 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13591 * [In 0.9.6g-engine release:]
13592 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13594 *Lynn Gazis <lgazis@rainbow.com>*
13596 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13598 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13599 and get fix the header length calculation.
13600 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13601 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13603 * Use proper error handling instead of 'assertions' in buffer
13604 overflow checks added in 0.9.6e. This prevents DoS (the
13605 assertions could call abort()).
13607 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13609 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13611 * Add various sanity checks to asn1_get_length() to reject
13612 the ASN1 length bytes if they exceed sizeof(long), will appear
13613 negative or the content length exceeds the length of the
13616 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13618 * Fix cipher selection routines: ciphers without encryption had no flags
13619 for the cipher strength set and where therefore not handled correctly
13620 by the selection routines (PR #130).
13624 * Fix EVP_dsa_sha macro.
13629 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13630 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13631 that was added in OpenSSL 0.9.6d.
13633 As the countermeasure turned out to be incompatible with some
13634 broken SSL implementations, the new option is part of SSL_OP_ALL.
13635 SSL_OP_ALL is usually employed when compatibility with weird SSL
13636 implementations is desired (e.g. '-bugs' option to 's_client' and
13637 's_server'), so the new option is automatically set in many
13642 * Changes in security patch:
13644 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13645 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13646 Air Force Materiel Command, USAF, under agreement number
13649 * Add various sanity checks to asn1_get_length() to reject
13650 the ASN1 length bytes if they exceed sizeof(long), will appear
13651 negative or the content length exceeds the length of the
13652 supplied buffer. ([CVE-2002-0659])
13654 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13656 * Assertions for various potential buffer overflows, not known to
13657 happen in practice.
13659 *Ben Laurie (CHATS)*
13661 * Various temporary buffers to hold ASCII versions of integers were
13662 too small for 64 bit platforms. ([CVE-2002-0655])
13663 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13665 * Remote buffer overflow in SSL3 protocol - an attacker could
13666 supply an oversized session ID to a client. ([CVE-2002-0656])
13668 *Ben Laurie (CHATS)*
13670 * Remote buffer overflow in SSL2 protocol - an attacker could
13671 supply an oversized client master key. ([CVE-2002-0656])
13673 *Ben Laurie (CHATS)*
13675 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13677 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13678 encoded as NULL) with id-dsa-with-sha1.
13680 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13682 * Check various `X509_...()` return values in apps/req.c.
13684 *Nils Larsch <nla@trustcenter.de>*
13686 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13687 an end-of-file condition would erroneously be flagged, when the CRLF
13688 was just at the end of a processed block. The bug was discovered when
13689 processing data through a buffering memory BIO handing the data to a
13690 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13691 <ptsekov@syntrex.com> and Nedelcho Stanev.
13695 * Implement a countermeasure against a vulnerability recently found
13696 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13697 before application data chunks to avoid the use of known IVs
13698 with data potentially chosen by the attacker.
13702 * Fix length checks in ssl3_get_client_hello().
13706 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13707 to prevent ssl3_read_internal() from incorrectly assuming that
13708 ssl3_read_bytes() found application data while handshake
13709 processing was enabled when in fact s->s3->in_read_app_data was
13710 merely automatically cleared during the initial handshake.
13712 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13714 * Fix object definitions for Private and Enterprise: they were not
13715 recognized in their shortname (=lowercase) representation. Extend
13716 obj_dat.pl to issue an error when using undefined keywords instead
13717 of silently ignoring the problem (Svenning Sorensen
13718 <sss@sss.dnsalias.net>).
13722 * Fix DH_generate_parameters() so that it works for 'non-standard'
13723 generators, i.e. generators other than 2 and 5. (Previously, the
13724 code did not properly initialise the 'add' and 'rem' values to
13725 BN_generate_prime().)
13727 In the new general case, we do not insist that 'generator' is
13728 actually a primitive root: This requirement is rather pointless;
13729 a generator of the order-q subgroup is just as good, if not
13734 * Map new X509 verification errors to alerts. Discovered and submitted by
13735 Tom Wu <tom@arcot.com>.
13739 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13740 returning non-zero before the data has been completely received
13741 when using non-blocking I/O.
13743 *Bodo Moeller; problem pointed out by John Hughes*
13745 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13747 *Ben Laurie, Lutz Jaenicke*
13749 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13750 Yoram Zahavi <YoramZ@gilian.com>).
13754 * Add information about CygWin 1.3 and on, and preserve proper
13755 configuration for the versions before that.
13757 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13759 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13760 check whether we deal with a copy of a session and do not delete from
13761 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13762 <izhar@checkpoint.com>.
13766 * Do not store session data into the internal session cache, if it
13767 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13768 flag is set). Proposed by Aslam <aslam@funk.com>.
13772 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13777 * [In 0.9.6d-engine release:]
13778 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13780 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13782 * Add the configuration target linux-s390x.
13784 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13786 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13787 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13788 variable as an indication that a ClientHello message has been
13789 received. As the flag value will be lost between multiple
13790 invocations of ssl3_accept when using non-blocking I/O, the
13791 function may not be aware that a handshake has actually taken
13792 place, thus preventing a new session from being added to the
13795 To avoid this problem, we now set s->new_session to 2 instead of
13796 using a local variable.
13798 *Lutz Jaenicke, Bodo Moeller*
13800 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13801 if the SSL_R_LENGTH_MISMATCH error is detected.
13803 *Geoff Thorpe, Bodo Moeller*
13805 * New 'shared_ldflag' column in Configure platform table.
13809 * Fix EVP_CIPHER_mode macro.
13811 *"Dan S. Camper" <dan@bti.net>*
13813 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13814 type, we must throw them away by setting rr->length to 0.
13816 *D P Chang <dpc@qualys.com>*
13818 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13820 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13821 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13822 worked incorrectly for those cases where range = `10..._2` and
13823 `3*range` is two bits longer than range.)
13827 * Only add signing time to PKCS7 structures if it is not already
13832 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13833 OBJ_ld_ce should be OBJ_id_ce.
13834 Also some ip-pda OIDs in crypto/objects/objects.txt were
13835 incorrect (cf. RFC 3039).
13837 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13839 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13840 returns early because it has nothing to do.
13842 *Andy Schneider <andy.schneider@bjss.co.uk>*
13844 * [In 0.9.6c-engine release:]
13845 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13847 *Andy Schneider <andy.schneider@bjss.co.uk>*
13849 * [In 0.9.6c-engine release:]
13850 Add support for Cryptographic Appliance's keyserver technology.
13851 (Use engine 'keyclient')
13853 *Cryptographic Appliances and Geoff Thorpe*
13855 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13856 is called via tools/c89.sh because arguments have to be
13857 rearranged (all '-L' options must appear before the first object
13860 *Richard Shapiro <rshapiro@abinitio.com>*
13862 * [In 0.9.6c-engine release:]
13863 Add support for Broadcom crypto accelerator cards, backported
13866 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13868 * [In 0.9.6c-engine release:]
13869 Add support for SureWare crypto accelerator cards from
13870 Baltimore Technologies. (Use engine 'sureware')
13872 *Baltimore Technologies and Mark Cox*
13874 * [In 0.9.6c-engine release:]
13875 Add support for crypto accelerator cards from Accelerated
13876 Encryption Processing, www.aep.ie. (Use engine 'aep')
13878 *AEP Inc. and Mark Cox*
13880 * Add a configuration entry for gcc on UnixWare.
13882 *Gary Benson <gbenson@redhat.com>*
13884 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13885 messages are stored in a single piece (fixed-length part and
13886 variable-length part combined) and fix various bugs found on the way.
13890 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13891 instead. BIO_gethostbyname() does not know what timeouts are
13892 appropriate, so entries would stay in cache even when they have
13894 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13896 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13897 faced with a pathologically small ClientHello fragment that does
13898 not contain client_version: Instead of aborting with an error,
13899 simply choose the highest available protocol version (i.e.,
13900 TLS 1.0 unless it is disabled). In practice, ClientHello
13901 messages are never sent like this, but this change gives us
13902 strictly correct behaviour at least for TLS.
13906 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13907 never resets s->method to s->ctx->method when called from within
13908 one of the SSL handshake functions.
13910 *Bodo Moeller; problem pointed out by Niko Baric*
13912 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13913 (sent using the client's version number) if client_version is
13914 smaller than the protocol version in use. Also change
13915 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13916 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13917 the client will at least see that alert.
13921 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13926 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13927 client receives HelloRequest while in a handshake.
13929 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13931 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13932 should end in 'break', not 'goto end' which circumvents various
13933 cleanups done in state SSL_ST_OK. But session related stuff
13934 must be disabled for SSL_ST_OK in the case that we just sent a
13937 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13938 before just sending a HelloRequest.
13940 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13942 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13943 reveal whether illegal block cipher padding was found or a MAC
13944 verification error occurred. (Neither SSLerr() codes nor alerts
13945 are directly visible to potential attackers, but the information
13946 may leak via logfiles.)
13948 Similar changes are not required for the SSL 2.0 implementation
13949 because the number of padding bytes is sent in clear for SSL 2.0,
13950 and the extra bytes are just ignored. However ssl/s2_pkt.c
13951 failed to verify that the purported number of padding bytes is in
13956 * Add OpenUNIX-8 support including shared libraries
13957 (Boyd Lynn Gerber <gerberb@zenez.com>).
13961 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13962 'wristwatch attack' using huge encoding parameters (cf.
13963 James H. Manger's CRYPTO 2001 paper). Note that the
13964 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13965 encoding parameters and hence was not vulnerable.
13969 * BN_sqr() bug fix.
13971 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13973 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13974 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13975 followed by modular reduction.
13977 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13979 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13980 equivalent based on BN_pseudo_rand() instead of BN_rand().
13984 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13985 This function was broken, as the check for a new client hello message
13986 to handle SGC did not allow these large messages.
13987 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
13991 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
13995 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
13996 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14000 * Rework the configuration and shared library support for Tru64 Unix.
14001 The configuration part makes use of modern compiler features and
14002 still retains old compiler behavior for those that run older versions
14003 of the OS. The shared library support part includes a variant that
14004 uses the RPATH feature, and is available through the special
14005 configuration target "alpha-cc-rpath", which will never be selected
14008 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14010 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14011 with the same message size as in ssl3_get_certificate_request().
14012 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14013 messages might inadvertently be reject as too long.
14015 *Petr Lampa <lampa@fee.vutbr.cz>*
14017 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14021 * Modified SSL library such that the verify_callback that has been set
14022 specifically for an SSL object with SSL_set_verify() is actually being
14023 used. Before the change, a verify_callback set with this function was
14024 ignored and the verify_callback() set in the SSL_CTX at the time of
14025 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14026 to allow the necessary settings.
14030 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14031 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14032 done automatically (in contradiction to the requirements of the C
14033 standard). This made problems when used from OpenSSH.
14037 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14038 dh->length and always used
14040 BN_rand_range(priv_key, dh->p).
14042 BN_rand_range() is not necessary for Diffie-Hellman, and this
14043 specific range makes Diffie-Hellman unnecessarily inefficient if
14044 dh->length (recommended exponent length) is much smaller than the
14045 length of dh->p. We could use BN_rand_range() if the order of
14046 the subgroup was stored in the DH structure, but we only have
14051 BN_rand(priv_key, l, ...)
14053 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14060 RSA_eay_public_encrypt
14061 RSA_eay_private_decrypt
14062 RSA_eay_private_encrypt (signing)
14063 RSA_eay_public_decrypt (signature verification)
14065 (default implementations for RSA_public_encrypt,
14066 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14067 always reject numbers >= n.
14071 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14072 to synchronize access to 'locking_thread'. This is necessary on
14073 systems where access to 'locking_thread' (an 'unsigned long'
14074 variable) is not atomic.
14078 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14079 *before* setting the 'crypto_lock_rand' flag. The previous code had
14080 a race condition if 0 is a valid thread ID.
14082 *Travis Vitek <vitek@roguewave.com>*
14084 * Add support for shared libraries under Irix.
14086 *Albert Chin-A-Young <china@thewrittenword.com>*
14088 * Add configuration option to build on Linux on both big-endian and
14089 little-endian MIPS.
14091 *Ralf Baechle <ralf@uni-koblenz.de>*
14093 * Add the possibility to create shared libraries on HP-UX.
14097 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14099 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14100 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14101 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14102 PRNG state recovery was possible based on the output of
14103 one PRNG request appropriately sized to gain knowledge on
14104 'md' followed by enough consecutive 1-byte PRNG requests
14105 to traverse all of 'state'.
14107 1. When updating 'md_local' (the current thread's copy of 'md')
14108 during PRNG output generation, hash all of the previous
14109 'md_local' value, not just the half used for PRNG output.
14111 2. Make the number of bytes from 'state' included into the hash
14112 independent from the number of PRNG bytes requested.
14114 The first measure alone would be sufficient to avoid
14115 Markku-Juhani's attack. (Actually it had never occurred
14116 to me that the half of 'md_local' used for chaining was the
14117 half from which PRNG output bytes were taken -- I had always
14118 assumed that the secret half would be used.) The second
14119 measure makes sure that additional data from 'state' is never
14120 mixed into 'md_local' in small portions; this heuristically
14121 further strengthens the PRNG.
14125 * Fix crypto/bn/asm/mips3.s.
14129 * When only the key is given to "enc", the IV is undefined. Print out
14130 an error message in this case.
14134 * Handle special case when X509_NAME is empty in X509 printing routines.
14138 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14139 positive and less than q.
14143 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14144 used: it isn't thread safe and the add_lock_callback should handle
14147 *Paul Rose <Paul.Rose@bridge.com>*
14149 * Verify that incoming data obeys the block size in
14150 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14156 *Ulf Möller, Bodo Möller*
14158 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14159 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14160 when fixing the server behaviour for backwards-compatible 'client
14161 hello' messages. (Note that the attack is impractical against
14162 SSL 3.0 and TLS 1.0 anyway because length and version checking
14163 means that the probability of guessing a valid ciphertext is
14164 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14167 Before 0.9.5, the countermeasure (hide the error by generating a
14168 random 'decryption result') did not work properly because
14169 ERR_clear_error() was missing, meaning that SSL_get_error() would
14170 detect the supposedly ignored error.
14172 Both problems are now fixed.
14176 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14177 (previously it was 1024).
14181 * Fix for compatibility mode trust settings: ignore trust settings
14182 unless some valid trust or reject settings are present.
14186 * Fix for blowfish EVP: its a variable length cipher.
14190 * Fix various bugs related to DSA S/MIME verification. Handle missing
14191 parameters in DSA public key structures and return an error in the
14192 DSA routines if parameters are absent.
14196 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14197 in the current directory if neither $RANDFILE nor $HOME was set.
14198 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14199 caused some confusion to Windows users who haven't defined $HOME.
14200 Thus RAND_file_name() is changed again: e_os.h can define a
14201 DEFAULT_HOME, which will be used if $HOME is not set.
14202 For Windows, we use "C:"; on other platforms, we still require
14203 environment variables.
14205 * Move 'if (!initialized) RAND_poll()' into regions protected by
14206 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14207 having multiple threads call RAND_poll() concurrently.
14211 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14212 combination of a flag and a thread ID variable.
14213 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14214 flag), *other* threads can enter ssleay_add_bytes without obeying
14215 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14216 that they do not hold after the first thread unsets add_do_not_lock).
14220 * Change bctest again: '-x' expressions are not available in all
14221 versions of 'test'.
14225 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14227 * Fix a couple of memory leaks in PKCS7_dataDecode()
14229 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14231 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14232 the default extension for executables, if any. Also, make the perl
14233 scripts that use symlink() to test if it really exists and use "cp"
14234 if it doesn't. All this made OpenSSL compilable and installable in
14239 * Fix for asn1_GetSequence() for indefinite length constructed data.
