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 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
27 to ignore unknown ciphers.
31 * The -cipher-commands and -digest-commands options of the command line
32 utility list has been deprecated.
33 Instead use the -cipher-algorithms and -digest-algorithms options.
37 * Deprecated all the libcrypto and libssl error string loading
38 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
39 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
40 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
41 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
42 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
43 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
44 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
45 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
46 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
47 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
49 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
50 now loads error strings automatically.
54 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
55 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
56 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
57 are to be used by servers requiring ephemeral DH keys. Instead applications
58 should consider using the built-in DH parameters that are available by
59 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
60 necessary then applications can use the alternative functions
61 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
62 replacement for the "callback" functions. The callback was originally useful
63 in order to have different parameters for export and non-export ciphersuites.
64 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
65 functions should be replaced by one of the other methods described above.
69 * The -crypt option to the passwd command line tool has been removed.
73 * The -C option to the x509, dhparam, dsaparam, and ecparam commands
78 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
80 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
81 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
82 The inverse ciphers use AES decryption for wrapping, and
83 AES encryption for unwrapping.
87 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
88 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
89 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
90 OpenSSL 3.0 these are replaced by the more generic functions
91 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
92 The old versions have been converted to deprecated macros that just call the
97 * The security callback, which can be customised by application code, supports
98 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
99 in the "other" parameter. In most places this is what is passed. All these
100 places occur server side. However there was one client side call of this
101 security operation and it passed a DH object instead. This is incorrect
102 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
103 of the other locations. Therefore this client side call has been changed to
104 pass an EVP_PKEY instead.
108 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
109 interface. Their functionality remains unchanged.
113 * Added new option for 'openssl list', '-providers', which will display the
114 list of loaded providers, their names, version and status. It optionally
115 displays their gettable parameters.
119 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
120 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
121 type is internally recognised so the workaround is no longer needed.
123 Functionality is still retained as it is, but will only work with
124 EVP_PKEYs with a legacy internal key.
128 * Deprecated EVP_PKEY_CTX_set_rsa_keygen_pubexp() & introduced
129 EVP_PKEY_CTX_set1_rsa_keygen_pubexp(), which is now preferred.
133 * Changed all "STACK" functions to be macros instead of inline functions. Macro
134 parameters are still checked for type safety at compile time via helper
139 * Remove the RAND_DRBG API
141 The RAND_DRBG API did not fit well into the new provider concept as
142 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
143 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
144 and some of its API calls are rather low-level. This holds in particular
145 for the callback mechanism (RAND_DRBG_set_callbacks()).
147 Adding a compatibility layer to continue supporting the RAND_DRBG API as
148 a legacy API for a regular deprecation period turned out to come at the
149 price of complicating the new provider API unnecessarily. Since the
150 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
153 *Paul Dale and Matthias St. Pierre*
155 * Allow SSL_set1_host() and SSL_add1_host() to take IP literal addresses
156 as well as actual hostnames.
160 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
161 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
162 conversely, silently ignore DTLS protocol version bounds when configuring
163 TLS-based contexts. The commands can be repeated to set bounds of both
164 types. The same applies with the corresponding "min_protocol" and
165 "max_protocol" command-line switches, in case some application uses both TLS
168 SSL_CTX instances that are created for a fixed protocol version (e.g.
169 TLSv1_server_method()) also silently ignore version bounds. Previously
170 attempts to apply bounds to these protocol versions would result in an
171 error. Now only the "version-flexible" SSL_CTX instances are subject to
172 limits in configuration files in command-line options.
176 * Deprecated the `ENGINE` API. Engines should be replaced with providers
181 * Reworked the recorded ERR codes to make better space for system errors.
182 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
183 given code is a system error (true) or an OpenSSL error (false).
187 * Reworked the test perl framework to better allow parallel testing.
189 *Nicola Tuveri and David von Oheimb*
191 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
192 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
196 * 'Configure' has been changed to figure out the configuration target if
197 none is given on the command line. Consequently, the 'config' script is
198 now only a mere wrapper. All documentation is changed to only mention
201 *Rich Salz and Richard Levitte*
203 * Added a library context `OSSL_LIB_CTX` that applications as well as
204 other libraries can use to form a separate context within which
205 libcrypto operations are performed.
207 There are two ways this can be used:
209 - Directly, by passing a library context to functions that take
210 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
212 - Indirectly, by creating a new library context and then assigning
213 it as the new default, with `OSSL_LIB_CTX_set0_default`.
215 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
216 apart from the functions directly related to `OSSL_LIB_CTX`, accept
217 NULL to indicate that the default library context should be used.
219 Library code that changes the default library context using
220 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
221 second call before returning to the caller.
223 _(Note: the library context was initially called `OPENSSL_CTX` and
224 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
228 * Handshake now fails if Extended Master Secret extension is dropped
233 * Dropped interactive mode from the 'openssl' program. From now on,
234 the `openssl` command without arguments is equivalent to `openssl
239 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
240 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
241 While the old function names have been retained for backward compatibility
242 they should not be used in new developments
243 because their return values are confusing: Unlike other `_cmp()` functions
244 they do not return 0 in case their arguments are equal.
248 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
249 EC_GROUP_get_field_type().
253 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
254 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
255 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
256 Applications should rely on the library automatically assigning a suitable
257 EC_METHOD internally upon EC_GROUP construction.
261 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
262 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
263 assigned internally without application intervention.
264 Users of EC_GROUP_new() should switch to a different suitable constructor.
268 * Add CAdES-BES signature verification support, mostly derived
269 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
271 *Filipe Raimundo da Silva*
273 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
277 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
278 functions are not widely used and now OpenSSL automatically perform this
279 conversion when needed.
283 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
284 EC_KEY_precompute_mult(). These functions are not widely used and
285 applications should instead switch to named curves which OpenSSL has
286 hardcoded lookup tables for.
290 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
291 should instead use the L<EC_POINT_mul(3)> function.
295 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
296 that are not applicable to the new provider model. Applications should
297 instead use EVP_default_properties_is_fips_enabled() and
298 EVP_default_properties_enable_fips().
302 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
303 is set, an unexpected EOF is ignored, it pretends a close notify was received
304 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
308 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
309 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
310 used and applications should instead use the
311 L<EC_POINT_set_affine_coordinates(3)> and
312 L<EC_POINT_get_affine_coordinates(3)> functions.
316 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
317 arrays to be more easily constructed via a series of utility functions.
318 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
319 the various push functions and finally convert to a passable OSSL_PARAM
320 array using OSSL_PARAM_BLD_to_param().
324 * The security strength of SHA1 and MD5 based signatures in TLS has been
325 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
326 working at the default security level of 1 and instead requires security
327 level 0. The security level can be changed either using the cipher string
328 with @SECLEVEL, or calling SSL_CTX_set_security_level().
332 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
333 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
334 internal keys, if they correspond to one of those built in types.
338 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
339 contain a provider side internal key.
343 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
344 They are old functions that we don't use, and that you could disable with
345 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
349 * Project text documents not yet having a proper file name extension
350 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
351 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
355 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
356 have been converted to Markdown with the goal to produce documents
357 which not only look pretty when viewed online in the browser, but
358 remain well readable inside a plain text editor.
360 To achieve this goal, a 'minimalistic' Markdown style has been applied
361 which avoids formatting elements that interfere too much with the
362 reading flow in the text file. For example, it
364 * avoids [ATX headings][] and uses [setext headings][] instead
365 (which works for `<h1>` and `<h2>` headings only).
366 * avoids [inline links][] and uses [reference links][] instead.
367 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
369 [ATX headings]: https://github.github.com/gfm/#atx-headings
370 [setext headings]: https://github.github.com/gfm/#setext-headings
371 [inline links]: https://github.github.com/gfm/#inline-link
372 [reference links]: https://github.github.com/gfm/#reference-link
373 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
374 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
376 *Matthias St. Pierre*
378 * The test suite is changed to preserve results of each test recipe.
379 A new directory test-runs/ with subdirectories named like the
380 test recipes are created in the build tree for this purpose.
384 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
385 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
386 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
388 *David von Oheimb, Martin Peylo*
390 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
391 The legacy OCSP-focused and only partly documented API is retained.
392 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
396 * Added `util/check-format.pl`, a tool for checking adherence to the
397 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
398 The checks performed are incomplete and yield some false positives.
399 Still the tool should be useful for detecting most typical glitches.
403 * BIO_do_connect and BIO_do_handshake have been extended:
404 If domain name resolution yields multiple IP addresses all of them are tried
405 after connect() failures.
409 * All of the low level RSA functions have been deprecated including:
411 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
412 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
413 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
414 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
415 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
416 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
417 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
418 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
419 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
420 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
421 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
422 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
423 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
424 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
425 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
426 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
427 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
428 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
429 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
430 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
431 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
432 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
433 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
434 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
435 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
436 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
437 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
438 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
440 Use of these low level functions has been informally discouraged for a long
441 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
442 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
443 L<EVP_PKEY_decrypt(3)>.
447 * X509 certificates signed using SHA1 are no longer allowed at security
449 In TLS/SSL the default security level is 1. It can be set either
450 using the cipher string with @SECLEVEL, or calling
451 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
452 a call to SSL_CTX_use_certificate() will fail if the security level is not
454 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
455 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
460 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
461 modified to use PKEY APIs. These commands are now in maintenance mode
462 and no new features will be added to them.
466 * The command line utility rsautl has been deprecated.
467 Instead use the pkeyutl program.
471 * The command line utilities genrsa and rsa have been modified to use PKEY
472 APIs. They now write PKCS#8 keys by default. These commands are now in
473 maintenance mode and no new features will be added to them.
477 * All of the low level DH functions have been deprecated including:
479 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
480 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
481 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
482 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
483 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
484 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
485 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
486 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
487 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
488 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
489 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
490 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
491 DH_meth_get_generate_params and DH_meth_set_generate_params.
493 Use of these low level functions has been informally discouraged for a long
494 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
495 and L<EVP_PKEY_derive(3)>.
497 Additionally functions that read and write DH objects such as d2i_DHparams,
498 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
499 functions have also been deprecated. Applications should instead use the
500 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
502 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
503 EVP_PKEY_assign_DH(), EVP_PKEY_get0_DH, EVP_PKEY_get1_DH, EVP_PKEY_set1_DH
504 are also deprecated. Applications should instead either read or write an
505 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs. Or load an
506 EVP_PKEY directly from DH data using EVP_PKEY_fromdata().
508 *Paul Dale and Matt Caswell*
510 * All of the low level DSA functions have been deprecated including:
512 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
513 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
514 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
515 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
516 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
517 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
518 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
519 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
520 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
521 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
522 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
523 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
524 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
525 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
526 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
527 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
528 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
530 Use of these low level functions has been informally discouraged for a long
531 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
532 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
536 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
537 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
538 This means that applications don't have to look at the curve NID and
539 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
540 However, they still can, that EVP_PKEY_set_alias_type() call acts as
541 a no-op when the EVP_PKEY is already of the given type.
543 Parameter and key generation is also reworked to make it possible
544 to generate EVP_PKEY_SM2 parameters and keys without having to go
545 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
546 However, code that does the latter will still work as before.
550 * Deprecated low level ECDH and ECDSA functions. These include:
552 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
553 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
556 Use of these low level functions has been informally discouraged for a long
557 time. Instead applications should use the EVP_PKEY_derive(3),
558 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
562 * Deprecated the EC_KEY_METHOD functions. These include:
564 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
565 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
566 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
567 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
568 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
569 EC_KEY_METHOD_get_verify.
571 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
575 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
576 and EVP_PKEY_decrypt() instead.
577 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
578 and EVP_PKEY_encrypt() instead.
582 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
583 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
584 a new formulation to include all the things it can be used for,
585 as well as words of caution.
589 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
590 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
594 * All of the low level HMAC functions have been deprecated including:
596 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
597 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
600 Use of these low level functions has been informally discouraged for a long
601 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
602 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
603 and L<EVP_MAC_final(3)>.
607 * Over two thousand fixes were made to the documentation, including:
608 - Common options (such as -rand/-writerand, TLS version control, etc)
609 were refactored and point to newly-enhanced descriptions in openssl.pod.
610 - Added style conformance for all options (with help from Richard Levitte),
611 documented all reported missing options, added a CI build to check
612 that all options are documented and that no unimplemented options
614 - Documented some internals, such as all use of environment variables.
615 - Addressed all internal broken L<> references.
619 * All of the low level CMAC functions have been deprecated including:
621 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
622 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
624 Use of these low level functions has been informally discouraged for a long
625 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
626 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
627 and L<EVP_MAC_final(3)>.
631 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
632 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
635 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
636 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
637 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
638 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
639 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
640 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
641 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
642 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
643 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
644 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
646 Use of these low level functions has been informally discouraged
647 for a long time. Applications should use the EVP_DigestInit_ex(3),
648 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
652 * Corrected the documentation of the return values from the `EVP_DigestSign*`
653 set of functions. The documentation mentioned negative values for some
654 errors, but this was never the case, so the mention of negative values
657 Code that followed the documentation and thereby check with something
658 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
662 * All of the low level cipher functions have been deprecated including:
664 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
665 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
666 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
667 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
668 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
669 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
670 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
671 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
672 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
673 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
674 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
675 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
676 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
677 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
678 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
679 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
680 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
681 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
682 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
683 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
684 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
685 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
686 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
687 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
688 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
689 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
690 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
691 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
692 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
694 Use of these low level functions has been informally discouraged for
695 a long time. Applications should use the high level EVP APIs, e.g.
696 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
697 equivalently named decrypt functions instead.
699 *Matt Caswell and Paul Dale*
701 * Removed include/openssl/opensslconf.h.in and replaced it with
702 include/openssl/configuration.h.in, which differs in not including
703 <openssl/macros.h>. A short header include/openssl/opensslconf.h
704 was added to include both.
706 This allows internal hacks where one might need to modify the set
707 of configured macros, for example this if deprecated symbols are
708 still supposed to be available internally:
710 #include <openssl/configuration.h>
712 #undef OPENSSL_NO_DEPRECATED
713 #define OPENSSL_SUPPRESS_DEPRECATED
715 #include <openssl/macros.h>
717 This should not be used by applications that use the exported
718 symbols, as that will lead to linking errors.
722 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
723 used in exponentiation with 512-bit moduli. No EC algorithms are
724 affected. Analysis suggests that attacks against 2-prime RSA1024,
725 3-prime RSA1536, and DSA1024 as a result of this defect would be very
726 difficult to perform and are not believed likely. Attacks against DH512
727 are considered just feasible. However, for an attack the target would
728 have to re-use the DH512 private key, which is not recommended anyway.
729 Also applications directly using the low level API BN_mod_exp may be
730 affected if they use BN_FLG_CONSTTIME.
735 * Most memory-debug features have been deprecated, and the functionality
736 replaced with no-ops.
740 * Added documentation for the STACK API.
744 * Introduced a new method type and API, OSSL_ENCODER, to
745 represent generic encoders. An implementation is expected to
746 be able to encode an object associated with a given name (such
747 as an algorithm name for an asymmetric key) into forms given by
748 implementation properties.
750 Encoders are primarily used from inside libcrypto, through
751 calls to functions like EVP_PKEY_print_private(),
752 PEM_write_bio_PrivateKey() and similar.
754 Encoders are specified in such a way that they can be made to
755 directly handle the provider side portion of an object, if this
756 provider side part comes from the same provider as the encoder
757 itself, but can also be made to handle objects in parametrized
758 form (as an OSSL_PARAM array of data). This allows a provider to
759 offer generic encoders as a service for any other provider.
763 * Added a .pragma directive to the syntax of configuration files, to
764 allow varying behavior in a supported and predictable manner.
765 Currently added pragma:
769 This allows dollar signs to be a keyword character unless it's
770 followed by a opening brace or parenthesis. This is useful for
771 platforms where dollar signs are commonly used in names, such as
772 volume names and system directory names on VMS.
776 * Added functionality to create an EVP_PKEY from user data. This
777 is effectively the same as creating a RSA, DH or DSA object and
778 then assigning them to an EVP_PKEY, but directly using algorithm
779 agnostic EVP functions. A benefit is that this should be future
780 proof for public key algorithms to come.
784 * Change the interpretation of the '--api' configuration option to
785 mean that this is a desired API compatibility level with no
786 further meaning. The previous interpretation, that this would
787 also mean to remove all deprecated symbols up to and including
788 the given version, no requires that 'no-deprecated' is also used
789 in the configuration.
791 When building applications, the desired API compatibility level
792 can be set with the OPENSSL_API_COMPAT macro like before. For
793 API compatibility version below 3.0, the old style numerical
794 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
795 For version 3.0 and on, the value is expected to be the decimal
796 value calculated from the major and minor version like this:
798 MAJOR * 10000 + MINOR * 100
802 -DOPENSSL_API_COMPAT=30000 For 3.0
803 -DOPENSSL_API_COMPAT=30200 For 3.2
805 To hide declarations that are deprecated up to and including the
806 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
807 given when building the application as well.
811 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
812 access to certificate and CRL stores via URIs and OSSL_STORE
815 This adds the following functions:
817 - X509_LOOKUP_store()
818 - X509_STORE_load_file()
819 - X509_STORE_load_path()
820 - X509_STORE_load_store()
821 - SSL_add_store_cert_subjects_to_stack()
822 - SSL_CTX_set_default_verify_store()
823 - SSL_CTX_load_verify_file()
824 - SSL_CTX_load_verify_dir()
825 - SSL_CTX_load_verify_store()
829 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
830 The presence of this system service is determined at run-time.
834 * Added functionality to create an EVP_PKEY context based on data
835 for methods from providers. This takes an algorithm name and a
836 property query string and simply stores them, with the intent
837 that any operation that uses this context will use those strings
838 to fetch the needed methods implicitly, thereby making the port
839 of application written for pre-3.0 OpenSSL easier.
843 * The undocumented function NCONF_WIN32() has been deprecated; for
844 conversion details see the HISTORY section of doc/man5/config.pod
848 * Introduced the new functions EVP_DigestSignInit_ex() and
849 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
850 EVP_DigestVerifyUpdate() have been converted to functions. See the man
851 pages for further details.
855 * Over two thousand fixes were made to the documentation, including:
856 adding missing command flags, better style conformance, documentation
859 *Rich Salz, Richard Levitte*
861 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
862 X25519, X448, Ed25519 and Ed448.
866 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
871 * Deprecated the public definition of ERR_STATE as well as the function
872 ERR_get_state(). This is done in preparation of making ERR_STATE an
877 * Added ERR functionality to give callers access to the stored function
878 names that have replaced the older function code based functions.
880 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
881 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
882 ERR_peek_error_all() and ERR_peek_last_error_all().
884 These functions have become deprecated: ERR_get_error_line(),
885 ERR_get_error_line_data(), ERR_peek_error_line_data(),
886 ERR_peek_last_error_line_data() and ERR_func_error_string().
888 Users are recommended to use ERR_get_error_all(), or to pick information
889 with ERR_peek functions and finish off with getting the error code by using
894 * Extended testing to be verbose for failing tests only. The make variables
895 VERBOSE_FAILURE or VF can be used to enable this:
897 $ make VF=1 test # Unix
898 $ mms /macro=(VF=1) test ! OpenVMS
899 $ nmake VF=1 test # Windows
903 * Added several checks to X509_verify_cert() according to requirements in
904 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
905 (which may be done by using the CLI option `-x509_strict`):
906 * The basicConstraints of CA certificates must be marked critical.
907 * CA certificates must explicitly include the keyUsage extension.
908 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
909 * The issuer name of any certificate must not be empty.
910 * The subject name of CA certs, certs with keyUsage crlSign,
911 and certs without subjectAlternativeName must not be empty.
912 * If a subjectAlternativeName extension is given it must not be empty.
913 * The signatureAlgorithm field and the cert signature must be consistent.
914 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
915 must not be marked critical.
916 * The authorityKeyIdentifier must be given for X.509v3 certs
917 unless they are self-signed.
918 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
922 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
923 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
927 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
928 used even when parsing explicit parameters, when loading a encoded key
929 or calling `EC_GROUP_new_from_ecpkparameters()`/
930 `EC_GROUP_new_from_ecparameters()`.
931 This prevents bypass of security hardening and performance gains,
932 especially for curves with specialized EC_METHODs.
933 By default, if a key encoded with explicit parameters is loaded and later
934 encoded, the output is still encoded with explicit parameters, even if
935 internally a "named" EC_GROUP is used for computation.
939 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
940 this change, EC_GROUP_set_generator would accept order and/or cofactor as
941 NULL. After this change, only the cofactor parameter can be NULL. It also
942 does some minimal sanity checks on the passed order.
947 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
948 An attack is simple, if the first CMS_recipientInfo is valid but the
949 second CMS_recipientInfo is chosen ciphertext. If the second
950 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
951 encryption key will be replaced by garbage, and the message cannot be
952 decoded, but if the RSA decryption fails, the correct encryption key is
953 used and the recipient will not notice the attack.
954 As a work around for this potential attack the length of the decrypted
955 key must be equal to the cipher default key length, in case the
956 certifiate is not given and all recipientInfo are tried out.
957 The old behaviour can be re-enabled in the CMS code by setting the
958 CMS_DEBUG_DECRYPT flag.
962 * Early start up entropy quality from the DEVRANDOM seed source has been
963 improved for older Linux systems. The RAND subsystem will wait for
964 /dev/random to be producing output before seeding from /dev/urandom.
965 The seeded state is stored for future library initialisations using
966 a system global shared memory segment. The shared memory identifier
967 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
968 the desired value. The default identifier is 114.
972 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
973 when primes for RSA keys are computed.
974 Since we previously always generated primes == 2 (mod 3) for RSA keys,
975 the 2-prime and 3-prime RSA modules were easy to distinguish, since
976 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
977 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
978 This avoids possible fingerprinting of newly generated RSA modules.
982 * Correct the extended master secret constant on EBCDIC systems. Without this
983 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
984 negotiate EMS will fail. Unfortunately this also means that TLS connections
985 between EBCDIC systems with this fix, and EBCDIC systems without this
986 fix will fail if they negotiate EMS.
990 * Changed the library initialisation so that the config file is now loaded
991 by default. This was already the case for libssl. It now occurs for both
992 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
993 OPENSSL_init_crypto() to suppress automatic loading of a config file.
997 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
998 where the former acts as a replacement for ERR_put_error(), and the
999 latter replaces the combination ERR_put_error()+ERR_add_error_data().
1000 ERR_raise_data() adds more flexibility by taking a format string and
1001 an arbitrary number of arguments following it, to be processed with
1006 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
1007 to check if a named provider is loaded and available. When called, it
1008 will also activate all fallback providers if such are still present.
1012 * Enforce a minimum DH modulus size of 512 bits.
1016 * Changed DH parameters to generate the order q subgroup instead of 2q.
1017 Previously generated DH parameters are still accepted by DH_check
1018 but DH_generate_key works around that by clearing bit 0 of the
1019 private key for those. This avoids leaking bit 0 of the private key.
1023 * Significantly reduce secure memory usage by the randomness pools.
1027 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1032 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1033 algorithms. An implementation of a key exchange algorithm can be obtained
1034 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1035 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1036 the older EVP_PKEY_derive_init() function. See the man pages for the new
1037 functions for further details.
1041 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1045 * Removed the function names from error messages and deprecated the
1048 * Removed NextStep support and the macro OPENSSL_UNISTD
1052 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1053 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1054 Also removed "export var as function" capability; we do not export
1055 variables, only functions.
1059 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1060 an error and 1 indicating success. In previous versions of OpenSSL this
1061 was a void type. If a key was set longer than the maximum possible this
1066 * Support SM2 signing and verification schemes with X509 certificate.
1070 * Use SHA256 as the default digest for TS query in the ts app.
1074 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1075 This checks that the salt length is at least 128 bits, the derived key
1076 length is at least 112 bits, and that the iteration count is at least 1000.
1077 For backwards compatibility these checks are disabled by default in the
1078 default provider, but are enabled by default in the fips provider.
1079 To enable or disable these checks use the control
1080 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1084 * Default cipher lists/suites are now available via a function, the
1085 #defines are deprecated.
1089 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1090 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1091 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1095 * Join the directories crypto/x509 and crypto/x509v3
1099 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1100 This changes the size when using the genpkey app when no size is given. It
1101 fixes an omission in earlier changes that changed all RSA, DSA and DH
1102 generation apps to use 2048 bits by default.
1106 * Added command 'openssl kdf' that uses the EVP_KDF API.
1110 * Added command 'openssl mac' that uses the EVP_MAC API.
1114 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1115 as default directories. Also added the command 'openssl info'
1116 for scripting purposes.
1120 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1121 deprecated. These undocumented functions were never integrated into the EVP
1122 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1123 Bi-directional IGE mode. These modes were never formally standardised and
1124 usage of these functions is believed to be very small. In particular
1125 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1126 is ever used. The security implications are believed to be minimal, but
1127 this issue was never fixed for backwards compatibility reasons. New code
1128 should not use these modes.
1132 * Add prediction resistance to the DRBG reseeding process.
1136 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1137 mandated by IEEE Std 1619-2018.
1141 * Added newline escaping functionality to a filename when using openssl dgst.
1142 This output format is to replicate the output format found in the `*sum`
1143 checksum programs. This aims to preserve backward compatibility.
1145 *Matt Eaton, Richard Levitte, and Paul Dale*
1147 * Removed the heartbeat message in DTLS feature, as it has very
1148 little usage and doesn't seem to fulfill a valuable purpose.
1149 The configuration option is now deprecated.
1153 * Changed the output of 'openssl {digestname} < file' to display the
1154 digest name in its output.
1158 * Added a new generic trace API which provides support for enabling
1159 instrumentation through trace output. This feature is mainly intended
1160 as an aid for developers and is disabled by default. To utilize it,
1161 OpenSSL needs to be configured with the `enable-trace` option.
1163 If the tracing API is enabled, the application can activate trace output
1164 by registering BIOs as trace channels for a number of tracing and debugging
1167 The 'openssl' application has been expanded to enable any of the types
1168 available via environment variables defined by the user, and serves as
1169 one possible example on how to use this functionality.
1171 *Richard Levitte & Matthias St. Pierre*
1173 * Added build tests for C++. These are generated files that only do one
1174 thing, to include one public OpenSSL head file each. This tests that
1175 the public header files can be usefully included in a C++ application.
1177 This test isn't enabled by default. It can be enabled with the option
1178 'enable-buildtest-c++'.
1182 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1186 * Add KMAC to EVP_MAC.
1190 * Added property based algorithm implementation selection framework to
1195 * Added SCA hardening for modular field inversion in EC_GROUP through
1196 a new dedicated field_inv() pointer in EC_METHOD.
1197 This also addresses a leakage affecting conversions from projective
1198 to affine coordinates.
1200 *Billy Bob Brumley, Nicola Tuveri*
1202 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1203 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1204 those algorithms that were already supported through the EVP_PKEY API
1205 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1206 and scrypt are now wrappers that call EVP_KDF.
1210 * Build devcrypto engine as a dynamic engine.
1212 *Eneas U de Queiroz*
1214 * Add keyed BLAKE2 to EVP_MAC.
1218 * Fix a bug in the computation of the endpoint-pair shared secret used
1219 by DTLS over SCTP. This breaks interoperability with older versions
1220 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1221 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1222 interoperability with such broken implementations. However, enabling
1223 this switch breaks interoperability with correct implementations.
1225 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1226 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1230 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1234 * Change the license to the Apache License v2.0.
1238 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1240 - Major releases (indicated by incrementing the MAJOR release number)
1241 may introduce incompatible API/ABI changes.
1242 - Minor releases (indicated by incrementing the MINOR release number)
1243 may introduce new features but retain API/ABI compatibility.
1244 - Patch releases (indicated by incrementing the PATCH number)
1245 are intended for bug fixes and other improvements of existing
1246 features only (like improving performance or adding documentation)
1247 and retain API/ABI compatibility.
1251 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1255 * Remove the 'dist' target and add a tarball building script. The
1256 'dist' target has fallen out of use, and it shouldn't be
1257 necessary to configure just to create a source distribution.
1261 * Recreate the OS390-Unix config target. It no longer relies on a
1262 special script like it did for OpenSSL pre-1.1.0.
1266 * Instead of having the source directories listed in Configure, add
1267 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1272 * Add GMAC to EVP_MAC.
1276 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1280 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1281 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1282 to facilitate the continued use of MACs through raw private keys in
1283 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1287 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1288 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1292 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1293 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1294 are retained for backwards compatibility.
1298 * AES-XTS mode now enforces that its two keys are different to mitigate
1299 the attacked described in "Efficient Instantiations of Tweakable
1300 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1301 Details of this attack can be obtained from:
1302 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1306 * Rename the object files, i.e. give them other names than in previous
1307 versions. Their names now include the name of the final product, as
1308 well as its type mnemonic (bin, lib, shlib).
1312 * Added new option for 'openssl list', '-objects', which will display the
1313 list of built in objects, i.e. OIDs with names.
1317 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1318 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1323 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1324 improves application performance by removing data copies and providing
1325 applications with zero-copy system calls such as sendfile and splice.
1329 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1330 option is set, openssl cleanses (zeroize) plaintext bytes from
1331 internal buffers after delivering them to the application. Note,
1332 the application is still responsible for cleansing other copies
1333 (e.g.: data received by SSL_read(3)).
1337 * `PKCS12_parse` now maintains the order of the parsed certificates
1338 when outputting them via `*ca` (rather than reversing it).
1342 * Deprecated pthread fork support methods. These were unused so no
1343 replacement is required.
1345 - OPENSSL_fork_prepare()
1346 - OPENSSL_fork_parent()
1347 - OPENSSL_fork_child()
1354 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1356 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1357 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1358 If an attacker can control both items being compared then this could lead
1359 to a possible denial of service attack. OpenSSL itself uses the
1360 GENERAL_NAME_cmp function for two purposes:
1361 1) Comparing CRL distribution point names between an available CRL and a
1362 CRL distribution point embedded in an X509 certificate
1363 2) When verifying that a timestamp response token signer matches the
1364 timestamp authority name (exposed via the API functions
1365 TS_RESP_verify_response and TS_RESP_verify_token)
1370 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1372 * Certificates with explicit curve parameters are now disallowed in
1373 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1377 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1378 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1379 conversely, silently ignore DTLS protocol version bounds when configuring
1380 TLS-based contexts. The commands can be repeated to set bounds of both
1381 types. The same applies with the corresponding "min_protocol" and
1382 "max_protocol" command-line switches, in case some application uses both TLS
1385 SSL_CTX instances that are created for a fixed protocol version (e.g.
1386 TLSv1_server_method()) also silently ignore version bounds. Previously
1387 attempts to apply bounds to these protocol versions would result in an
1388 error. Now only the "version-flexible" SSL_CTX instances are subject to
1389 limits in configuration files in command-line options.
1393 * Handshake now fails if Extended Master Secret extension is dropped
1398 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1400 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1402 * Fixed segmentation fault in SSL_check_chain()
1403 Server or client applications that call the SSL_check_chain() function
1404 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1405 dereference as a result of incorrect handling of the
1406 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1407 or unrecognised signature algorithm is received from the peer. This could
1408 be exploited by a malicious peer in a Denial of Service attack.
1413 * Added AES consttime code for no-asm configurations
1414 an optional constant time support for AES was added
1415 when building openssl for no-asm.
1416 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1417 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1418 At this time this feature is by default disabled.
1419 It will be enabled by default in 3.0.
1423 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1425 * Revert the change of EOF detection while reading in libssl to avoid
1426 regressions in applications depending on the current way of reporting
1427 the EOF. As the existing method is not fully accurate the change to
1428 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1429 branch and will be present in the 3.0 release.
1433 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1434 when primes for RSA keys are computed.
1435 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1436 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1437 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1438 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1439 This avoids possible fingerprinting of newly generated RSA modules.
1443 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1445 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1446 while reading in libssl then we would report an error back to the
1447 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1448 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1449 therefore give a hint as to what went wrong.
1453 * Check that ed25519 and ed448 are allowed by the security level. Previously
1454 signature algorithms not using an MD were not being checked that they were
1455 allowed by the security level.
1459 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1460 was not quite right. The behaviour was not consistent between resumption
1461 and normal handshakes, and also not quite consistent with historical
1462 behaviour. The behaviour in various scenarios has been clarified and
1463 it has been updated to make it match historical behaviour as closely as
1468 * *[VMS only]* The header files that the VMS compilers include automatically,
1469 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1470 that the C++ compiler doesn't understand. This is a shortcoming in the
1471 compiler, but can be worked around with `__cplusplus` guards.
1473 C++ applications that use OpenSSL libraries must be compiled using the
1474 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1475 functions. Otherwise, only functions with symbols of less than 31
1476 characters can be used, as the linker will not be able to successfully
1477 resolve symbols with longer names.
1481 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1482 The presence of this system service is determined at run-time.
1486 * Added newline escaping functionality to a filename when using openssl dgst.
1487 This output format is to replicate the output format found in the `*sum`
1488 checksum programs. This aims to preserve backward compatibility.
1490 *Matt Eaton, Richard Levitte, and Paul Dale*
1492 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1497 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1499 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1500 number generator (RNG). This was intended to include protection in the
1501 event of a fork() system call in order to ensure that the parent and child
1502 processes did not share the same RNG state. However this protection was not
1503 being used in the default case.
1505 A partial mitigation for this issue is that the output from a high
1506 precision timer is mixed into the RNG state so the likelihood of a parent
1507 and child process sharing state is significantly reduced.
1509 If an application already calls OPENSSL_init_crypto() explicitly using
1510 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1513 *Matthias St. Pierre*
1515 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1516 used even when parsing explicit parameters, when loading a encoded key
1517 or calling `EC_GROUP_new_from_ecpkparameters()`/
1518 `EC_GROUP_new_from_ecparameters()`.
1519 This prevents bypass of security hardening and performance gains,
1520 especially for curves with specialized EC_METHODs.
1521 By default, if a key encoded with explicit parameters is loaded and later
1522 encoded, the output is still encoded with explicit parameters, even if
1523 internally a "named" EC_GROUP is used for computation.
1527 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1528 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1529 NULL. After this change, only the cofactor parameter can be NULL. It also
1530 does some minimal sanity checks on the passed order.
1535 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1536 An attack is simple, if the first CMS_recipientInfo is valid but the
1537 second CMS_recipientInfo is chosen ciphertext. If the second
1538 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1539 encryption key will be replaced by garbage, and the message cannot be
1540 decoded, but if the RSA decryption fails, the correct encryption key is
1541 used and the recipient will not notice the attack.
1542 As a work around for this potential attack the length of the decrypted
1543 key must be equal to the cipher default key length, in case the
1544 certifiate is not given and all recipientInfo are tried out.
1545 The old behaviour can be re-enabled in the CMS code by setting the
1546 CMS_DEBUG_DECRYPT flag.
1551 * Early start up entropy quality from the DEVRANDOM seed source has been
1552 improved for older Linux systems. The RAND subsystem will wait for
1553 /dev/random to be producing output before seeding from /dev/urandom.
1554 The seeded state is stored for future library initialisations using
1555 a system global shared memory segment. The shared memory identifier
1556 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1557 the desired value. The default identifier is 114.
1561 * Correct the extended master secret constant on EBCDIC systems. Without this
1562 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1563 negotiate EMS will fail. Unfortunately this also means that TLS connections
1564 between EBCDIC systems with this fix, and EBCDIC systems without this
1565 fix will fail if they negotiate EMS.
1569 * Use Windows installation paths in the mingw builds
1571 Mingw isn't a POSIX environment per se, which means that Windows
1572 paths should be used for installation.
1577 * Changed DH_check to accept parameters with order q and 2q subgroups.
1578 With order 2q subgroups the bit 0 of the private key is not secret
1579 but DH_generate_key works around that by clearing bit 0 of the
1580 private key for those. This avoids leaking bit 0 of the private key.
1584 * Significantly reduce secure memory usage by the randomness pools.
1588 * Revert the DEVRANDOM_WAIT feature for Linux systems
1590 The DEVRANDOM_WAIT feature added a select() call to wait for the
1591 /dev/random device to become readable before reading from the
1592 /dev/urandom device.
1594 It turned out that this change had negative side effects on
1595 performance which were not acceptable. After some discussion it
1596 was decided to revert this feature and leave it up to the OS
1597 resp. the platform maintainer to ensure a proper initialization
1598 during early boot time.
1600 *Matthias St. Pierre*
1602 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1604 * Add build tests for C++. These are generated files that only do one
1605 thing, to include one public OpenSSL head file each. This tests that
1606 the public header files can be usefully included in a C++ application.
1608 This test isn't enabled by default. It can be enabled with the option
1609 'enable-buildtest-c++'.
1613 * Enable SHA3 pre-hashing for ECDSA and DSA.
1617 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1618 This changes the size when using the genpkey app when no size is given. It
1619 fixes an omission in earlier changes that changed all RSA, DSA and DH
1620 generation apps to use 2048 bits by default.