14240 If SEQUENCE is length is indefinite just set c->slen to the total
14241 amount of data available.
14243 *Steve Henson, reported by shige@FreeBSD.org*
14245 *This change does not apply to 0.9.7.*
14247 * Change bctest to avoid here-documents inside command substitution
14248 (workaround for FreeBSD /bin/sh bug).
14249 For compatibility with Ultrix, avoid shell functions (introduced
14250 in the bctest version that searches along $PATH).
14254 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14255 with des_encrypt() defined on some operating systems, like Solaris
14260 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14261 On the Importance of Eliminating Errors in Cryptographic
14262 Computations, J. Cryptology 14 (2001) 2, 101-119,
14263 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14267 * MIPS assembler BIGNUM division bug fix.
14271 * Disabled incorrect Alpha assembler code.
14275 * Fix PKCS#7 decode routines so they correctly update the length
14276 after reading an EOC for the EXPLICIT tag.
14280 *This change does not apply to 0.9.7.*
14282 * Fix bug in PKCS#12 key generation routines. This was triggered
14283 if a 3DES key was generated with a 0 initial byte. Include
14284 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14285 (but broken) behaviour.
14289 * Enhance bctest to search for a working bc along $PATH and print
14292 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14294 * Fix memory leaks in err.c: free err_data string if necessary;
14295 don't write to the wrong index in ERR_set_error_data.
14299 * Implement ssl23_peek (analogous to ssl23_read), which previously
14304 * Replace rdtsc with `_emit` statements for VC++ version 5.
14306 *Jeremy Cooper <jeremy@baymoo.org>*
14308 * Make it possible to reuse SSLv2 sessions.
14312 * In copy_email() check for >= 0 as a return value for
14313 X509_NAME_get_index_by_NID() since 0 is a valid index.
14315 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14317 * Avoid coredump with unsupported or invalid public keys by checking if
14318 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14319 PKCS7_verify() fails with non detached data.
14323 * Don't use getenv in library functions when run as setuid/setgid.
14324 New function OPENSSL_issetugid().
14328 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14329 due to incorrect handling of multi-threading:
14331 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14333 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14335 3. Count how many times MemCheck_off() has been called so that
14336 nested use can be treated correctly. This also avoids
14337 inband-signalling in the previous code (which relied on the
14338 assumption that thread ID 0 is impossible).
14342 * Add "-rand" option also to s_client and s_server.
14346 * Fix CPU detection on Irix 6.x.
14347 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14348 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14350 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14355 *This change does not apply to 0.9.7.*
14357 * Use the cached encoding of an X509_NAME structure rather than
14358 copying it. This is apparently the reason for the libsafe "errors"
14359 but the code is actually correct.
14363 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14364 Bleichenbacher's DSA attack.
14365 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14366 to be set and top=0 forces the highest bit to be set; top=-1 is new
14367 and leaves the highest bit random.
14369 *Ulf Moeller, Bodo Moeller*
14371 * In the `NCONF_...`-based implementations for `CONF_...` queries
14372 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14373 a temporary CONF structure with the data component set to NULL
14374 (which gives segmentation faults in lh_retrieve).
14375 Instead, use NULL for the CONF pointer in CONF_get_string and
14376 CONF_get_number (which may use environment variables) and directly
14377 return NULL from CONF_get_section.
14381 * Fix potential buffer overrun for EBCDIC.
14385 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14386 keyUsage if basicConstraints absent for a CA.
14390 * Make SMIME_write_PKCS7() write mail header values with a format that
14391 is more generally accepted (no spaces before the semicolon), since
14392 some programs can't parse those values properly otherwise. Also make
14393 sure BIO's that break lines after each write do not create invalid
14398 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14399 macros previously used would not encode an empty SEQUENCE OF
14400 and break the signature.
14404 *This change does not apply to 0.9.7.*
14406 * Zero the premaster secret after deriving the master secret in
14411 * Add some EVP_add_digest_alias registrations (as found in
14412 OpenSSL_add_all_digests()) to SSL_library_init()
14413 aka OpenSSL_add_ssl_algorithms(). This provides improved
14414 compatibility with peers using X.509 certificates
14415 with unconventional AlgorithmIdentifier OIDs.
14419 * Fix for Irix with NO_ASM.
14421 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14423 * ./config script fixes.
14425 *Ulf Moeller, Richard Levitte*
14427 * Fix 'openssl passwd -1'.
14431 * Change PKCS12_key_gen_asc() so it can cope with non null
14432 terminated strings whose length is passed in the passlen
14433 parameter, for example from PEM callbacks. This was done
14434 by adding an extra length parameter to asc2uni().
14436 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14438 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14439 call failed, free the DSA structure.
14443 * Fix to uni2asc() to cope with zero length Unicode strings.
14444 These are present in some PKCS#12 files.
14448 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14449 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14450 when writing a 32767 byte record.
14452 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14454 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14455 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14457 (RSA objects have a reference count access to which is protected
14458 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14459 so they are meant to be shared between threads.)
14460 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14461 "Reddie, Steven" <Steven.Reddie@ca.com>*
14463 * Fix a deadlock in CRYPTO_mem_leaks().
14467 * Use better test patterns in bntest.
14471 * rand_win.c fix for Borland C.
14475 * BN_rshift bugfix for n == 0.
14479 * Add a 'bctest' script that checks for some known 'bc' bugs
14480 so that 'make test' does not abort just because 'bc' is broken.
14484 * Store verify_result within SSL_SESSION also for client side to
14485 avoid potential security hole. (Re-used sessions on the client side
14486 always resulted in verify_result==X509_V_OK, not using the original
14487 result of the server certificate verification.)
14491 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14492 SSL3_RT_APPLICATION_DATA, return 0.
14493 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14498 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14499 releases, have been re-implemented by renaming the previous
14500 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14501 and ssl3_read_internal, respectively, and adding 'peek' parameters
14502 to them. The new ssl[23]_{read,peek} functions are calls to
14503 ssl[23]_read_internal with the 'peek' flag set appropriately.
14504 A 'peek' parameter has also been added to ssl3_read_bytes, which
14505 does the actual work for ssl3_read_internal.
14509 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14510 the method-specific "init()" handler. Also clean up ex_data after
14511 calling the method-specific "finish()" handler. Previously, this was
14512 happening the other way round.
14516 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14517 The previous value, 12, was not always sufficient for BN_mod_exp().
14521 * Make sure that shared libraries get the internal name engine with
14522 the full version number and not just 0. This should mark the
14523 shared libraries as not backward compatible. Of course, this should
14524 be changed again when we can guarantee backward binary compatibility.
14528 * Fix typo in get_cert_by_subject() in by_dir.c
14530 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14532 * Rework the system to generate shared libraries:
14534 - Make note of the expected extension for the shared libraries and
14535 if there is a need for symbolic links from for example libcrypto.so.0
14536 to libcrypto.so.0.9.7. There is extended info in Configure for
14539 - Make as few rebuilds of the shared libraries as possible.
14541 - Still avoid linking the OpenSSL programs with the shared libraries.
14543 - When installing, install the shared libraries separately from the
14548 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14550 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14551 and not in SSL_clear because the latter is also used by the
14552 accept/connect functions; previously, the settings made by
14553 SSL_set_read_ahead would be lost during the handshake.
14555 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14557 * Correct util/mkdef.pl to be selective about disabled algorithms.
14558 Previously, it would create entries for disabled algorithms no
14563 * Added several new manual pages for SSL_* function.
14567 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14569 * In ssl23_get_client_hello, generate an error message when faced
14570 with an initial SSL 3.0/TLS record that is too small to contain the
14571 first two bytes of the ClientHello message, i.e. client_version.
14572 (Note that this is a pathologic case that probably has never happened
14573 in real life.) The previous approach was to use the version number
14574 from the record header as a substitute; but our protocol choice
14575 should not depend on that one because it is not authenticated
14576 by the Finished messages.
14580 * More robust randomness gathering functions for Windows.
14582 *Jeffrey Altman <jaltman@columbia.edu>*
14584 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14585 not set then we don't setup the error code for issuer check errors
14586 to avoid possibly overwriting other errors which the callback does
14587 handle. If an application does set the flag then we assume it knows
14588 what it is doing and can handle the new informational codes
14593 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14594 a general "ANY" type, as such it should be able to decode anything
14595 including tagged types. However it didn't check the class so it would
14596 wrongly interpret tagged types in the same way as their universal
14597 counterpart and unknown types were just rejected. Changed so that the
14598 tagged and unknown types are handled in the same way as a SEQUENCE:
14599 that is the encoding is stored intact. There is also a new type
14600 "V_ASN1_OTHER" which is used when the class is not universal, in this
14601 case we have no idea what the actual type is so we just lump them all
14606 * On VMS, stdout may very well lead to a file that is written to
14607 in a record-oriented fashion. That means that every write() will
14608 write a separate record, which will be read separately by the
14609 programs trying to read from it. This can be very confusing.
14611 The solution is to put a BIO filter in the way that will buffer
14612 text until a linefeed is reached, and then write everything a
14613 line at a time, so every record written will be an actual line,
14614 not chunks of lines and not (usually doesn't happen, but I've
14615 seen it once) several lines in one record. BIO_f_linebuffer() is
14618 Currently, it's a VMS-only method, because that's where it has
14619 been tested well enough.
14623 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14624 it can return incorrect results.
14625 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14626 but it was in 0.9.6-beta[12].)
14630 * Disable the check for content being present when verifying detached
14631 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14632 include zero length content when signing messages.
14636 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14637 BIO_ctrl (for BIO pairs).
14641 * Add DSO method for VMS.
14645 * Bug fix: Montgomery multiplication could produce results with the
14650 * Add RPM specification openssl.spec and modify it to build three
14651 packages. The default package contains applications, application
14652 documentation and run-time libraries. The devel package contains
14653 include files, static libraries and function documentation. The
14654 doc package contains the contents of the doc directory. The original
14655 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14659 * Add a large number of documentation files for many SSL routines.
14661 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14663 * Add a configuration entry for Sony News 4.
14665 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14667 * Don't set the two most significant bits to one when generating a
14668 random number < q in the DSA library.
14672 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14673 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14674 the underlying transport is blocking) if a handshake took place.
14675 (The default behaviour is needed by applications such as s_client
14676 and s_server that use select() to determine when to use SSL_read;
14677 but for applications that know in advance when to expect data, it
14678 just makes things more complicated.)
14682 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14687 * Add a few more EBCDIC conditionals that make `req` and `x509`
14688 work better on such systems.
14690 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14692 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14693 Update PKCS12_parse() so it copies the friendlyName and the
14694 keyid to the certificates aux info.
14698 * Fix bug in PKCS7_verify() which caused an infinite loop
14699 if there was more than one signature.
14701 *Sven Uszpelkat <su@celocom.de>*
14703 * Major change in util/mkdef.pl to include extra information
14704 about each symbol, as well as presenting variables as well
14705 as functions. This change means that there's n more need
14706 to rebuild the .num files when some algorithms are excluded.
14710 * Allow the verify time to be set by an application,
14711 rather than always using the current time.
14715 * Phase 2 verify code reorganisation. The certificate
14716 verify code now looks up an issuer certificate by a
14717 number of criteria: subject name, authority key id
14718 and key usage. It also verifies self signed certificates
14719 by the same criteria. The main comparison function is
14720 X509_check_issued() which performs these checks.
14722 Lot of changes were necessary in order to support this
14723 without completely rewriting the lookup code.
14725 Authority and subject key identifier are now cached.
14727 The LHASH 'certs' is X509_STORE has now been replaced
14728 by a STACK_OF(X509_OBJECT). This is mainly because an
14729 LHASH can't store or retrieve multiple objects with
14730 the same hash value.
14732 As a result various functions (which were all internal
14733 use only) have changed to handle the new X509_STORE
14734 structure. This will break anything that messed round
14735 with X509_STORE internally.
14737 The functions X509_STORE_add_cert() now checks for an
14738 exact match, rather than just subject name.
14740 The X509_STORE API doesn't directly support the retrieval
14741 of multiple certificates matching a given criteria, however
14742 this can be worked round by performing a lookup first
14743 (which will fill the cache with candidate certificates)
14744 and then examining the cache for matches. This is probably
14745 the best we can do without throwing out X509_LOOKUP
14746 entirely (maybe later...).
14748 The X509_VERIFY_CTX structure has been enhanced considerably.
14750 All certificate lookup operations now go via a get_issuer()
14751 callback. Although this currently uses an X509_STORE it
14752 can be replaced by custom lookups. This is a simple way
14753 to bypass the X509_STORE hackery necessary to make this
14754 work and makes it possible to use more efficient techniques
14755 in future. A very simple version which uses a simple
14756 STACK for its trusted certificate store is also provided
14757 using X509_STORE_CTX_trusted_stack().
14759 The verify_cb() and verify() callbacks now have equivalents
14760 in the X509_STORE_CTX structure.
14762 X509_STORE_CTX also has a 'flags' field which can be used
14763 to customise the verify behaviour.
14767 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14768 excludes S/MIME capabilities.
14772 * When a certificate request is read in keep a copy of the
14773 original encoding of the signed data and use it when outputting
14774 again. Signatures then use the original encoding rather than
14775 a decoded, encoded version which may cause problems if the
14776 request is improperly encoded.
14780 * For consistency with other BIO_puts implementations, call
14781 buffer_write(b, ...) directly in buffer_puts instead of calling
14784 In BIO_puts, increment b->num_write as in BIO_write.
14786 *Peter.Sylvester@EdelWeb.fr*
14788 * Fix BN_mul_word for the case where the word is 0. (We have to use
14789 BN_zero, we may not return a BIGNUM with an array consisting of
14790 words set to zero.)
14794 * Avoid calling abort() from within the library when problems are
14795 detected, except if preprocessor symbols have been defined
14796 (such as REF_CHECK, BN_DEBUG etc.).
14800 * New openssl application 'rsautl'. This utility can be
14801 used for low level RSA operations. DER public key
14802 BIO/fp routines also added.
14806 * New Configure entry and patches for compiling on QNX 4.
14808 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14810 * A demo state-machine implementation was sponsored by
14811 Nuron (<http://www.nuron.com/>) and is now available in
14812 demos/state_machine.
14816 * New options added to the 'dgst' utility for signature
14817 generation and verification.
14821 * Unrecognized PKCS#7 content types are now handled via a
14822 catch all ASN1_TYPE structure. This allows unsupported
14823 types to be stored as a "blob" and an application can
14824 encode and decode it manually.
14828 * Fix various signed/unsigned issues to make a_strex.c
14829 compile under VC++.
14831 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14833 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14834 length if passed a buffer. ASN1_INTEGER_to_BN failed
14835 if passed a NULL BN and its argument was negative.
14837 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14839 * Modification to PKCS#7 encoding routines to output definite
14840 length encoding. Since currently the whole structures are in
14841 memory there's not real point in using indefinite length
14842 constructed encoding. However if OpenSSL is compiled with
14843 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14847 * Added BIO_vprintf() and BIO_vsnprintf().
14851 * Added more prefixes to parse for in the strings written
14852 through a logging bio, to cover all the levels that are available
14853 through syslog. The prefixes are now:
14855 PANIC, EMERG, EMR => LOG_EMERG
14856 ALERT, ALR => LOG_ALERT
14857 CRIT, CRI => LOG_CRIT
14858 ERROR, ERR => LOG_ERR
14859 WARNING, WARN, WAR => LOG_WARNING
14860 NOTICE, NOTE, NOT => LOG_NOTICE
14861 INFO, INF => LOG_INFO
14862 DEBUG, DBG => LOG_DEBUG
14864 and as before, if none of those prefixes are present at the
14865 beginning of the string, LOG_ERR is chosen.