1624 * Reorganize the manual pages to consistently have RETURN VALUES,
1625 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1626 util/fix-doc-nits accordingly.
1628 *Paul Yang, Joshua Lock*
1630 * Add the missing accessor EVP_PKEY_get0_engine()
1634 * Have apps like 's_client' and 's_server' output the signature scheme
1635 along with other cipher suite parameters when debugging.
1639 * Make OPENSSL_config() error agnostic again.
1643 * Do the error handling in RSA decryption constant time.
1647 * Prevent over long nonces in ChaCha20-Poly1305.
1649 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1650 for every encryption operation. RFC 7539 specifies that the nonce value
1651 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1652 and front pads the nonce with 0 bytes if it is less than 12
1653 bytes. However it also incorrectly allows a nonce to be set of up to 16
1654 bytes. In this case only the last 12 bytes are significant and any
1655 additional leading bytes are ignored.
1657 It is a requirement of using this cipher that nonce values are
1658 unique. Messages encrypted using a reused nonce value are susceptible to
1659 serious confidentiality and integrity attacks. If an application changes
1660 the default nonce length to be longer than 12 bytes and then makes a
1661 change to the leading bytes of the nonce expecting the new value to be a
1662 new unique nonce then such an application could inadvertently encrypt
1663 messages with a reused nonce.
1665 Additionally the ignored bytes in a long nonce are not covered by the
1666 integrity guarantee of this cipher. Any application that relies on the
1667 integrity of these ignored leading bytes of a long nonce may be further
1668 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1669 is safe because no such use sets such a long nonce value. However user
1670 applications that use this cipher directly and set a non-default nonce
1671 length to be longer than 12 bytes may be vulnerable.
1673 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1679 * Add DEVRANDOM_WAIT feature for Linux systems
1681 On older Linux systems where the getrandom() system call is not available,
1682 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1683 Contrary to getrandom(), the /dev/urandom device will not block during
1684 early boot when the kernel CSPRNG has not been seeded yet.
1686 To mitigate this known weakness, use select() to wait for /dev/random to
1687 become readable before reading from /dev/urandom.
1689 * Ensure that SM2 only uses SM3 as digest algorithm
1693 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1695 * Change the info callback signals for the start and end of a post-handshake
1696 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1697 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1698 confused by this and assume that a TLSv1.2 renegotiation has started. This
1699 can break KeyUpdate handling. Instead we no longer signal the start and end
1700 of a post handshake message exchange (although the messages themselves are
1701 still signalled). This could break some applications that were expecting
1702 the old signals. However without this KeyUpdate is not usable for many
1707 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1709 * Timing vulnerability in DSA signature generation
1711 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1712 timing side channel attack. An attacker could use variations in the signing
1713 algorithm to recover the private key.
1715 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1720 * Timing vulnerability in ECDSA signature generation
1722 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1723 timing side channel attack. An attacker could use variations in the signing
1724 algorithm to recover the private key.
1726 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1731 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1732 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1733 of two gigabytes and the error handling improved.
1735 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1736 categorized as a normal bug, not a security issue, because the DRBG reseeds
1737 automatically and is fully functional even without additional randomness
1738 provided by the application.
1740 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1742 * Add a new ClientHello callback. Provides a callback interface that gives
1743 the application the ability to adjust the nascent SSL object at the
1744 earliest stage of ClientHello processing, immediately after extensions have
1745 been collected but before they have been processed. In particular, this
1746 callback can adjust the supported TLS versions in response to the contents
1751 * Add SM2 base algorithm support.
1755 * s390x assembly pack: add (improved) hardware-support for the following
1756 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1757 aes-cfb/cfb8, aes-ecb.
1761 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1762 parameter is no longer accepted, as it leads to a corrupt table. NULL
1763 pem_str is reserved for alias entries only.
1767 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1768 step for prime curves. The new implementation is based on formulae from
1769 differential addition-and-doubling in homogeneous projective coordinates
1770 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1771 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1772 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1773 to work in projective coordinates.
1775 *Billy Bob Brumley, Nicola Tuveri*
1777 * Change generating and checking of primes so that the error rate of not
1778 being prime depends on the intended use based on the size of the input.
1779 For larger primes this will result in more rounds of Miller-Rabin.
1780 The maximal error rate for primes with more than 1080 bits is lowered
1783 *Kurt Roeckx, Annie Yousar*
1785 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1789 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1790 moving between systems, and to avoid confusion when a Windows build is
1791 done with mingw vs with MSVC. For POSIX installs, there's still a
1792 symlink or copy named 'tsget' to avoid that confusion as well.
1796 * Revert blinding in ECDSA sign and instead make problematic addition
1797 length-invariant. Switch even to fixed-length Montgomery multiplication.
1801 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1802 step for binary curves. The new implementation is based on formulae from
1803 differential addition-and-doubling in mixed Lopez-Dahab projective
1804 coordinates, modified to independently blind the operands.
1806 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1808 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1809 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1810 EC_METHODs to implement their own specialized "ladder step", to take
1811 advantage of more favorable coordinate systems or more efficient
1812 differential addition-and-doubling algorithms.
1814 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1816 * Modified the random device based seed sources to keep the relevant
1817 file descriptors open rather than reopening them on each access.
1818 This allows such sources to operate in a chroot() jail without
1819 the associated device nodes being available. This behaviour can be
1820 controlled using RAND_keep_random_devices_open().
1824 * Numerous side-channel attack mitigations have been applied. This may have
1825 performance impacts for some algorithms for the benefit of improved
1826 security. Specific changes are noted in this change log by their respective
1831 * AIX shared library support overhaul. Switch to AIX "natural" way of
1832 handling shared libraries, which means collecting shared objects of
1833 different versions and bitnesses in one common archive. This allows to
1834 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1835 doesn't affect the way 3rd party applications are linked, only how
1836 multi-version installation is managed.
1840 * Make ec_group_do_inverse_ord() more robust and available to other
1841 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1842 mitigations are applied to the fallback BN_mod_inverse().
1843 When using this function rather than BN_mod_inverse() directly, new
1844 EC cryptosystem implementations are then safer-by-default.
1848 * Add coordinate blinding for EC_POINT and implement projective
1849 coordinate blinding for generic prime curves as a countermeasure to
1850 chosen point SCA attacks.
1852 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1854 * Add blinding to ECDSA and DSA signatures to protect against side channel
1855 attacks discovered by Keegan Ryan (NCC Group).
1859 * Enforce checking in the pkeyutl command line app to ensure that the input
1860 length does not exceed the maximum supported digest length when performing
1861 a sign, verify or verifyrecover operation.
1865 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1866 I/O in combination with something like select() or poll() will hang. This
1867 can be turned off again using SSL_CTX_clear_mode().
1868 Many applications do not properly handle non-application data records, and
1869 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1870 around the problems in those applications, but can also break some.
1871 It's recommended to read the manpages about SSL_read(), SSL_write(),
1872 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1873 SSL_CTX_set_read_ahead() again.
1877 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1878 now allow empty (zero character) pass phrases.
1882 * Apply blinding to binary field modular inversion and remove patent
1883 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1887 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1888 binary and prime elliptic curves.
1892 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1893 constant time fixed point multiplication.
1897 * Revise elliptic curve scalar multiplication with timing attack
1898 defenses: ec_wNAF_mul redirects to a constant time implementation
1899 when computing fixed point and variable point multiplication (which
1900 in OpenSSL are mostly used with secret scalars in keygen, sign,
1901 ECDH derive operations).
1902 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1905 * Updated CONTRIBUTING
1909 * Updated DRBG / RAND to request nonce and additional low entropy
1910 randomness from the system.
1912 *Matthias St. Pierre*
1914 * Updated 'openssl rehash' to use OpenSSL consistent default.
1918 * Moved the load of the ssl_conf module to libcrypto, which helps
1919 loading engines that libssl uses before libssl is initialised.
1923 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1927 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1929 *Ingo Schwarze, Rich Salz*
1931 * Added output of accepting IP address and port for 'openssl s_server'
1935 * Added a new API for TLSv1.3 ciphersuites:
1936 SSL_CTX_set_ciphersuites()
1937 SSL_set_ciphersuites()
1941 * Memory allocation failures consistently add an error to the error
1946 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1947 in libcrypto when run as setuid/setgid.
1951 * Load any config file by default when libssl is used.
1955 * Added new public header file <openssl/rand_drbg.h> and documentation
1956 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1958 *Matthias St. Pierre*
1960 * QNX support removed (cannot find contributors to get their approval
1961 for the license change).
1965 * TLSv1.3 replay protection for early data has been implemented. See the
1966 SSL_read_early_data() man page for further details.
1970 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1971 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1972 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1973 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1974 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1975 configuration has been separated out. See the ciphers man page or the
1976 SSL_CTX_set_ciphersuites() man page for more information.
1980 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1981 in responder mode now supports the new "-multi" option, which
1982 spawns the specified number of child processes to handle OCSP
1983 requests. The "-timeout" option now also limits the OCSP
1984 responder's patience to wait to receive the full client request
1985 on a newly accepted connection. Child processes are respawned
1986 as needed, and the CA index file is automatically reloaded
1987 when changed. This makes it possible to run the "ocsp" responder
1988 as a long-running service, making the OpenSSL CA somewhat more
1989 feature-complete. In this mode, most diagnostic messages logged
1990 after entering the event loop are logged via syslog(3) rather than
1995 * Added support for X448 and Ed448. Heavily based on original work by
2000 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2001 objects loaded. This adds the functions OSSL_STORE_expect() and
2002 OSSL_STORE_find() as well as needed tools to construct searches and
2003 get the search data out of them.
2007 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2008 version of OpenSSL should review their configuration settings to ensure
2009 that they are still appropriate for TLSv1.3. For further information see:
2010 <https://wiki.openssl.org/index.php/TLS1.3>
2014 * Grand redesign of the OpenSSL random generator
2016 The default RAND method now utilizes an AES-CTR DRBG according to
2017 NIST standard SP 800-90Ar1. The new random generator is essentially
2018 a port of the default random generator from the OpenSSL FIPS 2.0
2019 object module. It is a hybrid deterministic random bit generator
2020 using an AES-CTR bit stream and which seeds and reseeds itself
2021 automatically using trusted system entropy sources.
2023 Some of its new features are:
2024 - Support for multiple DRBG instances with seed chaining.
2025 - The default RAND method makes use of a DRBG.
2026 - There is a public and private DRBG instance.
2027 - The DRBG instances are fork-safe.
2028 - Keep all global DRBG instances on the secure heap if it is enabled.
2029 - The public and private DRBG instance are per thread for lock free
2032 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2034 * Changed Configure so it only says what it does and doesn't dump
2035 so much data. Instead, ./configdata.pm should be used as a script
2036 to display all sorts of configuration data.
2040 * Added processing of "make variables" to Configure.
2044 * Added SHA512/224 and SHA512/256 algorithm support.
2048 * The last traces of Netware support, first removed in 1.1.0, have
2053 * Get rid of Makefile.shared, and in the process, make the processing
2054 of certain files (rc.obj, or the .def/.map/.opt files produced from
2055 the ordinal files) more visible and hopefully easier to trace and
2056 debug (or make silent).
2060 * Make it possible to have environment variable assignments as
2061 arguments to config / Configure.
2065 * Add multi-prime RSA (RFC 8017) support.
2069 * Add SM3 implemented according to GB/T 32905-2016
2070 *Jack Lloyd <jack.lloyd@ribose.com>,*
2071 *Ronald Tse <ronald.tse@ribose.com>,*
2072 *Erick Borsboom <erick.borsboom@ribose.com>*
2074 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2075 as documented in RFC6066.
2076 Based on a patch from Tomasz Moń
2078 *Filipe Raimundo da Silva*
2080 * Add SM4 implemented according to GB/T 32907-2016.
2081 *Jack Lloyd <jack.lloyd@ribose.com>,*
2082 *Ronald Tse <ronald.tse@ribose.com>,*
2083 *Erick Borsboom <erick.borsboom@ribose.com>*
2085 * Reimplement -newreq-nodes and ERR_error_string_n; the
2086 original author does not agree with the license change.
2090 * Add ARIA AEAD TLS support.
2094 * Some macro definitions to support VS6 have been removed. Visual
2095 Studio 6 has not worked since 1.1.0
2099 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2100 without clearing the errors.
2104 * Add "atfork" functions. If building on a system that without
2105 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2106 requirements. The RAND facility now uses/requires this.
2114 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2115 not possible to disable entirely. However, it's still possible to
2116 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2119 To disable, configure with 'no-ui-console'. 'no-ui' is still
2120 possible to use as an alias. Check at compile time with the
2121 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2122 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2126 * Add a STORE module, which implements a uniform and URI based reader of
2127 stores that can contain keys, certificates, CRLs and numerous other
2128 objects. The main API is loosely based on a few stdio functions,
2129 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2130 OSSL_STORE_error and OSSL_STORE_close.
2131 The implementation uses backends called "loaders" to implement arbitrary
2132 URI schemes. There is one built in "loader" for the 'file' scheme.
2136 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2137 then adjusted to work on FreeBSD 8.4 as well.
2138 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2139 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2143 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2144 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2145 error code calls like this:
2147 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2149 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2150 that can be encoded in C. For the foreseeable future, this will only
2153 *Richard Levitte and Tim Hudson*
2155 * Removed BSD cryptodev engine.
2159 * Add a build target 'build_all_generated', to build all generated files
2160 and only that. This can be used to prepare everything that requires
2161 things like perl for a system that lacks perl and then move everything
2162 to that system and do the rest of the build there.
2166 * In the UI interface, make it possible to duplicate the user data. This
2167 can be used by engines that need to retain the data for a longer time
2168 than just the call where this user data is passed.
2172 * Ignore the '-named_curve auto' value for compatibility of applications
2175 *Tomas Mraz <tmraz@fedoraproject.org>*
2177 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2178 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2179 alerts across multiple records (some of which could be empty). In practice
2180 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2181 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2182 support this at all. Supporting it adds significant complexity to the
2183 record layer, and its removal is unlikely to cause interoperability
2188 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2189 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2190 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2195 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2196 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2198 *Richard Levitte, Andy Polyakov*
2200 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2205 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2206 platform rather than 'mingw'.
2210 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2211 success if they are asked to add an object which already exists
2212 in the store. This change cascades to other functions which load
2213 certificates and CRLs.
2217 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2218 facilitate stack unwinding even from assembly subroutines.
2222 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2223 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2227 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2228 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2229 which is the minimum version we support.
2233 * Certificate time validation (X509_cmp_time) enforces stricter
2234 compliance with RFC 5280. Fractional seconds and timezone offsets
2235 are no longer allowed.
2239 * Add support for ARIA
2243 * s_client will now send the Server Name Indication (SNI) extension by
2244 default unless the new "-noservername" option is used. The server name is
2245 based on the host provided to the "-connect" option unless overridden by
2246 using "-servername".
2250 * Add support for SipHash
2254 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2255 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2256 prevent issues where no progress is being made and the peer continually
2257 sends unrecognised record types, using up resources processing them.
2261 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2262 using the algorithm defined in
2263 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2267 * Heartbeat support has been removed; the ABI is changed for now.
2269 *Richard Levitte, Rich Salz*
2271 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2275 * The RSA "null" method, which was partially supported to avoid patent
2276 issues, has been replaced to always returns NULL.
2283 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2285 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2286 used even when parsing explicit parameters, when loading a encoded key
2287 or calling `EC_GROUP_new_from_ecpkparameters()`/
2288 `EC_GROUP_new_from_ecparameters()`.
2289 This prevents bypass of security hardening and performance gains,
2290 especially for curves with specialized EC_METHODs.
2291 By default, if a key encoded with explicit parameters is loaded and later
2292 encoded, the output is still encoded with explicit parameters, even if
2293 internally a "named" EC_GROUP is used for computation.
2297 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2298 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2299 NULL. After this change, only the cofactor parameter can be NULL. It also
2300 does some minimal sanity checks on the passed order.
2305 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2306 An attack is simple, if the first CMS_recipientInfo is valid but the
2307 second CMS_recipientInfo is chosen ciphertext. If the second
2308 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2309 encryption key will be replaced by garbage, and the message cannot be
2310 decoded, but if the RSA decryption fails, the correct encryption key is
2311 used and the recipient will not notice the attack.
2312 As a work around for this potential attack the length of the decrypted
2313 key must be equal to the cipher default key length, in case the
2314 certifiate is not given and all recipientInfo are tried out.
2315 The old behaviour can be re-enabled in the CMS code by setting the
2316 CMS_DEBUG_DECRYPT flag.
2321 * Use Windows installation paths in the mingw builds
2323 Mingw isn't a POSIX environment per se, which means that Windows
2324 paths should be used for installation.
2329 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2331 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2332 This changes the size when using the genpkey app when no size is given. It
2333 fixes an omission in earlier changes that changed all RSA, DSA and DH
2334 generation apps to use 2048 bits by default.
2338 * Prevent over long nonces in ChaCha20-Poly1305.
2340 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2341 for every encryption operation. RFC 7539 specifies that the nonce value
2342 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2343 and front pads the nonce with 0 bytes if it is less than 12
2344 bytes. However it also incorrectly allows a nonce to be set of up to 16
2345 bytes. In this case only the last 12 bytes are significant and any
2346 additional leading bytes are ignored.
2348 It is a requirement of using this cipher that nonce values are
2349 unique. Messages encrypted using a reused nonce value are susceptible to
2350 serious confidentiality and integrity attacks. If an application changes
2351 the default nonce length to be longer than 12 bytes and then makes a
2352 change to the leading bytes of the nonce expecting the new value to be a
2353 new unique nonce then such an application could inadvertently encrypt
2354 messages with a reused nonce.
2356 Additionally the ignored bytes in a long nonce are not covered by the
2357 integrity guarantee of this cipher. Any application that relies on the
2358 integrity of these ignored leading bytes of a long nonce may be further
2359 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2360 is safe because no such use sets such a long nonce value. However user
2361 applications that use this cipher directly and set a non-default nonce
2362 length to be longer than 12 bytes may be vulnerable.
2364 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2370 * Added SCA hardening for modular field inversion in EC_GROUP through
2371 a new dedicated field_inv() pointer in EC_METHOD.
2372 This also addresses a leakage affecting conversions from projective
2373 to affine coordinates.
2375 *Billy Bob Brumley, Nicola Tuveri*
2377 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2378 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2382 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2386 * Remove the 'dist' target and add a tarball building script. The
2387 'dist' target has fallen out of use, and it shouldn't be
2388 necessary to configure just to create a source distribution.
2392 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2394 * Timing vulnerability in DSA signature generation
2396 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2397 timing side channel attack. An attacker could use variations in the signing
2398 algorithm to recover the private key.
2400 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2405 * Timing vulnerability in ECDSA signature generation
2407 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2408 timing side channel attack. An attacker could use variations in the signing
2409 algorithm to recover the private key.
2411 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2416 * Add coordinate blinding for EC_POINT and implement projective
2417 coordinate blinding for generic prime curves as a countermeasure to
2418 chosen point SCA attacks.
2420 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2422 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2424 * Client DoS due to large DH parameter
2426 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2427 malicious server can send a very large prime value to the client. This will
2428 cause the client to spend an unreasonably long period of time generating a
2429 key for this prime resulting in a hang until the client has finished. This
2430 could be exploited in a Denial Of Service attack.
2432 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2437 * Cache timing vulnerability in RSA Key Generation
2439 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2440 a cache timing side channel attack. An attacker with sufficient access to
2441 mount cache timing attacks during the RSA key generation process could
2442 recover the private key.
2444 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2445 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2450 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2451 parameter is no longer accepted, as it leads to a corrupt table. NULL
2452 pem_str is reserved for alias entries only.
2456 * Revert blinding in ECDSA sign and instead make problematic addition
2457 length-invariant. Switch even to fixed-length Montgomery multiplication.
2461 * Change generating and checking of primes so that the error rate of not
2462 being prime depends on the intended use based on the size of the input.
2463 For larger primes this will result in more rounds of Miller-Rabin.
2464 The maximal error rate for primes with more than 1080 bits is lowered
2467 *Kurt Roeckx, Annie Yousar*
2469 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2473 * Add blinding to ECDSA and DSA signatures to protect against side channel
2474 attacks discovered by Keegan Ryan (NCC Group).
2478 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2479 now allow empty (zero character) pass phrases.
2483 * Certificate time validation (X509_cmp_time) enforces stricter
2484 compliance with RFC 5280. Fractional seconds and timezone offsets
2485 are no longer allowed.
2489 * Fixed a text canonicalisation bug in CMS
2491 Where a CMS detached signature is used with text content the text goes
2492 through a canonicalisation process first prior to signing or verifying a
2493 signature. This process strips trailing space at the end of lines, converts
2494 line terminators to CRLF and removes additional trailing line terminators
2495 at the end of a file. A bug in the canonicalisation process meant that
2496 some characters, such as form-feed, were incorrectly treated as whitespace
2497 and removed. This is contrary to the specification (RFC5485). This fix
2498 could mean that detached text data signed with an earlier version of
2499 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2500 signed with a fixed OpenSSL may fail to verify with an earlier version of
2501 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2502 and use the "-binary" flag (for the "cms" command line application) or set
2503 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2507 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2509 * Constructed ASN.1 types with a recursive definition could exceed the stack
2511 Constructed ASN.1 types with a recursive definition (such as can be found
2512 in PKCS7) could eventually exceed the stack given malicious input with
2513 excessive recursion. This could result in a Denial Of Service attack. There
2514 are no such structures used within SSL/TLS that come from untrusted sources
2515 so this is considered safe.
2517 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2523 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2525 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2526 effectively reduced to only comparing the least significant bit of each
2527 byte. This allows an attacker to forge messages that would be considered as
2528 authenticated in an amount of tries lower than that guaranteed by the
2529 security claims of the scheme. The module can only be compiled by the
2530 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2532 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2538 * Add a build target 'build_all_generated', to build all generated files
2539 and only that. This can be used to prepare everything that requires
2540 things like perl for a system that lacks perl and then move everything
2541 to that system and do the rest of the build there.
2545 * Backport SSL_OP_NO_RENGOTIATION
2547 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2548 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2549 changes this is no longer possible in 1.1.0. Therefore the new
2550 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2551 1.1.0 to provide equivalent functionality.
2553 Note that if an application built against 1.1.0h headers (or above) is run
2554 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2555 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2559 * Removed the OS390-Unix config target. It relied on a script that doesn't
2564 * rsaz_1024_mul_avx2 overflow bug on x86_64
2566 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2567 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2568 Analysis suggests that attacks against RSA and DSA as a result of this
2569 defect would be very difficult to perform and are not believed likely.
2570 Attacks against DH1024 are considered just feasible, because most of the
2571 work necessary to deduce information about a private key may be performed
2572 offline. The amount of resources required for such an attack would be
2573 significant. However, for an attack on TLS to be meaningful, the server
2574 would have to share the DH1024 private key among multiple clients, which is
2575 no longer an option since CVE-2016-0701.
2577 This only affects processors that support the AVX2 but not ADX extensions
2578 like Intel Haswell (4th generation).
2580 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2581 was originally found via the OSS-Fuzz project.
2586 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2588 * bn_sqrx8x_internal carry bug on x86_64
2590 There is a carry propagating bug in the x86_64 Montgomery squaring
2591 procedure. No EC algorithms are affected. Analysis suggests that attacks
2592 against RSA and DSA as a result of this defect would be very difficult to
2593 perform and are not believed likely. Attacks against DH are considered just
2594 feasible (although very difficult) because most of the work necessary to
2595 deduce information about a private key may be performed offline. The amount
2596 of resources required for such an attack would be very significant and
2597 likely only accessible to a limited number of attackers. An attacker would
2598 additionally need online access to an unpatched system using the target
2599 private key in a scenario with persistent DH parameters and a private
2600 key that is shared between multiple clients.
2602 This only affects processors that support the BMI1, BMI2 and ADX extensions
2603 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2605 This issue was reported to OpenSSL by the OSS-Fuzz project.
2610 * Malformed X.509 IPAddressFamily could cause OOB read
2612 If an X.509 certificate has a malformed IPAddressFamily extension,
2613 OpenSSL could do a one-byte buffer overread. The most likely result
2614 would be an erroneous display of the certificate in text format.
2616 This issue was reported to OpenSSL by the OSS-Fuzz project.
2621 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2623 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2624 platform rather than 'mingw'.
2628 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2629 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2630 which is the minimum version we support.
2634 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2636 * Encrypt-Then-Mac renegotiation crash
2638 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2639 negotiated where it was not in the original handshake (or vice-versa) then
2640 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2641 and servers are affected.
2643 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2648 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2650 * Truncated packet could crash via OOB read
2652 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2653 cipher is being used, then a truncated packet can cause that host to
2654 perform an out-of-bounds read, usually resulting in a crash.
2656 This issue was reported to OpenSSL by Robert Święcki of Google.
2661 * Bad (EC)DHE parameters cause a client crash
2663 If a malicious server supplies bad parameters for a DHE or ECDHE key
2664 exchange then this can result in the client attempting to dereference a
2665 NULL pointer leading to a client crash. This could be exploited in a Denial
2668 This issue was reported to OpenSSL by Guido Vranken.
2673 * BN_mod_exp may produce incorrect results on x86_64
2675 There is a carry propagating bug in the x86_64 Montgomery squaring
2676 procedure. No EC algorithms are affected. Analysis suggests that attacks
2677 against RSA and DSA as a result of this defect would be very difficult to
2678 perform and are not believed likely. Attacks against DH are considered just
2679 feasible (although very difficult) because most of the work necessary to
2680 deduce information about a private key may be performed offline. The amount
2681 of resources required for such an attack would be very significant and
2682 likely only accessible to a limited number of attackers. An attacker would
2683 additionally need online access to an unpatched system using the target
2684 private key in a scenario with persistent DH parameters and a private
2685 key that is shared between multiple clients. For example this can occur by
2686 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2687 similar to CVE-2015-3193 but must be treated as a separate problem.
2689 This issue was reported to OpenSSL by the OSS-Fuzz project.
2694 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2696 * ChaCha20/Poly1305 heap-buffer-overflow
2698 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2699 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2700 crash. This issue is not considered to be exploitable beyond a DoS.
2702 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2707 * CMS Null dereference
2709 Applications parsing invalid CMS structures can crash with a NULL pointer
2710 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2711 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2712 structure callback if an attempt is made to free certain invalid encodings.
2713 Only CHOICE structures using a callback which do not handle NULL value are
2716 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2721 * Montgomery multiplication may produce incorrect results
2723 There is a carry propagating bug in the Broadwell-specific Montgomery
2724 multiplication procedure that handles input lengths divisible by, but
2725 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2726 and DH private keys are impossible. This is because the subroutine in
2727 question is not used in operations with the private key itself and an input
2728 of the attacker's direct choice. Otherwise the bug can manifest itself as
2729 transient authentication and key negotiation failures or reproducible
2730 erroneous outcome of public-key operations with specially crafted input.
2731 Among EC algorithms only Brainpool P-512 curves are affected and one
2732 presumably can attack ECDH key negotiation. Impact was not analyzed in
2733 detail, because pre-requisites for attack are considered unlikely. Namely
2734 multiple clients have to choose the curve in question and the server has to
2735 share the private key among them, neither of which is default behaviour.
2736 Even then only clients that chose the curve will be affected.
2738 This issue was publicly reported as transient failures and was not
2739 initially recognized as a security issue. Thanks to Richard Morgan for
2740 providing reproducible case.
2745 * Removed automatic addition of RPATH in shared libraries and executables,
2746 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2750 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2752 * Fix Use After Free for large message sizes
2754 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2755 message larger than approx 16k is received then the underlying buffer to
2756 store the incoming message is reallocated and moved. Unfortunately a
2757 dangling pointer to the old location is left which results in an attempt to
2758 write to the previously freed location. This is likely to result in a
2759 crash, however it could potentially lead to execution of arbitrary code.
2761 This issue only affects OpenSSL 1.1.0a.
2763 This issue was reported to OpenSSL by Robert Święcki.
2768 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2770 * OCSP Status Request extension unbounded memory growth
2772 A malicious client can send an excessively large OCSP Status Request
2773 extension. If that client continually requests renegotiation, sending a
2774 large OCSP Status Request extension each time, then there will be unbounded
2775 memory growth on the server. This will eventually lead to a Denial Of
2776 Service attack through memory exhaustion. Servers with a default
2777 configuration are vulnerable even if they do not support OCSP. Builds using
2778 the "no-ocsp" build time option are not affected.
2780 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2785 * SSL_peek() hang on empty record
2787 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2788 sends an empty record. This could be exploited by a malicious peer in a
2789 Denial Of Service attack.
2791 This issue was reported to OpenSSL by Alex Gaynor.
2796 * Excessive allocation of memory in tls_get_message_header() and
2797 dtls1_preprocess_fragment()
2799 A (D)TLS message includes 3 bytes for its length in the header for the
2800 message. This would allow for messages up to 16Mb in length. Messages of
2801 this length are excessive and OpenSSL includes a check to ensure that a
2802 peer is sending reasonably sized messages in order to avoid too much memory
2803 being consumed to service a connection. A flaw in the logic of version
2804 1.1.0 means that memory for the message is allocated too early, prior to
2805 the excessive message length check. Due to way memory is allocated in
2806 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2807 to service a connection. This could lead to a Denial of Service through
2808 memory exhaustion. However, the excessive message length check still takes
2809 place, and this would cause the connection to immediately fail. Assuming
2810 that the application calls SSL_free() on the failed connection in a timely
2811 manner then the 21Mb of allocated memory will then be immediately freed
2812 again. Therefore the excessive memory allocation will be transitory in
2813 nature. This then means that there is only a security impact if:
2815 1) The application does not call SSL_free() in a timely manner in the event
2816 that the connection fails
2818 2) The application is working in a constrained environment where there is
2819 very little free memory
2821 3) The attacker initiates multiple connection attempts such that there are
2822 multiple connections in a state where memory has been allocated for the
2823 connection; SSL_free() has not yet been called; and there is insufficient
2824 memory to service the multiple requests.
2826 Except in the instance of (1) above any Denial Of Service is likely to be
2827 transitory because as soon as the connection fails the memory is
2828 subsequently freed again in the SSL_free() call. However there is an
2829 increased risk during this period of application crashes due to the lack of
2830 memory - which would then mean a more serious Denial of Service.
2832 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2833 (CVE-2016-6307 and CVE-2016-6308)
2837 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2838 had to be removed. Primary reason is that vendor assembler can't
2839 assemble our modules with -KPIC flag. As result it, assembly
2840 support, was not even available as option. But its lack means
2841 lack of side-channel resistant code, which is incompatible with
2842 security by todays standards. Fortunately gcc is readily available
2843 prepackaged option, which we firmly point at...
2847 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2849 * Windows command-line tool supports UTF-8 opt-in option for arguments
2850 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2851 (to any value) allows Windows user to access PKCS#12 file generated
2852 with Windows CryptoAPI and protected with non-ASCII password, as well
2853 as files generated under UTF-8 locale on Linux also protected with
2858 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2859 have been disabled by default and removed from DEFAULT, just like RC4.
2860 See the RC4 item below to re-enable both.
2864 * The method for finding the storage location for the Windows RAND seed file
2865 has changed. First we check %RANDFILE%. If that is not set then we check
2866 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2867 all else fails we fall back to C:\.
2871 * The EVP_EncryptUpdate() function has had its return type changed from void
2872 to int. A return of 0 indicates and error while a return of 1 indicates
2877 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2878 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2879 off the constant time implementation for RSA, DSA and DH have been made
2880 no-ops and deprecated.
2884 * Windows RAND implementation was simplified to only get entropy by
2885 calling CryptGenRandom(). Various other RAND-related tickets
2888 *Joseph Wylie Yandle, Rich Salz*
2890 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2891 and `OPENSSL_LH_`, respectively. The old names are available
2892 with API compatibility. They new names are now completely documented.
2896 * Unify TYPE_up_ref(obj) methods signature.
2897 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2898 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2899 int (instead of void) like all others TYPE_up_ref() methods.
2900 So now these methods also check the return value of CRYPTO_atomic_add(),
2901 and the validity of object reference counter.
2903 *fdasilvayy@gmail.com*
2905 * With Windows Visual Studio builds, the .pdb files are installed
2906 alongside the installed libraries and executables. For a static
2907 library installation, ossl_static.pdb is the associate compiler
2908 generated .pdb file to be used when linking programs.
2912 * Remove openssl.spec. Packaging files belong with the packagers.
2916 * Automatic Darwin/OSX configuration has had a refresh, it will now
2917 recognise x86_64 architectures automatically. You can still decide
2918 to build for a different bitness with the environment variable
2919 KERNEL_BITS (can be 32 or 64), for example:
2921 KERNEL_BITS=32 ./config
2925 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2926 256 bit AES and HMAC with SHA256.
2930 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2934 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2938 * To enable users to have their own config files and build file templates,
2939 Configure looks in the directory indicated by the environment variable
2940 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2941 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2942 name and is used as is.
2946 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2947 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2948 X509_CERT_FILE_CTX was removed.
2952 * "shared" builds are now the default. To create only static libraries use
2953 the "no-shared" Configure option.
2957 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2958 All of these option have not worked for some while and are fundamental
2963 * Make various cleanup routines no-ops and mark them as deprecated. Most
2964 global cleanup functions are no longer required because they are handled
2965 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2966 Explicitly de-initing can cause problems (e.g. where a library that uses
2967 OpenSSL de-inits, but an application is still using it). The affected
2968 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2969 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2970 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2971 COMP_zlib_cleanup().
2975 * --strict-warnings no longer enables runtime debugging options
2976 such as REF_DEBUG. Instead, debug options are automatically
2977 enabled with '--debug' builds.
2979 *Andy Polyakov, Emilia Käsper*
2981 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2982 have been moved out of the public header files. New functions for managing
2983 these have been added.
2987 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
2988 objects have been moved out of the public header files. New
2989 functions for managing these have been added.
2993 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
2994 have been moved out of the public header files. New functions for managing
2995 these have been added.
2999 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3000 moved out of the public header files. New functions for managing these
3005 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3009 * Removed the mk1mf build scripts.
3013 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3014 it is always safe to #include a header now.
3018 * Removed the aged BC-32 config and all its supporting scripts
3022 * Removed support for Ultrix, Netware, and OS/2.
3026 * Add support for HKDF.
3028 *Alessandro Ghedini*
3030 * Add support for blake2b and blake2s
3034 * Added support for "pipelining". Ciphers that have the
3035 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3036 encryptions/decryptions simultaneously. There are currently no built-in
3037 ciphers with this property but the expectation is that engines will be able
3038 to offer it to significantly improve throughput. Support has been extended
3039 into libssl so that multiple records for a single connection can be
3040 processed in one go (for >=TLS 1.1).
3044 * Added the AFALG engine. This is an async capable engine which is able to
3045 offload work to the Linux kernel. In this initial version it only supports
3046 AES128-CBC. The kernel must be version 4.1.0 or greater.
3050 * OpenSSL now uses a new threading API. It is no longer necessary to
3051 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3052 are two supported threading models: pthreads and windows threads. It is
3053 also possible to configure OpenSSL at compile time for "no-threads". The
3054 old threading API should no longer be used. The functions have been
3055 replaced with "no-op" compatibility macros.
3057 *Alessandro Ghedini, Matt Caswell*
3059 * Modify behavior of ALPN to invoke callback after SNI/servername
3060 callback, such that updates to the SSL_CTX affect ALPN.
3064 * Add SSL_CIPHER queries for authentication and key-exchange.
3068 * Changes to the DEFAULT cipherlist:
3069 - Prefer (EC)DHE handshakes over plain RSA.
3070 - Prefer AEAD ciphers over legacy ciphers.
3071 - Prefer ECDSA over RSA when both certificates are available.
3072 - Prefer TLSv1.2 ciphers/PRF.
3073 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3078 * Change the ECC default curve list to be this, in order: x25519,
3079 secp256r1, secp521r1, secp384r1.
3083 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3084 disabled by default. They can be re-enabled using the
3085 enable-weak-ssl-ciphers option to Configure.
3089 * If the server has ALPN configured, but supports no protocols that the
3090 client advertises, send a fatal "no_application_protocol" alert.
3091 This behaviour is SHALL in RFC 7301, though it isn't universally
3092 implemented by other servers.
3096 * Add X25519 support.
3097 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3098 for public and private key encoding using the format documented in
3099 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3100 key generation and key derivation.
3102 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3107 * Deprecate SRP_VBASE_get_by_user.
3108 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3109 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3110 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3111 seed, even if the seed is configured.
3113 Users should use SRP_VBASE_get1_by_user instead. Note that in
3114 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3115 also that even though configuring the SRP seed attempts to hide
3116 invalid usernames by continuing the handshake with fake
3117 credentials, this behaviour is not constant time and no strong
3118 guarantees are made that the handshake is indistinguishable from
3119 that of a valid user.