14867 On Win32, the `LOG_*` levels are mapped according to this:
14869 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14870 LOG_WARNING => EVENTLOG_WARNING_TYPE
14871 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14875 * Made it possible to reconfigure with just the configuration
14876 argument "reconf" or "reconfigure". The command line arguments
14877 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14878 and are retrieved from there when reconfiguring.
14884 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14886 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14890 * The obj_dat.pl script was messing up the sorting of object
14891 names. The reason was that it compared the quoted version
14892 of strings as a result "OCSP" > "OCSP Signing" because
14893 " > SPACE. Changed script to store unquoted versions of
14894 names and add quotes on output. It was also omitting some
14895 names from the lookup table if they were given a default
14896 value (that is if SN is missing it is given the same
14897 value as LN and vice versa), these are now added on the
14898 grounds that if an object has a name we should be able to
14899 look it up. Finally added warning output when duplicate
14900 short or long names are found.
14904 * Changes needed for Tandem NSK.
14906 *Scott Uroff <scott@xypro.com>*
14908 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14909 RSA_padding_check_SSLv23(), special padding was never detected
14910 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14911 version rollback attacks was not effective.
14913 In s23_clnt.c, don't use special rollback-attack detection padding
14914 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14915 client; similarly, in s23_srvr.c, don't do the rollback check if
14916 SSL 2.0 is the only protocol enabled in the server.
14920 * Make it possible to get hexdumps of unprintable data with 'openssl
14921 asn1parse'. By implication, the functions ASN1_parse_dump() and
14922 BIO_dump_indent() are added.
14926 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14927 these print out strings and name structures based on various
14928 flags including RFC2253 support and proper handling of
14929 multibyte characters. Added options to the 'x509' utility
14930 to allow the various flags to be set.
14934 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14935 Also change the functions X509_cmp_current_time() and
14936 X509_gmtime_adj() work with an ASN1_TIME structure,
14937 this will enable certificates using GeneralizedTime in validity
14938 dates to be checked.
14942 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14943 negative public key encodings) on by default,
14944 NO_NEG_PUBKEY_BUG can be set to disable it.
14948 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14949 content octets. An i2c_ASN1_OBJECT is unnecessary because
14950 the encoding can be trivially obtained from the structure.
14954 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14955 not read locks (`CRYPTO_r_[un]lock`).
14959 * A first attempt at creating official support for shared
14960 libraries through configuration. I've kept it so the
14961 default is static libraries only, and the OpenSSL programs
14962 are always statically linked for now, but there are
14963 preparations for dynamic linking in place.
14964 This has been tested on Linux and Tru64.
14968 * Randomness polling function for Win9x, as described in:
14969 Peter Gutmann, Software Generation of Practically Strong
14974 * Fix so PRNG is seeded in req if using an already existing
14979 * New options to smime application. -inform and -outform
14980 allow alternative formats for the S/MIME message including
14981 PEM and DER. The -content option allows the content to be
14982 specified separately. This should allow things like Netscape
14983 form signing output easier to verify.
14987 * Fix the ASN1 encoding of tags using the 'long form'.
14991 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
14992 STRING types. These convert content octets to and from the
14993 underlying type. The actual tag and length octets are
14994 already assumed to have been read in and checked. These
14995 are needed because all other string types have virtually
14996 identical handling apart from the tag. By having versions
14997 of the ASN1 functions that just operate on content octets
14998 IMPLICIT tagging can be handled properly. It also allows
14999 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15000 and ASN1_INTEGER are identical apart from the tag.
15004 * Change the handling of OID objects as follows:
15006 - New object identifiers are inserted in objects.txt, following
15007 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15008 - objects.pl is used to process obj_mac.num and create a new
15010 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15013 This is currently kind of a hack, and the perl code in objects.pl
15014 isn't very elegant, but it works as I intended. The simplest way
15015 to check that it worked correctly is to look in obj_dat.h and
15016 check the array nid_objs and make sure the objects haven't moved
15017 around (this is important!). Additions are OK, as well as
15018 consistent name changes.
15022 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15026 * Addition of the command line parameter '-rand file' to 'openssl req'.
15027 The given file adds to whatever has already been seeded into the
15028 random pool through the RANDFILE configuration file option or
15029 environment variable, or the default random state file.
15033 * mkstack.pl now sorts each macro group into lexical order.
15034 Previously the output order depended on the order the files
15035 appeared in the directory, resulting in needless rewriting
15040 * Patches to make OpenSSL compile under Win32 again. Mostly
15041 work arounds for the VC++ problem that it treats func() as
15042 func(void). Also stripped out the parts of mkdef.pl that
15043 added extra typesafe functions: these no longer exist.
15047 * Reorganisation of the stack code. The macros are now all
15048 collected in safestack.h . Each macro is defined in terms of
15049 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15050 DEBUG_SAFESTACK is now handled in terms of function casts,
15051 this has the advantage of retaining type safety without the
15052 use of additional functions. If DEBUG_SAFESTACK is not defined
15053 then the non typesafe macros are used instead. Also modified the
15054 mkstack.pl script to handle the new form. Needs testing to see
15055 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15056 the default if no major problems. Similar behaviour for ASN1_SET_OF
15057 and PKCS12_STACK_OF.
15061 * When some versions of IIS use the 'NET' form of private key the
15062 key derivation algorithm is different. Normally MD5(password) is
15063 used as a 128 bit RC4 key. In the modified case
15064 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15065 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15066 as the old Netscape_RSA functions except they have an additional
15067 'sgckey' parameter which uses the modified algorithm. Also added
15068 an -sgckey command line option to the rsa utility. Thanks to
15069 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15070 algorithm to openssl-dev.
15074 * The evp_local.h macros were using 'c.##kname' which resulted in
15075 invalid expansion on some systems (SCO 5.0.5 for example).
15076 Corrected to 'c.kname'.
15078 *Phillip Porch <root@theporch.com>*
15080 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15081 a STACK of email addresses from a certificate or request, these look
15082 in the subject name and the subject alternative name extensions and
15083 omit any duplicate addresses.
15087 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15088 This makes DSA verification about 2 % faster.
15092 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15093 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15094 plus overhead for 1024 bit moduli).
15095 This makes exponentiations about 0.5 % faster for 1024 bit
15096 exponents (as measured by "openssl speed rsa2048").
15100 * Rename memory handling macros to avoid conflicts with other
15102 Malloc => OPENSSL_malloc
15103 Malloc_locked => OPENSSL_malloc_locked
15104 Realloc => OPENSSL_realloc
15105 Free => OPENSSL_free
15109 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15110 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15114 * CygWin32 support.
15116 *John Jarvie <jjarvie@newsguy.com>*
15118 * The type-safe stack code has been rejigged. It is now only compiled
15119 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15120 by default all type-specific stack functions are "#define"d back to
15121 standard stack functions. This results in more streamlined output
15122 but retains the type-safety checking possibilities of the original
15127 * The STACK code has been cleaned up, and certain type declarations
15128 that didn't make a lot of sense have been brought in line. This has
15129 also involved a cleanup of sorts in safestack.h to more correctly
15130 map type-safe stack functions onto their plain stack counterparts.
15131 This work has also resulted in a variety of "const"ifications of
15132 lots of the code, especially `_cmp` operations which should normally
15133 be prototyped with "const" parameters anyway.
15137 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15138 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15139 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15140 where all of 'md' is used each time the PRNG is used, but 'state'
15141 is used only indexed by a cyclic counter. As entropy may not be
15142 well distributed from the beginning, 'md' is important as a
15143 chaining variable. However, the output function chains only half
15144 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15145 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15146 in all of 'state' being rewritten, with the new values depending
15147 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15151 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15152 the handshake is continued after ssl_verify_cert_chain();
15153 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15154 can lead to 'unexplainable' connection aborts later.
15156 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15158 * Major EVP API cipher revision.
15159 Add hooks for extra EVP features. This allows various cipher
15160 parameters to be set in the EVP interface. Support added for variable
15161 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15162 setting of RC2 and RC5 parameters.
15164 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15167 Remove lots of duplicated code from the EVP library. For example *every*
15168 cipher init() function handles the 'iv' in the same way according to the
15169 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15170 for CFB and OFB modes they zero ctx->num.
15172 New functionality allows removal of S/MIME code RC2 hack.
15174 Most of the routines have the same form and so can be declared in terms
15177 By shifting this to the top level EVP_CipherInit() it can be removed from
15178 all individual ciphers. If the cipher wants to handle IVs or keys
15179 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15182 Change lots of functions like EVP_EncryptUpdate() to now return a
15183 value: although software versions of the algorithms cannot fail
15184 any installed hardware versions can.
15188 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15189 this option is set, tolerate broken clients that send the negotiated
15190 protocol version number instead of the requested protocol version
15195 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15196 i.e. non-zero for export ciphersuites, zero otherwise.
15197 Previous versions had this flag inverted, inconsistent with
15198 rsa_tmp_cb (..._TMP_RSA_CB).
15200 *Bodo Moeller; problem reported by Amit Chopra*
15202 * Add missing DSA library text string. Work around for some IIS
15203 key files with invalid SEQUENCE encoding.
15207 * Add a document (doc/standards.txt) that list all kinds of standards
15208 and so on that are implemented in OpenSSL.
15212 * Enhance c_rehash script. Old version would mishandle certificates
15213 with the same subject name hash and wouldn't handle CRLs at all.
15214 Added -fingerprint option to crl utility, to support new c_rehash
15219 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15223 * Fix for SSL server purpose checking. Server checking was
15224 rejecting certificates which had extended key usage present
15225 but no ssl client purpose.
15227 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15229 * Make PKCS#12 code work with no password. The PKCS#12 spec
15230 is a little unclear about how a blank password is handled.
15231 Since the password in encoded as a BMPString with terminating
15232 double NULL a zero length password would end up as just the
15233 double NULL. However no password at all is different and is
15234 handled differently in the PKCS#12 key generation code. NS
15235 treats a blank password as zero length. MSIE treats it as no
15236 password on export: but it will try both on import. We now do
15237 the same: PKCS12_parse() tries zero length and no password if
15238 the password is set to "" or NULL (NULL is now a valid password:
15239 it wasn't before) as does the pkcs12 application.
15243 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15244 perror when PEM_read_bio_X509_REQ fails, the error message must
15245 be obtained from the error queue.
15249 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15250 it in ERR_remove_state if appropriate, and change ERR_get_state
15251 accordingly to avoid race conditions (this is necessary because
15252 thread_hash is no longer constant once set).
15256 * Bugfix for linux-elf makefile.one.
15260 * RSA_get_default_method() will now cause a default
15261 RSA_METHOD to be chosen if one doesn't exist already.
15262 Previously this was only set during a call to RSA_new()
15263 or RSA_new_method(NULL) meaning it was possible for
15264 RSA_get_default_method() to return NULL.
15268 * Added native name translation to the existing DSO code
15269 that will convert (if the flag to do so is set) filenames
15270 that are sufficiently small and have no path information
15271 into a canonical native form. Eg. "blah" converted to
15272 "libblah.so" or "blah.dll" etc.
15276 * New function ERR_error_string_n(e, buf, len) which is like
15277 ERR_error_string(e, buf), but writes at most 'len' bytes
15278 including the 0 terminator. For ERR_error_string_n, 'buf'
15281 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15283 * CONF library reworked to become more general. A new CONF
15284 configuration file reader "class" is implemented as well as a
15285 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15286 old `CONF_*` functions are still there, but are reimplemented to
15287 work in terms of the new functions. Also, a set of functions
15288 to handle the internal storage of the configuration data is
15289 provided to make it easier to write new configuration file
15290 reader "classes" (I can definitely see something reading a
15291 configuration file in XML format, for example), called `_CONF_*`,
15292 or "the configuration storage API"...
15294 The new configuration file reading functions are:
15296 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15297 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15299 NCONF_default, NCONF_WIN32
15301 NCONF_dump_fp, NCONF_dump_bio
15303 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15304 NCONF_new creates a new CONF object. This works in the same way
15305 as other interfaces in OpenSSL, like the BIO interface.
15306 `NCONF_dump_*` dump the internal storage of the configuration file,
15307 which is useful for debugging. All other functions take the same
15308 arguments as the old `CONF_*` functions with the exception of the
15309 first that must be a `CONF *` instead of a `LHASH *`.
15311 To make it easier to use the new classes with the old `CONF_*` functions,
15312 the function CONF_set_default_method is provided.
15316 * Add '-tls1' option to 'openssl ciphers', which was already
15317 mentioned in the documentation but had not been implemented.
15318 (This option is not yet really useful because even the additional
15319 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15323 * Initial DSO code added into libcrypto for letting OpenSSL (and
15324 OpenSSL-based applications) load shared libraries and bind to
15325 them in a portable way.
15327 *Geoff Thorpe, with contributions from Richard Levitte*
15329 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15331 * Make sure _lrotl and _lrotr are only used with MSVC.
15333 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15334 (the default implementation of RAND_status).
15336 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15337 to '-clrext' (= clear extensions), as intended and documented.
15338 *Bodo Moeller; inconsistency pointed out by Michael Attili
15339 <attili@amaxo.com>*
15341 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15342 was larger than the MD block size.
15344 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15346 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15347 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15348 using the passed key: if the passed key was a private key the result
15349 of X509_print(), for example, would be to print out all the private key
15354 * des_quad_cksum() byte order bug fix.
15355 *Ulf Möller, using the problem description in krb4-0.9.7, where
15356 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15358 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15361 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15363 * For easily testing in shell scripts whether some command
15364 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15365 returns with exit code 0 iff no command of the given name is available.
15366 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15367 the output goes to stdout and nothing is printed to stderr.
15368 Additional arguments are always ignored.
15370 Since for each cipher there is a command of the same name,
15371 the 'no-cipher' compilation switches can be tested this way.
15373 ('openssl no-XXX' is not able to detect pseudo-commands such
15374 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15378 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15382 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15383 is set; it will be thrown away anyway because each handshake creates
15385 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15386 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15387 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15388 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15392 * New s_client option -ign_eof: EOF at stdin is ignored, and
15393 'Q' and 'R' lose their special meanings (quit/renegotiate).
15394 This is part of what -quiet does; unlike -quiet, -ign_eof
15395 does not suppress any output.
15399 * Add compatibility options to the purpose and trust code. The
15400 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15401 accepts a certificate or CA, this was the previous behaviour,
15402 with all the associated security issues.
15404 X509_TRUST_COMPAT is the old trust behaviour: only and
15405 automatically trust self signed roots in certificate store. A
15406 new trust setting X509_TRUST_DEFAULT is used to specify that
15407 a purpose has no associated trust setting and it should instead
15408 use the value in the default purpose.
15412 * Fix the PKCS#8 DSA private key code so it decodes keys again
15413 and fix a memory leak.
15417 * In util/mkerr.pl (which implements 'make errors'), preserve
15418 reason strings from the previous version of the .c file, as
15419 the default to have only downcase letters (and digits) in
15420 automatically generated reasons codes is not always appropriate.
15424 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15425 using strerror. Previously, ERR_reason_error_string() returned
15426 library names as reason strings for SYSerr; but SYSerr is a special
15427 case where small numbers are errno values, not library numbers.
15431 * Add '-dsaparam' option to 'openssl dhparam' application. This
15432 converts DSA parameters into DH parameters. (When creating parameters,
15433 DSA_generate_parameters is used.)
15437 * Include 'length' (recommended exponent length) in C code generated
15438 by 'openssl dhparam -C'.