3123 * Configuration change; it's now possible to build dynamic engines
3124 without having to build shared libraries and vice versa. This
3125 only applies to the engines in engines/, those in crypto/engine/
3126 will always be built into libcrypto (i.e. "static").
3128 Building dynamic engines is enabled by default; to disable, use
3129 the configuration option "disable-dynamic-engine".
3131 The only requirements for building dynamic engines are the
3132 presence of the DSO module and building with position independent
3133 code, so they will also automatically be disabled if configuring
3134 with "disable-dso" or "disable-pic".
3136 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3137 are also taken away from openssl/opensslconf.h, as they are
3142 * Configuration change; if there is a known flag to compile
3143 position independent code, it will always be applied on the
3144 libcrypto and libssl object files, and never on the application
3145 object files. This means other libraries that use routines from
3146 libcrypto / libssl can be made into shared libraries regardless
3147 of how OpenSSL was configured.
3149 If this isn't desirable, the configuration options "disable-pic"
3150 or "no-pic" can be used to disable the use of PIC. This will
3151 also disable building shared libraries and dynamic engines.
3155 * Removed JPAKE code. It was experimental and has no wide use.
3159 * The INSTALL_PREFIX Makefile variable has been renamed to
3160 DESTDIR. That makes for less confusion on what this variable
3161 is for. Also, the configuration option --install_prefix is
3166 * Heartbeat for TLS has been removed and is disabled by default
3167 for DTLS; configure with enable-heartbeats. Code that uses the
3168 old #define's might need to be updated.
3170 *Emilia Käsper, Rich Salz*
3172 * Rename REF_CHECK to REF_DEBUG.
3176 * New "unified" build system
3178 The "unified" build system is aimed to be a common system for all
3179 platforms we support. With it comes new support for VMS.
3181 This system builds supports building in a different directory tree
3182 than the source tree. It produces one Makefile (for unix family
3183 or lookalikes), or one descrip.mms (for VMS).
3185 The source of information to make the Makefile / descrip.mms is
3186 small files called 'build.info', holding the necessary
3187 information for each directory with source to compile, and a
3188 template in Configurations, like unix-Makefile.tmpl or
3191 With this change, the library names were also renamed on Windows
3192 and on VMS. They now have names that are closer to the standard
3193 on Unix, and include the major version number, and in certain
3194 cases, the architecture they are built for. See "Notes on shared
3195 libraries" in INSTALL.
3197 We rely heavily on the perl module Text::Template.
3201 * Added support for auto-initialisation and de-initialisation of the library.
3202 OpenSSL no longer requires explicit init or deinit routines to be called,
3203 except in certain circumstances. See the OPENSSL_init_crypto() and
3204 OPENSSL_init_ssl() man pages for further information.
3208 * The arguments to the DTLSv1_listen function have changed. Specifically the
3209 "peer" argument is now expected to be a BIO_ADDR object.
3211 * Rewrite of BIO networking library. The BIO library lacked consistent
3212 support of IPv6, and adding it required some more extensive
3213 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3214 which hold all types of addresses and chains of address information.
3215 It also introduces a new API, with functions like BIO_socket,
3216 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3217 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3218 have been adapted accordingly.
3222 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3227 * CRIME protection: disable compression by default, even if OpenSSL is
3228 compiled with zlib enabled. Applications can still enable compression
3229 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3230 using the SSL_CONF library to configure compression.
3234 * The signature of the session callback configured with
3235 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3236 was explicitly marked as `const unsigned char*` instead of
3241 * Always DPURIFY. Remove the use of uninitialized memory in the
3242 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3246 * Removed many obsolete configuration items, including
3247 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3248 MD2_CHAR, MD2_INT, MD2_LONG
3250 IDEA_SHORT, IDEA_LONG
3251 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3253 *Rich Salz, with advice from Andy Polyakov*
3255 * Many BN internals have been moved to an internal header file.
3257 *Rich Salz with help from Andy Polyakov*
3259 * Configuration and writing out the results from it has changed.
3260 Files such as Makefile include/openssl/opensslconf.h and are now
3261 produced through general templates, such as Makefile.in and
3262 crypto/opensslconf.h.in and some help from the perl module
3265 Also, the center of configuration information is no longer
3266 Makefile. Instead, Configure produces a perl module in
3267 configdata.pm which holds most of the config data (in the hash
3268 table %config), the target data that comes from the target
3269 configuration in one of the `Configurations/*.conf` files (in
3274 * To clarify their intended purposes, the Configure options
3275 --prefix and --openssldir change their semantics, and become more
3276 straightforward and less interdependent.
3278 --prefix shall be used exclusively to give the location INSTALLTOP
3279 where programs, scripts, libraries, include files and manuals are
3280 going to be installed. The default is now /usr/local.
3282 --openssldir shall be used exclusively to give the default
3283 location OPENSSLDIR where certificates, private keys, CRLs are
3284 managed. This is also where the default openssl.cnf gets
3286 If the directory given with this option is a relative path, the
3287 values of both the --prefix value and the --openssldir value will
3288 be combined to become OPENSSLDIR.
3289 The default for --openssldir is INSTALLTOP/ssl.
3291 Anyone who uses --openssldir to specify where OpenSSL is to be
3292 installed MUST change to use --prefix instead.
3296 * The GOST engine was out of date and therefore it has been removed. An up
3297 to date GOST engine is now being maintained in an external repository.
3298 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3299 support for GOST ciphersuites (these are only activated if a GOST engine
3304 * EGD is no longer supported by default; use enable-egd when
3307 *Ben Kaduk and Rich Salz*
3309 * The distribution now has Makefile.in files, which are used to
3310 create Makefile's when Configure is run. *Configure must be run
3311 before trying to build now.*
3315 * The return value for SSL_CIPHER_description() for error conditions
3320 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3322 Obtaining and performing DNSSEC validation of TLSA records is
3323 the application's responsibility. The application provides
3324 the TLSA records of its choice to OpenSSL, and these are then
3325 used to authenticate the peer.
3327 The TLSA records need not even come from DNS. They can, for
3328 example, be used to implement local end-entity certificate or
3329 trust-anchor "pinning", where the "pin" data takes the form
3330 of TLSA records, which can augment or replace verification
3331 based on the usual WebPKI public certification authorities.
3335 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3336 continues to support deprecated interfaces in default builds.
3337 However, applications are strongly advised to compile their
3338 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3339 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3340 or the 1.1.0 releases.
3342 In environments in which all applications have been ported to
3343 not use any deprecated interfaces OpenSSL's Configure script
3344 should be used with the --api=1.1.0 option to entirely remove
3345 support for the deprecated features from the library and
3346 unconditionally disable them in the installed headers.
3347 Essentially the same effect can be achieved with the "no-deprecated"
3348 argument to Configure, except that this will always restrict
3349 the build to just the latest API, rather than a fixed API
3352 As applications are ported to future revisions of the API,
3353 they should update their compile-time OPENSSL_API_COMPAT define
3354 accordingly, but in most cases should be able to continue to
3355 compile with later releases.
3357 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3358 0x10000000L and 0x00908000L, respectively. However those
3359 versions did not support the OPENSSL_API_COMPAT feature, and
3360 so applications are not typically tested for explicit support
3361 of just the undeprecated features of either release.
3365 * Add support for setting the minimum and maximum supported protocol.
3366 It can bet set via the SSL_set_min_proto_version() and
3367 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3368 MaxProtocol. It's recommended to use the new APIs to disable
3369 protocols instead of disabling individual protocols using
3370 SSL_set_options() or SSL_CONF's Protocol. This change also
3371 removes support for disabling TLS 1.2 in the OpenSSL TLS
3372 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3376 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3380 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3381 and integrates ECDSA and ECDH functionality into EC. Implementations can
3382 now redirect key generation and no longer need to convert to or from
3385 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3386 include the ec.h header file instead.
3390 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3391 ciphers who are no longer supported and drops support the ephemeral RSA key
3392 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3396 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3397 opaque. For HMAC_CTX, the following constructors and destructors
3400 HMAC_CTX *HMAC_CTX_new(void);
3401 void HMAC_CTX_free(HMAC_CTX *ctx);
3403 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3404 destroy such methods has been added. See EVP_MD_meth_new(3) and
3405 EVP_CIPHER_meth_new(3) for documentation.
3408 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3409 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3410 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3411 an already created structure.
3412 2) For consistency with the majority of our object creators and
3413 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3414 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3415 for deprecated builds.
3419 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3420 cryptographic operations to be performed asynchronously as long as an
3421 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3422 further details. Libssl has also had this capability integrated with the
3423 introduction of the new mode SSL_MODE_ASYNC and associated error
3424 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3425 pages. This work was developed in partnership with Intel Corp.
3429 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3430 always enabled now. If you want to disable the support you should
3431 exclude it using the list of supported ciphers. This also means that the
3432 "-no_ecdhe" option has been removed from s_server.
3436 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3437 SSL_{CTX_}set1_curves() which can set a list.
3441 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3442 curve you want to support using SSL_{CTX_}set1_curves().
3446 * State machine rewrite. The state machine code has been significantly
3447 refactored in order to remove much duplication of code and solve issues
3448 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3449 further details). This change does have some associated API changes.
3450 Notably the SSL_state() function has been removed and replaced by
3451 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3452 SSL_set_state() has been removed altogether. The previous handshake states
3453 defined in ssl.h and ssl3.h have also been removed.
3457 * All instances of the string "ssleay" in the public API were replaced
3458 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3459 Some error codes related to internal RSA_eay API's were renamed.
3463 * The demo files in crypto/threads were moved to demo/threads.
3467 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3470 *Matt Caswell, Rich Salz*
3472 * New ASN.1 embed macro.
3474 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3475 structure is not allocated: it is part of the parent. That is instead of
3483 This reduces memory fragmentation and make it impossible to accidentally
3484 set a mandatory field to NULL.
3486 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3487 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3488 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3493 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3497 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3498 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3499 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3500 DES and RC4 ciphersuites.
3504 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3505 This changes the decoding behaviour for some invalid messages,
3506 though the change is mostly in the more lenient direction, and
3507 legacy behaviour is preserved as much as possible.
3511 * Fix no-stdio build.
3512 *David Woodhouse <David.Woodhouse@intel.com> and also*
3513 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3515 * New testing framework
3516 The testing framework has been largely rewritten and is now using
3517 perl and the perl modules Test::Harness and an extended variant of
3518 Test::More called OpenSSL::Test to do its work. All test scripts in
3519 test/ have been rewritten into test recipes, and all direct calls to
3520 executables in test/Makefile have become individual recipes using the
3521 simplified testing OpenSSL::Test::Simple.
3523 For documentation on our testing modules, do:
3525 perldoc test/testlib/OpenSSL/Test/Simple.pm
3526 perldoc test/testlib/OpenSSL/Test.pm
3530 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3531 are used; the latter aborts on memory leaks (usually checked on exit).
3532 Some undocumented "set malloc, etc., hooks" functions were removed
3533 and others were changed. All are now documented.
3537 * In DSA_generate_parameters_ex, if the provided seed is too short,
3540 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3542 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3543 from RFC4279, RFC4785, RFC5487, RFC5489.
3545 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3546 original RSA_PSK patch.
3550 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3551 era flag was never set throughout the codebase (only read). Also removed
3552 SSL3_FLAGS_POP_BUFFER which was only used if
3553 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3557 * Changed the default name options in the "ca", "crl", "req" and "x509"
3558 to be "oneline" instead of "compat".
3562 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3563 not aware of clients that still exhibit this bug, and the workaround
3564 hasn't been working properly for a while.
3568 * The return type of BIO_number_read() and BIO_number_written() as well as
3569 the corresponding num_read and num_write members in the BIO structure has
3570 changed from unsigned long to uint64_t. On platforms where an unsigned
3571 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3576 * Given the pervasive nature of TLS extensions it is inadvisable to run
3577 OpenSSL without support for them. It also means that maintaining
3578 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3579 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3583 * Removed support for the two export grade static DH ciphersuites
3584 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3585 were newly added (along with a number of other static DH ciphersuites) to
3586 1.0.2. However the two export ones have *never* worked since they were
3587 introduced. It seems strange in any case to be adding new export
3588 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3592 * Version negotiation has been rewritten. In particular SSLv23_method(),
3593 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3594 and turned into macros which simply call the new preferred function names
3595 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3596 should use the new names instead. Also as part of this change the ssl23.h
3597 header file has been removed.
3601 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3602 code and the associated standard is no longer considered fit-for-purpose.
3606 * RT2547 was closed. When generating a private key, try to make the
3607 output file readable only by the owner. This behavior change might
3608 be noticeable when interacting with other software.
3610 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3615 * Added HTTP GET support to the ocsp command.
3619 * Changed default digest for the dgst and enc commands from MD5 to
3624 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3628 * Added support for TLS extended master secret from
3629 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3630 initial patch which was a great help during development.
3634 * All libssl internal structures have been removed from the public header
3635 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3636 now redundant). Users should not attempt to access internal structures
3637 directly. Instead they should use the provided API functions.
3641 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3642 Access to deprecated functions can be re-enabled by running config with
3643 "enable-deprecated". In addition applications wishing to use deprecated
3644 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3645 will, by default, disable some transitive includes that previously existed
3646 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3650 * Added support for OCB mode. OpenSSL has been granted a patent license
3651 compatible with the OpenSSL license for use of OCB. Details are available
3652 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3653 for OCB can be removed by calling config with no-ocb.
3657 * SSLv2 support has been removed. It still supports receiving a SSLv2
3658 compatible client hello.
3662 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3663 done while fixing the error code for the key-too-small case.
3665 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3667 * CA.sh has been removed; use CA.pl instead.
3671 * Removed old DES API.
3675 * Remove various unsupported platforms:
3681 Sinix/ReliantUNIX RM400
3686 16-bit platforms such as WIN16
3690 * Clean up OPENSSL_NO_xxx #define's
3691 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3692 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3693 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3694 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3695 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3696 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3697 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3698 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3699 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3700 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3704 * Cleaned up dead code
3705 Remove all but one '#ifdef undef' which is to be looked at.
3709 * Clean up calling of xxx_free routines.
3710 Just like free(), fix most of the xxx_free routines to accept
3711 NULL. Remove the non-null checks from callers. Save much code.
3715 * Add secure heap for storage of private keys (when possible).
3716 Add BIO_s_secmem(), CBIGNUM, etc.
3717 Contributed by Akamai Technologies under our Corporate CLA.
3721 * Experimental support for a new, fast, unbiased prime candidate generator,
3722 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3724 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3726 * New output format NSS in the sess_id command line tool. This allows
3727 exporting the session id and the master key in NSS keylog format.
3729 *Martin Kaiser <martin@kaiser.cx>*
3731 * Harmonize version and its documentation. -f flag is used to display
3734 *mancha <mancha1@zoho.com>*
3736 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3737 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3739 *mancha <mancha1@zoho.com>*
3741 * Fix some double frees. These are not thought to be exploitable.
3743 *mancha <mancha1@zoho.com>*
3745 * A missing bounds check in the handling of the TLS heartbeat extension
3746 can be used to reveal up to 64k of memory to a connected client or
3749 Thanks for Neel Mehta of Google Security for discovering this bug and to
3750 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3751 preparing the fix ([CVE-2014-0160])
3753 *Adam Langley, Bodo Moeller*
3755 * Fix for the attack described in the paper "Recovering OpenSSL
3756 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3757 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3758 <http://eprint.iacr.org/2014/140>
3760 Thanks to Yuval Yarom and Naomi Benger for discovering this
3761 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3763 *Yuval Yarom and Naomi Benger*
3765 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3766 this fixes a limitation in previous versions of OpenSSL.
3770 * Experimental encrypt-then-mac support.
3772 Experimental support for encrypt then mac from
3773 draft-gutmann-tls-encrypt-then-mac-02.txt
3775 To enable it set the appropriate extension number (0x42 for the test
3776 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3778 For non-compliant peers (i.e. just about everything) this should have no
3781 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3785 * Add EVP support for key wrapping algorithms, to avoid problems with
3786 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3787 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3788 algorithms and include tests cases.
3792 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3797 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3798 MGF1 digest and OAEP label.
3802 * Make openssl verify return errors.
3804 *Chris Palmer <palmer@google.com> and Ben Laurie*
3806 * New function ASN1_TIME_diff to calculate the difference between two
3807 ASN1_TIME structures or one structure and the current time.
3811 * Update fips_test_suite to support multiple command line options. New
3812 test to induce all self test errors in sequence and check expected
3817 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3818 sign or verify all in one operation.
3822 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3823 test programs and fips_test_suite. Includes functionality to parse
3824 the minimal script output of fipsalgest.pl directly.
3828 * Add authorisation parameter to FIPS_module_mode_set().
3832 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3836 * Use separate DRBG fields for internal and external flags. New function
3837 FIPS_drbg_health_check() to perform on demand health checking. Add
3838 generation tests to fips_test_suite with reduced health check interval to
3839 demonstrate periodic health checking. Add "nodh" option to
3840 fips_test_suite to skip very slow DH test.
3844 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3849 * More extensive health check for DRBG checking many more failure modes.
3850 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3851 combination: call this in fips_test_suite.
3855 * Add support for canonical generation of DSA parameter 'g'. See
3858 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3859 POST to handle HMAC cases.
3863 * Add functions FIPS_module_version() and FIPS_module_version_text()
3864 to return numerical and string versions of the FIPS module number.
3868 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3869 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3870 outside the validated module in the FIPS capable OpenSSL.
3874 * Minor change to DRBG entropy callback semantics. In some cases
3875 there is no multiple of the block length between min_len and
3876 max_len. Allow the callback to return more than max_len bytes
3877 of entropy but discard any extra: it is the callback's responsibility
3878 to ensure that the extra data discarded does not impact the
3879 requested amount of entropy.
3883 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3884 information in FIPS186-3, SP800-57 and SP800-131A.
3888 * CCM support via EVP. Interface is very similar to GCM case except we
3889 must supply all data in one chunk (i.e. no update, final) and the
3890 message length must be supplied if AAD is used. Add algorithm test
3895 * Initial version of POST overhaul. Add POST callback to allow the status
3896 of POST to be monitored and/or failures induced. Modify fips_test_suite
3897 to use callback. Always run all selftests even if one fails.
3901 * XTS support including algorithm test driver in the fips_gcmtest program.
3902 Note: this does increase the maximum key length from 32 to 64 bytes but
3903 there should be no binary compatibility issues as existing applications
3904 will never use XTS mode.
3908 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3909 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3910 performs algorithm blocking for unapproved PRNG types. Also do not
3911 set PRNG type in FIPS_mode_set(): leave this to the application.
3912 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3913 the standard OpenSSL PRNG: set additional data to a date time vector.
3917 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3918 This shouldn't present any incompatibility problems because applications
3919 shouldn't be using these directly and any that are will need to rethink
3920 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3924 * Extensive self tests and health checking required by SP800-90 DRBG.
3925 Remove strength parameter from FIPS_drbg_instantiate and always
3926 instantiate at maximum supported strength.
3930 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3934 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3938 * New function DH_compute_key_padded() to compute a DH key and pad with
3939 leading zeroes if needed: this complies with SP800-56A et al.
3943 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3944 anything, incomplete, subject to change and largely untested at present.
3948 * Modify fipscanisteronly build option to only build the necessary object
3949 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3953 * Add experimental option FIPSSYMS to give all symbols in
3954 fipscanister.o and FIPS or fips prefix. This will avoid
3955 conflicts with future versions of OpenSSL. Add perl script
3956 util/fipsas.pl to preprocess assembly language source files
3957 and rename any affected symbols.
3961 * Add selftest checks and algorithm block of non-fips algorithms in
3962 FIPS mode. Remove DES2 from selftests.
3966 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3967 return internal method without any ENGINE dependencies. Add new
3968 tiny fips sign and verify functions.
3972 * New build option no-ec2m to disable characteristic 2 code.
3976 * New build option "fipscanisteronly". This only builds fipscanister.o
3977 and (currently) associated fips utilities. Uses the file Makefile.fips
3978 instead of Makefile.org as the prototype.
3982 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3983 Update fips_gcmtest to use IV generator.
3987 * Initial, experimental EVP support for AES-GCM. AAD can be input by
3988 setting output buffer to NULL. The `*Final` function must be
3989 called although it will not retrieve any additional data. The tag
3990 can be set or retrieved with a ctrl. The IV length is by default 12
3991 bytes (96 bits) but can be set to an alternative value. If the IV
3992 length exceeds the maximum IV length (currently 16 bytes) it cannot be
3997 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
3998 underlying do_cipher function handles all cipher semantics itself
3999 including padding and finalisation. This is useful if (for example)
4000 an ENGINE cipher handles block padding itself. The behaviour of
4001 do_cipher is subtly changed if this flag is set: the return value
4002 is the number of characters written to the output buffer (zero is
4003 no longer an error code) or a negative error code. Also if the
4004 input buffer is NULL and length 0 finalisation should be performed.
4008 * If a candidate issuer certificate is already part of the constructed
4009 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4013 * Improve forward-security support: add functions
4015 void SSL_CTX_set_not_resumable_session_callback(
4016 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4017 void SSL_set_not_resumable_session_callback(
4018 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4020 for use by SSL/TLS servers; the callback function will be called whenever a
4021 new session is created, and gets to decide whether the session may be
4022 cached to make it resumable (return 0) or not (return 1). (As by the
4023 SSL/TLS protocol specifications, the session_id sent by the server will be
4024 empty to indicate that the session is not resumable; also, the server will
4025 not generate RFC 4507 (RFC 5077) session tickets.)
4027 A simple reasonable callback implementation is to return is_forward_secure.
4028 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4029 by the SSL/TLS server library, indicating whether it can provide forward
4032 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4034 * New -verify_name option in command line utilities to set verification
4039 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4040 Add CMAC pkey methods.
4044 * Experimental renegotiation in s_server -www mode. If the client
4045 browses /reneg connection is renegotiated. If /renegcert it is
4046 renegotiated requesting a certificate.
4050 * Add an "external" session cache for debugging purposes to s_server. This
4051 should help trace issues which normally are only apparent in deployed
4052 multi-process servers.
4056 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4057 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4058 BIO_set_cipher() and some obscure PEM functions were changed so they
4059 can now return an error. The RAND changes required a change to the
4060 RAND_METHOD structure.
4064 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4065 a gcc attribute to warn if the result of a function is ignored. This
4066 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4067 whose return value is often ignored.
4071 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4072 These allow SCTs (signed certificate timestamps) to be requested and
4073 validated when establishing a connection.
4075 *Rob Percival <robpercival@google.com>*
4080 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4082 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4083 used even when parsing explicit parameters, when loading a encoded key
4084 or calling `EC_GROUP_new_from_ecpkparameters()`/
4085 `EC_GROUP_new_from_ecparameters()`.
4086 This prevents bypass of security hardening and performance gains,
4087 especially for curves with specialized EC_METHODs.
4088 By default, if a key encoded with explicit parameters is loaded and later
4089 encoded, the output is still encoded with explicit parameters, even if
4090 internally a "named" EC_GROUP is used for computation.
4094 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4095 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4096 NULL. After this change, only the cofactor parameter can be NULL. It also
4097 does some minimal sanity checks on the passed order.
4102 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4103 An attack is simple, if the first CMS_recipientInfo is valid but the
4104 second CMS_recipientInfo is chosen ciphertext. If the second
4105 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4106 encryption key will be replaced by garbage, and the message cannot be
4107 decoded, but if the RSA decryption fails, the correct encryption key is
4108 used and the recipient will not notice the attack.
4109 As a work around for this potential attack the length of the decrypted
4110 key must be equal to the cipher default key length, in case the
4111 certifiate is not given and all recipientInfo are tried out.
4112 The old behaviour can be re-enabled in the CMS code by setting the
4113 CMS_DEBUG_DECRYPT flag.
4118 * Document issue with installation paths in diverse Windows builds
4120 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4121 binaries and run-time config file.
4126 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4128 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4129 This changes the size when using the genpkey app when no size is given. It
4130 fixes an omission in earlier changes that changed all RSA, DSA and DH
4131 generation apps to use 2048 bits by default.
4135 * Add FIPS support for Android Arm 64-bit
4137 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4138 Module in Version 2.0.10. For some reason, the corresponding target
4139 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4140 built with FIPS support on Android Arm 64-bit. This omission has been
4143 *Matthias St. Pierre*
4145 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4147 * 0-byte record padding oracle
4149 If an application encounters a fatal protocol error and then calls
4150 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4151 then OpenSSL can respond differently to the calling application if a 0 byte
4152 record is received with invalid padding compared to if a 0 byte record is
4153 received with an invalid MAC. If the application then behaves differently
4154 based on that in a way that is detectable to the remote peer, then this
4155 amounts to a padding oracle that could be used to decrypt data.
4157 In order for this to be exploitable "non-stitched" ciphersuites must be in
4158 use. Stitched ciphersuites are optimised implementations of certain
4159 commonly used ciphersuites. Also the application must call SSL_shutdown()
4160 twice even if a protocol error has occurred (applications should not do
4161 this but some do anyway).
4163 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4164 Aviram, with additional investigation by Steven Collison and Andrew
4165 Hourselt. It was reported to OpenSSL on 10th December 2018.
4170 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4174 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4176 * Microarchitecture timing vulnerability in ECC scalar multiplication
4178 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4179 shown to be vulnerable to a microarchitecture timing side channel attack.
4180 An attacker with sufficient access to mount local timing attacks during
4181 ECDSA signature generation could recover the private key.
4183 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4184 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4190 * Timing vulnerability in DSA signature generation
4192 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4193 timing side channel attack. An attacker could use variations in the signing
4194 algorithm to recover the private key.
4196 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4201 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4202 Module, accidentally introduced while backporting security fixes from the
4203 development branch and hindering the use of ECC in FIPS mode.
4207 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4209 * Client DoS due to large DH parameter
4211 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4212 malicious server can send a very large prime value to the client. This will
4213 cause the client to spend an unreasonably long period of time generating a
4214 key for this prime resulting in a hang until the client has finished. This
4215 could be exploited in a Denial Of Service attack.
4217 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4222 * Cache timing vulnerability in RSA Key Generation
4224 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4225 a cache timing side channel attack. An attacker with sufficient access to
4226 mount cache timing attacks during the RSA key generation process could
4227 recover the private key.
4229 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4230 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4235 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4236 parameter is no longer accepted, as it leads to a corrupt table. NULL
4237 pem_str is reserved for alias entries only.
4241 * Revert blinding in ECDSA sign and instead make problematic addition
4242 length-invariant. Switch even to fixed-length Montgomery multiplication.
4246 * Change generating and checking of primes so that the error rate of not
4247 being prime depends on the intended use based on the size of the input.
4248 For larger primes this will result in more rounds of Miller-Rabin.
4249 The maximal error rate for primes with more than 1080 bits is lowered
4252 *Kurt Roeckx, Annie Yousar*
4254 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4258 * Add blinding to ECDSA and DSA signatures to protect against side channel
4259 attacks discovered by Keegan Ryan (NCC Group).
4263 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4264 now allow empty (zero character) pass phrases.
4268 * Certificate time validation (X509_cmp_time) enforces stricter
4269 compliance with RFC 5280. Fractional seconds and timezone offsets
4270 are no longer allowed.
4274 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4276 * Constructed ASN.1 types with a recursive definition could exceed the stack
4278 Constructed ASN.1 types with a recursive definition (such as can be found
4279 in PKCS7) could eventually exceed the stack given malicious input with
4280 excessive recursion. This could result in a Denial Of Service attack. There
4281 are no such structures used within SSL/TLS that come from untrusted sources
4282 so this is considered safe.
4284 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4290 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4292 * Read/write after SSL object in error state
4294 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4295 mechanism. The intent was that if a fatal error occurred during a handshake
4296 then OpenSSL would move into the error state and would immediately fail if
4297 you attempted to continue the handshake. This works as designed for the
4298 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4299 SSL_connect()), however due to a bug it does not work correctly if
4300 SSL_read() or SSL_write() is called directly. In that scenario, if the
4301 handshake fails then a fatal error will be returned in the initial function
4302 call. If SSL_read()/SSL_write() is subsequently called by the application
4303 for the same SSL object then it will succeed and the data is passed without
4304 being decrypted/encrypted directly from the SSL/TLS record layer.
4306 In order to exploit this issue an application bug would have to be present
4307 that resulted in a call to SSL_read()/SSL_write() being issued after having
4308 already received a fatal error.
4310 This issue was reported to OpenSSL by David Benjamin (Google).
4315 * rsaz_1024_mul_avx2 overflow bug on x86_64
4317 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4318 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4319 Analysis suggests that attacks against RSA and DSA as a result of this
4320 defect would be very difficult to perform and are not believed likely.
4321 Attacks against DH1024 are considered just feasible, because most of the
4322 work necessary to deduce information about a private key may be performed
4323 offline. The amount of resources required for such an attack would be
4324 significant. However, for an attack on TLS to be meaningful, the server
4325 would have to share the DH1024 private key among multiple clients, which is
4326 no longer an option since CVE-2016-0701.
4328 This only affects processors that support the AVX2 but not ADX extensions
4329 like Intel Haswell (4th generation).
4331 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4332 was originally found via the OSS-Fuzz project.
4337 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4339 * bn_sqrx8x_internal carry bug on x86_64
4341 There is a carry propagating bug in the x86_64 Montgomery squaring
4342 procedure. No EC algorithms are affected. Analysis suggests that attacks
4343 against RSA and DSA as a result of this defect would be very difficult to
4344 perform and are not believed likely. Attacks against DH are considered just
4345 feasible (although very difficult) because most of the work necessary to
4346 deduce information about a private key may be performed offline. The amount
4347 of resources required for such an attack would be very significant and
4348 likely only accessible to a limited number of attackers. An attacker would
4349 additionally need online access to an unpatched system using the target
4350 private key in a scenario with persistent DH parameters and a private
4351 key that is shared between multiple clients.
4353 This only affects processors that support the BMI1, BMI2 and ADX extensions
4354 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4356 This issue was reported to OpenSSL by the OSS-Fuzz project.
4361 * Malformed X.509 IPAddressFamily could cause OOB read
4363 If an X.509 certificate has a malformed IPAddressFamily extension,
4364 OpenSSL could do a one-byte buffer overread. The most likely result
4365 would be an erroneous display of the certificate in text format.
4367 This issue was reported to OpenSSL by the OSS-Fuzz project.
4372 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4374 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4375 platform rather than 'mingw'.
4379 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4381 * Truncated packet could crash via OOB read
4383 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4384 cipher is being used, then a truncated packet can cause that host to
4385 perform an out-of-bounds read, usually resulting in a crash.
4387 This issue was reported to OpenSSL by Robert Święcki of Google.
4392 * BN_mod_exp may produce incorrect results on x86_64
4394 There is a carry propagating bug in the x86_64 Montgomery squaring
4395 procedure. No EC algorithms are affected. Analysis suggests that attacks
4396 against RSA and DSA as a result of this defect would be very difficult to
4397 perform and are not believed likely. Attacks against DH are considered just
4398 feasible (although very difficult) because most of the work necessary to
4399 deduce information about a private key may be performed offline. The amount
4400 of resources required for such an attack would be very significant and
4401 likely only accessible to a limited number of attackers. An attacker would
4402 additionally need online access to an unpatched system using the target
4403 private key in a scenario with persistent DH parameters and a private
4404 key that is shared between multiple clients. For example this can occur by
4405 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4406 similar to CVE-2015-3193 but must be treated as a separate problem.
4408 This issue was reported to OpenSSL by the OSS-Fuzz project.
4413 * Montgomery multiplication may produce incorrect results
4415 There is a carry propagating bug in the Broadwell-specific Montgomery
4416 multiplication procedure that handles input lengths divisible by, but
4417 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4418 and DH private keys are impossible. This is because the subroutine in
4419 question is not used in operations with the private key itself and an input
4420 of the attacker's direct choice. Otherwise the bug can manifest itself as
4421 transient authentication and key negotiation failures or reproducible
4422 erroneous outcome of public-key operations with specially crafted input.
4423 Among EC algorithms only Brainpool P-512 curves are affected and one
4424 presumably can attack ECDH key negotiation. Impact was not analyzed in
4425 detail, because pre-requisites for attack are considered unlikely. Namely
4426 multiple clients have to choose the curve in question and the server has to
4427 share the private key among them, neither of which is default behaviour.
4428 Even then only clients that chose the curve will be affected.
4430 This issue was publicly reported as transient failures and was not
4431 initially recognized as a security issue. Thanks to Richard Morgan for
4432 providing reproducible case.
4437 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4438 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4439 prevent issues where no progress is being made and the peer continually
4440 sends unrecognised record types, using up resources processing them.
4444 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4446 * Missing CRL sanity check
4448 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4449 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4450 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4452 This issue only affects the OpenSSL 1.0.2i
4457 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4459 * OCSP Status Request extension unbounded memory growth
4461 A malicious client can send an excessively large OCSP Status Request
4462 extension. If that client continually requests renegotiation, sending a
4463 large OCSP Status Request extension each time, then there will be unbounded
4464 memory growth on the server. This will eventually lead to a Denial Of
4465 Service attack through memory exhaustion. Servers with a default
4466 configuration are vulnerable even if they do not support OCSP. Builds using
4467 the "no-ocsp" build time option are not affected.
4469 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4474 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4477 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4483 * OOB write in MDC2_Update()
4485 An overflow can occur in MDC2_Update() either if called directly or
4486 through the EVP_DigestUpdate() function using MDC2. If an attacker
4487 is able to supply very large amounts of input data after a previous
4488 call to EVP_EncryptUpdate() with a partial block then a length check
4489 can overflow resulting in a heap corruption.
4491 The amount of data needed is comparable to SIZE_MAX which is impractical
4494 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4499 * Malformed SHA512 ticket DoS
4501 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4502 DoS attack where a malformed ticket will result in an OOB read which will
4505 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4506 a custom server callback and ticket lookup mechanism.
4508 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4513 * OOB write in BN_bn2dec()
4515 The function BN_bn2dec() does not check the return value of BN_div_word().
4516 This can cause an OOB write if an application uses this function with an
4517 overly large BIGNUM. This could be a problem if an overly large certificate
4518 or CRL is printed out from an untrusted source. TLS is not affected because
4519 record limits will reject an oversized certificate before it is parsed.
4521 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4526 * OOB read in TS_OBJ_print_bio()
4528 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4529 the total length the OID text representation would use and not the amount
4530 of data written. This will result in OOB reads when large OIDs are
4533 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4538 * Pointer arithmetic undefined behaviour
4540 Avoid some undefined pointer arithmetic
4542 A common idiom in the codebase is to check limits in the following manner:
4545 Where "p" points to some malloc'd data of SIZE bytes and
4548 "len" here could be from some externally supplied data (e.g. from a TLS
4551 The rules of C pointer arithmetic are such that "p + len" is only well
4552 defined where len <= SIZE. Therefore the above idiom is actually
4553 undefined behaviour.
4555 For example this could cause problems if some malloc implementation
4556 provides an address for "p" such that "p + len" actually overflows for
4557 values of len that are too big and therefore p + len < limit.
4559 This issue was reported to OpenSSL by Guido Vranken
4564 * Constant time flag not preserved in DSA signing
4566 Operations in the DSA signing algorithm should run in constant time in
4567 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4568 implementation means that a non-constant time codepath is followed for
4569 certain operations. This has been demonstrated through a cache-timing
4570 attack to be sufficient for an attacker to recover the private DSA key.
4572 This issue was reported by César Pereida (Aalto University), Billy Brumley
4573 (Tampere University of Technology), and Yuval Yarom (The University of
4574 Adelaide and NICTA).
4579 * DTLS buffered message DoS
4581 In a DTLS connection where handshake messages are delivered out-of-order
4582 those messages that OpenSSL is not yet ready to process will be buffered
4583 for later use. Under certain circumstances, a flaw in the logic means that
4584 those messages do not get removed from the buffer even though the handshake
4585 has been completed. An attacker could force up to approx. 15 messages to
4586 remain in the buffer when they are no longer required. These messages will
4587 be cleared when the DTLS connection is closed. The default maximum size for
4588 a message is 100k. Therefore the attacker could force an additional 1500k
4589 to be consumed per connection. By opening many simulataneous connections an
4590 attacker could cause a DoS attack through memory exhaustion.
4592 This issue was reported to OpenSSL by Quan Luo.
4597 * DTLS replay protection DoS
4599 A flaw in the DTLS replay attack protection mechanism means that records
4600 that arrive for future epochs update the replay protection "window" before
4601 the MAC for the record has been validated. This could be exploited by an
4602 attacker by sending a record for the next epoch (which does not have to
4603 decrypt or have a valid MAC), with a very large sequence number. This means
4604 that all subsequent legitimate packets are dropped causing a denial of
4605 service for a specific DTLS connection.