15442 * The second argument to set_label in perlasm was already being used
15443 so couldn't be used as a "file scope" flag. Moved to third argument
15448 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15449 instead of RAND_bytes for encryption IVs and salts.
15453 * Include RAND_status() into RAND_METHOD instead of implementing
15454 it only for md_rand.c Otherwise replacing the PRNG by calling
15455 RAND_set_rand_method would be impossible.
15459 * Don't let DSA_generate_key() enter an infinite loop if the random
15460 number generation fails.
15464 * New 'rand' application for creating pseudo-random output.
15468 * Added configuration support for Linux/IA64
15470 *Rolf Haberrecker <rolf@suse.de>*
15472 * Assembler module support for Mingw32.
15476 * Shared library support for HPUX (in shlib/).
15478 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15480 * Shared library support for Solaris gcc.
15482 *Lutz Behnke <behnke@trustcenter.de>*
15484 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15486 * PKCS7_encrypt() was adding text MIME headers twice because they
15487 were added manually and by SMIME_crlf_copy().
15491 * In bntest.c don't call BN_rand with zero bits argument.
15493 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15495 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15496 case was implemented. This caused BN_div_recp() to fail occasionally.
15500 * Add an optional second argument to the set_label() in the perl
15501 assembly language builder. If this argument exists and is set
15502 to 1 it signals that the assembler should use a symbol whose
15503 scope is the entire file, not just the current function. This
15504 is needed with MASM which uses the format label:: for this scope.
15506 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15508 * Change the ASN1 types so they are typedefs by default. Before
15509 almost all types were #define'd to ASN1_STRING which was causing
15510 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15515 * Change names of new functions to the new get1/get0 naming
15516 convention: After 'get1', the caller owns a reference count
15517 and has to call `..._free`; 'get0' returns a pointer to some
15518 data structure without incrementing reference counters.
15519 (Some of the existing 'get' functions increment a reference
15520 counter, some don't.)
15521 Similarly, 'set1' and 'add1' functions increase reference
15522 counters or duplicate objects.
15526 * Allow for the possibility of temp RSA key generation failure:
15527 the code used to assume it always worked and crashed on failure.
15531 * Fix potential buffer overrun problem in BIO_printf().
15532 *Ulf Möller, using public domain code by Patrick Powell; problem
15533 pointed out by David Sacerdote <das33@cornell.edu>*
15535 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15536 RAND_egd() and RAND_status(). In the command line application,
15537 the EGD socket can be specified like a seed file using RANDFILE
15542 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15543 Some CAs (e.g. Verisign) distribute certificates in this form.
15547 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15548 list to exclude them. This means that no special compilation option
15549 is needed to use anonymous DH: it just needs to be included in the
15554 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15555 EVP_MD_type. The old functionality is available in a new macro called
15556 EVP_MD_md(). Change code that uses it and update docs.
15560 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15561 where the `void *` argument is replaced by a function pointer argument.
15562 Previously `void *` was abused to point to functions, which works on
15563 many platforms, but is not correct. As these functions are usually
15564 called by macros defined in OpenSSL header files, most source code
15565 should work without changes.
15569 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15570 sections with information on -D... compiler switches used for
15571 compiling the library so that applications can see them. To enable
15572 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15573 must be defined. E.g.,
15574 #define OPENSSL_ALGORITHM_DEFINES
15575 #include <openssl/opensslconf.h>
15576 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15578 *Richard Levitte, Ulf and Bodo Möller*
15580 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15585 * Change the 'other' type in certificate aux info to a STACK_OF
15586 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15587 the required ASN1 format: arbitrary types determined by an OID.
15591 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15592 argument to 'req'. This is not because the function is newer or
15593 better than others it just uses the work 'NEW' in the certificate
15594 request header lines. Some software needs this.
15598 * Reorganise password command line arguments: now passwords can be
15599 obtained from various sources. Delete the PEM_cb function and make
15600 it the default behaviour: i.e. if the callback is NULL and the
15601 usrdata argument is not NULL interpret it as a null terminated pass
15602 phrase. If usrdata and the callback are NULL then the pass phrase
15603 is prompted for as usual.
15607 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15608 the support is automatically enabled. The resulting binaries will
15609 autodetect the card and use it if present.
15611 *Ben Laurie and Compaq Inc.*
15613 * Work around for Netscape hang bug. This sends certificate request
15614 and server done in one record. Since this is perfectly legal in the
15615 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15616 the bugs/SSLv3 entry for more info.
15620 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15624 * Add -rand argument to smime and pkcs12 applications and read/write
15629 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15633 * Add command line password options to the remaining applications.
15637 * Bug fix for BN_div_recp() for numerators with an even number of
15642 * More tests in bntest.c, and changed test_bn output.
15646 * ./config recognizes MacOS X now.
15650 * Bug fix for BN_div() when the first words of num and divisor are
15651 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15655 * Add support for various broken PKCS#8 formats, and command line
15656 options to produce them.
15660 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15661 get temporary BIGNUMs from a BN_CTX.
15665 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15670 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15671 include a #define from the old name to the new. The original intent
15672 was that statically linked binaries could for example just call
15673 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15674 link with digests. This never worked because SSLeay_add_all_digests()
15675 and SSLeay_add_all_ciphers() were in the same source file so calling
15676 one would link with the other. They are now in separate source files.
15680 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15684 * Use a less unusual form of the Miller-Rabin primality test (it used
15685 a binary algorithm for exponentiation integrated into the Miller-Rabin
15686 loop, our standard modexp algorithms are faster).
15690 * Support for the EBCDIC character set completed.
15692 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15694 * Source code cleanups: use const where appropriate, eliminate casts,
15695 use `void *` instead of `char *` in lhash.
15699 * Bugfix: ssl3_send_server_key_exchange was not restartable
15700 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15701 this the server could overwrite ephemeral keys that the client
15706 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15707 using 50 iterations of the Rabin-Miller test.
15709 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15710 iterations of the Rabin-Miller test as required by the appendix
15711 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15712 As BN_is_prime_fasttest includes trial division, DSA parameter
15713 generation becomes much faster.
15715 This implies a change for the callback functions in DSA_is_prime
15716 and DSA_generate_parameters: The callback function is called once
15717 for each positive witness in the Rabin-Miller test, not just
15718 occasionally in the inner loop; and the parameters to the
15719 callback function now provide an iteration count for the outer
15720 loop rather than for the current invocation of the inner loop.
15721 DSA_generate_parameters additionally can call the callback
15722 function with an 'iteration count' of -1, meaning that a
15723 candidate has passed the trial division test (when q is generated
15724 from an application-provided seed, trial division is skipped).
15728 * New function BN_is_prime_fasttest that optionally does trial
15729 division before starting the Rabin-Miller test and has
15730 an additional BN_CTX * argument (whereas BN_is_prime always
15731 has to allocate at least one BN_CTX).
15732 'callback(1, -1, cb_arg)' is called when a number has passed the
15733 trial division stage.
15737 * Fix for bug in CRL encoding. The validity dates weren't being handled
15742 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15746 * New function BN_pseudo_rand().
15750 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15751 bignum version of BN_from_montgomery() with the working code from
15752 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15757 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15758 made it impossible to use the same SSL_SESSION data structure in
15759 SSL2 clients in multiple threads.
15763 * The return value of RAND_load_file() no longer counts bytes obtained
15764 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15765 to seed the PRNG (previously an explicit byte count was required).
15767 *Ulf Möller, Bodo Möller*
15769 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15770 used `char *` instead of `void *` and had casts all over the place.
15774 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15778 * Retain source code compatibility for BN_prime_checks macro:
15779 BN_is_prime(..., BN_prime_checks, ...) now uses
15780 BN_prime_checks_for_size to determine the appropriate number of
15781 Rabin-Miller iterations.
15785 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15786 DH_CHECK_P_NOT_SAFE_PRIME.
15787 (Check if this is true? OpenPGP calls them "strong".)
15791 * Merge the functionality of "dh" and "gendh" programs into a new program
15792 "dhparam". The old programs are retained for now but will handle DH keys
15793 (instead of parameters) in future.
15797 * Make the ciphers, s_server and s_client programs check the return values
15798 when a new cipher list is set.
15802 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15803 ciphers. Before when the 56bit ciphers were enabled the sorting was
15806 The syntax for the cipher sorting has been extended to support sorting by
15807 cipher-strength (using the strength_bits hard coded in the tables).
15808 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15810 Fix a bug in the cipher-command parser: when supplying a cipher command
15811 string with an "undefined" symbol (neither command nor alphanumeric
15812 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15813 an error is flagged.
15815 Due to the strength-sorting extension, the code of the
15816 ssl_create_cipher_list() function was completely rearranged. I hope that
15817 the readability was also increased :-)
15819 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15821 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15822 for the first serial number and places 2 in the serial number file. This
15823 avoids problems when the root CA is created with serial number zero and
15824 the first user certificate has the same issuer name and serial number
15829 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15830 the new code. Add documentation for this stuff.
15834 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15835 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15836 structures and behave in an analogous way to the X509v3 functions:
15837 they shouldn't be called directly but wrapper functions should be used
15840 So we also now have some wrapper functions that call the X509at functions
15841 when passed certificate requests. (TO DO: similar things can be done with
15842 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15843 things. Some of these need some d2i or i2d and print functionality
15844 because they handle more complex structures.)
15848 * Add missing #ifndefs that caused missing symbols when building libssl
15849 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15850 NO_RSA in `ssl/s2*.c`.
15852 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15854 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15855 has a return value which indicates the quality of the random data
15856 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15857 error queue. New function RAND_pseudo_bytes() generates output that is
15858 guaranteed to be unique but not unpredictable. RAND_add is like
15859 RAND_seed, but takes an extra argument for an entropy estimate
15860 (RAND_seed always assumes full entropy).
15864 * Do more iterations of Rabin-Miller probable prime test (specifically,
15865 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15866 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15867 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15868 false-positive rate of at most 2^-80 for random input.
15872 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15876 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15877 in the 0.9.5 release), this returns the chain
15878 from an X509_CTX structure with a dup of the stack and all
15879 the X509 reference counts upped: so the stack will exist
15880 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15883 Also make SSL_SESSION_print() print out the verify return
15888 * Add manpage for the pkcs12 command. Also change the default
15889 behaviour so MAC iteration counts are used unless the new
15890 -nomaciter option is used. This improves file security and
15891 only older versions of MSIE (4.0 for example) need it.
15895 * Honor the no-xxx Configure options when creating .DEF files.
15899 * Add PKCS#10 attributes to field table: challengePassword,
15900 unstructuredName and unstructuredAddress. These are taken from
15901 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15902 international characters are used.
15904 More changes to X509_ATTRIBUTE code: allow the setting of types
15905 based on strings. Remove the 'loc' parameter when adding
15906 attributes because these will be a SET OF encoding which is sorted
15911 * Initial changes to the 'req' utility to allow request generation
15912 automation. This will allow an application to just generate a template
15913 file containing all the field values and have req construct the
15916 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15917 used all over the place including certificate requests and PKCS#7
15918 structures. They are currently handled manually where necessary with
15919 some primitive wrappers for PKCS#7. The new functions behave in a
15920 manner analogous to the X509 extension functions: they allow
15921 attributes to be looked up by NID and added.
15923 Later something similar to the X509V3 code would be desirable to
15924 automatically handle the encoding, decoding and printing of the
15925 more complex types. The string types like challengePassword can
15926 be handled by the string table functions.
15928 Also modified the multi byte string table handling. Now there is
15929 a 'global mask' which masks out certain types. The table itself
15930 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15931 is useful when for example there is only one permissible type
15932 (as in countryName) and using the mask might result in no valid
15937 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15938 SSL_get_peer_finished to allow applications to obtain the latest
15939 Finished messages sent to the peer or expected from the peer,
15940 respectively. (SSL_get_peer_finished is usually the Finished message
15941 actually received from the peer, otherwise the protocol will be aborted.)
15943 As the Finished message are message digests of the complete handshake
15944 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15945 be used for external authentication procedures when the authentication
15946 provided by SSL/TLS is not desired or is not enough.
15950 * Enhanced support for Alpha Linux is added. Now ./config checks if
15951 the host supports BWX extension and if Compaq C is present on the
15952 $PATH. Just exploiting of the BWX extension results in 20-30%
15953 performance kick for some algorithms, e.g. DES and RC4 to mention
15954 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15959 * Add support for MS "fast SGC". This is arguably a violation of the
15960 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15961 weak crypto and after checking the certificate is SGC a second one
15962 with strong crypto. MS SGC stops the first handshake after receiving
15963 the server certificate message and sends a second client hello. Since
15964 a server will typically do all the time consuming operations before
15965 expecting any further messages from the client (server key exchange
15966 is the most expensive) there is little difference between the two.
15968 To get OpenSSL to support MS SGC we have to permit a second client
15969 hello message after we have sent server done. In addition we have to
15970 reset the MAC if we do get this second client hello.
15974 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15975 if a DER encoded private key is RSA or DSA traditional format. Changed
15976 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15977 format DER encoded private key. Newer code should use PKCS#8 format which
15978 has the key type encoded in the ASN1 structure. Added DER private key
15979 support to pkcs8 application.
15983 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15984 ciphersuites has been selected (as required by the SSL 3/TLS 1
15985 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
15986 is set, we interpret this as a request to violate the specification
15987 (the worst that can happen is a handshake failure, and 'correct'
15988 behaviour would result in a handshake failure anyway).
15992 * In SSL_CTX_add_session, take into account that there might be multiple
15993 SSL_SESSION structures with the same session ID (e.g. when two threads
15994 concurrently obtain them from an external cache).
15995 The internal cache can handle only one SSL_SESSION with a given ID,
15996 so if there's a conflict, we now throw out the old one to achieve
16001 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16002 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16003 some routines that use cipher OIDs: some ciphers do not have OIDs
16004 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16009 * Simplify the trust setting structure and code. Now we just have
16010 two sequences of OIDs for trusted and rejected settings. These will
16011 typically have values the same as the extended key usage extension
16012 and any application specific purposes.
16014 The trust checking code now has a default behaviour: it will just
16015 check for an object with the same NID as the passed id. Functions can
16016 be provided to override either the default behaviour or the behaviour
16017 for a given id. SSL client, server and email already have functions
16018 in place for compatibility: they check the NID and also return "trusted"
16019 if the certificate is self signed.
16023 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16024 traditional format into an EVP_PKEY structure.
16028 * Add a password callback function PEM_cb() which either prompts for
16029 a password if usr_data is NULL or otherwise assumes it is a null
16030 terminated password. Allow passwords to be passed on command line
16031 environment or config files in a few more utilities.
16035 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16036 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16037 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16038 Update documentation.
16042 * Support for ASN1 "NULL" type. This could be handled before by using
16043 ASN1_TYPE but there wasn't any function that would try to read a NULL
16044 and produce an error if it couldn't. For compatibility we also have
16045 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16046 don't allocate anything because they don't need to.
16050 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16053 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16055 * Rebuild of the memory allocation routines used by OpenSSL code and
16056 possibly others as well. The purpose is to make an interface that
16057 provide hooks so anyone can build a separate set of allocation and
16058 deallocation routines to be used by OpenSSL, for example memory
16059 pool implementations, or something else, which was previously hard
16060 since Malloc(), Realloc() and Free() were defined as macros having
16061 the values malloc, realloc and free, respectively (except for Win32
16062 compilations). The same is provided for memory debugging code.
16063 OpenSSL already comes with functionality to find memory leaks, but
16064 this gives people a chance to debug other memory problems.
16066 With these changes, a new set of functions and macros have appeared:
16068 CRYPTO_set_mem_debug_functions() [F]
16069 CRYPTO_get_mem_debug_functions() [F]
16070 CRYPTO_dbg_set_options() [F]
16071 CRYPTO_dbg_get_options() [F]
16072 CRYPTO_malloc_debug_init() [M]
16074 The memory debug functions are NULL by default, unless the library
16075 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16076 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16077 gives the standard debugging functions that come with OpenSSL) or
16078 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16079 provided by the library user) must be used. When the standard
16080 debugging functions are used, CRYPTO_dbg_set_options can be used to
16081 request additional information:
16082 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16083 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16085 Also, things like CRYPTO_set_mem_functions will always give the
16086 expected result (the new set of functions is used for allocation
16087 and deallocation) at all times, regardless of platform and compiler
16090 To finish it up, some functions that were never use in any other
16091 way than through macros have a new API and new semantic:
16093 CRYPTO_dbg_malloc()
16094 CRYPTO_dbg_realloc()
16097 All macros of value have retained their old syntax.
16099 *Richard Levitte and Bodo Moeller*
16101 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16102 ordering of SMIMECapabilities wasn't in "strength order" and there
16103 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16108 * Some ASN1 types with illegal zero length encoding (INTEGER,
16109 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16111 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16113 * Merge in my S/MIME library for OpenSSL. This provides a simple
16114 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16115 functionality to handle multipart/signed properly) and a utility
16116 called 'smime' to call all this stuff. This is based on code I
16117 originally wrote for Celo who have kindly allowed it to be
16118 included in OpenSSL.
16122 * Add variants des_set_key_checked and des_set_key_unchecked of
16123 des_set_key (aka des_key_sched). Global variable des_check_key
16124 decides which of these is called by des_set_key; this way
16125 des_check_key behaves as it always did, but applications and
16126 the library itself, which was buggy for des_check_key == 1,
16127 have a cleaner way to pick the version they need.
16131 * New function PKCS12_newpass() which changes the password of a
16136 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16137 dynamic mix. In both cases the ids can be used as an index into the
16138 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16139 functions so they accept a list of the field values and the
16140 application doesn't need to directly manipulate the X509_TRUST
16145 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16150 * Modify the way the V3 extension code looks up extensions. This now
16151 works in a similar way to the object code: we have some "standard"
16152 extensions in a static table which is searched with OBJ_bsearch()
16153 and the application can add dynamic ones if needed. The file
16154 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16155 updated whenever a new extension is added to the core code and kept
16156 in ext_nid order. There is a simple program 'tabtest.c' which checks
16157 this. New extensions are not added too often so this file can readily
16158 be maintained manually.
16160 There are two big advantages in doing things this way. The extensions
16161 can be looked up immediately and no longer need to be "added" using
16162 X509V3_add_standard_extensions(): this function now does nothing.
16163 Side note: I get *lots* of email saying the extension code doesn't
16164 work because people forget to call this function.
16165 Also no dynamic allocation is done unless new extensions are added:
16166 so if we don't add custom extensions there is no need to call
16167 X509V3_EXT_cleanup().
16171 * Modify enc utility's salting as follows: make salting the default. Add a
16172 magic header, so unsalted files fail gracefully instead of just decrypting
16173 to garbage. This is because not salting is a big security hole, so people
16174 should be discouraged from doing it.
16178 * Fixes and enhancements to the 'x509' utility. It allowed a message
16179 digest to be passed on the command line but it only used this
16180 parameter when signing a certificate. Modified so all relevant
16181 operations are affected by the digest parameter including the
16182 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16183 DSA key was used because it didn't fix the digest.
16187 * Initial certificate chain verify code. Currently tests the untrusted
16188 certificates for consistency with the verify purpose (which is set
16189 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16191 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16192 this is because it will reject chains with invalid extensions whereas
16193 every previous version of OpenSSL and SSLeay made no checks at all.
16195 Trust code: checks the root CA for the relevant trust settings. Trust
16196 settings have an initial value consistent with the verify purpose: e.g.
16197 if the verify purpose is for SSL client use it expects the CA to be
16198 trusted for SSL client use. However the default value can be changed to
16199 permit custom trust settings: one example of this would be to only trust
16200 certificates from a specific "secure" set of CAs.
16202 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16203 which should be used for version portability: especially since the
16204 verify structure is likely to change more often now.
16206 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16207 to set them. If not set then assume SSL clients will verify SSL servers
16210 Two new options to the verify program: -untrusted allows a set of
16211 untrusted certificates to be passed in and -purpose which sets the
16212 intended purpose of the certificate. If a purpose is set then the
16213 new chain verify code is used to check extension consistency.
16217 * Support for the authority information access extension.
16221 * Modify RSA and DSA PEM read routines to transparently handle
16222 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16223 public keys in a format compatible with certificate
16224 SubjectPublicKeyInfo structures. Unfortunately there were already
16225 functions called *_PublicKey_* which used various odd formats so
16226 these are retained for compatibility: however the DSA variants were
16227 never in a public release so they have been deleted. Changed dsa/rsa
16228 utilities to handle the new format: note no releases ever handled public
16229 keys so we should be OK.
16231 The primary motivation for this change is to avoid the same fiasco
16232 that dogs private keys: there are several incompatible private key
16233 formats some of which are standard and some OpenSSL specific and
16234 require various evil hacks to allow partial transparent handling and
16235 even then it doesn't work with DER formats. Given the option anything
16236 other than PKCS#8 should be dumped: but the other formats have to
16237 stay in the name of compatibility.
16239 With public keys and the benefit of hindsight one standard format
16240 is used which works with EVP_PKEY, RSA or DSA structures: though
16241 it clearly returns an error if you try to read the wrong kind of key.
16243 Added a -pubkey option to the 'x509' utility to output the public key.
16244 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16245 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16246 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16247 that do the same as the `EVP_PKEY_assign_*()` except they up the
16248 reference count of the added key (they don't "swallow" the
16253 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16254 CRLs would fail if the file contained no certificates or no CRLs:
16255 added a new function to read in both types and return the number
16256 read: this means that if none are read it will be an error. The
16257 DER versions of the certificate and CRL reader would always fail
16258 because it isn't possible to mix certificates and CRLs in DER format
16259 without choking one or the other routine. Changed this to just read
16260 a certificate: this is the best we can do. Also modified the code
16261 in apps/verify.c to take notice of return codes: it was previously
16262 attempting to read in certificates from NULL pointers and ignoring
16263 any errors: this is one reason why the cert and CRL reader seemed
16264 to work. It doesn't check return codes from the default certificate
16265 routines: these may well fail if the certificates aren't installed.
16269 * Code to support otherName option in GeneralName.
16273 * First update to verify code. Change the verify utility
16274 so it warns if it is passed a self signed certificate:
16275 for consistency with the normal behaviour. X509_verify
16276 has been modified to it will now verify a self signed
16277 certificate if *exactly* the same certificate appears
16278 in the store: it was previously impossible to trust a
16279 single self signed certificate. This means that:
16280 openssl verify ss.pem
16281 now gives a warning about a self signed certificate but
16282 openssl verify -CAfile ss.pem ss.pem
16287 * For servers, store verify_result in SSL_SESSION data structure
16288 (and add it to external session representation).
16289 This is needed when client certificate verifications fails,
16290 but an application-provided verification callback (set by
16291 SSL_CTX_set_cert_verify_callback) allows accepting the session
16292 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16293 but returns 1): When the session is reused, we have to set
16294 ssl->verify_result to the appropriate error code to avoid
16297 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16299 * Fix a bug in the new PKCS#7 code: it didn't consider the
16300 case in PKCS7_dataInit() where the signed PKCS7 structure
16301 didn't contain any existing data because it was being created.
16303 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16305 * Add a salt to the key derivation routines in enc.c. This
16306 forms the first 8 bytes of the encrypted file. Also add a
16307 -S option to allow a salt to be input on the command line.
16311 * New function X509_cmp(). Oddly enough there wasn't a function
16312 to compare two certificates. We do this by working out the SHA1
16313 hash and comparing that. X509_cmp() will be needed by the trust
16318 * SSL_get1_session() is like SSL_get_session(), but increments
16319 the reference count in the SSL_SESSION returned.
16321 *Geoff Thorpe <geoff@eu.c2.net>*
16323 * Fix for 'req': it was adding a null to request attributes.
16324 Also change the X509_LOOKUP and X509_INFO code to handle
16325 certificate auxiliary information.
16329 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16334 * Add the possibility to add extra information to the memory leak
16335 detecting output, to form tracebacks, showing from where each
16336 allocation was originated: CRYPTO_push_info("constant string") adds
16337 the string plus current file name and line number to a per-thread
16338 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16339 is like calling CYRPTO_pop_info() until the stack is empty.
16340 Also updated memory leak detection code to be multi-thread-safe.
16344 * Add options -text and -noout to pkcs7 utility and delete the
16345 encryption options which never did anything. Update docs.
16349 * Add options to some of the utilities to allow the pass phrase
16350 to be included on either the command line (not recommended on
16351 OSes like Unix) or read from the environment. Update the
16352 manpages and fix a few bugs.
16356 * Add a few manpages for some of the openssl commands.
16360 * Fix the -revoke option in ca. It was freeing up memory twice,
16361 leaking and not finding already revoked certificates.
16365 * Extensive changes to support certificate auxiliary information.
16366 This involves the use of X509_CERT_AUX structure and X509_AUX
16367 functions. An X509_AUX function such as PEM_read_X509_AUX()
16368 can still read in a certificate file in the usual way but it
16369 will also read in any additional "auxiliary information". By
16370 doing things this way a fair degree of compatibility can be
16371 retained: existing certificates can have this information added
16372 using the new 'x509' options.
16374 Current auxiliary information includes an "alias" and some trust
16375 settings. The trust settings will ultimately be used in enhanced
16376 certificate chain verification routines: currently a certificate
16377 can only be trusted if it is self signed and then it is trusted
16382 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16383 The problem was that one of the replacement routines had not been working
16384 since SSLeay releases. For now the offending routine has been replaced
16385 with non-optimised assembler. Even so, this now gives around 95%
16386 performance improvement for 1024 bit RSA signs.
16390 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16391 handling. Most clients have the effective key size in bits equal to
16392 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16393 A few however don't do this and instead use the size of the decrypted key
16394 to determine the RC2 key length and the AlgorithmIdentifier to determine
16395 the effective key length. In this case the effective key length can still
16396 be 40 bits but the key length can be 168 bits for example. This is fixed
16397 by manually forcing an RC2 key into the EVP_PKEY structure because the
16398 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16399 the key length and effective key length are equal.
16403 * Add a bunch of functions that should simplify the creation of
16404 X509_NAME structures. Now you should be able to do:
16405 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16406 and have it automatically work out the correct field type and fill in
16407 the structures. The more adventurous can try:
16408 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16409 and it will (hopefully) work out the correct multibyte encoding.
16413 * Change the 'req' utility to use the new field handling and multibyte
16414 copy routines. Before the DN field creation was handled in an ad hoc
16415 way in req, ca, and x509 which was rather broken and didn't support
16416 BMPStrings or UTF8Strings. Since some software doesn't implement
16417 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16418 using the dirstring_type option. See the new comment in the default
16419 openssl.cnf for more info.
16423 * Make crypto/rand/md_rand.c more robust:
16424 - Assure unique random numbers after fork().
16425 - Make sure that concurrent threads access the global counter and
16426 md serializably so that we never lose entropy in them
16427 or use exactly the same state in multiple threads.
16428 Access to the large state is not always serializable because
16429 the additional locking could be a performance killer, and
16430 md should be large enough anyway.
16434 * New file apps/app_rand.c with commonly needed functionality
16435 for handling the random seed file.
16437 Use the random seed file in some applications that previously did not:
16439 dsaparam -genkey (which also ignored its '-rand' option),
16442 x509 (when signing).
16443 Except on systems with /dev/urandom, it is crucial to have a random
16444 seed file at least for key creation, DSA signing, and for DH exchanges;
16445 for RSA signatures we could do without one.
16447 gendh and gendsa (unlike genrsa) used to read only the first byte
16448 of each file listed in the '-rand' option. The function as previously
16449 found in genrsa is now in app_rand.c and is used by all programs
16450 that support '-rand'.
16454 * In RAND_write_file, use mode 0600 for creating files;
16455 don't just chmod when it may be too late.
16459 * Report an error from X509_STORE_load_locations
16460 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16464 * New function ASN1_mbstring_copy() this copies a string in either
16465 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16466 into an ASN1_STRING type. A mask of permissible types is passed
16467 and it chooses the "minimal" type to use or an error if not type
16472 * Add function equivalents to the various macros in asn1.h. The old
16473 macros are retained with an `M_` prefix. Code inside the library can
16474 use the `M_` macros. External code (including the openssl utility)
16475 should *NOT* in order to be "shared library friendly".
16479 * Add various functions that can check a certificate's extensions
16480 to see if it usable for various purposes such as SSL client,
16481 server or S/MIME and CAs of these types. This is currently
16482 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16483 verification. Also added a -purpose flag to x509 utility to
16484 print out all the purposes.
16488 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16493 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16494 for, obtain and decode and extension and obtain its critical flag.
16495 This allows all the necessary extension code to be handled in a
16496 single function call.
16500 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16501 platforms. See crypto/rc4/rc4_enc.c for further details.
16505 * New -noout option to asn1parse. This causes no output to be produced
16506 its main use is when combined with -strparse and -out to extract data
16507 from a file (which may not be in ASN.1 format).
16511 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16512 when producing the local key id.
16514 *Richard Levitte <levitte@stacken.kth.se>*
16516 * New option -dhparam in s_server. This allows a DH parameter file to be
16517 stated explicitly. If it is not stated then it tries the first server
16518 certificate file. The previous behaviour hard coded the filename
16523 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16524 a public key to be input or output. For example:
16525 openssl rsa -in key.pem -pubout -out pubkey.pem
16526 Also added necessary DSA public key functions to handle this.
16530 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16531 in the message. This was handled by allowing
16532 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16534 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16536 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16537 to the end of the strings whereas this didn't. This would cause problems
16538 if strings read with d2i_ASN1_bytes() were later modified.
16540 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16542 * Fix for base64 decode bug. When a base64 bio reads only one line of
16543 data and it contains EOF it will end up returning an error. This is
16544 caused by input 46 bytes long. The cause is due to the way base64
16545 BIOs find the start of base64 encoded data. They do this by trying a
16546 trial decode on each line until they find one that works. When they
16547 do a flag is set and it starts again knowing it can pass all the
16548 data directly through the decoder. Unfortunately it doesn't reset
16549 the context it uses. This means that if EOF is reached an attempt
16550 is made to pass two EOFs through the context and this causes the
16551 resulting error. This can also cause other problems as well. As is
16552 usual with these problems it takes *ages* to find and the fix is
16553 trivial: move one line.
16555 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16557 * Ugly workaround to get s_client and s_server working under Windows. The
16558 old code wouldn't work because it needed to select() on sockets and the
16559 tty (for keypresses and to see if data could be written). Win32 only
16560 supports select() on sockets so we select() with a 1s timeout on the
16561 sockets and then see if any characters are waiting to be read, if none
16562 are present then we retry, we also assume we can always write data to
16563 the tty. This isn't nice because the code then blocks until we've
16564 received a complete line of data and it is effectively polling the
16565 keyboard at 1s intervals: however it's quite a bit better than not
16566 working at all :-) A dedicated Windows application might handle this
16567 with an event loop for example.
16571 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16572 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16573 will be called when RSA_sign() and RSA_verify() are used. This is useful
16574 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16575 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16576 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16577 This necessitated the support of an extra signature type NID_md5_sha1
16578 for SSL signatures and modifications to the SSL library to use it instead
16579 of calling RSA_public_decrypt() and RSA_private_encrypt().