4607 This issue was reported to OpenSSL by the OCAP audit team.
4612 * Certificate message OOB reads
4614 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4615 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4616 theoretical DoS risk but this has not been observed in practice on common
4619 The messages affected are client certificate, client certificate request
4620 and server certificate. As a result the attack can only be performed
4621 against a client or a server which enables client authentication.
4623 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4628 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4630 * Prevent padding oracle in AES-NI CBC MAC check
4632 A MITM attacker can use a padding oracle attack to decrypt traffic
4633 when the connection uses an AES CBC cipher and the server support
4636 This issue was introduced as part of the fix for Lucky 13 padding
4637 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4638 constant time by making sure that always the same bytes are read and
4639 compared against either the MAC or padding bytes. But it no longer
4640 checked that there was enough data to have both the MAC and padding
4643 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4648 * Fix EVP_EncodeUpdate overflow
4650 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4651 Base64 encoding of binary data. If an attacker is able to supply very large
4652 amounts of input data then a length check can overflow resulting in a heap
4655 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4656 the `PEM_write_bio*` family of functions. These are mainly used within the
4657 OpenSSL command line applications, so any application which processes data
4658 from an untrusted source and outputs it as a PEM file should be considered
4659 vulnerable to this issue. User applications that call these APIs directly
4660 with large amounts of untrusted data may also be vulnerable.
4662 This issue was reported by Guido Vranken.
4667 * Fix EVP_EncryptUpdate overflow
4669 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4670 is able to supply very large amounts of input data after a previous call to
4671 EVP_EncryptUpdate() with a partial block then a length check can overflow
4672 resulting in a heap corruption. Following an analysis of all OpenSSL
4673 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4674 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4675 the first called function after an EVP_EncryptInit(), and therefore that
4676 specific call must be safe. The second form is where the length passed to
4677 EVP_EncryptUpdate() can be seen from the code to be some small value and
4678 therefore there is no possibility of an overflow. Since all instances are
4679 one of these two forms, it is believed that there can be no overflows in
4680 internal code due to this problem. It should be noted that
4681 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4682 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4683 of these calls have also been analysed too and it is believed there are no
4684 instances in internal usage where an overflow could occur.
4686 This issue was reported by Guido Vranken.
4691 * Prevent ASN.1 BIO excessive memory allocation
4693 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4694 a short invalid encoding can cause allocation of large amounts of memory
4695 potentially consuming excessive resources or exhausting memory.
4697 Any application parsing untrusted data through d2i BIO functions is
4698 affected. The memory based functions such as d2i_X509() are *not* affected.
4699 Since the memory based functions are used by the TLS library, TLS
4700 applications are not affected.
4702 This issue was reported by Brian Carpenter.
4709 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4710 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4711 in arbitrary stack data being returned in the buffer.
4713 This issue was reported by Guido Vranken.
4718 * Modify behavior of ALPN to invoke callback after SNI/servername
4719 callback, such that updates to the SSL_CTX affect ALPN.
4723 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4728 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4729 methods are enabled and ssl2 is disabled the methods return NULL.
4733 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4735 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4736 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4737 provide any "EXPORT" or "LOW" strength ciphers.
4741 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4742 is by default disabled at build-time. Builds that are not configured with
4743 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4744 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4745 will need to explicitly call either of:
4747 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4749 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4751 as appropriate. Even if either of those is used, or the application
4752 explicitly uses the version-specific SSLv2_method() or its client and
4753 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4754 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4755 ciphers, and SSLv2 56-bit DES are no longer available.
4760 * Fix a double-free in DSA code
4762 A double free bug was discovered when OpenSSL parses malformed DSA private
4763 keys and could lead to a DoS attack or memory corruption for applications
4764 that receive DSA private keys from untrusted sources. This scenario is
4767 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4773 * Disable SRP fake user seed to address a server memory leak.
4775 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4777 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4778 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4779 was changed to ignore the "fake user" SRP seed, even if the seed
4782 Users should use SRP_VBASE_get1_by_user instead. Note that in
4783 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4784 also that even though configuring the SRP seed attempts to hide
4785 invalid usernames by continuing the handshake with fake
4786 credentials, this behaviour is not constant time and no strong
4787 guarantees are made that the handshake is indistinguishable from
4788 that of a valid user.
4793 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4795 In the BN_hex2bn function the number of hex digits is calculated using an
4796 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4797 large values of `i` this can result in `bn_expand` not allocating any
4798 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4799 field as NULL leading to a subsequent NULL ptr deref. For very large values
4800 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4801 In this case memory is allocated to the internal BIGNUM data field, but it
4802 is insufficiently sized leading to heap corruption. A similar issue exists
4803 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4804 is ever called by user applications with very large untrusted hex/dec data.
4805 This is anticipated to be a rare occurrence.
4807 All OpenSSL internal usage of these functions use data that is not expected
4808 to be untrusted, e.g. config file data or application command line
4809 arguments. If user developed applications generate config file data based
4810 on untrusted data then it is possible that this could also lead to security
4811 consequences. This is also anticipated to be rare.
4813 This issue was reported to OpenSSL by Guido Vranken.
4818 * Fix memory issues in `BIO_*printf` functions
4820 The internal `fmtstr` function used in processing a "%s" format string in
4821 the `BIO_*printf` functions could overflow while calculating the length of a
4822 string and cause an OOB read when printing very long strings.
4824 Additionally the internal `doapr_outch` function can attempt to write to an
4825 OOB memory location (at an offset from the NULL pointer) in the event of a
4826 memory allocation failure. In 1.0.2 and below this could be caused where
4827 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4828 could be in processing a very long "%s" format string. Memory leaks can
4831 The first issue may mask the second issue dependent on compiler behaviour.
4832 These problems could enable attacks where large amounts of untrusted data
4833 is passed to the `BIO_*printf` functions. If applications use these functions
4834 in this way then they could be vulnerable. OpenSSL itself uses these
4835 functions when printing out human-readable dumps of ASN.1 data. Therefore
4836 applications that print this data could be vulnerable if the data is from
4837 untrusted sources. OpenSSL command line applications could also be
4838 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4839 as command line arguments.
4841 Libssl is not considered directly vulnerable. Additionally certificates etc
4842 received via remote connections via libssl are also unlikely to be able to
4843 trigger these issues because of message size limits enforced within libssl.
4845 This issue was reported to OpenSSL Guido Vranken.
4850 * Side channel attack on modular exponentiation
4852 A side-channel attack was found which makes use of cache-bank conflicts on
4853 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4854 of RSA keys. The ability to exploit this issue is limited as it relies on
4855 an attacker who has control of code in a thread running on the same
4856 hyper-threaded core as the victim thread which is performing decryptions.
4858 This issue was reported to OpenSSL by Yuval Yarom, The University of
4859 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4860 Nadia Heninger, University of Pennsylvania with more information at
4861 <http://cachebleed.info>.
4866 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4867 if no keysize is specified with default_bits. This fixes an
4868 omission in an earlier change that changed all RSA/DSA key generation
4869 apps to use 2048 bits by default.
4873 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4875 * DH small subgroups
4877 Historically OpenSSL only ever generated DH parameters based on "safe"
4878 primes. More recently (in version 1.0.2) support was provided for
4879 generating X9.42 style parameter files such as those required for RFC 5114
4880 support. The primes used in such files may not be "safe". Where an
4881 application is using DH configured with parameters based on primes that are
4882 not "safe" then an attacker could use this fact to find a peer's private
4883 DH exponent. This attack requires that the attacker complete multiple
4884 handshakes in which the peer uses the same private DH exponent. For example
4885 this could be used to discover a TLS server's private DH exponent if it's
4886 reusing the private DH exponent or it's using a static DH ciphersuite.
4888 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4889 TLS. It is not on by default. If the option is not set then the server
4890 reuses the same private DH exponent for the life of the server process and
4891 would be vulnerable to this attack. It is believed that many popular
4892 applications do set this option and would therefore not be at risk.
4894 The fix for this issue adds an additional check where a "q" parameter is
4895 available (as is the case in X9.42 based parameters). This detects the
4896 only known attack, and is the only possible defense for static DH
4897 ciphersuites. This could have some performance impact.
4899 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4900 default and cannot be disabled. This could have some performance impact.
4902 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4907 * SSLv2 doesn't block disabled ciphers
4909 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4910 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4911 been disabled, provided that the SSLv2 protocol was not also disabled via
4914 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4915 and Sebastian Schinzel.
4920 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4922 * BN_mod_exp may produce incorrect results on x86_64
4924 There is a carry propagating bug in the x86_64 Montgomery squaring
4925 procedure. No EC algorithms are affected. Analysis suggests that attacks
4926 against RSA and DSA as a result of this defect would be very difficult to
4927 perform and are not believed likely. Attacks against DH are considered just
4928 feasible (although very difficult) because most of the work necessary to
4929 deduce information about a private key may be performed offline. The amount
4930 of resources required for such an attack would be very significant and
4931 likely only accessible to a limited number of attackers. An attacker would
4932 additionally need online access to an unpatched system using the target
4933 private key in a scenario with persistent DH parameters and a private
4934 key that is shared between multiple clients. For example this can occur by
4935 default in OpenSSL DHE based SSL/TLS ciphersuites.
4937 This issue was reported to OpenSSL by Hanno Böck.
4942 * Certificate verify crash with missing PSS parameter
4944 The signature verification routines will crash with a NULL pointer
4945 dereference if presented with an ASN.1 signature using the RSA PSS
4946 algorithm and absent mask generation function parameter. Since these
4947 routines are used to verify certificate signature algorithms this can be
4948 used to crash any certificate verification operation and exploited in a
4949 DoS attack. Any application which performs certificate verification is
4950 vulnerable including OpenSSL clients and servers which enable client
4953 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4958 * X509_ATTRIBUTE memory leak
4960 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4961 memory. This structure is used by the PKCS#7 and CMS routines so any
4962 application which reads PKCS#7 or CMS data from untrusted sources is
4963 affected. SSL/TLS is not affected.
4965 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4971 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4972 This changes the decoding behaviour for some invalid messages,
4973 though the change is mostly in the more lenient direction, and
4974 legacy behaviour is preserved as much as possible.
4978 * In DSA_generate_parameters_ex, if the provided seed is too short,
4981 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4983 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4985 * Alternate chains certificate forgery
4987 During certificate verification, OpenSSL will attempt to find an
4988 alternative certificate chain if the first attempt to build such a chain
4989 fails. An error in the implementation of this logic can mean that an
4990 attacker could cause certain checks on untrusted certificates to be
4991 bypassed, such as the CA flag, enabling them to use a valid leaf
4992 certificate to act as a CA and "issue" an invalid certificate.
4994 This issue was reported to OpenSSL by Adam Langley/David Benjamin
4999 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5001 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5002 incompatibility in the handling of HMAC. The previous ABI has now been
5007 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5009 * Malformed ECParameters causes infinite loop
5011 When processing an ECParameters structure OpenSSL enters an infinite loop
5012 if the curve specified is over a specially malformed binary polynomial
5015 This can be used to perform denial of service against any
5016 system which processes public keys, certificate requests or
5017 certificates. This includes TLS clients and TLS servers with
5018 client authentication enabled.
5020 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5025 * Exploitable out-of-bounds read in X509_cmp_time
5027 X509_cmp_time does not properly check the length of the ASN1_TIME
5028 string and can read a few bytes out of bounds. In addition,
5029 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5032 An attacker can use this to craft malformed certificates and CRLs of
5033 various sizes and potentially cause a segmentation fault, resulting in
5034 a DoS on applications that verify certificates or CRLs. TLS clients
5035 that verify CRLs are affected. TLS clients and servers with client
5036 authentication enabled may be affected if they use custom verification
5039 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5040 independently by Hanno Böck.
5045 * PKCS7 crash with missing EnvelopedContent
5047 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5048 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5049 with missing content and trigger a NULL pointer dereference on parsing.
5051 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5052 structures from untrusted sources are affected. OpenSSL clients and
5053 servers are not affected.
5055 This issue was reported to OpenSSL by Michal Zalewski (Google).
5060 * CMS verify infinite loop with unknown hash function
5062 When verifying a signedData message the CMS code can enter an infinite loop
5063 if presented with an unknown hash function OID. This can be used to perform
5064 denial of service against any system which verifies signedData messages using
5066 This issue was reported to OpenSSL by Johannes Bauer.
5071 * Race condition handling NewSessionTicket
5073 If a NewSessionTicket is received by a multi-threaded client when attempting to
5074 reuse a previous ticket then a race condition can occur potentially leading to
5075 a double free of the ticket data.
5080 * Only support 256-bit or stronger elliptic curves with the
5081 'ecdh_auto' setting (server) or by default (client). Of supported
5082 curves, prefer P-256 (both).
5086 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5088 * ClientHello sigalgs DoS fix
5090 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5091 invalid signature algorithms extension a NULL pointer dereference will
5092 occur. This can be exploited in a DoS attack against the server.
5094 This issue was was reported to OpenSSL by David Ramos of Stanford
5098 *Stephen Henson and Matt Caswell*
5100 * Multiblock corrupted pointer fix
5102 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5103 feature only applies on 64 bit x86 architecture platforms that support AES
5104 NI instructions. A defect in the implementation of "multiblock" can cause
5105 OpenSSL's internal write buffer to become incorrectly set to NULL when
5106 using non-blocking IO. Typically, when the user application is using a
5107 socket BIO for writing, this will only result in a failed connection.
5108 However if some other BIO is used then it is likely that a segmentation
5109 fault will be triggered, thus enabling a potential DoS attack.
5111 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5116 * Segmentation fault in DTLSv1_listen fix
5118 The DTLSv1_listen function is intended to be stateless and processes the
5119 initial ClientHello from many peers. It is common for user code to loop
5120 over the call to DTLSv1_listen until a valid ClientHello is received with
5121 an associated cookie. A defect in the implementation of DTLSv1_listen means
5122 that state is preserved in the SSL object from one invocation to the next
5123 that can lead to a segmentation fault. Errors processing the initial
5124 ClientHello can trigger this scenario. An example of such an error could be
5125 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5128 This issue was reported to OpenSSL by Per Allansson.
5133 * Segmentation fault in ASN1_TYPE_cmp fix
5135 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5136 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5137 certificate signature algorithm consistency this can be used to crash any
5138 certificate verification operation and exploited in a DoS attack. Any
5139 application which performs certificate verification is vulnerable including
5140 OpenSSL clients and servers which enable client authentication.
5145 * Segmentation fault for invalid PSS parameters fix
5147 The signature verification routines will crash with a NULL pointer
5148 dereference if presented with an ASN.1 signature using the RSA PSS
5149 algorithm and invalid parameters. Since these routines are used to verify
5150 certificate signature algorithms this can be used to crash any
5151 certificate verification operation and exploited in a DoS attack. Any
5152 application which performs certificate verification is vulnerable including
5153 OpenSSL clients and servers which enable client authentication.
5155 This issue was was reported to OpenSSL by Brian Carpenter.
5160 * ASN.1 structure reuse memory corruption fix
5162 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5163 memory corruption via an invalid write. Such reuse is and has been
5164 strongly discouraged and is believed to be rare.
5166 Applications that parse structures containing CHOICE or ANY DEFINED BY
5167 components may be affected. Certificate parsing (d2i_X509 and related
5168 functions) are however not affected. OpenSSL clients and servers are
5174 * PKCS7 NULL pointer dereferences fix
5176 The PKCS#7 parsing code does not handle missing outer ContentInfo
5177 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5178 missing content and trigger a NULL pointer dereference on parsing.
5180 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5181 otherwise parse PKCS#7 structures from untrusted sources are
5182 affected. OpenSSL clients and servers are not affected.
5184 This issue was reported to OpenSSL by Michal Zalewski (Google).
5189 * DoS via reachable assert in SSLv2 servers fix
5191 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5192 servers that both support SSLv2 and enable export cipher suites by sending
5193 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5195 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5196 (OpenSSL development team).
5201 * Empty CKE with client auth and DHE fix
5203 If client auth is used then a server can seg fault in the event of a DHE
5204 ciphersuite being selected and a zero length ClientKeyExchange message
5205 being sent by the client. This could be exploited in a DoS attack.
5210 * Handshake with unseeded PRNG fix
5212 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5213 with an unseeded PRNG. The conditions are:
5214 - The client is on a platform where the PRNG has not been seeded
5215 automatically, and the user has not seeded manually
5216 - A protocol specific client method version has been used (i.e. not
5217 SSL_client_methodv23)
5218 - A ciphersuite is used that does not require additional random data from
5219 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5221 If the handshake succeeds then the client random that has been used will
5222 have been generated from a PRNG with insufficient entropy and therefore the
5223 output may be predictable.
5225 For example using the following command with an unseeded openssl will
5226 succeed on an unpatched platform:
5228 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5233 * Use After Free following d2i_ECPrivatekey error fix
5235 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5236 could cause a use after free condition. This, in turn, could cause a double
5237 free in several private key parsing functions (such as d2i_PrivateKey
5238 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5239 for applications that receive EC private keys from untrusted
5240 sources. This scenario is considered rare.
5242 This issue was discovered by the BoringSSL project and fixed in their
5248 * X509_to_X509_REQ NULL pointer deref fix
5250 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5251 the certificate key is invalid. This function is rarely used in practice.
5253 This issue was discovered by Brian Carpenter.
5258 * Removed the export ciphers from the DEFAULT ciphers
5262 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5264 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5265 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5266 So far those who have to target multiple platforms would compromise
5267 and argue that binary targeting say ARMv5 would still execute on
5268 ARMv8. "Universal" build resolves this compromise by providing
5269 near-optimal performance even on newer platforms.
5273 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5274 (other platforms pending).
5276 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5278 * Add support for the SignedCertificateTimestampList certificate and
5279 OCSP response extensions from RFC6962.
5283 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5284 for corner cases. (Certain input points at infinity could lead to
5285 bogus results, with non-infinity inputs mapped to infinity too.)
5289 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5290 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5291 common cases are optimized and there still is room for further
5292 improvements. Vector Permutation AES for Altivec is also added.
5296 * Add support for little-endian ppc64 Linux target.
5298 *Marcelo Cerri (IBM)*
5300 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5301 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5302 are optimized and there still is room for further improvements.
5303 Both 32- and 64-bit modes are supported.
5305 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5307 * Improved ARMv7 NEON support.
5311 * Support for SPARC Architecture 2011 crypto extensions, first
5312 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5313 SHA256/512, MD5, GHASH and modular exponentiation.
5315 *Andy Polyakov, David Miller*
5317 * Accelerated modular exponentiation for Intel processors, a.k.a.
5320 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5322 * Support for new and upcoming Intel processors, including AVX2,
5323 BMI and SHA ISA extensions. This includes additional "stitched"
5324 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5327 This work was sponsored by Intel Corp.
5331 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5332 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5333 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5337 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5338 this fixes a limitation in previous versions of OpenSSL.
5342 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5343 MGF1 digest and OAEP label.
5347 * Add EVP support for key wrapping algorithms, to avoid problems with
5348 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5349 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5350 algorithms and include tests cases.
5354 * Add functions to allocate and set the fields of an ECDSA_METHOD
5357 *Douglas E. Engert, Steve Henson*
5359 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5360 difference in days and seconds between two tm or ASN1_TIME structures.
5364 * Add -rev test option to s_server to just reverse order of characters
5365 received by client and send back to server. Also prints an abbreviated
5366 summary of the connection parameters.
5370 * New option -brief for s_client and s_server to print out a brief summary
5371 of connection parameters.
5375 * Add callbacks for arbitrary TLS extensions.
5377 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5379 * New option -crl_download in several openssl utilities to download CRLs
5380 from CRLDP extension in certificates.
5384 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5388 * New function X509_CRL_diff to generate a delta CRL from the difference
5389 of two full CRLs. Add support to "crl" utility.
5393 * New functions to set lookup_crls function and to retrieve
5394 X509_STORE from X509_STORE_CTX.
5398 * Print out deprecated issuer and subject unique ID fields in
5403 * Extend OCSP I/O functions so they can be used for simple general purpose
5404 HTTP as well as OCSP. New wrapper function which can be used to download
5405 CRLs using the OCSP API.
5409 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5413 * `SSL_CONF*` functions. These provide a common framework for application
5414 configuration using configuration files or command lines.
5418 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5419 message callback and prints the results. Needs compile time option
5420 "enable-ssl-trace". New options to s_client and s_server to enable
5425 * New ctrl and macro to retrieve supported points extensions.
5426 Print out extension in s_server and s_client.
5430 * New functions to retrieve certificate signature and signature
5435 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5440 * New Suite B modes for TLS code. These use and enforce the requirements
5441 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5442 only use Suite B curves. The Suite B modes can be set by using the
5443 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5447 * New chain verification flags for Suite B levels of security. Check
5448 algorithms are acceptable when flags are set in X509_verify_cert.
5452 * Make tls1_check_chain return a set of flags indicating checks passed
5453 by a certificate chain. Add additional tests to handle client
5454 certificates: checks for matching certificate type and issuer name
5459 * If an attempt is made to use a signature algorithm not in the peer
5460 preference list abort the handshake. If client has no suitable
5461 signature algorithms in response to a certificate request do not
5462 use the certificate.
5466 * If server EC tmp key is not in client preference list abort handshake.
5470 * Add support for certificate stores in CERT structure. This makes it
5471 possible to have different stores per SSL structure or one store in
5472 the parent SSL_CTX. Include distinct stores for certificate chain
5473 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5474 to build and store a certificate chain in CERT structure: returning
5475 an error if the chain cannot be built: this will allow applications
5476 to test if a chain is correctly configured.
5478 Note: if the CERT based stores are not set then the parent SSL_CTX
5479 store is used to retain compatibility with existing behaviour.
5483 * New function ssl_set_client_disabled to set a ciphersuite disabled
5484 mask based on the current session, check mask when sending client
5485 hello and checking the requested ciphersuite.
5489 * New ctrls to retrieve and set certificate types in a certificate
5490 request message. Print out received values in s_client. If certificate
5491 types is not set with custom values set sensible values based on
5492 supported signature algorithms.
5496 * Support for distinct client and server supported signature algorithms.
5500 * Add certificate callback. If set this is called whenever a certificate
5501 is required by client or server. An application can decide which
5502 certificate chain to present based on arbitrary criteria: for example
5503 supported signature algorithms. Add very simple example to s_server.
5504 This fixes many of the problems and restrictions of the existing client
5505 certificate callback: for example you can now clear an existing
5506 certificate and specify the whole chain.
5510 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5511 the certificate can be used for (if anything). Set valid_flags field
5512 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5513 to have similar checks in it.
5515 Add new "cert_flags" field to CERT structure and include a "strict mode".
5516 This enforces some TLS certificate requirements (such as only permitting
5517 certificate signature algorithms contained in the supported algorithms
5518 extension) which some implementations ignore: this option should be used
5519 with caution as it could cause interoperability issues.
5523 * Update and tidy signature algorithm extension processing. Work out
5524 shared signature algorithms based on preferences and peer algorithms
5525 and print them out in s_client and s_server. Abort handshake if no
5526 shared signature algorithms.
5530 * Add new functions to allow customised supported signature algorithms
5531 for SSL and SSL_CTX structures. Add options to s_client and s_server
5536 * New function SSL_certs_clear() to delete all references to certificates
5537 from an SSL structure. Before this once a certificate had been added
5538 it couldn't be removed.
5542 * Integrate hostname, email address and IP address checking with certificate
5543 verification. New verify options supporting checking in openssl utility.
5547 * Fixes and wildcard matching support to hostname and email checking
5548 functions. Add manual page.
5550 *Florian Weimer (Red Hat Product Security Team)*
5552 * New functions to check a hostname email or IP address against a
5553 certificate. Add options x509 utility to print results of checks against
5558 * Fix OCSP checking.
5560 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5562 * Initial experimental support for explicitly trusted non-root CAs.
5563 OpenSSL still tries to build a complete chain to a root but if an
5564 intermediate CA has a trust setting included that is used. The first
5565 setting is used: whether to trust (e.g., -addtrust option to the x509
5570 * Add -trusted_first option which attempts to find certificates in the
5571 trusted store even if an untrusted chain is also supplied.
5575 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5576 platform support for Linux and Android.
5580 * Support for linux-x32, ILP32 environment in x86_64 framework.
5584 * Experimental multi-implementation support for FIPS capable OpenSSL.
5585 When in FIPS mode the approved implementations are used as normal,
5586 when not in FIPS mode the internal unapproved versions are used instead.
5587 This means that the FIPS capable OpenSSL isn't forced to use the
5588 (often lower performance) FIPS implementations outside FIPS mode.
5592 * Transparently support X9.42 DH parameters when calling
5593 PEM_read_bio_DHparameters. This means existing applications can handle
5594 the new parameter format automatically.
5598 * Initial experimental support for X9.42 DH parameter format: mainly
5599 to support use of 'q' parameter for RFC5114 parameters.
5603 * Add DH parameters from RFC5114 including test data to dhtest.
5607 * Support for automatic EC temporary key parameter selection. If enabled
5608 the most preferred EC parameters are automatically used instead of
5609 hardcoded fixed parameters. Now a server just has to call:
5610 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5611 support ECDH and use the most appropriate parameters.
5615 * Enhance and tidy EC curve and point format TLS extension code. Use
5616 static structures instead of allocation if default values are used.
5617 New ctrls to set curves we wish to support and to retrieve shared curves.
5618 Print out shared curves in s_server. New options to s_server and s_client
5619 to set list of supported curves.
5623 * New ctrls to retrieve supported signature algorithms and
5624 supported curve values as an array of NIDs. Extend openssl utility
5625 to print out received values.
5629 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5630 between NIDs and the more common NIST names such as "P-256". Enhance
5631 ecparam utility and ECC method to recognise the NIST names for curves.
5635 * Enhance SSL/TLS certificate chain handling to support different
5636 chains for each certificate instead of one chain in the parent SSL_CTX.
5640 * Support for fixed DH ciphersuite client authentication: where both
5641 server and client use DH certificates with common parameters.
5645 * Support for fixed DH ciphersuites: those requiring DH server
5650 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5652 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5653 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5654 X509_CINF_get_signature were reverted post internal team review.
5659 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5661 * OCSP Status Request extension unbounded memory growth
5663 A malicious client can send an excessively large OCSP Status Request
5664 extension. If that client continually requests renegotiation, sending a
5665 large OCSP Status Request extension each time, then there will be unbounded
5666 memory growth on the server. This will eventually lead to a Denial Of
5667 Service attack through memory exhaustion. Servers with a default
5668 configuration are vulnerable even if they do not support OCSP. Builds using
5669 the "no-ocsp" build time option are not affected.
5671 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5676 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5679 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5685 * OOB write in MDC2_Update()
5687 An overflow can occur in MDC2_Update() either if called directly or
5688 through the EVP_DigestUpdate() function using MDC2. If an attacker
5689 is able to supply very large amounts of input data after a previous
5690 call to EVP_EncryptUpdate() with a partial block then a length check
5691 can overflow resulting in a heap corruption.
5693 The amount of data needed is comparable to SIZE_MAX which is impractical
5696 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5701 * Malformed SHA512 ticket DoS
5703 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5704 DoS attack where a malformed ticket will result in an OOB read which will
5707 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5708 a custom server callback and ticket lookup mechanism.
5710 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5715 * OOB write in BN_bn2dec()
5717 The function BN_bn2dec() does not check the return value of BN_div_word().
5718 This can cause an OOB write if an application uses this function with an
5719 overly large BIGNUM. This could be a problem if an overly large certificate
5720 or CRL is printed out from an untrusted source. TLS is not affected because
5721 record limits will reject an oversized certificate before it is parsed.
5723 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5728 * OOB read in TS_OBJ_print_bio()
5730 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5731 the total length the OID text representation would use and not the amount
5732 of data written. This will result in OOB reads when large OIDs are
5735 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5740 * Pointer arithmetic undefined behaviour
5742 Avoid some undefined pointer arithmetic
5744 A common idiom in the codebase is to check limits in the following manner:
5747 Where "p" points to some malloc'd data of SIZE bytes and
5750 "len" here could be from some externally supplied data (e.g. from a TLS
5753 The rules of C pointer arithmetic are such that "p + len" is only well
5754 defined where len <= SIZE. Therefore the above idiom is actually
5755 undefined behaviour.
5757 For example this could cause problems if some malloc implementation
5758 provides an address for "p" such that "p + len" actually overflows for
5759 values of len that are too big and therefore p + len < limit.
5761 This issue was reported to OpenSSL by Guido Vranken
5766 * Constant time flag not preserved in DSA signing
5768 Operations in the DSA signing algorithm should run in constant time in
5769 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5770 implementation means that a non-constant time codepath is followed for
5771 certain operations. This has been demonstrated through a cache-timing
5772 attack to be sufficient for an attacker to recover the private DSA key.
5774 This issue was reported by César Pereida (Aalto University), Billy Brumley
5775 (Tampere University of Technology), and Yuval Yarom (The University of
5776 Adelaide and NICTA).
5781 * DTLS buffered message DoS
5783 In a DTLS connection where handshake messages are delivered out-of-order
5784 those messages that OpenSSL is not yet ready to process will be buffered
5785 for later use. Under certain circumstances, a flaw in the logic means that
5786 those messages do not get removed from the buffer even though the handshake
5787 has been completed. An attacker could force up to approx. 15 messages to
5788 remain in the buffer when they are no longer required. These messages will
5789 be cleared when the DTLS connection is closed. The default maximum size for
5790 a message is 100k. Therefore the attacker could force an additional 1500k
5791 to be consumed per connection. By opening many simulataneous connections an
5792 attacker could cause a DoS attack through memory exhaustion.
5794 This issue was reported to OpenSSL by Quan Luo.
5799 * DTLS replay protection DoS
5801 A flaw in the DTLS replay attack protection mechanism means that records
5802 that arrive for future epochs update the replay protection "window" before
5803 the MAC for the record has been validated. This could be exploited by an
5804 attacker by sending a record for the next epoch (which does not have to
5805 decrypt or have a valid MAC), with a very large sequence number. This means
5806 that all subsequent legitimate packets are dropped causing a denial of
5807 service for a specific DTLS connection.
5809 This issue was reported to OpenSSL by the OCAP audit team.
5814 * Certificate message OOB reads
5816 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5817 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5818 theoretical DoS risk but this has not been observed in practice on common
5821 The messages affected are client certificate, client certificate request
5822 and server certificate. As a result the attack can only be performed
5823 against a client or a server which enables client authentication.
5825 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5830 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5832 * Prevent padding oracle in AES-NI CBC MAC check
5834 A MITM attacker can use a padding oracle attack to decrypt traffic
5835 when the connection uses an AES CBC cipher and the server support
5838 This issue was introduced as part of the fix for Lucky 13 padding
5839 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5840 constant time by making sure that always the same bytes are read and
5841 compared against either the MAC or padding bytes. But it no longer
5842 checked that there was enough data to have both the MAC and padding
5845 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5850 * Fix EVP_EncodeUpdate overflow
5852 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5853 Base64 encoding of binary data. If an attacker is able to supply very large
5854 amounts of input data then a length check can overflow resulting in a heap
5857 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5858 the `PEM_write_bio*` family of functions. These are mainly used within the
5859 OpenSSL command line applications, so any application which processes data
5860 from an untrusted source and outputs it as a PEM file should be considered
5861 vulnerable to this issue. User applications that call these APIs directly
5862 with large amounts of untrusted data may also be vulnerable.
5864 This issue was reported by Guido Vranken.
5869 * Fix EVP_EncryptUpdate overflow
5871 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5872 is able to supply very large amounts of input data after a previous call to
5873 EVP_EncryptUpdate() with a partial block then a length check can overflow
5874 resulting in a heap corruption. Following an analysis of all OpenSSL
5875 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5876 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5877 the first called function after an EVP_EncryptInit(), and therefore that
5878 specific call must be safe. The second form is where the length passed to
5879 EVP_EncryptUpdate() can be seen from the code to be some small value and
5880 therefore there is no possibility of an overflow. Since all instances are
5881 one of these two forms, it is believed that there can be no overflows in
5882 internal code due to this problem. It should be noted that
5883 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5884 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5885 of these calls have also been analysed too and it is believed there are no
5886 instances in internal usage where an overflow could occur.
5888 This issue was reported by Guido Vranken.
5893 * Prevent ASN.1 BIO excessive memory allocation
5895 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5896 a short invalid encoding can casuse allocation of large amounts of memory
5897 potentially consuming excessive resources or exhausting memory.
5899 Any application parsing untrusted data through d2i BIO functions is
5900 affected. The memory based functions such as d2i_X509() are *not* affected.
5901 Since the memory based functions are used by the TLS library, TLS
5902 applications are not affected.
5904 This issue was reported by Brian Carpenter.
5911 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5912 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5913 in arbitrary stack data being returned in the buffer.
5915 This issue was reported by Guido Vranken.
5920 * Modify behavior of ALPN to invoke callback after SNI/servername
5921 callback, such that updates to the SSL_CTX affect ALPN.
5925 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5930 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5931 methods are enabled and ssl2 is disabled the methods return NULL.
5935 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5937 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5938 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5939 provide any "EXPORT" or "LOW" strength ciphers.
5943 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5944 is by default disabled at build-time. Builds that are not configured with
5945 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5946 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5947 will need to explicitly call either of:
5949 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5951 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5953 as appropriate. Even if either of those is used, or the application
5954 explicitly uses the version-specific SSLv2_method() or its client and
5955 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5956 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5957 ciphers, and SSLv2 56-bit DES are no longer available.
5962 * Fix a double-free in DSA code
5964 A double free bug was discovered when OpenSSL parses malformed DSA private
5965 keys and could lead to a DoS attack or memory corruption for applications
5966 that receive DSA private keys from untrusted sources. This scenario is
5969 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5975 * Disable SRP fake user seed to address a server memory leak.
5977 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5979 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5980 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5981 was changed to ignore the "fake user" SRP seed, even if the seed
5984 Users should use SRP_VBASE_get1_by_user instead. Note that in
5985 SRP_VBASE_get1_by_user, caller must free the returned value. Note
5986 also that even though configuring the SRP seed attempts to hide
5987 invalid usernames by continuing the handshake with fake
5988 credentials, this behaviour is not constant time and no strong
5989 guarantees are made that the handshake is indistinguishable from
5990 that of a valid user.
5995 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
5997 In the BN_hex2bn function the number of hex digits is calculated using an
5998 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
5999 large values of `i` this can result in `bn_expand` not allocating any
6000 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6001 field as NULL leading to a subsequent NULL ptr deref. For very large values
6002 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6003 In this case memory is allocated to the internal BIGNUM data field, but it
6004 is insufficiently sized leading to heap corruption. A similar issue exists
6005 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6006 is ever called by user applications with very large untrusted hex/dec data.
6007 This is anticipated to be a rare occurrence.
6009 All OpenSSL internal usage of these functions use data that is not expected
6010 to be untrusted, e.g. config file data or application command line
6011 arguments. If user developed applications generate config file data based
6012 on untrusted data then it is possible that this could also lead to security
6013 consequences. This is also anticipated to be rare.
6015 This issue was reported to OpenSSL by Guido Vranken.
6020 * Fix memory issues in `BIO_*printf` functions
6022 The internal `fmtstr` function used in processing a "%s" format string in
6023 the `BIO_*printf` functions could overflow while calculating the length of a
6024 string and cause an OOB read when printing very long strings.
6026 Additionally the internal `doapr_outch` function can attempt to write to an
6027 OOB memory location (at an offset from the NULL pointer) in the event of a
6028 memory allocation failure. In 1.0.2 and below this could be caused where
6029 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6030 could be in processing a very long "%s" format string. Memory leaks can
6033 The first issue may mask the second issue dependent on compiler behaviour.
6034 These problems could enable attacks where large amounts of untrusted data
6035 is passed to the `BIO_*printf` functions. If applications use these functions
6036 in this way then they could be vulnerable. OpenSSL itself uses these
6037 functions when printing out human-readable dumps of ASN.1 data. Therefore
6038 applications that print this data could be vulnerable if the data is from
6039 untrusted sources. OpenSSL command line applications could also be
6040 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6041 as command line arguments.
6043 Libssl is not considered directly vulnerable. Additionally certificates etc
6044 received via remote connections via libssl are also unlikely to be able to
6045 trigger these issues because of message size limits enforced within libssl.
6047 This issue was reported to OpenSSL Guido Vranken.
6052 * Side channel attack on modular exponentiation
6054 A side-channel attack was found which makes use of cache-bank conflicts on
6055 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6056 of RSA keys. The ability to exploit this issue is limited as it relies on
6057 an attacker who has control of code in a thread running on the same
6058 hyper-threaded core as the victim thread which is performing decryptions.
6060 This issue was reported to OpenSSL by Yuval Yarom, The University of
6061 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6062 Nadia Heninger, University of Pennsylvania with more information at
6063 <http://cachebleed.info>.
6068 * Change the req app to generate a 2048-bit RSA/DSA key by default,
6069 if no keysize is specified with default_bits. This fixes an
6070 omission in an earlier change that changed all RSA/DSA key generation
6071 apps to use 2048 bits by default.