16583 * Add new -verify -CAfile and -CApath options to the crl program, these
16584 will lookup a CRL issuers certificate and verify the signature in a
16585 similar way to the verify program. Tidy up the crl program so it
16586 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16587 less strict. It will now permit CRL extensions even if it is not
16588 a V2 CRL: this will allow it to tolerate some broken CRLs.
16592 * Initialize all non-automatic variables each time one of the openssl
16593 sub-programs is started (this is necessary as they may be started
16594 multiple times from the "OpenSSL>" prompt).
16596 *Lennart Bang, Bodo Moeller*
16598 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16599 removing all other RSA functionality (this is what NO_RSA does). This
16600 is so (for example) those in the US can disable those operations covered
16601 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16606 * Non-copying interface to BIO pairs.
16607 (still largely untested)
16611 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16612 ASCII string. This was handled independently in various places before.
16616 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16617 UTF8 strings a character at a time.
16621 * Use client_version from client hello to select the protocol
16622 (s23_srvr.c) and for RSA client key exchange verification
16623 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16627 * Add various utility functions to handle SPKACs, these were previously
16628 handled by poking round in the structure internals. Added new function
16629 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16630 print, verify and generate SPKACs. Based on an original idea from
16631 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16635 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16639 * Allow the config file extension section to be overwritten on the
16640 command line. Based on an original idea from Massimiliano Pala
16641 <madwolf@comune.modena.it>. The new option is called -extensions
16642 and can be applied to ca, req and x509. Also -reqexts to override
16643 the request extensions in req and -crlexts to override the crl extensions
16648 * Add new feature to the SPKAC handling in ca. Now you can include
16649 the same field multiple times by preceding it by "XXXX." for example:
16652 this is the same syntax as used in the req config file.
16656 * Allow certificate extensions to be added to certificate requests. These
16657 are specified in a 'req_extensions' option of the req section of the
16658 config file. They can be printed out with the -text option to req but
16659 are otherwise ignored at present.
16663 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16664 data read consists of only the final block it would not decrypted because
16665 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16666 A misplaced 'break' also meant the decrypted final block might not be
16667 copied until the next read.
16671 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16672 a few extra parameters to the DH structure: these will be useful if
16673 for example we want the value of 'q' or implement X9.42 DH.
16677 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16678 provides hooks that allow the default DSA functions or functions on a
16679 "per key" basis to be replaced. This allows hardware acceleration and
16680 hardware key storage to be handled without major modification to the
16681 library. Also added low level modexp hooks and CRYPTO_EX structure and
16682 associated functions.
16686 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16687 as "read only": it can't be written to and the buffer it points to will
16688 not be freed. Reading from a read only BIO is much more efficient than
16689 a normal memory BIO. This was added because there are several times when
16690 an area of memory needs to be read from a BIO. The previous method was
16691 to create a memory BIO and write the data to it, this results in two
16692 copies of the data and an O(n^2) reading algorithm. There is a new
16693 function BIO_new_mem_buf() which creates a read only memory BIO from
16694 an area of memory. Also modified the PKCS#7 routines to use read only
16699 * Bugfix: ssl23_get_client_hello did not work properly when called in
16700 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16701 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16702 but a retry condition occurred while trying to read the rest.
16706 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16707 NID_pkcs7_encrypted by default: this was wrong since this should almost
16708 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16709 the encrypted data type: this is a more sensible place to put it and it
16710 allows the PKCS#12 code to be tidied up that duplicated this
16715 * Changed obj_dat.pl script so it takes its input and output files on
16716 the command line. This should avoid shell escape redirection problems
16721 * Initial support for certificate extension requests, these are included
16722 in things like Xenroll certificate requests. Included functions to allow
16723 extensions to be obtained and added.
16727 * -crlf option to s_client and s_server for sending newlines as
16728 CRLF (as required by many protocols).
16732 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16734 * Install libRSAglue.a when OpenSSL is built with RSAref.
16736 *Ralf S. Engelschall*
16738 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16740 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16742 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16747 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16748 DH parameters/keys (q is lost during that conversion, but the resulting
16749 DH parameters contain its length).
16751 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16752 much faster than DH_generate_parameters (which creates parameters
16753 where `p = 2*q + 1`), and also the smaller q makes DH computations
16754 much more efficient (160-bit exponentiation instead of 1024-bit
16755 exponentiation); so this provides a convenient way to support DHE
16756 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16757 utter importance to use
16758 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16760 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16761 when such DH parameters are used, because otherwise small subgroup
16762 attacks may become possible!
16766 * Avoid memory leak in i2d_DHparams.
16770 * Allow the -k option to be used more than once in the enc program:
16771 this allows the same encrypted message to be read by multiple recipients.
16775 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16776 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16777 it will always use the numerical form of the OID, even if it has a short
16782 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16783 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16784 otherwise bn_mod_exp was called. In the case of hardware keys for example
16785 no private key components need be present and it might store extra data
16786 in the RSA structure, which cannot be accessed from bn_mod_exp.
16787 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16788 private key operations.
16792 * Added support for SPARC Linux.
16796 * pem_password_cb function type incompatibly changed from
16797 typedef int pem_password_cb(char *buf, int size, int rwflag);
16799 ....(char *buf, int size, int rwflag, void *userdata);
16800 so that applications can pass data to their callbacks:
16801 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16802 additional void * argument, which is just handed through whenever
16803 the password callback is called.
16805 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16807 New function SSL_CTX_set_default_passwd_cb_userdata.
16809 Compatibility note: As many C implementations push function arguments
16810 onto the stack in reverse order, the new library version is likely to
16811 interoperate with programs that have been compiled with the old
16812 pem_password_cb definition (PEM_whatever takes some data that
16813 happens to be on the stack as its last argument, and the callback
16814 just ignores this garbage); but there is no guarantee whatsoever that
16817 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16818 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16819 problems not only on Windows, but also on some Unix platforms.
16820 To avoid problematic command lines, these definitions are now in an
16821 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16822 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16826 * MIPS III/IV assembler module is reimplemented.
16830 * More DES library cleanups: remove references to srand/rand and
16831 delete an unused file.
16835 * Add support for the free Netwide assembler (NASM) under Win32,
16836 since not many people have MASM (ml) and it can be hard to obtain.
16837 This is currently experimental but it seems to work OK and pass all
16838 the tests. Check out INSTALL.W32 for info.
16842 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16843 without temporary keys kept an extra copy of the server key,
16844 and connections with temporary keys did not free everything in case
16849 * New function RSA_check_key and new openssl rsa option -check
16850 for verifying the consistency of RSA keys.
16852 *Ulf Moeller, Bodo Moeller*
16854 * Various changes to make Win32 compile work:
16855 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16856 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16857 comparison" warnings.
16858 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16862 * Add a debugging option to PKCS#5 v2 key generation function: when
16863 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16864 derived keys are printed to stderr.
16868 * Copy the flags in ASN1_STRING_dup().
16870 *Roman E. Pavlov <pre@mo.msk.ru>*
16872 * The x509 application mishandled signing requests containing DSA
16873 keys when the signing key was also DSA and the parameters didn't match.
16875 It was supposed to omit the parameters when they matched the signing key:
16876 the verifying software was then supposed to automatically use the CA's
16877 parameters if they were absent from the end user certificate.
16879 Omitting parameters is no longer recommended. The test was also
16880 the wrong way round! This was probably due to unusual behaviour in
16881 EVP_cmp_parameters() which returns 1 if the parameters match.
16882 This meant that parameters were omitted when they *didn't* match and
16883 the certificate was useless. Certificates signed with 'ca' didn't have
16886 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16888 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16889 The interface is as follows:
16890 Applications can use
16891 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16892 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16893 "off" is now the default.
16894 The library internally uses
16895 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16896 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16897 to disable memory-checking temporarily.
16899 Some inconsistent states that previously were possible (and were
16900 even the default) are now avoided.
16902 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16903 with each memory chunk allocated; this is occasionally more helpful
16904 than just having a counter.
16906 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16908 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16913 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16914 which largely parallels "options", but is for changing API behaviour,
16915 whereas "options" are about protocol behaviour.
16916 Initial "mode" flags are:
16918 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16919 a single record has been written.
16920 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16921 retries use the same buffer location.
16922 (But all of the contents must be
16927 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16930 * Fix problems with no-hmac etc.
16932 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16934 * New functions RSA_get_default_method(), RSA_set_method() and
16935 RSA_get_method(). These allows replacement of RSA_METHODs without having
16936 to mess around with the internals of an RSA structure.
16940 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16941 Also really enable memory leak checks in openssl.c and in some
16944 *Chad C. Mulligan, Bodo Moeller*
16946 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16947 up the length of negative integers. This has now been simplified to just
16948 store the length when it is first determined and use it later, rather
16949 than trying to keep track of where data is copied and updating it to
16951 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16953 * Add a new function PKCS7_signatureVerify. This allows the verification
16954 of a PKCS#7 signature but with the signing certificate passed to the
16955 function itself. This contrasts with PKCS7_dataVerify which assumes the
16956 certificate is present in the PKCS#7 structure. This isn't always the
16957 case: certificates can be omitted from a PKCS#7 structure and be
16958 distributed by "out of band" means (such as a certificate database).
16962 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16963 function prototypes in pem.h, also change util/mkdef.pl to add the
16964 necessary function names.
16968 * mk1mf.pl (used by Windows builds) did not properly read the
16969 options set by Configure in the top level Makefile, and Configure
16970 was not even able to write more than one option correctly.
16971 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16975 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16976 file to be loaded from a BIO or FILE pointer. The BIO version will
16977 for example allow memory BIOs to contain config info.
16981 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16982 Whoever hopes to achieve shared-library compatibility across versions
16983 must use this, not the compile-time macro.
16984 (Exercise 0.9.4: Which is the minimum library version required by
16986 Note: All this applies only to multi-threaded programs, others don't
16991 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
16992 through a BIO pair triggered the default case, i.e.
16993 SSLerr(...,SSL_R_UNKNOWN_STATE).
16997 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
16998 can use the SSL library even if none of the specific BIOs is
17003 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17004 for the encoded length.
17006 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17008 * Add initial documentation of the X509V3 functions.
17012 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17013 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17014 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17015 secure PKCS#8 private key format with a high iteration count.
17019 * Fix determination of Perl interpreter: A perl or perl5
17020 *directory* in $PATH was also accepted as the interpreter.
17022 *Ralf S. Engelschall*
17024 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17025 wrong with it but it was very old and did things like calling
17026 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17027 unusual formatting.
17031 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17032 to use the new extension code.
17036 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17037 with macros. This should make it easier to change their form, add extra
17038 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17043 * Add to configuration table a new entry that can specify an alternative
17044 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17045 according to Mark Crispin <MRC@Panda.COM>.
17049 * DES CBC did not update the IV. Weird.
17053 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17054 Changing the behaviour of the former might break existing programs --
17055 where IV updating is needed, des_ncbc_encrypt can be used.
17058 * When bntest is run from "make test" it drives bc to check its
17059 calculations, as well as internally checking them. If an internal check
17060 fails, it needs to cause bc to give a non-zero result or make test carries
17061 on without noticing the failure. Fixed.
17065 * DES library cleanups.
17069 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17070 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17071 ciphers. NOTE: although the key derivation function has been verified
17072 against some published test vectors it has not been extensively tested
17073 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17078 * Instead of "mkdir -p", which is not fully portable, use new
17079 Perl script "util/mkdir-p.pl".
17083 * Rewrite the way password based encryption (PBE) is handled. It used to
17084 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17085 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17086 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17087 the 'parameter' field of the AlgorithmIdentifier is passed to the
17088 underlying key generation function so it must do its own ASN1 parsing.
17089 This has also changed the EVP_PBE_CipherInit() function which now has a
17090 'parameter' argument instead of literal salt and iteration count values
17091 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17095 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17096 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17097 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17098 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17099 value was just used as a "magic string" and not used directly its
17100 value doesn't matter.
17104 * Introduce some semblance of const correctness to BN. Shame C doesn't
17109 * "linux-sparc64" configuration (ultrapenguin).
17111 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17112 "linux-sparc" configuration.
17114 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17116 * config now generates no-xxx options for missing ciphers.
17120 * Support the EBCDIC character set (work in progress).
17121 File ebcdic.c not yet included because it has a different license.
17123 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17125 * Support BS2000/OSD-POSIX.
17127 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17129 * Make callbacks for key generation use `void *` instead of `char *`.
17133 * Make S/MIME samples compile (not yet tested).
17137 * Additional typesafe stacks.
17141 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17145 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17147 * New configuration variant "sco5-gcc".
17149 * Updated some demos.
17151 *Sean O Riordain, Wade Scholine*
17153 * Add missing BIO_free at exit of pkcs12 application.
17157 * Fix memory leak in conf.c.
17161 * Updates for Win32 to assembler version of MD5.
17165 * Set #! path to perl in apps/der_chop to where we found it
17166 instead of using a fixed path.
17170 * SHA library changes for irix64-mips4-cc.
17174 * Improvements for VMS support.
17178 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17180 * Bignum library bug fix. IRIX 6 passes "make test" now!
17181 This also avoids the problems with SC4.2 and unpatched SC5.
17183 *Andy Polyakov <appro@fy.chalmers.se>*
17185 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17186 These are required because of the typesafe stack would otherwise break
17187 existing code. If old code used a structure member which used to be STACK
17188 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17189 sk_num or sk_value it would produce an error because the num, data members
17190 are not present in STACK_OF. Now it just produces a warning. sk_set
17191 replaces the old method of assigning a value to sk_value
17192 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17193 that does this will no longer work (and should use sk_set instead) but
17194 this could be regarded as a "questionable" behaviour anyway.
17198 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17199 correctly handle encrypted S/MIME data.
17203 * Change type of various DES function arguments from des_cblock
17204 (which means, in function argument declarations, pointer to char)
17205 to des_cblock * (meaning pointer to array with 8 char elements),
17206 which allows the compiler to do more typechecking; it was like
17207 that back in SSLeay, but with lots of ugly casts.
17209 Introduce new type const_des_cblock.
17213 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17214 problems: find RecipientInfo structure that matches recipient certificate
17215 and initialise the ASN1 structures properly based on passed cipher.
17219 * Belatedly make the BN tests actually check the results.
17223 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17224 to and from BNs: it was completely broken. New compilation option
17225 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17226 key elements as negative integers.
17230 * Reorganize and speed up MD5.
17232 *Andy Polyakov <appro@fy.chalmers.se>*
17236 *Richard Levitte <richard@levitte.org>*
17238 * New option -out to asn1parse to allow the parsed structure to be
17239 output to a file. This is most useful when combined with the -strparse
17240 option to examine the output of things like OCTET STRINGS.
17244 * Make SSL library a little more fool-proof by not requiring any longer
17245 that `SSL_set_{accept,connect}_state` be called before
17246 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17247 in many applications because usually everything *appeared* to work as
17248 intended anyway -- now it really works as intended).
17252 * Move openssl.cnf out of lib/.
17256 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17257 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17258 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17260 *Ralf S. Engelschall*
17262 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17263 handle PKCS#7 enveloped data properly.
17265 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17267 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17268 copying pointers. The cert_st handling is changed by this in
17269 various ways (and thus what used to be known as ctx->default_cert
17270 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17271 any longer when s->cert does not give us what we need).
17272 ssl_cert_instantiate becomes obsolete by this change.
17273 As soon as we've got the new code right (possibly it already is?),
17274 we have solved a couple of bugs of the earlier code where s->cert
17275 was used as if it could not have been shared with other SSL structures.
17277 Note that using the SSL API in certain dirty ways now will result
17278 in different behaviour than observed with earlier library versions:
17279 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17280 does not influence s as it used to.
17282 In order to clean up things more thoroughly, inside SSL_SESSION
17283 we don't use CERT any longer, but a new structure SESS_CERT
17284 that holds per-session data (if available); currently, this is
17285 the peer's certificate chain and, for clients, the server's certificate
17286 and temporary key. CERT holds only those values that can have
17287 meaningful defaults in an SSL_CTX.
17291 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17292 from the internal representation. Various PKCS#7 fixes: remove some
17293 evil casts and set the enc_dig_alg field properly based on the signing
17298 * Allow PKCS#12 password to be set from the command line or the
17299 environment. Let 'ca' get its config file name from the environment
17300 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17305 * Allow certificate policies extension to use an IA5STRING for the
17306 organization field. This is contrary to the PKIX definition but
17307 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17312 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17313 without disallowing inline assembler and the like for non-pedantic builds.
17317 * Support Borland C++ builder.
17319 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17325 * SHA-1 cleanups and performance enhancements.
17327 *Andy Polyakov <appro@fy.chalmers.se>*
17329 * Sparc v8plus assembler for the bignum library.
17331 *Andy Polyakov <appro@fy.chalmers.se>*
17333 * Accept any -xxx and +xxx compiler options in Configure.
17337 * Update HPUX configuration.
17341 * Add missing `sk_<type>_unshift()` function to safestack.h
17343 *Ralf S. Engelschall*
17345 * New function SSL_CTX_use_certificate_chain_file that sets the
17346 "extra_cert"s in addition to the certificate. (This makes sense
17347 only for "PEM" format files, as chains as a whole are not
17352 * Support verify_depth from the SSL API.
17353 x509_vfy.c had what can be considered an off-by-one-error:
17354 Its depth (which was not part of the external interface)
17355 was actually counting the number of certificates in a chain;
17356 now it really counts the depth.
17360 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17361 instead of X509err, which often resulted in confusing error
17362 messages since the error codes are not globally unique
17363 (e.g. an alleged error in ssl3_accept when a certificate
17364 didn't match the private key).
17366 * New function SSL_CTX_set_session_id_context that allows to set a default
17367 value (so that you don't need SSL_set_session_id_context for each
17368 connection using the SSL_CTX).
17372 * OAEP decoding bug fix.
17376 * Support INSTALL_PREFIX for package builders, as proposed by
17381 * New Configure options "threads" and "no-threads". For systems
17382 where the proper compiler options are known (currently Solaris
17383 and Linux), "threads" is the default.
17387 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17391 * Install various scripts to $(OPENSSLDIR)/misc, not to
17392 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17393 such as /usr/local/bin.
17397 * "make linux-shared" to build shared libraries.
17399 *Niels Poppe <niels@netbox.org>*
17401 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17405 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17406 extension adding in x509 utility.
17410 * Remove NOPROTO sections and error code comments.
17414 * Partial rewrite of the DEF file generator to now parse the ANSI
17419 * New Configure options --prefix=DIR and --openssldir=DIR.
17423 * Complete rewrite of the error code script(s). It is all now handled
17424 by one script at the top level which handles error code gathering,
17425 header rewriting and C source file generation. It should be much better
17426 than the old method: it now uses a modified version of Ulf's parser to
17427 read the ANSI prototypes in all header files (thus the old K&R definitions
17428 aren't needed for error creation any more) and do a better job of
17429 translating function codes into names. The old 'ASN1 error code embedded
17430 in a comment' is no longer necessary and it doesn't use .err files which
17431 have now been deleted. Also the error code call doesn't have to appear all
17432 on one line (which resulted in some large lines...).
17436 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17440 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17441 0 (which usually indicates a closed connection), but continue reading.
17445 * Fix some race conditions.
17449 * Add support for CRL distribution points extension. Add Certificate
17450 Policies and CRL distribution points documentation.
17454 * Move the autogenerated header file parts to crypto/opensslconf.h.
17458 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17459 8 of keying material. Merlin has also confirmed interop with this fix
17460 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17462 *Merlin Hughes <merlin@baltimore.ie>*
17464 * Fix lots of warnings.
17466 *Richard Levitte <levitte@stacken.kth.se>*
17468 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17469 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17471 *Richard Levitte <levitte@stacken.kth.se>*
17473 * Fix problems with sizeof(long) == 8.
17475 *Andy Polyakov <appro@fy.chalmers.se>*
17477 * Change functions to ANSI C.
17481 * Fix typos in error codes.
17483 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17485 * Remove defunct assembler files from Configure.
17489 * SPARC v8 assembler BIGNUM implementation.
17491 *Andy Polyakov <appro@fy.chalmers.se>*
17493 * Support for Certificate Policies extension: both print and set.
17494 Various additions to support the r2i method this uses.
17498 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17499 return a const string when you are expecting an allocated buffer.
17503 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17504 types DirectoryString and DisplayText.
17508 * Add code to allow r2i extensions to access the configuration database,
17509 add an LHASH database driver and add several ctx helper functions.
17513 * Fix an evil bug in bn_expand2() which caused various BN functions to
17514 fail when they extended the size of a BIGNUM.
17518 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17519 support typesafe stack.
17523 * Fix typo in SSL_[gs]et_options().
17525 *Nils Frostberg <nils@medcom.se>*
17527 * Delete various functions and files that belonged to the (now obsolete)
17528 old X509V3 handling code.
17532 * New Configure option "rsaref".
17536 * Don't auto-generate pem.h.
17540 * Introduce type-safe ASN.1 SETs.
17544 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17546 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17548 * Introduce type-safe STACKs. This will almost certainly break lots of code
17549 that links with OpenSSL (well at least cause lots of warnings), but fear
17550 not: the conversion is trivial, and it eliminates loads of evil casts. A
17551 few STACKed things have been converted already. Feel free to convert more.
17552 In the fullness of time, I'll do away with the STACK type altogether.
17556 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17557 specified in `<certfile>` by updating the entry in the index.txt file.
17558 This way one no longer has to edit the index.txt file manually for
17559 revoking a certificate. The -revoke option does the gory details now.
17561 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17563 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17564 `-text` option at all and this way the `-noout -text` combination was
17565 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17567 *Ralf S. Engelschall*
17569 * Make sure a corresponding plain text error message exists for the
17570 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17571 verify callback function determined that a certificate was revoked.
17573 *Ralf S. Engelschall*
17575 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17576 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17577 all available ciphers including rc5, which was forgotten until now.
17578 In order to let the testing shell script know which algorithms
17579 are available, a new (up to now undocumented) command
17580 `openssl list-cipher-commands` is used.
17584 * Bugfix: s_client occasionally would sleep in select() when
17585 it should have checked SSL_pending() first.
17589 * New functions DSA_do_sign and DSA_do_verify to provide access to
17590 the raw DSA values prior to ASN.1 encoding.
17594 * Tweaks to Configure
17596 *Niels Poppe <niels@netbox.org>*
17598 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17603 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17607 * New config option to avoid instructions that are illegal on the 80386.
17608 The default code is faster, but requires at least a 486.
17612 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17613 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17614 same as SSL2_VERSION anyway.
17618 * New "-showcerts" option for s_client.
17622 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17623 application. Various cleanups and fixes.
17627 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17628 modify error routines to work internally. Add error codes and PBE init
17629 to library startup routines.
17633 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17634 packing functions to asn1 and evp. Changed function names and error
17635 codes along the way.
17639 * PKCS12 integration: and so it begins... First of several patches to
17640 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17641 objects to objects.h
17645 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17646 and display support for Thawte strong extranet extension.
17650 * Add LinuxPPC support.
17652 *Jeff Dubrule <igor@pobox.org>*
17654 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17655 bn_div_words in alpha.s.
17657 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17659 * Make sure the RSA OAEP test is skipped under -DRSAref because
17660 OAEP isn't supported when OpenSSL is built with RSAref.
17662 *Ulf Moeller <ulf@fitug.de>*
17664 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17665 so they no longer are missing under -DNOPROTO.
17667 *Soren S. Jorvang <soren@t.dk>*
17669 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17671 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17672 doesn't work when the session is reused. Coming soon!
17676 * Fix a security hole, that allows sessions to be reused in the wrong
17677 context thus bypassing client cert protection! All software that uses
17678 client certs and session caches in multiple contexts NEEDS PATCHING to
17679 allow session reuse! A fuller solution is in the works.
17681 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17683 * Some more source tree cleanups (removed obsolete files
17684 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17685 permission on "config" script to be executable) and a fix for the INSTALL
17688 *Ulf Moeller <ulf@fitug.de>*
17690 * Remove some legacy and erroneous uses of malloc, free instead of
17693 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17695 * Make rsa_oaep_test return non-zero on error.
17697 *Ulf Moeller <ulf@fitug.de>*
17699 * Add support for native Solaris shared libraries. Configure
17700 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17701 if someone would make that last step automatic.
17703 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17705 * ctx_size was not built with the right compiler during "make links". Fixed.
17709 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17710 except NULL ciphers". This means the default cipher list will no longer
17711 enable NULL ciphers. They need to be specifically enabled e.g. with
17712 the string "DEFAULT:eNULL".
17716 * Fix to RSA private encryption routines: if p < q then it would
17717 occasionally produce an invalid result. This will only happen with
17718 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17722 * Be less restrictive and allow also `perl util/perlpath.pl
17723 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17724 because this way one can also use an interpreter named `perl5` (which is
17725 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17726 installed as `perl`).
17728 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17730 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17732 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17734 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17735 advapi32.lib to Win32 build and change the pem test comparison
17736 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17737 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17738 and crypto/des/ede_cbcm_enc.c.
17742 * DES quad checksum was broken on big-endian architectures. Fixed.
17746 * Comment out two functions in bio.h that aren't implemented. Fix up the
17747 Win32 test batch file so it (might) work again. The Win32 test batch file
17748 is horrible: I feel ill....
17752 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17753 in e_os.h. Audit of header files to check ANSI and non ANSI
17754 sections: 10 functions were absent from non ANSI section and not exported
17755 from Windows DLLs. Fixed up libeay.num for new functions.
17759 * Make `openssl version` output lines consistent.
17761 *Ralf S. Engelschall*
17763 * Fix Win32 symbol export lists for BIO functions: Added
17764 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17765 to ms/libeay{16,32}.def.
17767 *Ralf S. Engelschall*
17769 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17770 fine under Unix and passes some trivial tests I've now added. But the
17771 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17772 added to make sure no one expects that this stuff really works in the
17773 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17774 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17777 *Ralf S. Engelschall*
17779 * Fix the generation of two part addresses in perl.
17781 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17783 * Add config entry for Linux on MIPS.
17785 *John Tobey <jtobey@channel1.com>*
17787 * Make links whenever Configure is run, unless we are on Windoze.
17791 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17792 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17797 * Add a useful kludge to allow package maintainers to specify compiler and
17798 other platforms details on the command line without having to patch the
17799 Configure script every time: One now can use
17800 `perl Configure <id>:<details>`,
17801 i.e. platform ids are allowed to have details appended
17802 to them (separated by colons). This is treated as there would be a static
17803 pre-configured entry in Configure's %table under key `<id>` with value
17804 `<details>` and `perl Configure <id>` is called. So, when you want to
17805 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17806 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17807 now, which overrides the FreeBSD-elf entry on-the-fly.
17809 *Ralf S. Engelschall*
17811 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17815 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17816 on the `perl Configure ...` command line. This way one can compile
17817 OpenSSL libraries with Position Independent Code (PIC) which is needed
17818 for linking it into DSOs.
17820 *Ralf S. Engelschall*
17822 * Remarkably, export ciphers were totally broken and no-one had noticed!
17827 * Cleaned up the LICENSE document: The official contact for any license
17828 questions now is the OpenSSL core team under openssl-core@openssl.org.
17829 And add a paragraph about the dual-license situation to make sure people
17830 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17831 to the OpenSSL toolkit.
17833 *Ralf S. Engelschall*
17835 * General source tree makefile cleanups: Made `making xxx in yyy...`
17836 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17837 Additionally cleaned up the `make links` target: Remove unnecessary
17838 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17839 to speed processing and no longer clutter the display with confusing
17840 stuff. Instead only the actually done links are displayed.
17842 *Ralf S. Engelschall*
17844 * Permit null encryption ciphersuites, used for authentication only. It used
17845 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17846 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17851 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17852 signed attributes when verifying signatures (this would break them),
17853 the detached data encoding was wrong and public keys obtained using
17854 X509_get_pubkey() weren't freed.
17858 * Add text documentation for the BUFFER functions. Also added a work around
17859 to a Win95 console bug. This was triggered by the password read stuff: the
17860 last character typed gets carried over to the next fread(). If you were
17861 generating a new cert request using 'req' for example then the last
17862 character of the passphrase would be CR which would then enter the first
17867 * Added the new 'Includes OpenSSL Cryptography Software' button as
17868 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17869 button and can be used by applications based on OpenSSL to show the
17870 relationship to the OpenSSL project.
17872 *Ralf S. Engelschall*
17874 * Remove confusing variables in function signatures in files
17875 ssl/ssl_lib.c and ssl/ssl.h.
17877 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17879 * Don't install bss_file.c under PREFIX/include/
17881 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17883 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17884 functions that return function pointers and has support for NT specific
17885 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17886 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17887 unsigned to signed types: this was killing the Win32 compile.
17891 * Add new certificate file to stack functions,
17892 SSL_add_dir_cert_subjects_to_stack() and
17893 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17894 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17895 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17896 This means that Apache-SSL and similar packages don't have to mess around
17897 to add as many CAs as they want to the preferred list.
17901 * Experiment with doxygen documentation. Currently only partially applied to
17903 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17904 openssl.doxy as the configuration file.
17908 * Get rid of remaining C++-style comments which strict C compilers hate.
17910 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17912 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17913 compiled in by default: it has problems with large keys.
17917 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17918 DH private keys and/or callback functions which directly correspond to
17919 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17920 is needed for applications which have to configure certificates on a
17921 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17923 For the RSA certificate situation is makes no difference, but
17924 for the DSA certificate situation this fixes the "no shared cipher"
17925 problem where the OpenSSL cipher selection procedure failed because the
17926 temporary keys were not overtaken from the context and the API provided
17927 no way to reconfigure them.
17928 The new functions now let applications reconfigure the stuff and they
17929 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17930 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17931 non-public-API function ssl_cert_instantiate() is used as a helper
17932 function and also to reduce code redundancy inside ssl_rsa.c.
17934 *Ralf S. Engelschall*
17936 * Move s_server -dcert and -dkey options out of the undocumented feature
17937 area because they are useful for the DSA situation and should be
17938 recognized by the users.
17940 *Ralf S. Engelschall*
17942 * Fix the cipher decision scheme for export ciphers: the export bits are
17943 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17944 SSL_EXP_MASK. So, the original variable has to be used instead of the
17945 already masked variable.
17947 *Richard Levitte <levitte@stacken.kth.se>*
17949 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17951 *Richard Levitte <levitte@stacken.kth.se>*
17953 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17954 from `int` to `unsigned int` because it is a length and initialized by
17955 EVP_DigestFinal() which expects an `unsigned int *`.
17957 *Richard Levitte <levitte@stacken.kth.se>*
17959 * Don't hard-code path to Perl interpreter on shebang line of Configure
17960 script. Instead use the usual Shell->Perl transition trick.
17962 *Ralf S. Engelschall*
17964 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17965 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17966 -noout -modulus` as it's already the case for `openssl rsa -noout
17967 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17968 currently the public key is printed (a decision which was already done by
17969 `openssl dsa -modulus` in the past) which serves a similar purpose.