6075 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6077 * Protection for DH small subgroup attacks
6079 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6080 switched on by default and cannot be disabled. This could have some
6085 * SSLv2 doesn't block disabled ciphers
6087 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6088 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6089 been disabled, provided that the SSLv2 protocol was not also disabled via
6092 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6093 and Sebastian Schinzel.
6098 * Reject DH handshakes with parameters shorter than 1024 bits.
6102 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6104 * Certificate verify crash with missing PSS parameter
6106 The signature verification routines will crash with a NULL pointer
6107 dereference if presented with an ASN.1 signature using the RSA PSS
6108 algorithm and absent mask generation function parameter. Since these
6109 routines are used to verify certificate signature algorithms this can be
6110 used to crash any certificate verification operation and exploited in a
6111 DoS attack. Any application which performs certificate verification is
6112 vulnerable including OpenSSL clients and servers which enable client
6115 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6120 * X509_ATTRIBUTE memory leak
6122 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6123 memory. This structure is used by the PKCS#7 and CMS routines so any
6124 application which reads PKCS#7 or CMS data from untrusted sources is
6125 affected. SSL/TLS is not affected.
6127 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6133 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6134 This changes the decoding behaviour for some invalid messages,
6135 though the change is mostly in the more lenient direction, and
6136 legacy behaviour is preserved as much as possible.
6140 * In DSA_generate_parameters_ex, if the provided seed is too short,
6141 use a random seed, as already documented.
6143 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6145 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6147 * Alternate chains certificate forgery
6149 During certificate verfification, OpenSSL will attempt to find an
6150 alternative certificate chain if the first attempt to build such a chain
6151 fails. An error in the implementation of this logic can mean that an
6152 attacker could cause certain checks on untrusted certificates to be
6153 bypassed, such as the CA flag, enabling them to use a valid leaf
6154 certificate to act as a CA and "issue" an invalid certificate.
6156 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6162 * Race condition handling PSK identify hint
6164 If PSK identity hints are received by a multi-threaded client then
6165 the values are wrongly updated in the parent SSL_CTX structure. This can
6166 result in a race condition potentially leading to a double free of the
6172 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6174 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6175 incompatibility in the handling of HMAC. The previous ABI has now been
6178 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6180 * Malformed ECParameters causes infinite loop
6182 When processing an ECParameters structure OpenSSL enters an infinite loop
6183 if the curve specified is over a specially malformed binary polynomial
6186 This can be used to perform denial of service against any
6187 system which processes public keys, certificate requests or
6188 certificates. This includes TLS clients and TLS servers with
6189 client authentication enabled.
6191 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6196 * Exploitable out-of-bounds read in X509_cmp_time
6198 X509_cmp_time does not properly check the length of the ASN1_TIME
6199 string and can read a few bytes out of bounds. In addition,
6200 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6203 An attacker can use this to craft malformed certificates and CRLs of
6204 various sizes and potentially cause a segmentation fault, resulting in
6205 a DoS on applications that verify certificates or CRLs. TLS clients
6206 that verify CRLs are affected. TLS clients and servers with client
6207 authentication enabled may be affected if they use custom verification
6210 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6211 independently by Hanno Böck.
6216 * PKCS7 crash with missing EnvelopedContent
6218 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6219 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6220 with missing content and trigger a NULL pointer dereference on parsing.
6222 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6223 structures from untrusted sources are affected. OpenSSL clients and
6224 servers are not affected.
6226 This issue was reported to OpenSSL by Michal Zalewski (Google).
6231 * CMS verify infinite loop with unknown hash function
6233 When verifying a signedData message the CMS code can enter an infinite loop
6234 if presented with an unknown hash function OID. This can be used to perform
6235 denial of service against any system which verifies signedData messages using
6237 This issue was reported to OpenSSL by Johannes Bauer.
6242 * Race condition handling NewSessionTicket
6244 If a NewSessionTicket is received by a multi-threaded client when attempting to
6245 reuse a previous ticket then a race condition can occur potentially leading to
6246 a double free of the ticket data.
6251 * Reject DH handshakes with parameters shorter than 768 bits.
6253 *Kurt Roeckx and Emilia Kasper*
6255 * dhparam: generate 2048-bit parameters by default.
6257 *Kurt Roeckx and Emilia Kasper*
6259 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6261 * Segmentation fault in ASN1_TYPE_cmp fix
6263 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6264 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6265 certificate signature algorithm consistency this can be used to crash any
6266 certificate verification operation and exploited in a DoS attack. Any
6267 application which performs certificate verification is vulnerable including
6268 OpenSSL clients and servers which enable client authentication.
6273 * ASN.1 structure reuse memory corruption fix
6275 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6276 memory corruption via an invalid write. Such reuse is and has been
6277 strongly discouraged and is believed to be rare.
6279 Applications that parse structures containing CHOICE or ANY DEFINED BY
6280 components may be affected. Certificate parsing (d2i_X509 and related
6281 functions) are however not affected. OpenSSL clients and servers are
6287 * PKCS7 NULL pointer dereferences fix
6289 The PKCS#7 parsing code does not handle missing outer ContentInfo
6290 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6291 missing content and trigger a NULL pointer dereference on parsing.
6293 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6294 otherwise parse PKCS#7 structures from untrusted sources are
6295 affected. OpenSSL clients and servers are not affected.
6297 This issue was reported to OpenSSL by Michal Zalewski (Google).
6302 * DoS via reachable assert in SSLv2 servers fix
6304 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6305 servers that both support SSLv2 and enable export cipher suites by sending
6306 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6308 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6309 (OpenSSL development team).
6314 * Use After Free following d2i_ECPrivatekey error fix
6316 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6317 could cause a use after free condition. This, in turn, could cause a double
6318 free in several private key parsing functions (such as d2i_PrivateKey
6319 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6320 for applications that receive EC private keys from untrusted
6321 sources. This scenario is considered rare.
6323 This issue was discovered by the BoringSSL project and fixed in their
6329 * X509_to_X509_REQ NULL pointer deref fix
6331 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6332 the certificate key is invalid. This function is rarely used in practice.
6334 This issue was discovered by Brian Carpenter.
6339 * Removed the export ciphers from the DEFAULT ciphers
6343 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6345 * Build fixes for the Windows and OpenVMS platforms
6347 *Matt Caswell and Richard Levitte*
6349 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6351 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6352 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6353 dereference. This could lead to a Denial Of Service attack. Thanks to
6354 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6359 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6360 dtls1_buffer_record function under certain conditions. In particular this
6361 could occur if an attacker sent repeated DTLS records with the same
6362 sequence number but for the next epoch. The memory leak could be exploited
6363 by an attacker in a Denial of Service attack through memory exhaustion.
6364 Thanks to Chris Mueller for reporting this issue.
6369 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6370 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6371 method would be set to NULL which could later result in a NULL pointer
6372 dereference. Thanks to Frank Schmirler for reporting this issue.
6377 * Abort handshake if server key exchange message is omitted for ephemeral
6380 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6381 reporting this issue.
6386 * Remove non-export ephemeral RSA code on client and server. This code
6387 violated the TLS standard by allowing the use of temporary RSA keys in
6388 non-export ciphersuites and could be used by a server to effectively
6389 downgrade the RSA key length used to a value smaller than the server
6390 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6391 INRIA or reporting this issue.
6396 * Fixed issue where DH client certificates are accepted without verification.
6397 An OpenSSL server will accept a DH certificate for client authentication
6398 without the certificate verify message. This effectively allows a client to
6399 authenticate without the use of a private key. This only affects servers
6400 which trust a client certificate authority which issues certificates
6401 containing DH keys: these are extremely rare and hardly ever encountered.
6402 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6408 * Ensure that the session ID context of an SSL is updated when its
6409 SSL_CTX is updated via SSL_set_SSL_CTX.
6411 The session ID context is typically set from the parent SSL_CTX,
6412 and can vary with the CTX.
6416 * Fix various certificate fingerprint issues.
6418 By using non-DER or invalid encodings outside the signed portion of a
6419 certificate the fingerprint can be changed without breaking the signature.
6420 Although no details of the signed portion of the certificate can be changed
6421 this can cause problems with some applications: e.g. those using the
6422 certificate fingerprint for blacklists.
6424 1. Reject signatures with non zero unused bits.
6426 If the BIT STRING containing the signature has non zero unused bits reject
6427 the signature. All current signature algorithms require zero unused bits.
6429 2. Check certificate algorithm consistency.
6431 Check the AlgorithmIdentifier inside TBS matches the one in the
6432 certificate signature. NB: this will result in signature failure
6433 errors for some broken certificates.
6435 Thanks to Konrad Kraszewski from Google for reporting this issue.
6437 3. Check DSA/ECDSA signatures use DER.
6439 Re-encode DSA/ECDSA signatures and compare with the original received
6440 signature. Return an error if there is a mismatch.
6442 This will reject various cases including garbage after signature
6443 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6444 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6445 (negative or with leading zeroes).
6447 Further analysis was conducted and fixes were developed by Stephen Henson
6448 of the OpenSSL core team.
6454 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6455 results on some platforms, including x86_64. This bug occurs at random
6456 with a very low probability, and is not known to be exploitable in any
6457 way, though its exact impact is difficult to determine. Thanks to Pieter
6458 Wuille (Blockstream) who reported this issue and also suggested an initial
6459 fix. Further analysis was conducted by the OpenSSL development team and
6460 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6461 the OpenSSL core team.
6466 * Do not resume sessions on the server if the negotiated protocol
6467 version does not match the session's version. Resuming with a different
6468 version, while not strictly forbidden by the RFC, is of questionable
6469 sanity and breaks all known clients.
6471 *David Benjamin, Emilia Käsper*
6473 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6474 early CCS messages during renegotiation. (Note that because
6475 renegotiation is encrypted, this early CCS was not exploitable.)
6479 * Tighten client-side session ticket handling during renegotiation:
6480 ensure that the client only accepts a session ticket if the server sends
6481 the extension anew in the ServerHello. Previously, a TLS client would
6482 reuse the old extension state and thus accept a session ticket if one was
6483 announced in the initial ServerHello.
6485 Similarly, ensure that the client requires a session ticket if one
6486 was advertised in the ServerHello. Previously, a TLS client would
6487 ignore a missing NewSessionTicket message.
6491 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6495 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6496 sends a carefully crafted handshake message, to cause OpenSSL to fail
6497 to free up to 64k of memory causing a memory leak. This could be
6498 exploited in a Denial Of Service attack. This issue affects OpenSSL
6499 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6500 whether SRTP is used or configured. Implementations of OpenSSL that
6501 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6503 The fix was developed by the OpenSSL team.
6508 * Session Ticket Memory Leak.
6510 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6511 integrity of that ticket is first verified. In the event of a session
6512 ticket integrity check failing, OpenSSL will fail to free memory
6513 causing a memory leak. By sending a large number of invalid session
6514 tickets an attacker could exploit this issue in a Denial Of Service
6520 * Build option no-ssl3 is incomplete.
6522 When OpenSSL is configured with "no-ssl3" as a build option, servers
6523 could accept and complete a SSL 3.0 handshake, and clients could be
6524 configured to send them.
6527 *Akamai and the OpenSSL team*
6529 * Add support for TLS_FALLBACK_SCSV.
6530 Client applications doing fallback retries should call
6531 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6534 *Adam Langley, Bodo Moeller*
6536 * Add additional DigestInfo checks.
6538 Re-encode DigestInto in DER and check against the original when
6539 verifying RSA signature: this will reject any improperly encoded
6540 DigestInfo structures.
6542 Note: this is a precautionary measure and no attacks are currently known.
6546 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6548 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6549 SRP code can be overrun an internal buffer. Add sanity check that
6550 g, A, B < N to SRP code.
6552 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6553 Group for discovering this issue.
6558 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6559 TLS 1.0 instead of higher protocol versions when the ClientHello message
6560 is badly fragmented. This allows a man-in-the-middle attacker to force a
6561 downgrade to TLS 1.0 even if both the server and the client support a
6562 higher protocol version, by modifying the client's TLS records.
6564 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6565 researching this issue.
6570 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6571 to a denial of service attack. A malicious server can crash the client
6572 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6573 ciphersuite and sending carefully crafted handshake messages.
6575 Thanks to Felix Gröbert (Google) for discovering and researching this
6581 * By sending carefully crafted DTLS packets an attacker could cause openssl
6582 to leak memory. This can be exploited through a Denial of Service attack.
6583 Thanks to Adam Langley for discovering and researching this issue.
6588 * An attacker can force openssl to consume large amounts of memory whilst
6589 processing DTLS handshake messages. This can be exploited through a
6590 Denial of Service attack.
6591 Thanks to Adam Langley for discovering and researching this issue.
6596 * An attacker can force an error condition which causes openssl to crash
6597 whilst processing DTLS packets due to memory being freed twice. This
6598 can be exploited through a Denial of Service attack.
6599 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6605 * If a multithreaded client connects to a malicious server using a resumed
6606 session and the server sends an ec point format extension it could write
6607 up to 255 bytes to freed memory.
6609 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6615 * A malicious server can crash an OpenSSL client with a null pointer
6616 dereference (read) by specifying an SRP ciphersuite even though it was not
6617 properly negotiated with the client. This can be exploited through a
6618 Denial of Service attack.
6620 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6621 discovering and researching this issue.
6626 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6627 X509_name_oneline, X509_name_print_ex et al. to leak some information
6628 from the stack. Applications may be affected if they echo pretty printing
6629 output to the attacker.
6631 Thanks to Ivan Fratric (Google) for discovering this issue.
6634 *Emilia Käsper, and Steve Henson*
6636 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6637 for corner cases. (Certain input points at infinity could lead to
6638 bogus results, with non-infinity inputs mapped to infinity too.)
6642 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6644 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6645 handshake can force the use of weak keying material in OpenSSL
6646 SSL/TLS clients and servers.
6648 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6649 researching this issue. ([CVE-2014-0224])
6651 *KIKUCHI Masashi, Steve Henson*
6653 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6654 OpenSSL DTLS client the code can be made to recurse eventually crashing
6657 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6660 *Imre Rad, Steve Henson*
6662 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6663 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6664 client or server. This is potentially exploitable to run arbitrary
6665 code on a vulnerable client or server.
6667 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6669 *Jüri Aedla, Steve Henson*
6671 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6672 are subject to a denial of service attack.
6674 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6675 this issue. ([CVE-2014-3470])
6677 *Felix Gröbert, Ivan Fratric, Steve Henson*
6679 * Harmonize version and its documentation. -f flag is used to display
6682 *mancha <mancha1@zoho.com>*
6684 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6685 in i2d_ECPrivateKey.
6687 *mancha <mancha1@zoho.com>*
6689 * Fix some double frees. These are not thought to be exploitable.
6691 *mancha <mancha1@zoho.com>*
6693 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6695 * A missing bounds check in the handling of the TLS heartbeat extension
6696 can be used to reveal up to 64k of memory to a connected client or
6699 Thanks for Neel Mehta of Google Security for discovering this bug and to
6700 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6701 preparing the fix ([CVE-2014-0160])
6703 *Adam Langley, Bodo Moeller*
6705 * Fix for the attack described in the paper "Recovering OpenSSL
6706 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6707 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6708 <http://eprint.iacr.org/2014/140>
6710 Thanks to Yuval Yarom and Naomi Benger for discovering this
6711 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6713 *Yuval Yarom and Naomi Benger*
6715 * TLS pad extension: draft-agl-tls-padding-03
6717 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6718 TLS client Hello record length value would otherwise be > 255 and
6719 less that 512 pad with a dummy extension containing zeroes so it
6720 is at least 512 bytes long.
6722 *Adam Langley, Steve Henson*
6724 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6726 * Fix for TLS record tampering bug. A carefully crafted invalid
6727 handshake could crash OpenSSL with a NULL pointer exception.
6728 Thanks to Anton Johansson for reporting this issues.
6731 * Keep original DTLS digest and encryption contexts in retransmission
6732 structures so we can use the previous session parameters if they need
6733 to be resent. ([CVE-2013-6450])
6737 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6738 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6739 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6740 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6741 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6742 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6744 *Rob Stradling, Adam Langley*
6746 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6748 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6749 supporting platforms or when small records were transferred.
6751 *Andy Polyakov, Steve Henson*
6753 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6755 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6757 This addresses the flaw in CBC record processing discovered by
6758 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6759 at: <http://www.isg.rhul.ac.uk/tls/>
6761 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6762 Security Group at Royal Holloway, University of London
6763 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6764 Emilia Käsper for the initial patch.
6767 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6769 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6770 ciphersuites which can be exploited in a denial of service attack.
6771 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6772 and detecting this bug and to Wolfgang Ettlinger
6773 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6778 * Return an error when checking OCSP signatures when key is NULL.
6779 This fixes a DoS attack. ([CVE-2013-0166])
6783 * Make openssl verify return errors.
6785 *Chris Palmer <palmer@google.com> and Ben Laurie*
6787 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6788 the right response is stapled. Also change SSL_get_certificate()
6789 so it returns the certificate actually sent.
6790 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6792 *Rob Stradling <rob.stradling@comodo.com>*
6794 * Fix possible deadlock when decoding public keys.
6798 * Don't use TLS 1.0 record version number in initial client hello
6803 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6805 * Sanity check record length before skipping explicit IV in TLS
6806 1.2, 1.1 and DTLS to fix DoS attack.
6808 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6809 fuzzing as a service testing platform.
6814 * Initialise tkeylen properly when encrypting CMS messages.
6815 Thanks to Solar Designer of Openwall for reporting this issue.
6819 * In FIPS mode don't try to use composite ciphers as they are not
6824 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6826 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6827 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6828 mean any application compiled against OpenSSL 1.0.0 headers setting
6829 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6830 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6831 0x10000000L Any application which was previously compiled against
6832 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6833 will need to be recompiled as a result. Letting be results in
6834 inability to disable specifically TLS 1.1 and in client context,
6835 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6839 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6840 disable just protocol X, but all protocols above X *if* there are
6841 protocols *below* X still enabled. In more practical terms it means
6842 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6843 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6844 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6849 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6851 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6852 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6853 in CRYPTO_realloc_clean.
6855 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6856 issue and to Adam Langley <agl@chromium.org> for fixing it.
6859 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6861 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6865 * Workarounds for some broken servers that "hang" if a client hello
6866 record length exceeds 255 bytes.
6868 1. Do not use record version number > TLS 1.0 in initial client
6869 hello: some (but not all) hanging servers will now work.
6870 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6871 the number of ciphers sent in the client hello. This should be
6872 set to an even number, such as 50, for example by passing:
6873 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6874 Most broken servers should now work.
6875 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6876 TLS 1.2 client support entirely.
6880 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6884 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6886 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6887 STRING form instead of a DigestInfo.
6891 * The format used for MDC2 RSA signatures is inconsistent between EVP
6892 and the RSA_sign/RSA_verify functions. This was made more apparent when
6893 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6894 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6895 the correct format in RSA_verify so both forms transparently work.
6899 * Some servers which support TLS 1.0 can choke if we initially indicate
6900 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6901 encrypted premaster secret. As a workaround use the maximum permitted
6902 client version in client hello, this should keep such servers happy
6903 and still work with previous versions of OpenSSL.
6907 * Add support for TLS/DTLS heartbeats.
6909 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6911 * Add support for SCTP.
6913 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6915 * Improved PRNG seeding for VOS.
6917 *Paul Green <Paul.Green@stratus.com>*
6919 * Extensive assembler packs updates, most notably:
6921 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6922 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6923 - x86_64: bit-sliced AES implementation;
6924 - ARM: NEON support, contemporary platforms optimizations;
6925 - s390x: z196 support;
6926 - `*`: GHASH and GF(2^m) multiplication implementations;
6930 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6931 (removal of unnecessary code)
6933 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6935 * Add TLS key material exporter from RFC 5705.
6939 * Add DTLS-SRTP negotiation from RFC 5764.
6943 * Add Next Protocol Negotiation,
6944 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6945 disabled with a no-npn flag to config or Configure. Code donated
6948 *Adam Langley <agl@google.com> and Ben Laurie*
6950 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6951 NIST-P256, NIST-P521, with constant-time single point multiplication on
6952 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6953 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6954 Code made available under Apache License version 2.0.
6956 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6957 line to include this in your build of OpenSSL, and run "make depend" (or
6958 "make update"). This enables the following EC_METHODs:
6960 EC_GFp_nistp224_method()
6961 EC_GFp_nistp256_method()
6962 EC_GFp_nistp521_method()
6964 EC_GROUP_new_by_curve_name() will automatically use these (while
6965 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6968 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6970 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6971 all platforms. Move ssize_t definition from e_os.h to the public
6972 header file e_os2.h as it now appears in public header file cms.h
6976 * New -sigopt option to the ca, req and x509 utilities. Additional
6977 signature parameters can be passed using this option and in
6982 * Add RSA PSS signing function. This will generate and set the
6983 appropriate AlgorithmIdentifiers for PSS based on those in the
6984 corresponding EVP_MD_CTX structure. No application support yet.
6988 * Support for companion algorithm specific ASN1 signing routines.
6989 New function ASN1_item_sign_ctx() signs a pre-initialised
6990 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
6991 the appropriate parameters.
6995 * Add new algorithm specific ASN1 verification initialisation function
6996 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
6997 handling will be the same no matter what EVP_PKEY_METHOD is used.
6998 Add a PSS handler to support verification of PSS signatures: checked
6999 against a number of sample certificates.
7003 * Add signature printing for PSS. Add PSS OIDs.
7005 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7007 * Add algorithm specific signature printing. An individual ASN1 method
7008 can now print out signatures instead of the standard hex dump.
7010 More complex signatures (e.g. PSS) can print out more meaningful
7011 information. Include DSA version that prints out the signature
7016 * Password based recipient info support for CMS library: implementing
7021 * Split password based encryption into PBES2 and PBKDF2 functions. This
7022 neatly separates the code into cipher and PBE sections and is required
7023 for some algorithms that split PBES2 into separate pieces (such as
7024 password based CMS).
7028 * Session-handling fixes:
7029 - Fix handling of connections that are resuming with a session ID,
7030 but also support Session Tickets.
7031 - Fix a bug that suppressed issuing of a new ticket if the client
7032 presented a ticket with an expired session.
7033 - Try to set the ticket lifetime hint to something reasonable.
7034 - Make tickets shorter by excluding irrelevant information.
7035 - On the client side, don't ignore renewed tickets.
7037 *Adam Langley, Bodo Moeller (Google)*
7039 * Fix PSK session representation.
7043 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7045 This work was sponsored by Intel.
7049 * Add GCM support to TLS library. Some custom code is needed to split
7050 the IV between the fixed (from PRF) and explicit (from TLS record)
7051 portions. This adds all GCM ciphersuites supported by RFC5288 and
7052 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7053 add a special AESGCM string for GCM only.
7057 * Expand range of ctrls for AES GCM. Permit setting invocation
7058 field on decrypt and retrieval of invocation field only on encrypt.
7062 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7063 As required by RFC5289 these ciphersuites cannot be used if for
7064 versions of TLS earlier than 1.2.
7068 * For FIPS capable OpenSSL interpret a NULL default public key method
7069 as unset and return the appropriate default but do *not* set the default.
7070 This means we can return the appropriate method in applications that
7071 switch between FIPS and non-FIPS modes.
7075 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7076 ENGINE is used then we cannot handle that in the FIPS module so we
7077 keep original code iff non-FIPS operations are allowed.
7081 * Add -attime option to openssl utilities.
7083 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7085 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7089 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7090 FIPS EC methods unconditionally for now.
7094 * New build option no-ec2m to disable characteristic 2 code.
7098 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7099 all cases can be covered as some introduce binary incompatibilities.
7103 * Redirect RSA operations to FIPS module including keygen,
7104 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7108 * Add similar low level API blocking to ciphers.
7112 * Low level digest APIs are not approved in FIPS mode: any attempt
7113 to use these will cause a fatal error. Applications that *really* want
7114 to use them can use the `private_*` version instead.
7118 * Redirect cipher operations to FIPS module for FIPS builds.
7122 * Redirect digest operations to FIPS module for FIPS builds.
7126 * Update build system to add "fips" flag which will link in fipscanister.o
7127 for static and shared library builds embedding a signature if needed.
7131 * Output TLS supported curves in preference order instead of numerical
7132 order. This is currently hardcoded for the highest order curves first.
7133 This should be configurable so applications can judge speed vs strength.
7137 * Add TLS v1.2 server support for client authentication.
7141 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7146 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7147 FIPS modules versions.
7151 * Add TLS v1.2 client side support for client authentication. Keep cache
7152 of handshake records longer as we don't know the hash algorithm to use
7153 until after the certificate request message is received.
7157 * Initial TLS v1.2 client support. Add a default signature algorithms
7158 extension including all the algorithms we support. Parse new signature
7159 format in client key exchange. Relax some ECC signing restrictions for
7160 TLS v1.2 as indicated in RFC5246.
7164 * Add server support for TLS v1.2 signature algorithms extension. Switch
7165 to new signature format when needed using client digest preference.
7166 All server ciphersuites should now work correctly in TLS v1.2. No client
7167 support yet and no support for client certificates.
7171 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7172 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7173 ciphersuites. At present only RSA key exchange ciphersuites work with
7174 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7175 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7176 and version checking.
7180 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7181 with this defined it will not be affected by any changes to ssl internal
7182 structures. Add several utility functions to allow openssl application
7183 to work with OPENSSL_NO_SSL_INTERN defined.
7187 * A long standing patch to add support for SRP from EdelWeb (Peter
7188 Sylvester and Christophe Renou) was integrated.
7189 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7190 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7193 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7197 * Permit abbreviated handshakes when renegotiating using the function
7198 SSL_renegotiate_abbreviated().
7200 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7202 * Add call to ENGINE_register_all_complete() to
7203 ENGINE_load_builtin_engines(), so some implementations get used
7204 automatically instead of needing explicit application support.
7208 * Add support for TLS key exporter as described in RFC5705.
7210 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7212 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7213 a few changes are required:
7215 Add SSL_OP_NO_TLSv1_1 flag.
7216 Add TLSv1_1 methods.
7217 Update version checking logic to handle version 1.1.
7218 Add explicit IV handling (ported from DTLS code).
7219 Add command line options to s_client/s_server.
7226 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7228 * X509_ATTRIBUTE memory leak
7230 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7231 memory. This structure is used by the PKCS#7 and CMS routines so any
7232 application which reads PKCS#7 or CMS data from untrusted sources is
7233 affected. SSL/TLS is not affected.
7235 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7241 * Race condition handling PSK identify hint
7243 If PSK identity hints are received by a multi-threaded client then
7244 the values are wrongly updated in the parent SSL_CTX structure. This can
7245 result in a race condition potentially leading to a double free of the
7251 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7253 * Malformed ECParameters causes infinite loop
7255 When processing an ECParameters structure OpenSSL enters an infinite loop
7256 if the curve specified is over a specially malformed binary polynomial
7259 This can be used to perform denial of service against any
7260 system which processes public keys, certificate requests or
7261 certificates. This includes TLS clients and TLS servers with
7262 client authentication enabled.
7264 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7269 * Exploitable out-of-bounds read in X509_cmp_time
7271 X509_cmp_time does not properly check the length of the ASN1_TIME
7272 string and can read a few bytes out of bounds. In addition,
7273 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7276 An attacker can use this to craft malformed certificates and CRLs of
7277 various sizes and potentially cause a segmentation fault, resulting in
7278 a DoS on applications that verify certificates or CRLs. TLS clients
7279 that verify CRLs are affected. TLS clients and servers with client
7280 authentication enabled may be affected if they use custom verification
7283 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7284 independently by Hanno Böck.
7289 * PKCS7 crash with missing EnvelopedContent
7291 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7292 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7293 with missing content and trigger a NULL pointer dereference on parsing.
7295 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7296 structures from untrusted sources are affected. OpenSSL clients and
7297 servers are not affected.
7299 This issue was reported to OpenSSL by Michal Zalewski (Google).
7304 * CMS verify infinite loop with unknown hash function
7306 When verifying a signedData message the CMS code can enter an infinite loop
7307 if presented with an unknown hash function OID. This can be used to perform
7308 denial of service against any system which verifies signedData messages using
7310 This issue was reported to OpenSSL by Johannes Bauer.
7315 * Race condition handling NewSessionTicket
7317 If a NewSessionTicket is received by a multi-threaded client when attempting to
7318 reuse a previous ticket then a race condition can occur potentially leading to
7319 a double free of the ticket data.
7324 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7326 * Segmentation fault in ASN1_TYPE_cmp fix
7328 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7329 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7330 certificate signature algorithm consistency this can be used to crash any
7331 certificate verification operation and exploited in a DoS attack. Any
7332 application which performs certificate verification is vulnerable including
7333 OpenSSL clients and servers which enable client authentication.
7338 * ASN.1 structure reuse memory corruption fix
7340 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7341 memory corruption via an invalid write. Such reuse is and has been
7342 strongly discouraged and is believed to be rare.
7344 Applications that parse structures containing CHOICE or ANY DEFINED BY
7345 components may be affected. Certificate parsing (d2i_X509 and related
7346 functions) are however not affected. OpenSSL clients and servers are
7352 * PKCS7 NULL pointer dereferences fix
7354 The PKCS#7 parsing code does not handle missing outer ContentInfo
7355 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7356 missing content and trigger a NULL pointer dereference on parsing.
7358 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7359 otherwise parse PKCS#7 structures from untrusted sources are
7360 affected. OpenSSL clients and servers are not affected.
7362 This issue was reported to OpenSSL by Michal Zalewski (Google).
7367 * DoS via reachable assert in SSLv2 servers fix
7369 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7370 servers that both support SSLv2 and enable export cipher suites by sending
7371 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7373 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7374 (OpenSSL development team).
7379 * Use After Free following d2i_ECPrivatekey error fix
7381 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7382 could cause a use after free condition. This, in turn, could cause a double
7383 free in several private key parsing functions (such as d2i_PrivateKey
7384 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7385 for applications that receive EC private keys from untrusted
7386 sources. This scenario is considered rare.
7388 This issue was discovered by the BoringSSL project and fixed in their
7394 * X509_to_X509_REQ NULL pointer deref fix
7396 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7397 the certificate key is invalid. This function is rarely used in practice.
7399 This issue was discovered by Brian Carpenter.
7404 * Removed the export ciphers from the DEFAULT ciphers
7408 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7410 * Build fixes for the Windows and OpenVMS platforms
7412 *Matt Caswell and Richard Levitte*
7414 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7416 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7417 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7418 dereference. This could lead to a Denial Of Service attack. Thanks to
7419 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7424 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7425 dtls1_buffer_record function under certain conditions. In particular this
7426 could occur if an attacker sent repeated DTLS records with the same
7427 sequence number but for the next epoch. The memory leak could be exploited
7428 by an attacker in a Denial of Service attack through memory exhaustion.
7429 Thanks to Chris Mueller for reporting this issue.
7434 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7435 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7436 method would be set to NULL which could later result in a NULL pointer
7437 dereference. Thanks to Frank Schmirler for reporting this issue.
7442 * Abort handshake if server key exchange message is omitted for ephemeral
7445 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7446 reporting this issue.
7451 * Remove non-export ephemeral RSA code on client and server. This code
7452 violated the TLS standard by allowing the use of temporary RSA keys in
7453 non-export ciphersuites and could be used by a server to effectively
7454 downgrade the RSA key length used to a value smaller than the server
7455 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7456 INRIA or reporting this issue.
7461 * Fixed issue where DH client certificates are accepted without verification.
7462 An OpenSSL server will accept a DH certificate for client authentication
7463 without the certificate verify message. This effectively allows a client to
7464 authenticate without the use of a private key. This only affects servers
7465 which trust a client certificate authority which issues certificates
7466 containing DH keys: these are extremely rare and hardly ever encountered.
7467 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7473 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7474 results on some platforms, including x86_64. This bug occurs at random
7475 with a very low probability, and is not known to be exploitable in any
7476 way, though its exact impact is difficult to determine. Thanks to Pieter
7477 Wuille (Blockstream) who reported this issue and also suggested an initial
7478 fix. Further analysis was conducted by the OpenSSL development team and
7479 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7480 the OpenSSL core team.
7485 * Fix various certificate fingerprint issues.
7487 By using non-DER or invalid encodings outside the signed portion of a
7488 certificate the fingerprint can be changed without breaking the signature.
7489 Although no details of the signed portion of the certificate can be changed
7490 this can cause problems with some applications: e.g. those using the
7491 certificate fingerprint for blacklists.
7493 1. Reject signatures with non zero unused bits.
7495 If the BIT STRING containing the signature has non zero unused bits reject
7496 the signature. All current signature algorithms require zero unused bits.
7498 2. Check certificate algorithm consistency.
7500 Check the AlgorithmIdentifier inside TBS matches the one in the
7501 certificate signature. NB: this will result in signature failure
7502 errors for some broken certificates.
7504 Thanks to Konrad Kraszewski from Google for reporting this issue.
7506 3. Check DSA/ECDSA signatures use DER.
7508 Reencode DSA/ECDSA signatures and compare with the original received
7509 signature. Return an error if there is a mismatch.
7511 This will reject various cases including garbage after signature
7512 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7513 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7514 (negative or with leading zeroes).
7516 Further analysis was conducted and fixes were developed by Stephen Henson
7517 of the OpenSSL core team.
7523 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7525 * Session Ticket Memory Leak.
7527 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7528 integrity of that ticket is first verified. In the event of a session
7529 ticket integrity check failing, OpenSSL will fail to free memory
7530 causing a memory leak. By sending a large number of invalid session
7531 tickets an attacker could exploit this issue in a Denial Of Service
7537 * Build option no-ssl3 is incomplete.
7539 When OpenSSL is configured with "no-ssl3" as a build option, servers
7540 could accept and complete a SSL 3.0 handshake, and clients could be
7541 configured to send them.
7544 *Akamai and the OpenSSL team*
7546 * Add support for TLS_FALLBACK_SCSV.
7547 Client applications doing fallback retries should call
7548 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7551 *Adam Langley, Bodo Moeller*
7553 * Add additional DigestInfo checks.
7555 Reencode DigestInto in DER and check against the original when
7556 verifying RSA signature: this will reject any improperly encoded
7557 DigestInfo structures.
7559 Note: this is a precautionary measure and no attacks are currently known.
7563 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7565 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7566 to a denial of service attack. A malicious server can crash the client
7567 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7568 ciphersuite and sending carefully crafted handshake messages.
7570 Thanks to Felix Gröbert (Google) for discovering and researching this
7576 * By sending carefully crafted DTLS packets an attacker could cause openssl
7577 to leak memory. This can be exploited through a Denial of Service attack.
7578 Thanks to Adam Langley for discovering and researching this issue.
7583 * An attacker can force openssl to consume large amounts of memory whilst
7584 processing DTLS handshake messages. This can be exploited through a
7585 Denial of Service attack.
7586 Thanks to Adam Langley for discovering and researching this issue.
7591 * An attacker can force an error condition which causes openssl to crash
7592 whilst processing DTLS packets due to memory being freed twice. This
7593 can be exploited through a Denial of Service attack.
7594 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7600 * If a multithreaded client connects to a malicious server using a resumed
7601 session and the server sends an ec point format extension it could write
7602 up to 255 bytes to freed memory.
7604 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7610 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7611 X509_name_oneline, X509_name_print_ex et al. to leak some information
7612 from the stack. Applications may be affected if they echo pretty printing
7613 output to the attacker.
7615 Thanks to Ivan Fratric (Google) for discovering this issue.
7618 *Emilia Käsper, and Steve Henson*
7620 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7621 for corner cases. (Certain input points at infinity could lead to
7622 bogus results, with non-infinity inputs mapped to infinity too.)
7626 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7628 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7629 handshake can force the use of weak keying material in OpenSSL
7630 SSL/TLS clients and servers.
7632 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7633 researching this issue. ([CVE-2014-0224])
7635 *KIKUCHI Masashi, Steve Henson*
7637 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7638 OpenSSL DTLS client the code can be made to recurse eventually crashing
7641 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7644 *Imre Rad, Steve Henson*
7646 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7647 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7648 client or server. This is potentially exploitable to run arbitrary
7649 code on a vulnerable client or server.
7651 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7653 *Jüri Aedla, Steve Henson*
7655 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7656 are subject to a denial of service attack.
7658 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7659 this issue. ([CVE-2014-3470])
7661 *Felix Gröbert, Ivan Fratric, Steve Henson*
7663 * Harmonize version and its documentation. -f flag is used to display
7666 *mancha <mancha1@zoho.com>*
7668 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7669 in i2d_ECPrivateKey.
7671 *mancha <mancha1@zoho.com>*
7673 * Fix some double frees. These are not thought to be exploitable.