17970 Additionally the NO_RSA no longer completely removes the whole -modulus
17971 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17974 *Ralf S. Engelschall*
17976 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17977 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17979 *Arne Ansper <arne@ats.cyber.ee>*
17981 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17982 to be added. Now both 'req' and 'ca' can use new objects defined in the
17987 * Add cool BIO that does syslog (or event log on NT).
17989 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
17991 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
17992 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
17993 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
17994 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
17998 * Add preliminary config info for new extension code.
18002 * Make RSA_NO_PADDING really use no padding.
18004 *Ulf Moeller <ulf@fitug.de>*
18006 * Generate errors when private/public key check is done.
18010 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18011 for some CRL extensions and new objects added.
18015 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18016 key usage extension and fuller support for authority key id.
18020 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18021 padding method for RSA, which is recommended for new applications in PKCS
18022 #1 v2.0 (RFC 2437, October 1998).
18023 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18024 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18025 against Bleichbacher's attack on RSA.
18026 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18029 * Updates to the new SSL compression code
18031 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18033 * Fix so that the version number in the master secret, when passed
18034 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18035 (because the server will not accept higher), that the version number
18036 is 0x03,0x01, not 0x03,0x00
18038 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18040 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
18041 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
18042 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
18046 * Support for RAW extensions where an arbitrary extension can be
18047 created by including its DER encoding. See apps/openssl.cnf for
18052 * Make sure latest Perl versions don't interpret some generated C array
18053 code as Perl array code in the crypto/err/err_genc.pl script.
18055 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18057 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18058 not many people have the assembler. Various Win32 compilation fixes and
18059 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18060 build instructions.
18064 * Modify configure script 'Configure' to automatically create crypto/date.h
18065 file under Win32 and also build pem.h from pem.org. New script
18066 util/mkfiles.pl to create the MINFO file on environments that can't do a
18067 'make files': perl util/mkfiles.pl >MINFO should work.
18071 * Major rework of DES function declarations, in the pursuit of correctness
18072 and purity. As a result, many evil casts evaporated, and some weirdness,
18073 too. You may find this causes warnings in your code. Zapping your evil
18074 casts will probably fix them. Mostly.
18078 * Fix for a typo in asn1.h. Bug fix to object creation script
18079 obj_dat.pl. It considered a zero in an object definition to mean
18080 "end of object": none of the objects in objects.h have any zeros
18081 so it wasn't spotted.
18083 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18085 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18086 Masking (CBCM). In the absence of test vectors, the best I have been able
18087 to do is check that the decrypt undoes the encrypt, so far. Send me test
18088 vectors if you have them.
18092 * Correct calculation of key length for export ciphers (too much space was
18093 allocated for null ciphers). This has not been tested!
18097 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18098 message is now correct (it understands "crypto" and "ssl" on its
18099 command line). There is also now an "update" option. This will update
18100 the util/ssleay.num and util/libeay.num files with any new functions.
18102 perl util/mkdef.pl crypto ssl update
18103 it will update them.
18107 * Overhauled the Perl interface:
18108 - ported BN stuff to OpenSSL's different BN library
18109 - made the perl/ source tree CVS-aware
18110 - renamed the package from SSLeay to OpenSSL (the files still contain
18111 their history because I've copied them in the repository)
18112 - removed obsolete files (the test scripts will be replaced
18113 by better Test::Harness variants in the future)
18115 *Ralf S. Engelschall*
18117 * First cut for a very conservative source tree cleanup:
18118 1. merge various obsolete readme texts into doc/ssleay.txt
18119 where we collect the old documents and readme texts.
18120 2. remove the first part of files where I'm already sure that we no
18121 longer need them because of three reasons: either they are just temporary
18122 files which were left by Eric or they are preserved original files where
18123 I've verified that the diff is also available in the CVS via "cvs diff
18124 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18125 the crypto/md/ stuff).
18127 *Ralf S. Engelschall*
18129 * More extension code. Incomplete support for subject and issuer alt
18130 name, issuer and authority key id. Change the i2v function parameters
18131 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18132 what that's for :-) Fix to ASN1 macro which messed up
18133 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18137 * Preliminary support for ENUMERATED type. This is largely copied from the
18142 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18144 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18146 * Make sure `make rehash` target really finds the `openssl` program.
18148 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18150 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18151 like to hear about it if this slows down other processors.
18155 * Add CygWin32 platform information to Configure script.
18157 *Alan Batie <batie@aahz.jf.intel.com>*
18159 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18161 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18163 * New program nseq to manipulate netscape certificate sequences
18167 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18172 * Fixes to BN code. Previously the default was to define BN_RECURSION
18173 but the BN code had some problems that would cause failures when
18174 doing certificate verification and some other functions.
18176 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18178 * Add ASN1 and PEM code to support netscape certificate sequences.
18182 * Add ASN1 and PEM code to support netscape certificate sequences.
18186 * Add several PKIX and private extended key usage OIDs.
18190 * Modify the 'ca' program to handle the new extension code. Modify
18191 openssl.cnf for new extension format, add comments.
18195 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18196 and add a sample to openssl.cnf so req -x509 now adds appropriate
18201 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18202 error code, add initial support to X509_print() and x509 application.
18206 * Takes a deep breath and start adding X509 V3 extension support code. Add
18207 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18208 stuff is currently isolated and isn't even compiled yet.
18212 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18213 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18214 Removed the versions check from X509 routines when loading extensions:
18215 this allows certain broken certificates that don't set the version
18216 properly to be processed.
18220 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18221 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18222 can still be regenerated with "make depend".
18226 * Spelling mistake in C version of CAST-128.
18228 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18230 * Changes to the error generation code. The perl script err-code.pl
18231 now reads in the old error codes and retains the old numbers, only
18232 adding new ones if necessary. It also only changes the .err files if new
18233 codes are added. The makefiles have been modified to only insert errors
18234 when needed (to avoid needlessly modifying header files). This is done
18235 by only inserting errors if the .err file is newer than the auto generated
18236 C file. To rebuild all the error codes from scratch (the old behaviour)
18237 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18238 or delete all the .err files.
18242 * CAST-128 was incorrectly implemented for short keys. The C version has
18243 been fixed, but is untested. The assembler versions are also fixed, but
18244 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18245 to regenerate it if needed.
18246 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18247 Hagino <itojun@kame.net>*
18249 * File was opened incorrectly in randfile.c.
18251 *Ulf Möller <ulf@fitug.de>*
18253 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18254 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18255 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18256 al: it's just almost always a UTCTime. Note this patch adds new error
18257 codes so do a "make errors" if there are problems.
18261 * Correct Linux 1 recognition in config.
18263 *Ulf Möller <ulf@fitug.de>*
18265 * Remove pointless MD5 hash when using DSA keys in ca.
18267 *Anonymous <nobody@replay.com>*
18269 * Generate an error if given an empty string as a cert directory. Also
18270 generate an error if handed NULL (previously returned 0 to indicate an
18271 error, but didn't set one).
18273 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18275 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18279 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18280 parameters. This was causing a warning which killed off the Win32 compile.
18284 * Remove C++ style comments from crypto/bn/bn_local.h.
18286 *Neil Costigan <neil.costigan@celocom.com>*
18288 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18289 based on a text string, looking up short and long names and finally
18290 "dot" format. The "dot" format stuff didn't work. Added new function
18291 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18292 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18293 OID is not part of the table.
18297 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18298 X509_LOOKUP_by_alias().
18302 * Sort openssl functions by name.
18306 * Get the gendsa program working (hopefully) and add it to app list. Remove
18307 encryption from sample DSA keys (in case anyone is interested the password
18312 * Make *all* `*_free` functions accept a NULL pointer.
18314 *Frans Heymans <fheymans@isaserver.be>*
18316 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18319 *Anonymous <nobody@replay.com>*
18321 * s_server should send the CAfile as acceptable CAs, not its own cert.
18323 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18325 * Don't blow it for numeric -newkey arguments to apps/req.
18327 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18329 * Temp key "for export" tests were wrong in s3_srvr.c.
18331 *Anonymous <nobody@replay.com>*
18333 * Add prototype for temp key callback functions
18334 SSL_CTX_set_tmp_{rsa,dh}_callback().
18338 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18339 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18343 * X509_name_add_entry() freed the wrong thing after an error.
18345 *Arne Ansper <arne@ats.cyber.ee>*
18347 * rsa_eay.c would attempt to free a NULL context.
18349 *Arne Ansper <arne@ats.cyber.ee>*
18351 * BIO_s_socket() had a broken should_retry() on Windoze.
18353 *Arne Ansper <arne@ats.cyber.ee>*
18355 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18357 *Arne Ansper <arne@ats.cyber.ee>*
18359 * Make sure the already existing X509_STORE->depth variable is initialized
18360 in X509_STORE_new(), but document the fact that this variable is still
18361 unused in the certificate verification process.
18363 *Ralf S. Engelschall*
18365 * Fix the various library and apps files to free up pkeys obtained from
18366 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18370 * Fix reference counting in X509_PUBKEY_get(). This makes
18371 demos/maurice/example2.c work, amongst others, probably.
18373 *Steve Henson and Ben Laurie*
18375 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18376 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18377 are no longer created. This way we have a single and consistent command
18378 line interface `openssl <command>`, similar to `cvs <command>`.
18380 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18382 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18383 BIT STRING wrapper always have zero unused bits.
18387 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18391 * Make the top-level INSTALL documentation easier to understand.
18395 * Makefiles updated to exit if an error occurs in a sub-directory
18396 make (including if user presses ^C) [Paul Sutton]
18398 * Make Montgomery context stuff explicit in RSA data structure.
18402 * Fix build order of pem and err to allow for generated pem.h.
18406 * Fix renumbering bug in X509_NAME_delete_entry().
18410 * Enhanced the err-ins.pl script so it makes the error library number
18411 global and can add a library name. This is needed for external ASN1 and
18412 other error libraries.
18416 * Fixed sk_insert which never worked properly.
18420 * Fix ASN1 macros so they can handle indefinite length constructed
18421 EXPLICIT tags. Some non standard certificates use these: they can now
18426 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18427 into a single doc/ssleay.txt bundle. This way the information is still
18428 preserved but no longer messes up this directory. Now it's new room for
18429 the new set of documentation files.
18431 *Ralf S. Engelschall*
18433 * SETs were incorrectly DER encoded. This was a major pain, because they
18434 shared code with SEQUENCEs, which aren't coded the same. This means that
18435 almost everything to do with SETs or SEQUENCEs has either changed name or
18436 number of arguments.
18438 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18440 * Fix test data to work with the above.
18444 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18445 was already fixed by Eric for 0.9.1 it seems.
18447 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18449 * Autodetect FreeBSD3.
18453 * Fix various bugs in Configure. This affects the following platforms:
18457 unixware-2.0-pentium
18462 * Eliminate generated files from CVS. Reorder tests to regenerate files
18463 before they are needed.
18467 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18471 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18473 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18474 changed SSLeay to OpenSSL in version strings.
18476 *Ralf S. Engelschall*
18478 * Some fixups to the top-level documents.
18482 * Fixed the nasty bug where rsaref.h was not found under compile-time
18483 because the symlink to include/ was missing.
18485 *Ralf S. Engelschall*
18487 * Incorporated the popular no-RSA/DSA-only patches
18488 which allow to compile a RSA-free SSLeay.
18490 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18492 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18493 when "ssleay" is still not found.
18495 *Ralf S. Engelschall*
18497 * Added more platforms to Configure: Cray T3E, HPUX 11,
18499 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18501 * Updated the README file.
18503 *Ralf S. Engelschall*
18505 * Added various .cvsignore files in the CVS repository subdirs
18506 to make a "cvs update" really silent.
18508 *Ralf S. Engelschall*
18510 * Recompiled the error-definition header files and added
18511 missing symbols to the Win32 linker tables.
18513 *Ralf S. Engelschall*
18515 * Cleaned up the top-level documents;
18516 o new files: CHANGES and LICENSE
18517 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18518 o merged COPYRIGHT into LICENSE
18519 o removed obsolete TODO file
18520 o renamed MICROSOFT to INSTALL.W32
18522 *Ralf S. Engelschall*
18524 * Removed dummy files from the 0.9.1b source tree:
18525 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18526 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18527 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18528 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18529 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18531 *Ralf S. Engelschall*
18533 * Added various platform portability fixes.
18537 * The Genesis of the OpenSSL rpject:
18538 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18539 Young and Tim J. Hudson created while they were working for C2Net until
18542 *The OpenSSL Project*
18544 ### Changes between 0.9.0b and 0.9.1b [not released]
18546 * Updated a few CA certificates under certs/
18550 * Changed some BIGNUM api stuff.
18554 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18555 DGUX x86, Linux Alpha, etc.
18559 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18560 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18565 * Add -strparse option to asn1pars program which parses nested
18568 *Dr Stephen Henson <shenson@bigfoot.com>*
18570 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18574 * DSA fix for "ca" program.
18578 * Added "-genkey" option to "dsaparam" program.
18582 * Added RIPE MD160 (rmd160) message digest.
18586 * Added -a (all) option to "ssleay version" command.
18590 * Added PLATFORM define which is the id given to Configure.
18594 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18598 * Extended the ASN.1 parser routines.
18602 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18606 * Added a BN_CTX to the BN library.
18610 * Fixed the weak key values in DES library
18614 * Changed API in EVP library for cipher aliases.
18618 * Added support for RC2/64bit cipher.
18622 * Converted the lhash library to the crypto/mem.c functions.
18626 * Added more recognized ASN.1 object ids.
18630 * Added more RSA padding checks for SSL/TLS.
18634 * Added BIO proxy/filter functionality.
18638 * Added extra_certs to SSL_CTX which can be used
18639 send extra CA certificates to the client in the CA cert chain sending
18640 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18644 * Now Fortezza is denied in the authentication phase because
18645 this is key exchange mechanism is not supported by SSLeay at all.
18649 * Additional PKCS1 checks.
18653 * Support the string "TLSv1" for all TLS v1 ciphers.
18657 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18658 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18662 * Fixed a few memory leaks.
18666 * Fixed various code and comment typos.
18670 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18671 bytes sent in the client random.
18673 *Edward Bishop <ebishop@spyglass.com>*
18677 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18678 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18679 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18680 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18681 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18682 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18683 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18684 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18685 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18686 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18687 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18688 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18689 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18690 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18691 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18692 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18693 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18694 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18695 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18696 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18697 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18698 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18699 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18700 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18701 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18702 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18703 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18704 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18705 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18706 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18707 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18708 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18709 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18710 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18711 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18712 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18713 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18714 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18715 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18716 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18717 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18718 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18719 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18720 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18721 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18722 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18723 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18724 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18725 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18726 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18727 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18728 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18729 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18730 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18731 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18732 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18733 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18734 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18735 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18736 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18737 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18738 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18739 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18740 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18741 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18742 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18743 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18744 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18745 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18746 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18747 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18748 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18749 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18750 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18751 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18752 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18753 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18754 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18755 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18756 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18757 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18758 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18759 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18760 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18761 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18762 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18763 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18764 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18765 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18766 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18767 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18768 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18769 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18770 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18771 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18772 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18773 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18774 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18775 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18776 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18777 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18778 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18779 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18780 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18781 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18782 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18783 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18784 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18785 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18786 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18787 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18788 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18789 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18790 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18791 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18792 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18793 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18794 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18795 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18796 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18797 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18798 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18799 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18800 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18801 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18802 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18803 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18804 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18805 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18806 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18807 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18808 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18809 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18810 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18811 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18812 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18813 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18814 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18815 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18816 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18817 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18818 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18819 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18820 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18821 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18822 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18823 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18824 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18825 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18826 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18827 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18828 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18829 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18830 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18831 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18832 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18833 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18834 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18835 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18836 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18837 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18838 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655