7675 *mancha <mancha1@zoho.com>*
7677 * Fix for the attack described in the paper "Recovering OpenSSL
7678 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7679 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7680 <http://eprint.iacr.org/2014/140>
7682 Thanks to Yuval Yarom and Naomi Benger for discovering this
7683 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7685 *Yuval Yarom and Naomi Benger*
7687 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7689 * Keep original DTLS digest and encryption contexts in retransmission
7690 structures so we can use the previous session parameters if they need
7691 to be resent. ([CVE-2013-6450])
7695 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7696 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7697 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7698 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7699 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7700 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7702 *Rob Stradling, Adam Langley*
7704 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7706 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7708 This addresses the flaw in CBC record processing discovered by
7709 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7710 at: <http://www.isg.rhul.ac.uk/tls/>
7712 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7713 Security Group at Royal Holloway, University of London
7714 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7715 Emilia Käsper for the initial patch.
7718 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7720 * Return an error when checking OCSP signatures when key is NULL.
7721 This fixes a DoS attack. ([CVE-2013-0166])
7725 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7726 the right response is stapled. Also change SSL_get_certificate()
7727 so it returns the certificate actually sent.
7728 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7729 (This is a backport)
7731 *Rob Stradling <rob.stradling@comodo.com>*
7733 * Fix possible deadlock when decoding public keys.
7737 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7739 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7742 * Sanity check record length before skipping explicit IV in DTLS
7745 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7746 fuzzing as a service testing platform.
7751 * Initialise tkeylen properly when encrypting CMS messages.
7752 Thanks to Solar Designer of Openwall for reporting this issue.
7756 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7758 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7759 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7760 in CRYPTO_realloc_clean.
7762 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7763 issue and to Adam Langley <agl@chromium.org> for fixing it.
7766 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7768 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7770 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7771 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7772 content decryption and always return the same error. Note: this attack
7773 needs on average 2^20 messages so it only affects automated senders. The
7774 old behaviour can be re-enabled in the CMS code by setting the
7775 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7776 an MMA defence is not necessary.
7777 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7778 this issue. ([CVE-2012-0884])
7782 * Fix CVE-2011-4619: make sure we really are receiving a
7783 client hello before rejecting multiple SGC restarts. Thanks to
7784 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7788 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7790 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7791 Thanks to Antonio Martin, Enterprise Secure Access Research and
7792 Development, Cisco Systems, Inc. for discovering this bug and
7793 preparing a fix. ([CVE-2012-0050])
7797 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7799 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7800 of the Vaudenay padding oracle attack on CBC mode encryption
7801 which enables an efficient plaintext recovery attack against
7802 the OpenSSL implementation of DTLS. Their attack exploits timing
7803 differences arising during decryption processing. A research
7804 paper describing this attack can be found at:
7805 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7806 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7807 Security Group at Royal Holloway, University of London
7808 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7809 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7810 for preparing the fix. ([CVE-2011-4108])
7812 *Robin Seggelmann, Michael Tuexen*
7814 * Clear bytes used for block padding of SSL 3.0 records.
7817 *Adam Langley (Google)*
7819 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7820 Kadianakis <desnacked@gmail.com> for discovering this issue and
7821 Adam Langley for preparing the fix. ([CVE-2011-4619])
7823 *Adam Langley (Google)*
7825 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7827 *Andrey Kulikov <amdeich@gmail.com>*
7829 * Prevent malformed RFC3779 data triggering an assertion failure.
7830 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7831 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7833 *Rob Austein <sra@hactrn.net>*
7835 * Improved PRNG seeding for VOS.
7837 *Paul Green <Paul.Green@stratus.com>*
7839 * Fix ssl_ciph.c set-up race.
7841 *Adam Langley (Google)*
7843 * Fix spurious failures in ecdsatest.c.
7845 *Emilia Käsper (Google)*
7847 * Fix the BIO_f_buffer() implementation (which was mixing different
7848 interpretations of the `..._len` fields).
7850 *Adam Langley (Google)*
7852 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7853 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7854 threads won't reuse the same blinding coefficients.
7856 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7857 lock to call BN_BLINDING_invert_ex, and avoids one use of
7858 BN_BLINDING_update for each BN_BLINDING structure (previously,
7859 the last update always remained unused).
7861 *Emilia Käsper (Google)*
7863 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7865 *Bob Buckholz (Google)*
7867 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7869 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7870 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7872 *Kaspar Brand <ossl@velox.ch>*
7874 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7875 for multi-threaded use of ECDH. ([CVE-2011-3210])
7877 *Adam Langley (Google)*
7879 * Fix x509_name_ex_d2i memory leak on bad inputs.
7883 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7884 signature public key algorithm by using OID xref utilities instead.
7885 Before this you could only use some ECC ciphersuites with SHA1 only.
7889 * Add protection against ECDSA timing attacks as mentioned in the paper
7890 by Billy Bob Brumley and Nicola Tuveri, see:
7891 <http://eprint.iacr.org/2011/232.pdf>
7893 *Billy Bob Brumley and Nicola Tuveri*
7895 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7897 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7899 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7901 * Fix bug in string printing code: if *any* escaping is enabled we must
7902 escape the escape character (backslash) or the resulting string is
7907 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7909 * Disable code workaround for ancient and obsolete Netscape browsers
7910 and servers: an attacker can use it in a ciphersuite downgrade attack.
7911 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7915 * Fixed J-PAKE implementation error, originally discovered by
7916 Sebastien Martini, further info and confirmation from Stefan
7917 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7921 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7923 * Fix extension code to avoid race conditions which can result in a buffer
7924 overrun vulnerability: resumed sessions must not be modified as they can
7925 be shared by multiple threads. CVE-2010-3864
7929 * Fix WIN32 build system to correctly link an ENGINE directory into
7934 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7936 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7939 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7941 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7943 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7944 context. The operation can be customised via the ctrl mechanism in
7945 case ENGINEs want to include additional functionality.
7949 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7953 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7954 output hashes compatible with older versions of OpenSSL.
7956 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7958 * Fix compression algorithm handling: if resuming a session use the
7959 compression algorithm of the resumed session instead of determining
7960 it from client hello again. Don't allow server to change algorithm.
7964 * Add load_crls() function to apps tidying load_certs() too. Add option
7965 to verify utility to allow additional CRLs to be included.
7969 * Update OCSP request code to permit adding custom headers to the request:
7970 some responders need this.
7974 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7977 *Julia Lawall <julia@diku.dk>*
7979 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7980 needlessly dereferenced structures, used obsolete functions and
7981 didn't handle all updated verify codes correctly.
7985 * Disable MD2 in the default configuration.
7989 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
7990 indicate the initial BIO being pushed or popped. This makes it possible
7991 to determine whether the BIO is the one explicitly called or as a result
7992 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
7993 it handles reference counts correctly and doesn't zero out the I/O bio
7994 when it is not being explicitly popped. WARNING: applications which
7995 included workarounds for the old buggy behaviour will need to be modified
7996 or they could free up already freed BIOs.
8000 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8001 renaming to all platforms (within the 0.9.8 branch, this was
8002 done conditionally on Netware platforms to avoid a name clash).
8004 *Guenter <lists@gknw.net>*
8006 * Add ECDHE and PSK support to DTLS.
8008 *Michael Tuexen <tuexen@fh-muenster.de>*
8010 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8015 * Add "missing" function EVP_MD_flags() (without this the only way to
8016 retrieve a digest flags is by accessing the structure directly. Update
8017 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8018 or cipher is registered as in the "from" argument. Print out all
8019 registered digests in the dgst usage message instead of manually
8020 attempting to work them out.
8024 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8025 this allows the use of compression and extensions. Change default cipher
8026 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8027 by default unless an application cipher string requests it.
8031 * Alter match criteria in PKCS12_parse(). It used to try to use local
8032 key ids to find matching certificates and keys but some PKCS#12 files
8033 don't follow the (somewhat unwritten) rules and this strategy fails.
8034 Now just gather all certificates together and the first private key
8035 then look for the first certificate that matches the key.
8039 * Support use of registered digest and cipher names for dgst and cipher
8040 commands instead of having to add each one as a special case. So now
8047 openssl dgst -sha256 foo
8049 and this works for ENGINE based algorithms too.
8053 * Update Gost ENGINE to support parameter files.
8055 *Victor B. Wagner <vitus@cryptocom.ru>*
8057 * Support GeneralizedTime in ca utility.
8059 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8061 * Enhance the hash format used for certificate directory links. The new
8062 form uses the canonical encoding (meaning equivalent names will work
8063 even if they aren't identical) and uses SHA1 instead of MD5. This form
8064 is incompatible with the older format and as a result c_rehash should
8065 be used to rebuild symbolic links.
8069 * Make PKCS#8 the default write format for private keys, replacing the
8070 traditional format. This form is standardised, more secure and doesn't
8071 include an implicit MD5 dependency.
8075 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8076 committed to OpenSSL should pass this lot as a minimum.
8080 * Add session ticket override functionality for use by EAP-FAST.
8082 *Jouni Malinen <j@w1.fi>*
8084 * Modify HMAC functions to return a value. Since these can be implemented
8085 in an ENGINE errors can occur.
8089 * Type-checked OBJ_bsearch_ex.
8093 * Type-checked OBJ_bsearch. Also some constification necessitated
8094 by type-checking. Still to come: TXT_DB, bsearch(?),
8095 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8100 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8101 seconds to a tm structure directly, instead of going through OS
8102 specific date routines. This avoids any issues with OS routines such
8103 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8104 and X509_time_adj_ex() to cover the extended range. The existing
8105 X509_time_adj() is still usable and will no longer have any date issues.
8109 * Delta CRL support. New use deltas option which will attempt to locate
8110 and search any appropriate delta CRLs available.
8112 This work was sponsored by Google.
8116 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8117 code and add additional score elements. Validate alternate CRL paths
8118 as part of the CRL checking and indicate a new error "CRL path validation
8119 error" in this case. Applications wanting additional details can use
8120 the verify callback and check the new "parent" field. If this is not
8121 NULL CRL path validation is taking place. Existing applications won't
8122 see this because it requires extended CRL support which is off by
8125 This work was sponsored by Google.
8129 * Support for freshest CRL extension.
8131 This work was sponsored by Google.
8135 * Initial indirect CRL support. Currently only supported in the CRLs
8136 passed directly and not via lookup. Process certificate issuer
8137 CRL entry extension and lookup CRL entries by bother issuer name
8138 and serial number. Check and process CRL issuer entry in IDP extension.
8140 This work was sponsored by Google.
8144 * Add support for distinct certificate and CRL paths. The CRL issuer
8145 certificate is validated separately in this case. Only enabled if
8146 an extended CRL support flag is set: this flag will enable additional
8147 CRL functionality in future.
8149 This work was sponsored by Google.
8153 * Add support for policy mappings extension.
8155 This work was sponsored by Google.
8159 * Fixes to pathlength constraint, self issued certificate handling,
8160 policy processing to align with RFC3280 and PKITS tests.
8162 This work was sponsored by Google.
8166 * Support for name constraints certificate extension. DN, email, DNS
8167 and URI types are currently supported.
8169 This work was sponsored by Google.
8173 * To cater for systems that provide a pointer-based thread ID rather
8174 than numeric, deprecate the current numeric thread ID mechanism and
8175 replace it with a structure and associated callback type. This
8176 mechanism allows a numeric "hash" to be extracted from a thread ID in
8177 either case, and on platforms where pointers are larger than 'long',
8178 mixing is done to help ensure the numeric 'hash' is usable even if it
8179 can't be guaranteed unique. The default mechanism is to use "&errno"
8180 as a pointer-based thread ID to distinguish between threads.
8182 Applications that want to provide their own thread IDs should now use
8183 CRYPTO_THREADID_set_callback() to register a callback that will call
8184 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8186 Note that ERR_remove_state() is now deprecated, because it is tied
8187 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8188 to free the current thread's error state should be replaced by
8189 ERR_remove_thread_state(NULL).
8191 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8192 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8193 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8194 application was previously providing a numeric thread callback that
8195 was inappropriate for distinguishing threads, then uniqueness might
8196 have been obtained with &errno that happened immediately in the
8197 intermediate development versions of OpenSSL; this is no longer the
8198 case, the numeric thread callback will now override the automatic use
8201 *Geoff Thorpe, with help from Bodo Moeller*
8203 * Initial support for different CRL issuing certificates. This covers a
8204 simple case where the self issued certificates in the chain exist and
8205 the real CRL issuer is higher in the existing chain.
8207 This work was sponsored by Google.
8211 * Removed effectively defunct crypto/store from the build.
8215 * Revamp of STACK to provide stronger type-checking. Still to come:
8216 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8217 ASN1_STRING, CONF_VALUE.
8221 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8222 RAM on SSL connections. This option can save about 34k per idle SSL.
8226 * Revamp of LHASH to provide stronger type-checking. Still to come:
8227 STACK, TXT_DB, bsearch, qsort.
8231 * Initial support for Cryptographic Message Syntax (aka CMS) based
8232 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8233 support for data, signedData, compressedData, digestedData and
8234 encryptedData, envelopedData types included. Scripts to check against
8235 RFC4134 examples draft and interop and consistency checks of many
8236 content types and variants.
8240 * Add options to enc utility to support use of zlib compression BIO.
8244 * Extend mk1mf to support importing of options and assembly language
8245 files from Configure script, currently only included in VC-WIN32.
8246 The assembly language rules can now optionally generate the source
8247 files from the associated perl scripts.
8251 * Implement remaining functionality needed to support GOST ciphersuites.
8252 Interop testing has been performed using CryptoPro implementations.
8254 *Victor B. Wagner <vitus@cryptocom.ru>*
8256 * s390x assembler pack.
8260 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8265 * Implement Opaque PRF Input TLS extension as specified in
8266 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8267 official specification yet and no extension type assignment by
8268 IANA exists, this extension (for now) will have to be explicitly
8269 enabled when building OpenSSL by providing the extension number
8270 to use. For example, specify an option
8272 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8274 to the "config" or "Configure" script to enable the extension,
8275 assuming extension number 0x9527 (which is a completely arbitrary
8276 and unofficial assignment based on the MD5 hash of the Internet
8277 Draft). Note that by doing so, you potentially lose
8278 interoperability with other TLS implementations since these might
8279 be using the same extension number for other purposes.
8281 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8282 opaque PRF input value to use in the handshake. This will create
8283 an internal copy of the length-'len' string at 'src', and will
8284 return non-zero for success.
8286 To get more control and flexibility, provide a callback function
8289 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8290 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8294 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8297 Callback function 'cb' will be called in handshakes, and is
8298 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8299 Argument 'arg' is for application purposes (the value as given to
8300 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8301 be provided to the callback function). The callback function
8302 has to return non-zero to report success: usually 1 to use opaque
8303 PRF input just if possible, or 2 to enforce use of the opaque PRF
8304 input. In the latter case, the library will abort the handshake
8305 if opaque PRF input is not successfully negotiated.
8307 Arguments 'peerinput' and 'len' given to the callback function
8308 will always be NULL and 0 in the case of a client. A server will
8309 see the client's opaque PRF input through these variables if
8310 available (NULL and 0 otherwise). Note that if the server
8311 provides an opaque PRF input, the length must be the same as the
8312 length of the client's opaque PRF input.
8314 Note that the callback function will only be called when creating
8315 a new session (session resumption can resume whatever was
8316 previously negotiated), and will not be called in SSL 2.0
8317 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8318 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8319 for applications that need to enforce opaque PRF input.
8323 * Update ssl code to support digests other than SHA1+MD5 for handshake
8326 *Victor B. Wagner <vitus@cryptocom.ru>*
8328 * Add RFC4507 support to OpenSSL. This includes the corrections in
8329 RFC4507bis. The encrypted ticket format is an encrypted encoded
8330 SSL_SESSION structure, that way new session features are automatically
8333 If a client application caches session in an SSL_SESSION structure
8334 support is transparent because tickets are now stored in the encoded
8337 The SSL_CTX structure automatically generates keys for ticket
8338 protection in servers so again support should be possible
8339 with no application modification.
8341 If a client or server wishes to disable RFC4507 support then the option
8342 SSL_OP_NO_TICKET can be set.
8344 Add a TLS extension debugging callback to allow the contents of any client
8345 or server extensions to be examined.
8347 This work was sponsored by Google.
8351 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8352 OpenSSL should now compile cleanly on gcc 4.2
8354 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8356 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8357 support including streaming MAC support: this is required for GOST
8358 ciphersuite support.
8360 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8362 * Add option -stream to use PKCS#7 streaming in smime utility. New
8363 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8364 to output in BER and PEM format.
8368 * Experimental support for use of HMAC via EVP_PKEY interface. This
8369 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8370 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8371 ENGINE support for HMAC keys which are unextractable. New -mac and
8372 -macopt options to dgst utility.
8376 * New option -sigopt to dgst utility. Update dgst to use
8377 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8378 alternative signing parameters such as X9.31 or PSS in the dgst
8383 * Change ssl_cipher_apply_rule(), the internal function that does
8384 the work each time a ciphersuite string requests enabling
8385 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8386 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8387 the order of disabled ciphersuites such that those ciphersuites
8388 that most recently went from enabled to disabled not only stay
8389 in order with respect to each other, but also have higher priority
8390 than other disabled ciphersuites the next time ciphersuites are
8393 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8394 the same ciphersuites as with "HIGH" alone, but in a specific
8395 order where the PSK ciphersuites come first (since they are the
8396 most recently disabled ciphersuites when "HIGH" is parsed).
8398 Also, change ssl_create_cipher_list() (using this new
8399 functionality) such that between otherwise identical
8400 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8405 * Change ssl_create_cipher_list() so that it automatically
8406 arranges the ciphersuites in reasonable order before starting
8407 to process the rule string. Thus, the definition for "DEFAULT"
8408 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8409 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8410 This makes it much easier to arrive at a reasonable default order
8411 in applications for which anonymous ciphers are OK (meaning
8412 that you can't actually use DEFAULT).
8414 *Bodo Moeller; suggested by Victor Duchovni*
8416 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8417 processing) into multiple integers instead of setting
8418 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8419 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8420 (These masks as well as the individual bit definitions are hidden
8421 away into the non-exported interface ssl/ssl_locl.h, so this
8422 change to the definition of the SSL_CIPHER structure shouldn't
8423 affect applications.) This give us more bits for each of these
8424 categories, so there is no longer a need to coagulate AES128 and
8425 AES256 into a single algorithm bit, and to coagulate Camellia128
8426 and Camellia256 into a single algorithm bit, which has led to all
8429 Thus, among other things, the kludge introduced in 0.9.7m and
8430 0.9.8e for masking out AES256 independently of AES128 or masking
8431 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8433 With the change, we also introduce new ciphersuite aliases that
8434 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8439 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8440 Use the leftmost N bytes of the signature input if the input is
8441 larger than the prime q (with N being the size in bytes of q).
8445 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8446 it yet and it is largely untested.
8450 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8454 * Initial incomplete changes to avoid need for function casts in OpenSSL
8455 some compilers (gcc 4.2 and later) reject their use. Safestack is
8456 reimplemented. Update ASN1 to avoid use of legacy functions.
8460 * Win32/64 targets are linked with Winsock2.
8464 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8465 to external functions. This can be used to increase CRL handling
8466 efficiency especially when CRLs are very large by (for example) storing
8467 the CRL revoked certificates in a database.
8471 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8472 new CRLs added to a directory can be used. New command line option
8473 -verify_return_error to s_client and s_server. This causes real errors
8474 to be returned by the verify callback instead of carrying on no matter
8475 what. This reflects the way a "real world" verify callback would behave.
8479 * GOST engine, supporting several GOST algorithms and public key formats.
8480 Kindly donated by Cryptocom.
8484 * Partial support for Issuing Distribution Point CRL extension. CRLs
8485 partitioned by DP are handled but no indirect CRL or reason partitioning
8486 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8487 selected via a scoring technique which handles IDP and AKID in CRLs.
8491 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8492 will ultimately be used for all verify operations: this will remove the
8493 X509_STORE dependency on certificate verification and allow alternative
8494 lookup methods. X509_STORE based implementations of these two callbacks.
8498 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8499 Modify get_crl() to find a valid (unexpired) CRL if possible.
8503 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8504 this would be called X509_CRL_cmp() but that name is already used by
8505 a function that just compares CRL issuer names. Cache several CRL
8506 extensions in X509_CRL structure and cache CRLDP in X509.
8510 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8511 this maps equivalent X509_NAME structures into a consistent structure.
8512 Name comparison can then be performed rapidly using memcmp().
8516 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8521 * Allow digests to supply their own micalg string for S/MIME type using
8522 the ctrl EVP_MD_CTRL_MICALG.
8526 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8527 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8528 ctrl. It can then customise the structure before and/or after signing
8533 * New function OBJ_add_sigid() to allow application defined signature OIDs
8534 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8535 to free up any added signature OIDs.
8539 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8540 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8541 digest and cipher tables. New options added to openssl utility:
8542 list-message-digest-algorithms and list-cipher-algorithms.
8546 * Change the array representation of binary polynomials: the list
8547 of degrees of non-zero coefficients is now terminated with -1.
8548 Previously it was terminated with 0, which was also part of the
8549 value; thus, the array representation was not applicable to
8550 polynomials where t^0 has coefficient zero. This change makes
8551 the array representation useful in a more general context.
8555 * Various modifications and fixes to SSL/TLS cipher string
8556 handling. For ECC, the code now distinguishes between fixed ECDH
8557 with RSA certificates on the one hand and with ECDSA certificates
8558 on the other hand, since these are separate ciphersuites. The
8559 unused code for Fortezza ciphersuites has been removed.
8561 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8562 (not "ECDHE"). For consistency with the code for DH
8563 certificates, use of ECDH certificates is now considered ECDH
8564 authentication, not RSA or ECDSA authentication (the latter is
8565 merely the CA's signing algorithm and not actively used in the
8568 The temporary ciphersuite alias "ECCdraft" is no longer
8569 available, and ECC ciphersuites are no longer excluded from "ALL"
8570 and "DEFAULT". The following aliases now exist for RFC 4492
8571 ciphersuites, most of these by analogy with the DH case:
8573 kECDHr - ECDH cert, signed with RSA
8574 kECDHe - ECDH cert, signed with ECDSA
8575 kECDH - ECDH cert (signed with either RSA or ECDSA)
8576 kEECDH - ephemeral ECDH
8577 ECDH - ECDH cert or ephemeral ECDH
8583 AECDH - anonymous ECDH
8584 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8588 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8589 Use correct micalg parameters depending on digest(s) in signed message.
8593 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8594 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8598 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8599 an engine to register a method. Add ENGINE lookups for methods and
8600 functional reference processing.
8604 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8605 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8610 * New -resign option to smime utility. This adds one or more signers
8611 to an existing PKCS#7 signedData structure. Also -md option to use an
8612 alternative message digest algorithm for signing.
8616 * Tidy up PKCS#7 routines and add new functions to make it easier to
8617 create PKCS7 structures containing multiple signers. Update smime
8618 application to support multiple signers.
8622 * New -macalg option to pkcs12 utility to allow setting of an alternative
8627 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8628 Reorganize PBE internals to lookup from a static table using NIDs,
8629 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8630 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8631 PRF which will be automatically used with PBES2.
8635 * Replace the algorithm specific calls to generate keys in "req" with the
8640 * Update PKCS#7 enveloped data routines to use new API. This is now
8641 supported by any public key method supporting the encrypt operation. A
8642 ctrl is added to allow the public key algorithm to examine or modify
8643 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8648 * Add a ctrl to asn1 method to allow a public key algorithm to express
8649 a default digest type to use. In most cases this will be SHA1 but some
8650 algorithms (such as GOST) need to specify an alternative digest. The
8651 return value indicates how strong the preference is 1 means optional and
8652 2 is mandatory (that is it is the only supported type). Modify
8653 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8654 use the default md. Update openssl utilities to use the default digest
8655 type for signing if it is not explicitly indicated.
8659 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8660 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8661 signing method from the key type. This effectively removes the link
8662 between digests and public key types.
8666 * Add an OID cross reference table and utility functions. Its purpose is to
8667 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8668 rsaEncryption. This will allow some of the algorithm specific hackery
8669 needed to use the correct OID to be removed.
8673 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8674 structures for PKCS7_sign(). They are now set up by the relevant public
8679 * Add provisional EC pkey method with support for ECDSA and ECDH.
8683 * Add support for key derivation (agreement) in the API, DH method and
8688 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8689 public and private key formats. As a side effect these add additional
8690 command line functionality not previously available: DSA signatures can be
8691 generated and verified using pkeyutl and DH key support and generation in
8698 *Oliver Tappe <zooey@hirschkaefer.de>*
8700 * New make target "install_html_docs" installs HTML renditions of the
8703 *Oliver Tappe <zooey@hirschkaefer.de>*
8705 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8706 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8707 support key and parameter generation and add initial key generation
8708 functionality for RSA.
8712 * Add functions for main EVP_PKEY_method operations. The undocumented
8713 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8714 `EVP_PKEY_{encrypt,decrypt}_old`.
8718 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8719 key API, doesn't do much yet.
8723 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8724 public key algorithms. New option to openssl utility:
8725 "list-public-key-algorithms" to print out info.
8729 * Implement the Supported Elliptic Curves Extension for
8730 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8734 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8735 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8739 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8740 utilities such as rsa, dsa, dsaparam etc except they process any key
8745 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8746 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8747 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8752 * Initial support for pluggable public key ASN1.
8753 De-spaghettify the public key ASN1 handling. Move public and private
8754 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8755 algorithm specific handling to a single module within the relevant
8756 algorithm directory. Add functions to allow (near) opaque processing
8757 of public and private key structures.
8761 * Implement the Supported Point Formats Extension for
8762 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8766 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8767 for the psk identity [hint] and the psk callback functions to the
8768 SSL_SESSION, SSL and SSL_CTX structure.
8771 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8775 SSL_CTX_use_psk_identity_hint
8776 SSL_get_psk_identity_hint
8777 SSL_get_psk_identity
8778 SSL_use_psk_identity_hint
8780 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8782 * Add RFC 3161 compliant time stamp request creation, response generation
8783 and response verification functionality.
8785 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8787 * Add initial support for TLS extensions, specifically for the server_name
8788 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8789 have new members for a host name. The SSL data structure has an
8790 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8791 stored in that context to allow for session resumption, even after the
8792 SSL has been switched to a new SSL_CTX in reaction to a client's
8793 server_name extension.
8795 New functions (subject to change):
8797 SSL_get_servername()
8798 SSL_get_servername_type()
8801 New CTRL codes and macros (subject to change):
8803 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8804 - SSL_CTX_set_tlsext_servername_callback()
8805 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8806 - SSL_CTX_set_tlsext_servername_arg()
8807 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8809 openssl s_client has a new '-servername ...' option.
8811 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8812 '-key2 ...', '-servername_fatal' (subject to change). This allows
8813 testing the HostName extension for a specific single host name ('-cert'
8814 and '-key' remain fallbacks for handshakes without HostName
8815 negotiation). If the unrecognized_name alert has to be sent, this by
8816 default is a warning; it becomes fatal with the '-servername_fatal'
8819 *Peter Sylvester, Remy Allais, Christophe Renou*
8821 * Whirlpool hash implementation is added.
8825 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8826 bn(64,32). Because of instruction set limitations it doesn't have
8827 any negative impact on performance. This was done mostly in order
8828 to make it possible to share assembler modules, such as bn_mul_mont
8829 implementations, between 32- and 64-bit builds without hassle.
8833 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8834 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8839 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8840 dedicated Montgomery multiplication procedure, is introduced.
8841 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8842 "64-bit" performance on certain 32-bit targets.
8846 * New option SSL_OP_NO_COMP to disable use of compression selectively
8847 in SSL structures. New SSL ctrl to set maximum send fragment size.
8848 Save memory by setting the I/O buffer sizes dynamically instead of
8849 using the maximum available value.
8853 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8854 in addition to the text details.
8858 * Very, very preliminary EXPERIMENTAL support for printing of general
8859 ASN1 structures. This currently produces rather ugly output and doesn't
8860 handle several customised structures at all.
8864 * Integrated support for PVK file format and some related formats such
8865 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8866 these in the 'rsa' and 'dsa' utilities.
8870 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8874 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8875 place for the (very old) "NETSCAPE" format certificates which are now
8876 handled using new ASN1 code equivalents.
8880 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8881 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8882 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8886 * Modify CRL distribution points extension code to print out previously
8887 unsupported fields. Enhance extension setting code to allow setting of
8892 * Add print and set support for Issuing Distribution Point CRL extension.
8896 * Change 'Configure' script to enable Camellia by default.
8903 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8905 * When rejecting SSL/TLS records due to an incorrect version number, never
8906 update s->server with a new major version number. As of
8907 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8908 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8909 the previous behavior could result in a read attempt at NULL when
8910 receiving specific incorrect SSL/TLS records once record payload
8911 protection is active. ([CVE-2010-0740])
8913 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8915 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8916 could be crashed if the relevant tables were not present (e.g. chrooted).
8918 *Tomas Hoger <thoger@redhat.com>*
8920 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8922 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8924 *Martin Olsson, Neel Mehta*
8926 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8927 accommodate for stack sorting, always a write lock!).
8931 * On some versions of WIN32 Heap32Next is very slow. This can cause
8932 excessive delays in the RAND_poll(): over a minute. As a workaround
8933 include a time check in the inner Heap32Next loop too.
8937 * The code that handled flushing of data in SSL/TLS originally used the
8938 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8939 the problem outlined in PR#1949. The fix suggested there however can
8940 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8941 of Apache). So instead simplify the code to flush unconditionally.
8942 This should be fine since flushing with no data to flush is a no op.
8946 * Handle TLS versions 2.0 and later properly and correctly use the
8947 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8948 off ancient servers have a habit of sticking around for a while...
8952 * Modify compression code so it frees up structures without using the
8953 ex_data callbacks. This works around a problem where some applications
8954 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8955 restarting) then use compression (e.g. SSL with compression) later.
8956 This results in significant per-connection memory leaks and
8957 has caused some security issues including CVE-2008-1678 and
8962 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8963 change when encrypting or decrypting.
8967 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8968 connect and renegotiate with servers which do not support RI.
8969 Until RI is more widely deployed this option is enabled by default.
8973 * Add "missing" ssl ctrls to clear options and mode.
8977 * If client attempts to renegotiate and doesn't support RI respond with
8978 a no_renegotiation alert as required by RFC5746. Some renegotiating
8979 TLS clients will continue a connection gracefully when they receive
8980 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8981 waiting for a server hello which it will never receive. Now we treat a
8982 received no_renegotiation alert as a fatal error. This is because
8983 applications requesting a renegotiation might well expect it to succeed
8984 and would have no code in place to handle the server denying it so the
8985 only safe thing to do is to terminate the connection.
8989 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
8990 peer supports secure renegotiation and 0 otherwise. Print out peer
8991 renegotiation support in s_client/s_server.
8995 * Replace the highly broken and deprecated SPKAC certification method with
8996 the updated NID creation version. This should correctly handle UTF8.
9000 * Implement RFC5746. Re-enable renegotiation but require the extension
9001 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9002 turns out to be a bad idea. It has been replaced by
9003 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9004 SSL_CTX_set_options(). This is really not recommended unless you
9005 know what you are doing.
9007 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9009 * Fixes to stateless session resumption handling. Use initial_ctx when
9010 issuing and attempting to decrypt tickets in case it has changed during
9011 servername handling. Use a non-zero length session ID when attempting
9012 stateless session resumption: this makes it possible to determine if
9013 a resumption has occurred immediately after receiving server hello
9014 (several places in OpenSSL subtly assume this) instead of later in
9019 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9020 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9021 fixes for a few places where the return code is not checked
9024 *Julia Lawall <julia@diku.dk>*
9026 * Add --strict-warnings option to Configure script to include devteam
9027 warnings in other configurations.
9031 * Add support for --libdir option and LIBDIR variable in makefiles. This
9032 makes it possible to install openssl libraries in locations which
9033 have names other than "lib", for example "/usr/lib64" which some
9036 *Steve Henson, based on patch from Jeremy Utley*
9038 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9039 X690 8.9.12 and can produce some misleading textual output of OIDs.
9041 *Steve Henson, reported by Dan Kaminsky*
9043 * Delete MD2 from algorithm tables. This follows the recommendation in
9044 several standards that it is not used in new applications due to
9045 several cryptographic weaknesses. For binary compatibility reasons
9046 the MD2 API is still compiled in by default.
9050 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9055 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9056 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9059 *Guenter <lists@gknw.net>*
9061 * Fix the server certificate chain building code to use X509_verify_cert(),
9062 it used to have an ad-hoc builder which was unable to cope with anything
9063 other than a simple chain.
9065 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9067 * Don't check self signed certificate signatures in X509_verify_cert()
9068 by default (a flag can override this): it just wastes time without
9069 adding any security. As a useful side effect self signed root CAs
9070 with non-FIPS digests are now usable in FIPS mode.
9074 * In dtls1_process_out_of_seq_message() the check if the current message
9075 is already buffered was missing. For every new message was memory
9076 allocated, allowing an attacker to perform an denial of service attack
9077 with sending out of seq handshake messages until there is no memory
9078 left. Additionally every future message was buffered, even if the
9079 sequence number made no sense and would be part of another handshake.
9080 So only messages with sequence numbers less than 10 in advance will be
9081 buffered. ([CVE-2009-1378])
9083 *Robin Seggelmann, discovered by Daniel Mentz*
9085 * Records are buffered if they arrive with a future epoch to be
9086 processed after finishing the corresponding handshake. There is
9087 currently no limitation to this buffer allowing an attacker to perform
9088 a DOS attack with sending records with future epochs until there is no
9089 memory left. This patch adds the pqueue_size() function to determine
9090 the size of a buffer and limits the record buffer to 100 entries.
9093 *Robin Seggelmann, discovered by Daniel Mentz*
9095 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9096 parent structure is freed. ([CVE-2009-1379])
9100 * Handle non-blocking I/O properly in SSL_shutdown() call.
9102 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9104 * Add `2.5.4.*` OIDs
9106 *Ilya O. <vrghost@gmail.com>*
9108 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9110 * Disable renegotiation completely - this fixes a severe security
9111 problem ([CVE-2009-3555]) at the cost of breaking all
9112 renegotiation. Renegotiation can be re-enabled by setting
9113 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9114 run-time. This is really not recommended unless you know what
9119 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9121 * Don't set val to NULL when freeing up structures, it is freed up by
9122 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9123 zeroing past the valid field. ([CVE-2009-0789])
9125 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9127 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9128 checked correctly. This would allow some invalid signed attributes to
9129 appear to verify correctly. ([CVE-2009-0591])
9131 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9133 * Reject UniversalString and BMPString types with invalid lengths. This
9134 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9135 a legal length. ([CVE-2009-0590])
9139 * Set S/MIME signing as the default purpose rather than setting it
9140 unconditionally. This allows applications to override it at the store
9145 * Permit restricted recursion of ASN1 strings. This is needed in practice
9146 to handle some structures.
9150 * Improve efficiency of mem_gets: don't search whole buffer each time
9153 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9155 * New -hex option for openssl rand.
9159 * Print out UTF8String and NumericString when parsing ASN1.
9163 * Support NumericString type for name components.
9167 * Allow CC in the environment to override the automatically chosen
9168 compiler. Note that nothing is done to ensure flags work with the
9173 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9175 * Properly check EVP_VerifyFinal() and similar return values
9178 *Ben Laurie, Bodo Moeller, Google Security Team*
9180 * Enable TLS extensions by default.
9184 * Allow the CHIL engine to be loaded, whether the application is
9185 multithreaded or not. (This does not release the developer from the
9186 obligation to set up the dynamic locking callbacks.)
9188 *Sander Temme <sander@temme.net>*
9190 * Use correct exit code if there is an error in dgst command.
9192 *Steve Henson; problem pointed out by Roland Dirlewanger*
9194 * Tweak Configure so that you need to say "experimental-jpake" to enable
9195 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9199 * Add experimental JPAKE support, including demo authentication in
9200 s_client and s_server.
9204 * Set the comparison function in v3_addr_canonize().
9206 *Rob Austein <sra@hactrn.net>*
9208 * Add support for XMPP STARTTLS in s_client.
9210 *Philip Paeps <philip@freebsd.org>*
9212 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9213 to ensure that even with this option, only ciphersuites in the
9214 server's preference list will be accepted. (Note that the option
9215 applies only when resuming a session, so the earlier behavior was
9216 just about the algorithm choice for symmetric cryptography.)
9220 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9222 * Fix NULL pointer dereference if a DTLS server received
9223 ChangeCipherSpec as first record ([CVE-2009-1386]).
9227 * Fix a state transition in s3_srvr.c and d1_srvr.c
9228 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9232 * The fix in 0.9.8c that supposedly got rid of unsafe
9233 double-checked locking was incomplete for RSA blinding,
9234 addressing just one layer of what turns out to have been
9235 doubly unsafe triple-checked locking.
9237 So now fix this for real by retiring the MONT_HELPER macro
9238 in crypto/rsa/rsa_eay.c.
9240 *Bodo Moeller; problem pointed out by Marius Schilder*
9242 * Various precautionary measures:
9244 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9246 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9247 (NB: This would require knowledge of the secret session ticket key
9248 to exploit, in which case you'd be SOL either way.)
9250 - Change bn_nist.c so that it will properly handle input BIGNUMs
9251 outside the expected range.
9253 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9256 *Neel Mehta, Bodo Moeller*
9258 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9259 the load fails. Useful for distros.
9261 *Ben Laurie and the FreeBSD team*
9263 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9267 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9271 * Expand ENGINE to support engine supplied SSL client certificate functions.
9273 This work was sponsored by Logica.
9277 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9278 keystores. Support for SSL/TLS client authentication too.
9279 Not compiled unless enable-capieng specified to Configure.
9281 This work was sponsored by Logica.
9285 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9286 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9287 attribute creation routines such as certificate requests and PKCS#12
9292 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9294 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9295 handshake which could lead to a client crash as found using the
9296 Codenomicon TLS test suite ([CVE-2008-1672])
9298 *Steve Henson, Mark Cox*
9300 * Fix double free in TLS server name extensions which could lead to
9301 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9305 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9307 Clear the error queue to ensure that error entries left from
9308 older function calls do not interfere with the correct operation.
9310 *Lutz Jaenicke, Erik de Castro Lopo*
9312 * Remove root CA certificates of commercial CAs:
9314 The OpenSSL project does not recommend any specific CA and does not
9315 have any policy with respect to including or excluding any CA.
9316 Therefore it does not make any sense to ship an arbitrary selection
9317 of root CA certificates with the OpenSSL software.
9321 * RSA OAEP patches to fix two separate invalid memory reads.
9322 The first one involves inputs when 'lzero' is greater than
9323 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9324 before the beginning of from). The second one involves inputs where
9325 the 'db' section contains nothing but zeroes (there is a one-byte
9326 invalid read after the end of 'db').
9328 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9330 * Partial backport from 0.9.9-dev:
9332 Introduce bn_mul_mont (dedicated Montgomery multiplication
9333 procedure) as a candidate for BIGNUM assembler implementation.
9334 While 0.9.9-dev uses assembler for various architectures, only
9335 x86_64 is available by default here in the 0.9.8 branch, and
9336 32-bit x86 is available through a compile-time setting.
9338 To try the 32-bit x86 assembler implementation, use Configure
9339 option "enable-montasm" (which exists only for this backport).
9341 As "enable-montasm" for 32-bit x86 disclaims code stability
9342 anyway, in this constellation we activate additional code
9343 backported from 0.9.9-dev for further performance improvements,
9344 namely BN_from_montgomery_word. (To enable this otherwise,
9345 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9347 *Andy Polyakov (backport partially by Bodo Moeller)*
9349 * Add TLS session ticket callback. This allows an application to set
9350 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9351 values. This is useful for key rollover for example where several key
9352 sets may exist with different names.
9356 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9357 This was broken until now in 0.9.8 releases, such that the only way
9358 a registered ENGINE could be used (assuming it initialises
9359 successfully on the host) was to explicitly set it as the default
9360 for the relevant algorithms. This is in contradiction with 0.9.7
9361 behaviour and the documentation. With this fix, when an ENGINE is
9362 registered into a given algorithm's table of implementations, the
9363 'uptodate' flag is reset so that auto-discovery will be used next
9364 time a new context for that algorithm attempts to select an
9367 *Ian Lister (tweaked by Geoff Thorpe)*
9369 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9370 implementation in the following ways:
9372 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9375 Lack of BER streaming support means one pass streaming processing is
9376 only supported if data is detached: setting the streaming flag is
9377 ignored for embedded content.
9379 CMS support is disabled by default and must be explicitly enabled
9380 with the enable-cms configuration option.
9384 * Update the GMP engine glue to do direct copies between BIGNUM and
9385 mpz_t when openssl and GMP use the same limb size. Otherwise the
9386 existing "conversion via a text string export" trick is still used.
9388 *Paul Sheer <paulsheer@gmail.com>*
9390 * Zlib compression BIO. This is a filter BIO which compressed and
9391 uncompresses any data passed through it.
9395 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9396 RFC3394 compatible AES key wrapping.
9400 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9401 sets string data without copying. X509_ALGOR_set0() and
9402 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9403 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9404 from an X509_ATTRIBUTE structure optionally checking it occurs only
9405 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9410 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9411 to get the expected BN_FLG_CONSTTIME behavior.
9413 *Bodo Moeller (Google)*
9417 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9418 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9419 - added some more tests to do_tests.pl
9420 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9421 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9422 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9423 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9424 - various changes to netware.pl to enable gcc-cross builds on Win32
9426 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9427 - various changes to fix missing prototype warnings
9428 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9429 - added AES, WHIRLPOOL and CPUID assembler code to build files
9430 - added missing AES assembler make rules to mk1mf.pl
9431 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9433 *Guenter Knauf <eflash@gmx.net>*
9435 * Implement certificate status request TLS extension defined in RFC3546.
9436 A client can set the appropriate parameters and receive the encoded
9437 OCSP response via a callback. A server can query the supplied parameters
9438 and set the encoded OCSP response in the callback. Add simplified examples
9439 to s_client and s_server.
9443 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9446 + Binary incompatibility of ssl_ctx_st structure
9447 + DTLS interoperation with non-compliant servers
9448 + Don't call get_session_cb() without proposed session
9449 + Fix ia64 assembler code
9451 *Andy Polyakov, Steve Henson*
9453 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9455 * DTLS Handshake overhaul. There were longstanding issues with
9456 OpenSSL DTLS implementation, which were making it impossible for
9457 RFC 4347 compliant client to communicate with OpenSSL server.
9458 Unfortunately just fixing these incompatibilities would "cut off"
9459 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9460 server keeps tolerating non RFC compliant syntax. The opposite is
9461 not true, 0.9.8f client can not communicate with earlier server.
9462 This update even addresses CVE-2007-4995.
9466 * Changes to avoid need for function casts in OpenSSL: some compilers
9467 (gcc 4.2 and later) reject their use.
9468 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9471 * Add RFC4507 support to OpenSSL. This includes the corrections in
9472 RFC4507bis. The encrypted ticket format is an encrypted encoded
9473 SSL_SESSION structure, that way new session features are automatically
9476 If a client application caches session in an SSL_SESSION structure
9477 support is transparent because tickets are now stored in the encoded
9480 The SSL_CTX structure automatically generates keys for ticket
9481 protection in servers so again support should be possible
9482 with no application modification.
9484 If a client or server wishes to disable RFC4507 support then the option
9485 SSL_OP_NO_TICKET can be set.
9487 Add a TLS extension debugging callback to allow the contents of any client
9488 or server extensions to be examined.
9490 This work was sponsored by Google.
9494 * Add initial support for TLS extensions, specifically for the server_name
9495 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9496 have new members for a host name. The SSL data structure has an
9497 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9498 stored in that context to allow for session resumption, even after the
9499 SSL has been switched to a new SSL_CTX in reaction to a client's
9500 server_name extension.
9502 New functions (subject to change):
9504 SSL_get_servername()
9505 SSL_get_servername_type()
9508 New CTRL codes and macros (subject to change):
9510 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9511 - SSL_CTX_set_tlsext_servername_callback()
9512 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9513 - SSL_CTX_set_tlsext_servername_arg()
9514 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9516 openssl s_client has a new '-servername ...' option.
9518 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9519 '-key2 ...', '-servername_fatal' (subject to change). This allows
9520 testing the HostName extension for a specific single host name ('-cert'
9521 and '-key' remain fallbacks for handshakes without HostName
9522 negotiation). If the unrecognized_name alert has to be sent, this by
9523 default is a warning; it becomes fatal with the '-servername_fatal'
9526 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9528 * Add AES and SSE2 assembly language support to VC++ build.
9532 * Mitigate attack on final subtraction in Montgomery reduction.
9536 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9537 (which previously caused an internal error).
9541 * Squeeze another 10% out of IGE mode when in != out.
9545 * AES IGE mode speedup.
9547 *Dean Gaudet (Google)*
9549 * Add the Korean symmetric 128-bit cipher SEED (see
9550 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9551 add SEED ciphersuites from RFC 4162:
9553 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9554 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9555 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9556 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9558 To minimize changes between patchlevels in the OpenSSL 0.9.8
9559 series, SEED remains excluded from compilation unless OpenSSL
9560 is configured with 'enable-seed'.
9562 *KISA, Bodo Moeller*
9564 * Mitigate branch prediction attacks, which can be practical if a
9565 single processor is shared, allowing a spy process to extract
9566 information. For detailed background information, see
9567 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9568 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9569 and Necessary Software Countermeasures"). The core of the change
9570 are new versions BN_div_no_branch() and
9571 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9572 respectively, which are slower, but avoid the security-relevant
9573 conditional branches. These are automatically called by BN_div()
9574 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9575 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9576 remove a conditional branch.
9578 BN_FLG_CONSTTIME is the new name for the previous
9579 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9580 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9581 in the exponent causes BN_mod_exp_mont() to use the alternative
9582 implementation in BN_mod_exp_mont_consttime().) The old name
9583 remains as a deprecated alias.
9585 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9586 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9587 constant-time implementations for more than just exponentiation.
9588 Here too the old name is kept as a deprecated alias.
9590 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9591 the BN_BLINDING structure gets an independent copy of the
9592 modulus. This means that the previous `BIGNUM *m` argument to
9593 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9594 essentially becomes `const BIGNUM *m`, although we can't actually
9595 change this in the header file before 0.9.9. It allows
9596 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9597 enable BN_FLG_CONSTTIME.
9599 *Matthew D Wood (Intel Corp)*
9601 * In the SSL/TLS server implementation, be strict about session ID
9602 context matching (which matters if an application uses a single
9603 external cache for different purposes). Previously,
9604 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9605 set. This did ensure strict client verification, but meant that,
9606 with applications using a single external cache for quite
9607 different requirements, clients could circumvent ciphersuite
9608 restrictions for a given session ID context by starting a session
9609 in a different context.
9613 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9614 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9615 authentication-only ciphersuites.
9619 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9620 not complete and could lead to a possible single byte overflow
9621 ([CVE-2007-5135]) [Ben Laurie]
9623 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9625 * Since AES128 and AES256 (and similarly Camellia128 and
9626 Camellia256) share a single mask bit in the logic of
9627 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9628 kludge to work properly if AES128 is available and AES256 isn't
9629 (or if Camellia128 is available and Camellia256 isn't).
9633 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9634 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9635 When a point or a seed is encoded in a BIT STRING, we need to
9636 prevent the removal of trailing zero bits to get the proper DER
9637 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9638 of a NamedBitList, for which trailing 0 bits need to be removed.)
9642 * Have SSL/TLS server implementation tolerate "mismatched" record
9643 protocol version while receiving ClientHello even if the
9644 ClientHello is fragmented. (The server can't insist on the
9645 particular protocol version it has chosen before the ServerHello
9646 message has informed the client about his choice.)
9650 * Add RFC 3779 support.
9652 *Rob Austein for ARIN, Ben Laurie*
9654 * Load error codes if they are not already present instead of using a
9655 static variable. This allows them to be cleanly unloaded and reloaded.
9656 Improve header file function name parsing.
9660 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9661 or CAPABILITY handshake as required by RFCs.
9665 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9667 * Introduce limits to prevent malicious keys being able to
9668 cause a denial of service. ([CVE-2006-2940])
9670 *Steve Henson, Bodo Moeller*
9672 * Fix ASN.1 parsing of certain invalid structures that can result
9673 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9675 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9676 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9678 * Fix SSL client code which could crash if connecting to a
9679 malicious SSLv2 server. ([CVE-2006-4343])
9681 *Tavis Ormandy and Will Drewry, Google Security Team*
9683 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9684 match only those. Before that, "AES256-SHA" would be interpreted
9685 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9686 the same strength classification in 0.9.7h) as we currently only
9687 have a single AES bit in the ciphersuite description bitmap.
9688 That change, however, also applied to ciphersuite strings such as
9689 "RC4-MD5" that intentionally matched multiple ciphersuites --
9690 namely, SSL 2.0 ciphersuites in addition to the more common ones
9691 from SSL 3.0/TLS 1.0.
9693 So we change the selection algorithm again: Naming an explicit
9694 ciphersuite selects this one ciphersuite, and any other similar
9695 ciphersuite (same bitmap) from *other* protocol versions.
9696 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9697 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9699 Since SSL 2.0 does not have any ciphersuites for which the
9700 128/256 bit distinction would be relevant, this works for now.
9701 The proper fix will be to use different bits for AES128 and
9702 AES256, which would have avoided the problems from the beginning;
9703 however, bits are scarce, so we can only do this in a new release
9704 (not just a patchlevel) when we can change the SSL_CIPHER
9705 definition to split the single 'unsigned long mask' bitmap into
9706 multiple values to extend the available space.
9710 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9712 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9713 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9715 * Add AES IGE and biIGE modes.
9719 * Change the Unix randomness entropy gathering to use poll() when
9720 possible instead of select(), since the latter has some
9721 undesirable limitations.
9723 *Darryl Miles via Richard Levitte and Bodo Moeller*
9725 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9726 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9727 cannot be implicitly activated as part of, e.g., the "AES" alias.
9728 However, please upgrade to OpenSSL 0.9.9[-dev] for
9729 non-experimental use of the ECC ciphersuites to get TLS extension
9730 support, which is required for curve and point format negotiation
9731 to avoid potential handshake problems.
9735 * Disable rogue ciphersuites:
9737 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9738 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9739 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9741 The latter two were purportedly from
9742 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9745 Also deactivate the remaining ciphersuites from
9746 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9747 unofficial, and the ID has long expired.
9751 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9752 dual-core machines) and other potential thread-safety issues.
9756 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9757 versions), which is now available for royalty-free use
9758 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9759 Also, add Camellia TLS ciphersuites from RFC 4132.
9761 To minimize changes between patchlevels in the OpenSSL 0.9.8
9762 series, Camellia remains excluded from compilation unless OpenSSL
9763 is configured with 'enable-camellia'.
9767 * Disable the padding bug check when compression is in use. The padding
9768 bug check assumes the first packet is of even length, this is not
9769 necessarily true if compression is enabled and can result in false
9770 positives causing handshake failure. The actual bug test is ancient
9771 code so it is hoped that implementations will either have fixed it by
9772 now or any which still have the bug do not support compression.
9776 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9778 * When applying a cipher rule check to see if string match is an explicit
9779 cipher suite and only match that one cipher suite if it is.
9783 * Link in manifests for VC++ if needed.
9785 *Austin Ziegler <halostatue@gmail.com>*
9787 * Update support for ECC-based TLS ciphersuites according to
9788 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9789 TLS extensions, which are supported starting with the 0.9.9
9790 branch, not in the OpenSSL 0.9.8 branch).
9794 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9795 opaque EVP_CIPHER_CTX handling.
9799 * Fixes and enhancements to zlib compression code. We now only use
9800 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9801 to conform with the standards mentioned here:
9802 <http://www.zlib.net/DLL_FAQ.txt>
9803 Static zlib linking now works on Windows and the new --with-zlib-include
9804 --with-zlib-lib options to Configure can be used to supply the location
9805 of the headers and library. Gracefully handle case where zlib library
9810 * Several fixes and enhancements to the OID generation code. The old code
9811 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9812 handle numbers larger than ULONG_MAX, truncated printing and had a
9813 non standard OBJ_obj2txt() behaviour.
9817 * Add support for building of engines under engine/ as shared libraries
9818 under VC++ build system.
9822 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9823 Hopefully, we will not see any false combination of paths any more.
9827 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9829 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9830 (part of SSL_OP_ALL). This option used to disable the
9831 countermeasure against man-in-the-middle protocol-version
9832 rollback in the SSL 2.0 server implementation, which is a bad
9833 idea. ([CVE-2005-2969])
9835 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9836 for Information Security, National Institute of Advanced Industrial
9837 Science and Technology [AIST], Japan)*
9839 * Add two function to clear and return the verify parameter flags.
9843 * Keep cipherlists sorted in the source instead of sorting them at
9844 runtime, thus removing the need for a lock.
9848 * Avoid some small subgroup attacks in Diffie-Hellman.
9850 *Nick Mathewson and Ben Laurie*
9852 * Add functions for well-known primes.
9856 * Extended Windows CE support.
9858 *Satoshi Nakamura and Andy Polyakov*
9860 * Initialize SSL_METHOD structures at compile time instead of during
9861 runtime, thus removing the need for a lock.
9865 * Make PKCS7_decrypt() work even if no certificate is supplied by
9866 attempting to decrypt each encrypted key in turn. Add support to
9871 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9873 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9876 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9880 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9881 key into the same file any more.
9885 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9889 * Add -utf8 command line and config file option to 'ca'.
9891 *Stefan <stf@udoma.org*
9893 * Removed the macro des_crypt(), as it seems to conflict with some
9894 libraries. Use DES_crypt().
9898 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9899 involves renaming the source and generated shared-libs for
9900 both. The engines will accept the corrected or legacy ids
9901 ('ncipher' and '4758_cca' respectively) when binding. NB,
9902 this only applies when building 'shared'.
9904 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9906 * Add attribute functions to EVP_PKEY structure. Modify
9907 PKCS12_create() to recognize a CSP name attribute and
9908 use it. Make -CSP option work again in pkcs12 utility.
9912 * Add new functionality to the bn blinding code:
9913 - automatic re-creation of the BN_BLINDING parameters after
9914 a fixed number of uses (currently 32)
9915 - add new function for parameter creation
9916 - introduce flags to control the update behaviour of the
9917 BN_BLINDING parameters
9918 - hide BN_BLINDING structure
9919 Add a second BN_BLINDING slot to the RSA structure to improve
9920 performance when a single RSA object is shared among several
9925 * Add support for DTLS.
9927 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9929 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9930 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9934 * Remove buggy and incomplete DH cert support from
9935 ssl/ssl_rsa.c and ssl/s3_both.c
9939 * Use SHA-1 instead of MD5 as the default digest algorithm for
9940 the apps/openssl applications.
9944 * Compile clean with "-Wall -Wmissing-prototypes
9945 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9946 DEBUG_SAFESTACK must also be set.
9950 * Change ./Configure so that certain algorithms can be disabled by default.
9951 The new counterpiece to "no-xxx" is "enable-xxx".
9953 The patented RC5 and MDC2 algorithms will now be disabled unless
9954 "enable-rc5" and "enable-mdc2", respectively, are specified.
9956 (IDEA remains enabled despite being patented. This is because IDEA
9957 is frequently required for interoperability, and there is no license
9958 fee for non-commercial use. As before, "no-idea" can be used to
9959 avoid this algorithm.)
9963 * Add processing of proxy certificates (see RFC 3820). This work was
9964 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9965 EGEE (Enabling Grids for E-science in Europe).
9969 * RC4 performance overhaul on modern architectures/implementations, such
9970 as Intel P4, IA-64 and AMD64.
9974 * New utility extract-section.pl. This can be used specify an alternative
9975 section number in a pod file instead of having to treat each file as
9976 a separate case in Makefile. This can be done by adding two lines to the
9979 =for comment openssl_section:XXX
9981 The blank line is mandatory.
9985 * New arguments -certform, -keyform and -pass for s_client and s_server
9986 to allow alternative format key and certificate files and passphrase
9991 * New structure X509_VERIFY_PARAM which combines current verify parameters,
9992 update associated structures and add various utility functions.
9994 Add new policy related verify parameters, include policy checking in
9995 standard verify code. Enhance 'smime' application with extra parameters
9996 to support policy checking and print out.
10000 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10001 Nehemiah processors. These extensions support AES encryption in hardware
10002 as well as RNG (though RNG support is currently disabled).
10004 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10006 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10010 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10012 *Andy Polyakov and a number of other people*
10014 * Improved PowerPC platform support. Most notably BIGNUM assembler
10015 implementation contributed by IBM.
10017 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10019 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10020 exponent rather than 'unsigned long'. There is a corresponding change to
10021 the new 'rsa_keygen' element of the RSA_METHOD structure.
10023 *Jelte Jansen, Geoff Thorpe*
10025 * Functionality for creating the initial serial number file is now
10026 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10028 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10029 number file to 1, which is bound to cause problems. To avoid
10030 the problems while respecting compatibility between different 0.9.7
10031 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10032 CA.pl for serial number initialization. With the new release 0.9.8,
10033 we can fix the problem directly in the 'ca' utility.)
10037 * Reduced header interdependencies by declaring more opaque objects in
10038 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10039 give fewer recursive includes, which could break lazy source code - so
10040 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10041 developers should define this symbol when building and using openssl to
10042 ensure they track the recommended behaviour, interfaces, [etc], but
10043 backwards-compatible behaviour prevails when this isn't defined.
10047 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10051 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10052 This will generate a random key of the appropriate length based on the
10053 cipher context. The EVP_CIPHER can provide its own random key generation
10054 routine to support keys of a specific form. This is used in the des and
10055 3des routines to generate a key of the correct parity. Update S/MIME
10056 code to use new functions and hence generate correct parity DES keys.
10057 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10058 valid (weak or incorrect parity).
10062 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10063 as looking them up. This is useful when the verified structure may contain
10064 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10065 present unless the new PKCS7_NO_CRL flag is asserted.
10069 * Extend ASN1 oid configuration module. It now additionally accepts the
10072 shortName = some long name, 1.2.3.4
10076 * Reimplemented the BN_CTX implementation. There is now no more static
10077 limitation on the number of variables it can handle nor the depth of the
10078 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10079 information can now expand as required, and rather than having a single
10080 static array of bignums, BN_CTX now uses a linked-list of such arrays
10081 allowing it to expand on demand whilst maintaining the usefulness of
10082 BN_CTX's "bundling".
10086 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10087 to allow all RSA operations to function using a single BN_CTX.
10091 * Preliminary support for certificate policy evaluation and checking. This
10092 is initially intended to pass the tests outlined in "Conformance Testing
10093 of Relying Party Client Certificate Path Processing Logic" v1.07.
10097 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10098 remained unused and not that useful. A variety of other little bignum
10099 tweaks and fixes have also been made continuing on from the audit (see
10104 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10105 associated ASN1, EVP and SSL functions and old ASN1 macros.
10109 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10110 and this should never fail. So the return value from the use of
10111 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10112 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10116 * BN_CTX_get() should return zero-valued bignums, providing the same
10117 initialised value as BN_new().
10119 *Geoff Thorpe, suggested by Ulf Möller*
10121 * Support for inhibitAnyPolicy certificate extension.
10125 * An audit of the BIGNUM code is underway, for which debugging code is
10126 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10127 is considered valid when processing BIGNUMs, and causes execution to
10128 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10129 further steps are taken to deliberately pollute unused data in BIGNUM
10130 structures to try and expose faulty code further on. For now, openssl will
10131 (in its default mode of operation) continue to tolerate the inconsistent
10132 forms that it has tolerated in the past, but authors and packagers should
10133 consider trying openssl and their own applications when compiled with
10134 these debugging symbols defined. It will help highlight potential bugs in
10135 their own code, and will improve the test coverage for OpenSSL itself. At
10136 some point, these tighter rules will become openssl's default to improve
10137 maintainability, though the assert()s and other overheads will remain only
10138 in debugging configurations. See bn.h for more details.
10140 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10142 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10143 that can only be obtained through BN_CTX_new() (which implicitly
10144 initialises it). The presence of this function only made it possible
10145 to overwrite an existing structure (and cause memory leaks).
10149 * Because of the callback-based approach for implementing LHASH as a
10150 template type, lh_insert() adds opaque objects to hash-tables and
10151 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10152 to clean up those corresponding objects before destroying the hash table
10153 (and losing the object pointers). So some over-zealous constifications in
10154 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10155 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10156 prototyped to have "const" restrictions on the object pointers they are
10157 given (and so aren't required to cast them away any more).
10161 * The tmdiff.h API was so ugly and minimal that our own timing utility
10162 (speed) prefers to use its own implementation. The two implementations
10163 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10164 its object type properly exposed (MS_TM) instead of casting to/from
10165 `char *`. This may still change yet if someone realises MS_TM and
10167 aren't necessarily the greatest nomenclatures - but this is what was used
10168 internally to the implementation so I've used that for now.
10172 * Ensure that deprecated functions do not get compiled when
10173 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10174 the self-tests were still using deprecated key-generation functions so
10175 these have been updated also.
10179 * Reorganise PKCS#7 code to separate the digest location functionality
10180 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10181 New function PKCS7_set_digest() to set the digest type for PKCS#7
10182 digestedData type. Add additional code to correctly generate the
10183 digestedData type and add support for this type in PKCS7 initialization
10188 * New function PKCS7_set0_type_other() this initializes a PKCS7
10189 structure of type "other".
10193 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10194 sure the loop does correctly stop and breaking ("division by zero")
10195 modulus operations are not performed. The (pre-generated) prime
10196 table crypto/bn/bn_prime.h was already correct, but it could not be
10197 re-generated on some platforms because of the "division by zero"
10198 situation in the script.
10200 *Ralf S. Engelschall*
10202 * Update support for ECC-based TLS ciphersuites according to
10203 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10204 SHA-1 now is only used for "small" curves (where the
10205 representation of a field element takes up to 24 bytes); for
10206 larger curves, the field element resulting from ECDH is directly
10207 used as premaster secret.
10209 *Douglas Stebila (Sun Microsystems Laboratories)*
10211 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10212 curve secp160r1 to the tests.
10214 *Douglas Stebila (Sun Microsystems Laboratories)*
10216 * Add the possibility to load symbols globally with DSO.
10218 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10220 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10221 control of the error stack.
10225 * Add support for STORE in ENGINE.
10229 * Add the STORE type. The intention is to provide a common interface
10230 to certificate and key stores, be they simple file-based stores, or
10231 HSM-type store, or LDAP stores, or...
10232 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10236 * Add a generic structure called OPENSSL_ITEM. This can be used to
10237 pass a list of arguments to any function as well as provide a way
10238 for a function to pass data back to the caller.
10242 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10243 works like BUF_strdup() but can be used to duplicate a portion of
10244 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10249 * Add the function sk_find_ex() which works like sk_find(), but will
10250 return an index to an element even if an exact match couldn't be
10251 found. The index is guaranteed to point at the element where the
10252 searched-for key would be inserted to preserve sorting order.
10256 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10257 takes an extra flags argument for optional functionality. Currently,
10258 the following flags are defined:
10260 OBJ_BSEARCH_VALUE_ON_NOMATCH
10261 This one gets OBJ_bsearch_ex() to return a pointer to the first
10262 element where the comparing function returns a negative or zero
10265 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10266 This one gets OBJ_bsearch_ex() to return a pointer to the first
10267 element where the comparing function returns zero. This is useful
10268 if there are more than one element where the comparing function
10273 * Make it possible to create self-signed certificates with 'openssl ca'
10274 in such a way that the self-signed certificate becomes part of the
10275 CA database and uses the same mechanisms for serial number generation
10276 as all other certificate signing. The new flag '-selfsign' enables
10277 this functionality. Adapt CA.sh and CA.pl.in.
10281 * Add functionality to check the public key of a certificate request
10282 against a given private. This is useful to check that a certificate
10283 request can be signed by that key (self-signing).
10287 * Make it possible to have multiple active certificates with the same
10288 subject in the CA index file. This is done only if the keyword
10289 'unique_subject' is set to 'no' in the main CA section (default
10290 if 'CA_default') of the configuration file. The value is saved
10291 with the database itself in a separate index attribute file,
10292 named like the index file with '.attr' appended to the name.
10296 * Generate multi-valued AVAs using '+' notation in config files for
10301 * Support for nameConstraints certificate extension.
10305 * Support for policyConstraints certificate extension.
10309 * Support for policyMappings certificate extension.
10313 * Make sure the default DSA_METHOD implementation only uses its
10314 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10315 and change its own handlers to be NULL so as to remove unnecessary
10316 indirection. This lets alternative implementations fallback to the
10317 default implementation more easily.
10321 * Support for directoryName in GeneralName related extensions
10326 * Make it possible to link applications using Makefile.shared.
10327 Make that possible even when linking against static libraries!
10331 * Support for single pass processing for S/MIME signing. This now
10332 means that S/MIME signing can be done from a pipe, in addition
10333 cleartext signing (multipart/signed type) is effectively streaming
10334 and the signed data does not need to be all held in memory.
10336 This is done with a new flag PKCS7_STREAM. When this flag is set
10337 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10338 is done after the data is output (and digests calculated) in
10339 SMIME_write_PKCS7().
10343 * Add full support for -rpath/-R, both in shared libraries and
10344 applications, at least on the platforms where it's known how
10349 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10350 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10351 will now compute a table of multiples of the generator that
10352 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10353 faster (notably in the case of a single point multiplication,
10354 scalar * generator).
10356 *Nils Larsch, Bodo Moeller*
10358 * IPv6 support for certificate extensions. The various extensions
10359 which use the IP:a.b.c.d can now take IPv6 addresses using the
10360 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10365 * Added an ENGINE that implements RSA by performing private key
10366 exponentiations with the GMP library. The conversions to and from
10367 GMP's mpz_t format aren't optimised nor are any montgomery forms
10368 cached, and on x86 it appears OpenSSL's own performance has caught up.
10369 However there are likely to be other architectures where GMP could
10370 provide a boost. This ENGINE is not built in by default, but it can be
10371 specified at Configure time and should be accompanied by the necessary
10372 linker additions, eg;
10373 ./config -DOPENSSL_USE_GMP -lgmp
10377 * "openssl engine" will not display ENGINE/DSO load failure errors when
10378 testing availability of engines with "-t" - the old behaviour is
10379 produced by increasing the feature's verbosity with "-tt".
10383 * ECDSA routines: under certain error conditions uninitialized BN objects
10384 could be freed. Solution: make sure initialization is performed early
10385 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10390 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10391 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10392 software implementations. For DSA and DH, parameter generation can
10393 also be overridden by providing the appropriate method callbacks.
10397 * Change the "progress" mechanism used in key-generation and
10398 primality testing to functions that take a new BN_GENCB pointer in
10399 place of callback/argument pairs. The new API functions have `_ex`
10400 postfixes and the older functions are reimplemented as wrappers for
10401 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10402 declarations of the old functions to help (graceful) attempts to
10403 migrate to the new functions. Also, the new key-generation API
10404 functions operate on a caller-supplied key-structure and return
10405 success/failure rather than returning a key or NULL - this is to
10406 help make "keygen" another member function of RSA_METHOD etc.
10408 Example for using the new callback interface:
10410 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10411 void *my_arg = ...;
10414 BN_GENCB_set(&my_cb, my_callback, my_arg);
10416 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10417 /* For the meaning of a, b in calls to my_callback(), see the
10418 * documentation of the function that calls the callback.
10419 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10420 * my_callback should return 1 if it wants BN_is_prime_ex()
10421 * to continue, or 0 to stop.
10426 * Change the ZLIB compression method to be stateful, and make it
10427 available to TLS with the number defined in
10428 draft-ietf-tls-compression-04.txt.
10432 * Add the ASN.1 structures and functions for CertificatePair, which
10433 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10435 CertificatePair ::= SEQUENCE {
10436 forward [0] Certificate OPTIONAL,
10437 reverse [1] Certificate OPTIONAL,
10438 -- at least one of the pair shall be present -- }
10440 Also implement the PEM functions to read and write certificate
10441 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10443 This needed to be defined, mostly for the sake of the LDAP
10444 attribute crossCertificatePair, but may prove useful elsewhere as
10449 * Make it possible to inhibit symlinking of shared libraries in
10450 Makefile.shared, for Cygwin's sake.
10454 * Extend the BIGNUM API by creating a function
10455 void BN_set_negative(BIGNUM *a, int neg);
10456 and a macro that behave like
10457 int BN_is_negative(const BIGNUM *a);
10459 to avoid the need to access 'a->neg' directly in applications.
10463 * Implement fast modular reduction for pseudo-Mersenne primes
10464 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10465 EC_GROUP_new_curve_GFp() will now automatically use this
10468 *Nils Larsch <nla@trustcenter.de>*
10470 * Add new lock type (CRYPTO_LOCK_BN).
10474 * Change the ENGINE framework to automatically load engines
10475 dynamically from specific directories unless they could be
10476 found to already be built in or loaded. Move all the
10477 current engines except for the cryptodev one to a new
10478 directory engines/.
10479 The engines in engines/ are built as shared libraries if
10480 the "shared" options was given to ./Configure or ./config.
10481 Otherwise, they are inserted in libcrypto.a.
10482 /usr/local/ssl/engines is the default directory for dynamic
10483 engines, but that can be overridden at configure time through
10484 the usual use of --prefix and/or --openssldir, and at run
10485 time with the environment variable OPENSSL_ENGINES.
10487 *Geoff Thorpe and Richard Levitte*
10489 * Add Makefile.shared, a helper makefile to build shared
10490 libraries. Adapt Makefile.org.
10494 * Add version info to Win32 DLLs.
10496 *Peter 'Luna' Runestig" <peter@runestig.com>*
10498 * Add new 'medium level' PKCS#12 API. Certificates and keys
10499 can be added using this API to created arbitrary PKCS#12
10500 files while avoiding the low level API.
10502 New options to PKCS12_create(), key or cert can be NULL and
10503 will then be omitted from the output file. The encryption
10504 algorithm NIDs can be set to -1 for no encryption, the mac
10505 iteration count can be set to 0 to omit the mac.
10507 Enhance pkcs12 utility by making the -nokeys and -nocerts
10508 options work when creating a PKCS#12 file. New option -nomac
10509 to omit the mac, NONE can be set for an encryption algorithm.
10510 New code is modified to use the enhanced PKCS12_create()
10511 instead of the low level API.
10515 * Extend ASN1 encoder to support indefinite length constructed
10516 encoding. This can output sequences tags and octet strings in
10517 this form. Modify pk7_asn1.c to support indefinite length
10518 encoding. This is experimental and needs additional code to
10519 be useful, such as an ASN1 bio and some enhanced streaming
10522 Extend template encode functionality so that tagging is passed
10523 down to the template encoder.
10527 * Let 'openssl req' fail if an argument to '-newkey' is not
10528 recognized instead of using RSA as a default.
10532 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10533 As these are not official, they are not included in "ALL";
10534 the "ECCdraft" ciphersuite group alias can be used to select them.
10536 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10538 * Add ECDH engine support.
10540 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10542 * Add ECDH in new directory crypto/ecdh/.
10544 *Douglas Stebila (Sun Microsystems Laboratories)*
10546 * Let BN_rand_range() abort with an error after 100 iterations
10547 without success (which indicates a broken PRNG).
10551 * Change BN_mod_sqrt() so that it verifies that the input value
10552 is really the square of the return value. (Previously,
10553 BN_mod_sqrt would show GIGO behaviour.)
10557 * Add named elliptic curves over binary fields from X9.62, SECG,
10558 and WAP/WTLS; add OIDs that were still missing.
10560 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10562 * Extend the EC library for elliptic curves over binary fields
10563 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10566 EC_GF2m_simple_method
10570 EC_GROUP_new_curve_GF2m
10571 EC_GROUP_set_curve_GF2m
10572 EC_GROUP_get_curve_GF2m
10573 EC_POINT_set_affine_coordinates_GF2m
10574 EC_POINT_get_affine_coordinates_GF2m
10575 EC_POINT_set_compressed_coordinates_GF2m
10577 Point compression for binary fields is disabled by default for
10578 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10581 As binary polynomials are represented as BIGNUMs, various members
10582 of the EC_GROUP and EC_POINT data structures can be shared
10583 between the implementations for prime fields and binary fields;
10584 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10585 are essentially identical to their `..._GFp` counterparts.
10586 (For simplicity, the `..._GFp` prefix has been dropped from
10587 various internal method names.)
10589 An internal 'field_div' method (similar to 'field_mul' and
10590 'field_sqr') has been added; this is used only for binary fields.
10592 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10594 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10595 through methods ('mul', 'precompute_mult').
10597 The generic implementations (now internally called 'ec_wNAF_mul'
10598 and 'ec_wNAF_precomputed_mult') remain the default if these
10599 methods are undefined.
10601 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10603 * New function EC_GROUP_get_degree, which is defined through
10604 EC_METHOD. For curves over prime fields, this returns the bit
10605 length of the modulus.
10607 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10609 * New functions EC_GROUP_dup, EC_POINT_dup.
10610 (These simply call ..._new and ..._copy).
10612 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10614 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10615 Polynomials are represented as BIGNUMs (where the sign bit is not
10616 used) in the following functions [macros]:
10619 BN_GF2m_sub [= BN_GF2m_add]
10620 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10621 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10622 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10624 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10625 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10626 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10627 BN_GF2m_cmp [= BN_ucmp]
10629 (Note that only the 'mod' functions are actually for fields GF(2^m).
10630 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10632 For some functions, an the irreducible polynomial defining a
10633 field can be given as an 'unsigned int[]' with strictly
10634 decreasing elements giving the indices of those bits that are set;
10635 i.e., p[] represents the polynomial
10636 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10638 p[0] > p[1] > ... > p[k] = 0.
10639 This applies to the following functions:
10642 BN_GF2m_mod_mul_arr
10643 BN_GF2m_mod_sqr_arr
10644 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10645 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10646 BN_GF2m_mod_exp_arr
10647 BN_GF2m_mod_sqrt_arr
10648 BN_GF2m_mod_solve_quad_arr
10652 Conversion can be performed by the following functions:
10657 bntest.c has additional tests for binary polynomial arithmetic.
10659 Two implementations for BN_GF2m_mod_div() are available.
10660 The default algorithm simply uses BN_GF2m_mod_inv() and
10661 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10662 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10663 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10665 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10667 * Add new error code 'ERR_R_DISABLED' that can be used when some
10668 functionality is disabled at compile-time.
10670 *Douglas Stebila <douglas.stebila@sun.com>*
10672 * Change default behaviour of 'openssl asn1parse' so that more
10673 information is visible when viewing, e.g., a certificate:
10675 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10676 mode the content of non-printable OCTET STRINGs is output in a
10677 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10678 avoid the appearance of a printable string.
10680 *Nils Larsch <nla@trustcenter.de>*
10682 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10684 EC_GROUP_set_asn1_flag()
10685 EC_GROUP_get_asn1_flag()
10686 EC_GROUP_set_point_conversion_form()
10687 EC_GROUP_get_point_conversion_form()
10688 These control ASN1 encoding details:
10689 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10690 has been set to OPENSSL_EC_NAMED_CURVE.
10691 - Points are encoded in uncompressed form by default; options for
10692 asn1_for are as for point2oct, namely
10693 POINT_CONVERSION_COMPRESSED
10694 POINT_CONVERSION_UNCOMPRESSED
10695 POINT_CONVERSION_HYBRID
10697 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10699 EC_GROUP_set_seed()
10700 EC_GROUP_get0_seed()
10701 EC_GROUP_get_seed_len()
10702 This is used only for ASN1 purposes (so far).
10704 *Nils Larsch <nla@trustcenter.de>*
10706 * Add 'field_type' member to EC_METHOD, which holds the NID
10707 of the appropriate field type OID. The new function
10708 EC_METHOD_get_field_type() returns this value.
10710 *Nils Larsch <nla@trustcenter.de>*
10713 EC_POINT_point2bn()
10714 EC_POINT_bn2point()
10715 EC_POINT_point2hex()
10716 EC_POINT_hex2point()
10717 providing useful interfaces to EC_POINT_point2oct() and
10718 EC_POINT_oct2point().
10720 *Nils Larsch <nla@trustcenter.de>*
10722 * Change internals of the EC library so that the functions
10723 EC_GROUP_set_generator()
10724 EC_GROUP_get_generator()
10725 EC_GROUP_get_order()
10726 EC_GROUP_get_cofactor()
10727 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10728 to methods, which would lead to unnecessary code duplication when
10729 adding different types of curves.
10731 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10733 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10734 arithmetic, and such that modified wNAFs are generated
10735 (which avoid length expansion in many cases).
10739 * Add a function EC_GROUP_check_discriminant() (defined via
10740 EC_METHOD) that verifies that the curve discriminant is non-zero.
10742 Add a function EC_GROUP_check() that makes some sanity tests
10743 on a EC_GROUP, its generator and order. This includes
10744 EC_GROUP_check_discriminant().
10746 *Nils Larsch <nla@trustcenter.de>*
10748 * Add ECDSA in new directory crypto/ecdsa/.
10750 Add applications 'openssl ecparam' and 'openssl ecdsa'
10751 (these are based on 'openssl dsaparam' and 'openssl dsa').
10753 ECDSA support is also included in various other files across the
10754 library. Most notably,
10755 - 'openssl req' now has a '-newkey ecdsa:file' option;
10756 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10757 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10758 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10759 them suitable for ECDSA where domain parameters must be
10760 extracted before the specific public key;
10761 - ECDSA engine support has been added.
10763 *Nils Larsch <nla@trustcenter.de>*
10765 * Include some named elliptic curves, and add OIDs from X9.62,
10766 SECG, and WAP/WTLS. Each curve can be obtained from the new
10768 EC_GROUP_new_by_curve_name(),
10769 and the list of available named curves can be obtained with
10770 EC_get_builtin_curves().
10771 Also add a 'curve_name' member to EC_GROUP objects, which can be
10773 EC_GROUP_set_curve_name()
10774 EC_GROUP_get_curve_name()
10776 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10778 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10779 was actually never needed) and in BN_mul(). The removal in BN_mul()
10780 required a small change in bn_mul_part_recursive() and the addition
10781 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10782 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10783 bn_sub_words() and bn_add_words() except they take arrays with
10788 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10790 * Cleanse PEM buffers before freeing them since they may contain
10793 *Benjamin Bennett <ben@psc.edu>*
10795 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10796 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10797 authentication-only ciphersuites.
10801 * Since AES128 and AES256 share a single mask bit in the logic of
10802 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10803 kludge to work properly if AES128 is available and AES256 isn't.
10807 * Expand security boundary to match 1.1.1 module.
10811 * Remove redundant features: hash file source, editing of test vectors
10812 modify fipsld to use external fips_premain.c signature.
10816 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10817 run algorithm test programs.
10821 * Make algorithm test programs more tolerant of whitespace.
10825 * Have SSL/TLS server implementation tolerate "mismatched" record
10826 protocol version while receiving ClientHello even if the
10827 ClientHello is fragmented. (The server can't insist on the
10828 particular protocol version it has chosen before the ServerHello
10829 message has informed the client about his choice.)
10833 * Load error codes if they are not already present instead of using a
10834 static variable. This allows them to be cleanly unloaded and reloaded.
10838 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10840 * Introduce limits to prevent malicious keys being able to
10841 cause a denial of service. ([CVE-2006-2940])
10843 *Steve Henson, Bodo Moeller*
10845 * Fix ASN.1 parsing of certain invalid structures that can result
10846 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10848 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10849 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10851 * Fix SSL client code which could crash if connecting to a
10852 malicious SSLv2 server. ([CVE-2006-4343])
10854 *Tavis Ormandy and Will Drewry, Google Security Team*
10856 * Change ciphersuite string processing so that an explicit
10857 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10858 will no longer include "AES128-SHA"), and any other similar
10859 ciphersuite (same bitmap) from *other* protocol versions (so that
10860 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10861 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10862 changes from 0.9.8b and 0.9.8d.
10866 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10868 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10869 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10871 * Change the Unix randomness entropy gathering to use poll() when
10872 possible instead of select(), since the latter has some
10873 undesirable limitations.
10875 *Darryl Miles via Richard Levitte and Bodo Moeller*
10877 * Disable rogue ciphersuites:
10879 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10880 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10881 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10883 The latter two were purportedly from
10884 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10887 Also deactivate the remaining ciphersuites from
10888 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10889 unofficial, and the ID has long expired.
10893 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10894 dual-core machines) and other potential thread-safety issues.
10898 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10900 * Adapt fipsld and the build system to link against the validated FIPS
10901 module in FIPS mode.
10905 * Fixes for VC++ 2005 build under Windows.
10909 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10910 from a Windows bash shell such as MSYS. It is autodetected from the
10911 "config" script when run from a VC++ environment. Modify standard VC++
10912 build to use fipscanister.o from the GNU make build.
10916 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10918 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10919 The value now differs depending on if you build for FIPS or not.
10920 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10921 safely run with a non-FIPSed libcrypto, as it may crash because of
10922 the difference induced by this change.
10926 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10928 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10929 (part of SSL_OP_ALL). This option used to disable the
10930 countermeasure against man-in-the-middle protocol-version
10931 rollback in the SSL 2.0 server implementation, which is a bad
10932 idea. ([CVE-2005-2969])
10934 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10935 for Information Security, National Institute of Advanced Industrial
10936 Science and Technology [AIST, Japan)]*
10938 * Minimal support for X9.31 signatures and PSS padding modes. This is
10939 mainly for FIPS compliance and not fully integrated at this stage.
10943 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10944 the exponentiation using a fixed-length exponent. (Otherwise,
10945 the information leaked through timing could expose the secret key
10946 after many signatures; cf. Bleichenbacher's attack on DSA with
10951 * Make a new fixed-window mod_exp implementation the default for
10952 RSA, DSA, and DH private-key operations so that the sequence of
10953 squares and multiplies and the memory access pattern are
10954 independent of the particular secret key. This will mitigate
10955 cache-timing and potential related attacks.
10957 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10958 and this is automatically used by BN_mod_exp_mont() if the new flag
10959 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10960 will use this BN flag for private exponents unless the flag
10961 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10962 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10964 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10966 * Change the client implementation for SSLv23_method() and
10967 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10968 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10969 (Previously, the SSL 2.0 backwards compatible Client Hello
10970 message format would be used even with SSL_OP_NO_SSLv2.)
10974 * Add support for smime-type MIME parameter in S/MIME messages which some
10979 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10980 a threadsafe manner. Modify rsa code to use new function and add calls
10981 to dsa and dh code (which had race conditions before).
10985 * Include the fixed error library code in the C error file definitions
10986 instead of fixing them up at runtime. This keeps the error code
10987 structures constant.
10991 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
10993 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
10996 * Fixes for newer kerberos headers. NB: the casts are needed because
10997 the 'length' field is signed on one version and unsigned on another
10998 with no (?) obvious way to tell the difference, without these VC++
10999 complains. Also the "definition" of FAR (blank) is no longer included
11000 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11001 some needed definitions.
11005 * Undo Cygwin change.
11009 * Added support for proxy certificates according to RFC 3820.
11010 Because they may be a security thread to unaware applications,
11011 they must be explicitly allowed in run-time. See
11012 docs/HOWTO/proxy_certificates.txt for further information.
11016 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11018 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11019 server and client random values. Previously
11020 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11021 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11023 This change has negligible security impact because:
11025 1. Server and client random values still have 24 bytes of pseudo random
11028 2. Server and client random values are sent in the clear in the initial
11031 3. The master secret is derived using the premaster secret (48 bytes in
11032 size for static RSA ciphersuites) as well as client server and random
11035 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11038 *Stephen Henson, reported by UK NISCC*
11040 * Use Windows randomness collection on Cygwin.
11044 * Fix hang in EGD/PRNGD query when communication socket is closed
11045 prematurely by EGD/PRNGD.
11047 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11049 * Prompt for pass phrases when appropriate for PKCS12 input format.
11053 * Back-port of selected performance improvements from development
11054 branch, as well as improved support for PowerPC platforms.
11058 * Add lots of checks for memory allocation failure, error codes to indicate
11059 failure and freeing up memory if a failure occurs.
11061 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11063 * Add new -passin argument to dgst.
11067 * Perform some character comparisons of different types in X509_NAME_cmp:
11068 this is needed for some certificates that re-encode DNs into UTF8Strings
11069 (in violation of RFC3280) and can't or won't issue name rollover
11074 * Make an explicit check during certificate validation to see that
11075 the CA setting in each certificate on the chain is correct. As a
11076 side effect always do the following basic checks on extensions,
11077 not just when there's an associated purpose to the check:
11079 - if there is an unhandled critical extension (unless the user
11080 has chosen to ignore this fault)
11081 - if the path length has been exceeded (if one is set at all)
11082 - that certain extensions fit the associated purpose (if one has
11087 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11089 * Avoid a race condition when CRLs are checked in a multi threaded
11090 environment. This would happen due to the reordering of the revoked
11091 entries during signature checking and serial number lookup. Now the
11092 encoding is cached and the serial number sort performed under a lock.
11093 Add new STACK function sk_is_sorted().
11097 * Add Delta CRL to the extension code.
11101 * Various fixes to s3_pkt.c so alerts are sent properly.
11103 *David Holmes <d.holmes@f5.com>*
11105 * Reduce the chances of duplicate issuer name and serial numbers (in
11106 violation of RFC3280) using the OpenSSL certificate creation utilities.
11107 This is done by creating a random 64 bit value for the initial serial
11108 number when a serial number file is created or when a self signed
11109 certificate is created using 'openssl req -x509'. The initial serial
11110 number file is created using 'openssl x509 -next_serial' in CA.pl
11111 rather than being initialized to 1.
11115 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11117 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11118 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11120 *Joe Orton, Steve Henson*
11122 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11125 *Joe Orton, Steve Henson*
11127 * Make it possible to have multiple active certificates with the same
11128 subject in the CA index file. This is done only if the keyword
11129 'unique_subject' is set to 'no' in the main CA section (default
11130 if 'CA_default') of the configuration file. The value is saved
11131 with the database itself in a separate index attribute file,
11132 named like the index file with '.attr' appended to the name.
11136 * X509 verify fixes. Disable broken certificate workarounds when
11137 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11138 keyUsage extension present. Don't accept CRLs with unhandled critical
11139 extensions: since verify currently doesn't process CRL extensions this
11140 rejects a CRL with *any* critical extensions. Add new verify error codes
11145 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11146 A clarification of RFC2560 will require the use of OCTET STRINGs and
11147 some implementations cannot handle the current raw format. Since OpenSSL
11148 copies and compares OCSP nonces as opaque blobs without any attempt at
11149 parsing them this should not create any compatibility issues.
11153 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11154 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11155 this HMAC (and other) operations are several times slower than OpenSSL
11160 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11162 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11164 * Use the correct content when signing type "other".
11168 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11170 * Fix various bugs revealed by running the NISCC test suite:
11172 Stop out of bounds reads in the ASN1 code when presented with
11173 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11175 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11177 If verify callback ignores invalid public key errors don't try to check
11178 certificate signature with the NULL public key.
11182 * New -ignore_err option in ocsp application to stop the server
11183 exiting on the first error in a request.
11187 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11188 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11193 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11194 extra data after the compression methods not only for TLS 1.0
11195 but also for SSL 3.0 (as required by the specification).
11197 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11199 * Change X509_certificate_type() to mark the key as exported/exportable
11200 when it's 512 *bits* long, not 512 bytes.
11204 * Change AES_cbc_encrypt() so it outputs exact multiple of
11205 blocks during encryption.
11209 * Various fixes to base64 BIO and non blocking I/O. On write
11210 flushes were not handled properly if the BIO retried. On read
11211 data was not being buffered properly and had various logic bugs.
11212 This also affects blocking I/O when the data being decoded is a
11217 * Various S/MIME bugfixes and compatibility changes:
11218 output correct application/pkcs7 MIME type if
11219 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11220 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11221 of files as .eml work). Correctly handle very long lines in MIME
11226 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11228 * Countermeasure against the Klima-Pokorny-Rosa extension of
11229 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11230 a protocol version number mismatch like a decryption error
11231 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11235 * Turn on RSA blinding by default in the default implementation
11236 to avoid a timing attack. Applications that don't want it can call
11237 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11238 They would be ill-advised to do so in most cases.
11240 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11242 * Change RSA blinding code so that it works when the PRNG is not
11243 seeded (in this case, the secret RSA exponent is abused as
11244 an unpredictable seed -- if it is not unpredictable, there
11245 is no point in blinding anyway). Make RSA blinding thread-safe
11246 by remembering the creator's thread ID in rsa->blinding and
11247 having all other threads use local one-time blinding factors
11248 (this requires more computation than sharing rsa->blinding, but
11249 avoids excessive locking; and if an RSA object is not shared
11250 between threads, blinding will still be very fast).
11254 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11255 ENGINE as defaults for all supported algorithms irrespective of
11256 the 'flags' parameter. 'flags' is now honoured, so applications
11257 should make sure they are passing it correctly.
11261 * Target "mingw" now allows native Windows code to be generated in
11262 the Cygwin environment as well as with the MinGW compiler.
11266 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11268 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11269 via timing by performing a MAC computation even if incorrect
11270 block cipher padding has been found. This is a countermeasure
11271 against active attacks where the attacker has to distinguish
11272 between bad padding and a MAC verification error. ([CVE-2003-0078])
11274 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11275 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11276 Martin Vuagnoux (EPFL, Ilion)*
11278 * Make the no-err option work as intended. The intention with no-err
11279 is not to have the whole error stack handling routines removed from
11280 libcrypto, it's only intended to remove all the function name and
11281 reason texts, thereby removing some of the footprint that may not
11282 be interesting if those errors aren't displayed anyway.
11284 NOTE: it's still possible for any application or module to have its
11285 own set of error texts inserted. The routines are there, just not
11286 used by default when no-err is given.
11290 * Add support for FreeBSD on IA64.
11292 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11294 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11295 Kerberos function mit_des_cbc_cksum(). Before this change,
11296 the value returned by DES_cbc_cksum() was like the one from
11297 mit_des_cbc_cksum(), except the bytes were swapped.
11299 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11301 * Allow an application to disable the automatic SSL chain building.
11302 Before this a rather primitive chain build was always performed in
11303 ssl3_output_cert_chain(): an application had no way to send the
11304 correct chain if the automatic operation produced an incorrect result.
11306 Now the chain builder is disabled if either:
11308 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11310 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11312 The reasoning behind this is that an application would not want the
11313 auto chain building to take place if extra chain certificates are
11314 present and it might also want a means of sending no additional
11315 certificates (for example the chain has two certificates and the
11320 * Add the possibility to build without the ENGINE framework.
11322 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11324 * Under Win32 gmtime() can return NULL: check return value in
11325 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11329 * DSA routines: under certain error conditions uninitialized BN objects
11330 could be freed. Solution: make sure initialization is performed early
11331 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11332 Nils Larsch <nla@trustcenter.de> via PR#459)
11336 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11337 checked on reconnect on the client side, therefore session resumption
11338 could still fail with a "ssl session id is different" error. This
11339 behaviour is masked when SSL_OP_ALL is used due to
11340 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11341 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11342 followup to PR #377.
11346 * IA-32 assembler support enhancements: unified ELF targets, support
11347 for SCO/Caldera platforms, fix for Cygwin shared build.
11351 * Add support for FreeBSD on sparc64. As a consequence, support for
11352 FreeBSD on non-x86 processors is separate from x86 processors on
11353 the config script, much like the NetBSD support.
11355 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11357 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11359 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11362 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11363 code (06) was taken as the first octet of the session ID and the last
11364 octet was ignored consequently. As a result SSLv2 client side session
11365 caching could not have worked due to the session ID mismatch between
11367 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11372 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11373 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11378 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11379 seems that in spite of existing for more than a year, many application
11380 author have done nothing to provide the necessary callbacks, which
11381 means that this particular engine will not work properly anywhere.
11382 This is a very unfortunate situation which forces us, in the name
11383 of usability, to give the hw_ncipher.c a static lock, which is part
11385 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11386 appear in 0.9.8 or later. We EXPECT application authors to have
11387 dealt properly with this when 0.9.8 is released (unless we actually
11388 make such changes in the libcrypto locking code that changes will
11389 have to be made anyway).
11393 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11394 octets have been read, EOF or an error occurs. Without this change
11395 some truncated ASN1 structures will not produce an error.
11399 * Disable Heimdal support, since it hasn't been fully implemented.
11400 Still give the possibility to force the use of Heimdal, but with
11401 warnings and a request that patches get sent to openssl-dev.
11405 * Add the VC-CE target, introduce the WINCE sysname, and add
11406 INSTALL.WCE and appropriate conditionals to make it build.
11408 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11410 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11411 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11412 edit numbers of the version.
11414 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11416 * Introduce safe string copy and catenation functions
11417 (BUF_strlcpy() and BUF_strlcat()).
11419 *Ben Laurie (CHATS) and Richard Levitte*
11421 * Avoid using fixed-size buffers for one-line DNs.
11423 *Ben Laurie (CHATS)*
11425 * Add BUF_MEM_grow_clean() to avoid information leakage when
11426 resizing buffers containing secrets, and use where appropriate.
11428 *Ben Laurie (CHATS)*
11430 * Avoid using fixed size buffers for configuration file location.
11432 *Ben Laurie (CHATS)*
11434 * Avoid filename truncation for various CA files.
11436 *Ben Laurie (CHATS)*
11438 * Use sizeof in preference to magic numbers.
11440 *Ben Laurie (CHATS)*
11442 * Avoid filename truncation in cert requests.
11444 *Ben Laurie (CHATS)*
11446 * Add assertions to check for (supposedly impossible) buffer
11449 *Ben Laurie (CHATS)*
11451 * Don't cache truncated DNS entries in the local cache (this could
11452 potentially lead to a spoofing attack).
11454 *Ben Laurie (CHATS)*
11456 * Fix various buffers to be large enough for hex/decimal
11457 representations in a platform independent manner.
11459 *Ben Laurie (CHATS)*
11461 * Add CRYPTO_realloc_clean() to avoid information leakage when
11462 resizing buffers containing secrets, and use where appropriate.
11464 *Ben Laurie (CHATS)*
11466 * Add BIO_indent() to avoid much slightly worrying code to do
11469 *Ben Laurie (CHATS)*
11471 * Convert sprintf()/BIO_puts() to BIO_printf().
11473 *Ben Laurie (CHATS)*
11475 * buffer_gets() could terminate with the buffer only half
11478 *Ben Laurie (CHATS)*
11480 * Add assertions to prevent user-supplied crypto functions from
11481 overflowing internal buffers by having large block sizes, etc.
11483 *Ben Laurie (CHATS)*
11485 * New OPENSSL_assert() macro (similar to assert(), but enabled
11488 *Ben Laurie (CHATS)*
11490 * Eliminate unused copy of key in RC4.
11492 *Ben Laurie (CHATS)*
11494 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11496 *Ben Laurie (CHATS)*
11498 * Fix off-by-one error in EGD path.
11500 *Ben Laurie (CHATS)*
11502 * If RANDFILE path is too long, ignore instead of truncating.
11504 *Ben Laurie (CHATS)*
11506 * Eliminate unused and incorrectly sized X.509 structure
11509 *Ben Laurie (CHATS)*
11511 * Eliminate unused and dangerous function knumber().
11513 *Ben Laurie (CHATS)*
11515 * Eliminate unused and dangerous structure, KSSL_ERR.
11517 *Ben Laurie (CHATS)*
11519 * Protect against overlong session ID context length in an encoded
11520 session object. Since these are local, this does not appear to be
11523 *Ben Laurie (CHATS)*
11525 * Change from security patch (see 0.9.6e below) that did not affect
11526 the 0.9.6 release series:
11528 Remote buffer overflow in SSL3 protocol - an attacker could
11529 supply an oversized master key in Kerberos-enabled versions.
11532 *Ben Laurie (CHATS)*
11534 * Change the SSL kerb5 codes to match RFC 2712.
11538 * Make -nameopt work fully for req and add -reqopt switch.
11540 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11542 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11544 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11546 * Make sure tests can be performed even if the corresponding algorithms
11547 have been removed entirely. This was also the last step to make
11548 OpenSSL compilable with DJGPP under all reasonable conditions.
11550 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11552 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11553 to allow version independent disabling of normally unselected ciphers,
11554 which may be activated as a side-effect of selecting a single cipher.
11556 (E.g., cipher list string "RSA" enables ciphersuites that are left
11557 out of "ALL" because they do not provide symmetric encryption.
11558 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11560 *Lutz Jaenicke, Bodo Moeller*
11562 * Add appropriate support for separate platform-dependent build
11563 directories. The recommended way to make a platform-dependent
11564 build directory is the following (tested on Linux), maybe with
11567 # Place yourself outside of the OpenSSL source tree. In
11568 # this example, the environment variable OPENSSL_SOURCE
11569 # is assumed to contain the absolute OpenSSL source directory.
11570 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11571 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11572 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11573 mkdir -p `dirname $F`
11574 ln -s $OPENSSL_SOURCE/$F $F
11577 To be absolutely sure not to disturb the source tree, a "make clean"
11578 is a good thing. If it isn't successful, don't worry about it,
11579 it probably means the source directory is very clean.
11583 * Make sure any ENGINE control commands make local copies of string
11584 pointers passed to them whenever necessary. Otherwise it is possible
11585 the caller may have overwritten (or deallocated) the original string
11586 data when a later ENGINE operation tries to use the stored values.
11588 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11590 * Improve diagnostics in file reading and command-line digests.
11592 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11594 * Add AES modes CFB and OFB to the object database. Correct an
11595 error in AES-CFB decryption.
11599 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11600 allows existing EVP_CIPHER_CTX structures to be reused after
11601 calling `EVP_*Final()`. This behaviour is used by encryption
11602 BIOs and some applications. This has the side effect that
11603 applications must explicitly clean up cipher contexts with
11604 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11608 * Check the values of dna and dnb in bn_mul_recursive before calling
11609 bn_mul_comba (a non zero value means the a or b arrays do not contain
11610 n2 elements) and fallback to bn_mul_normal if either is not zero.
11614 * Fix escaping of non-ASCII characters when using the -subj option
11615 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11619 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11620 form for "surname", serialNumber has no short form.
11621 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11622 therefore remove "mail" short name for "internet 7".
11623 The OID for unique identifiers in X509 certificates is
11624 x500UniqueIdentifier, not uniqueIdentifier.
11625 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11629 * Add an "init" command to the ENGINE config module and auto initialize
11630 ENGINEs. Without any "init" command the ENGINE will be initialized
11631 after all ctrl commands have been executed on it. If init=1 the
11632 ENGINE is initialized at that point (ctrls before that point are run
11633 on the uninitialized ENGINE and after on the initialized one). If
11634 init=0 then the ENGINE will not be initialized at all.
11638 * Fix the 'app_verify_callback' interface so that the user-defined
11639 argument is actually passed to the callback: In the
11640 SSL_CTX_set_cert_verify_callback() prototype, the callback
11641 declaration has been changed from
11644 int (*cb)(X509_STORE_CTX *,void *);
11645 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11646 i=s->ctx->app_verify_callback(&ctx)
11647 has been changed into
11648 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11650 To update applications using SSL_CTX_set_cert_verify_callback(),
11651 a dummy argument can be added to their callback functions.
11653 *D. K. Smetters <smetters@parc.xerox.com>*
11655 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11657 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11659 * Add and OPENSSL_LOAD_CONF define which will cause
11660 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11661 This allows older applications to transparently support certain
11662 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11663 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11664 load the config file and OPENSSL_add_all_algorithms_conf() which will
11665 always load it have also been added.
11669 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11670 Adjust NIDs and EVP layer.
11672 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11674 * Config modules support in openssl utility.
11676 Most commands now load modules from the config file,
11677 though in a few (such as version) this isn't done
11678 because it couldn't be used for anything.
11680 In the case of ca and req the config file used is
11681 the same as the utility itself: that is the -config
11682 command line option can be used to specify an
11687 * Move default behaviour from OPENSSL_config(). If appname is NULL
11688 use "openssl_conf" if filename is NULL use default openssl config file.
11692 * Add an argument to OPENSSL_config() to allow the use of an alternative
11693 config section name. Add a new flag to tolerate a missing config file
11694 and move code to CONF_modules_load_file().
11698 * Support for crypto accelerator cards from Accelerated Encryption
11699 Processing, www.aep.ie. (Use engine 'aep')
11700 The support was copied from 0.9.6c [engine] and adapted/corrected
11701 to work with the new engine framework.
11703 *AEP Inc. and Richard Levitte*
11705 * Support for SureWare crypto accelerator cards from Baltimore
11706 Technologies. (Use engine 'sureware')
11707 The support was copied from 0.9.6c [engine] and adapted
11708 to work with the new engine framework.
11712 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11713 make the newer ENGINE framework commands for the CHIL engine work.
11715 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11717 * Make it possible to produce shared libraries on ReliantUNIX.
11719 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11721 * Add the configuration target debug-linux-ppro.
11722 Make 'openssl rsa' use the general key loading routines
11723 implemented in apps.c, and make those routines able to
11724 handle the key format FORMAT_NETSCAPE and the variant
11727 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11729 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11731 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11733 * Add -keyform to rsautl, and document -engine.
11735 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11737 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11738 BIO_R_NO_SUCH_FILE error code rather than the generic
11739 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11743 * Add new functions
11744 ERR_peek_last_error
11745 ERR_peek_last_error_line
11746 ERR_peek_last_error_line_data.
11747 These are similar to
11749 ERR_peek_error_line
11750 ERR_peek_error_line_data,
11751 but report on the latest error recorded rather than the first one
11752 still in the error queue.
11754 *Ben Laurie, Bodo Moeller*
11756 * default_algorithms option in ENGINE config module. This allows things
11758 default_algorithms = ALL
11759 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11763 * Preliminary ENGINE config module.
11767 * New experimental application configuration code.
11771 * Change the AES code to follow the same name structure as all other
11772 symmetric ciphers, and behave the same way. Move everything to
11773 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11775 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11777 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11779 *Ben Laurie and Theo de Raadt*
11781 * Add option to output public keys in req command.
11783 *Massimiliano Pala madwolf@openca.org*
11785 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11786 (up to about 10% better than before for P-192 and P-224).
11790 * New functions/macros
11792 SSL_CTX_set_msg_callback(ctx, cb)
11793 SSL_CTX_set_msg_callback_arg(ctx, arg)
11794 SSL_set_msg_callback(ssl, cb)
11795 SSL_set_msg_callback_arg(ssl, arg)
11797 to request calling a callback function
11799 void cb(int write_p, int version, int content_type,
11800 const void *buf, size_t len, SSL *ssl, void *arg)
11802 whenever a protocol message has been completely received
11803 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11804 protocol version according to which the SSL library interprets
11805 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11806 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11807 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11808 specification (change_cipher_spec(20), alert(21), handshake(22)).
11809 'buf' and 'len' point to the actual message, 'ssl' to the
11810 SSL object, and 'arg' is the application-defined value set by
11811 SSL[_CTX]_set_msg_callback_arg().
11813 'openssl s_client' and 'openssl s_server' have new '-msg' options
11814 to enable a callback that displays all protocol messages.
11818 * Change the shared library support so shared libraries are built as
11819 soon as the corresponding static library is finished, and thereby get
11820 openssl and the test programs linked against the shared library.
11821 This still only happens when the keyword "shard" has been given to
11822 the configuration scripts.
11824 NOTE: shared library support is still an experimental thing, and
11825 backward binary compatibility is still not guaranteed.
11827 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11829 * Add support for Subject Information Access extension.
11831 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11833 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11834 additional bytes when new memory had to be allocated, not just
11835 when reusing an existing buffer.
11839 * New command line and configuration option 'utf8' for the req command.
11840 This allows field values to be specified as UTF8 strings.
11844 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11845 runs for the former and machine-readable output for the latter.
11849 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11850 of the e-mail address in the DN (i.e., it will go into a certificate
11851 extension only). The new configuration file option 'email_in_dn = no'
11852 has the same effect.
11854 *Massimiliano Pala madwolf@openca.org*
11856 * Change all functions with names starting with `des_` to be starting
11857 with `DES_` instead. Add wrappers that are compatible with libdes,
11858 but are named `_ossl_old_des_*`. Finally, add macros that map the
11859 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11860 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11861 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11864 Since we provide two compatibility mappings, the user needs to
11865 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11866 compatibility is desired. The default (i.e., when that macro
11867 isn't defined) is OpenSSL 0.9.6c compatibility.
11869 There are also macros that enable and disable the support of old
11870 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11871 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11872 are defined, the default will apply: to support the old des routines.
11874 In either case, one must include openssl/des.h to get the correct
11875 definitions. Do not try to just include openssl/des_old.h, that
11878 NOTE: This is a major break of an old API into a new one. Software
11879 authors are encouraged to switch to the `DES_` style functions. Some
11880 time in the future, des_old.h and the libdes compatibility functions
11881 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11882 default), and then completely removed.
11886 * Test for certificates which contain unsupported critical extensions.
11887 If such a certificate is found during a verify operation it is
11888 rejected by default: this behaviour can be overridden by either
11889 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11890 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11891 X509_supported_extension() has also been added which returns 1 if a
11892 particular extension is supported.
11896 * Modify the behaviour of EVP cipher functions in similar way to digests
11897 to retain compatibility with existing code.
11901 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11902 compatibility with existing code. In particular the 'ctx' parameter does
11903 not have to be to be initialized before the call to EVP_DigestInit() and
11904 it is tidied up after a call to EVP_DigestFinal(). New function
11905 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11906 EVP_MD_CTX_copy() changed to not require the destination to be
11907 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11908 requires the destination to be valid.
11910 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11911 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11915 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11916 so that complete 'Handshake' protocol structures are kept in memory
11917 instead of overwriting 'msg_type' and 'length' with 'body' data.
11921 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11923 *Massimo Santin via Richard Levitte*
11925 * Major restructuring to the underlying ENGINE code. This includes
11926 reduction of linker bloat, separation of pure "ENGINE" manipulation
11927 (initialisation, etc) from functionality dealing with implementations
11928 of specific crypto interfaces. This change also introduces integrated
11929 support for symmetric ciphers and digest implementations - so ENGINEs
11930 can now accelerate these by providing EVP_CIPHER and EVP_MD
11931 implementations of their own. This is detailed in
11932 [crypto/engine/README.md](crypto/engine/README.md)
11933 as it couldn't be adequately described here. However, there are a few
11934 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11935 were changed in the original introduction of ENGINE code have now
11936 reverted back - the hooking from this code to ENGINE is now a good
11937 deal more passive and at run-time, operations deal directly with
11938 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11939 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11940 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11941 they were not being used by the framework as there is no concept of a
11942 BIGNUM_METHOD and they could not be generalised to the new
11943 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11944 ENGINE_cpy() has been removed as it cannot be consistently defined in
11949 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11953 * Change mkdef.pl to sort symbols that get the same entry number,
11954 and make sure the automatically generated functions `ERR_load_*`
11955 become part of libeay.num as well.
11959 * New function SSL_renegotiate_pending(). This returns true once
11960 renegotiation has been requested (either SSL_renegotiate() call
11961 or HelloRequest/ClientHello received from the peer) and becomes
11962 false once a handshake has been completed.
11963 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11964 sends a HelloRequest, but does not ensure that a handshake takes
11965 place. SSL_renegotiate_pending() is useful for checking if the
11966 client has followed the request.)
11970 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11971 By default, clients may request session resumption even during
11972 renegotiation (if session ID contexts permit); with this option,
11973 session resumption is possible only in the first handshake.
11975 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11976 more bits available for options that should not be part of
11977 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11981 * Add some demos for certificate and certificate request creation.
11985 * Make maximum certificate chain size accepted from the peer application
11986 settable (`SSL*_get/set_max_cert_list()`), as proposed by
11987 "Douglas E. Engert" <deengert@anl.gov>.
11991 * Add support for shared libraries for Unixware-7
11992 (Boyd Lynn Gerber <gerberb@zenez.com>).
11996 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
11997 be done prior to destruction. Use this to unload error strings from
11998 ENGINEs that load their own error strings. NB: This adds two new API
11999 functions to "get" and "set" this destroy handler in an ENGINE.
12003 * Alter all existing ENGINE implementations (except "openssl" and
12004 "openbsd") to dynamically instantiate their own error strings. This
12005 makes them more flexible to be built both as statically-linked ENGINEs
12006 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12007 Also, add stub code to each that makes building them as self-contained
12008 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12012 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12013 implementations into applications that are completely implemented in
12014 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12015 commands that can be used to configure what shared-library to load and
12016 to control aspects of the way it is handled. Also, made an update to
12017 the [README-Engine.md](README-Engine.md) file
12018 that brings its information up-to-date and
12019 provides some information and instructions on the "dynamic" ENGINE
12020 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12024 * Make it possible to unload ranges of ERR strings with a new
12025 "ERR_unload_strings" function.
12029 * Add a copy() function to EVP_MD.
12033 * Make EVP_MD routines take a context pointer instead of just the
12034 md_data void pointer.
12038 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12039 that the digest can only process a single chunk of data
12040 (typically because it is provided by a piece of
12041 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12042 is only going to provide a single chunk of data, and hence the
12043 framework needn't accumulate the data for oneshot drivers.
12047 * As with "ERR", make it possible to replace the underlying "ex_data"
12048 functions. This change also alters the storage and management of global
12049 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12050 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12051 index counters. The API functions that use this state have been changed
12052 to take a "class_index" rather than pointers to the class's local STACK
12053 and counter, and there is now an API function to dynamically create new
12054 classes. This centralisation allows us to (a) plug a lot of the
12055 thread-safety problems that existed, and (b) makes it possible to clean
12056 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12057 such data would previously have always leaked in application code and
12058 workarounds were in place to make the memory debugging turn a blind eye
12059 to it. Application code that doesn't use this new function will still
12060 leak as before, but their memory debugging output will announce it now
12061 rather than letting it slide.
12063 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12064 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12065 has a return value to indicate success or failure.
12069 * Make it possible to replace the underlying "ERR" functions such that the
12070 global state (2 LHASH tables and 2 locks) is only used by the "default"
12071 implementation. This change also adds two functions to "get" and "set"
12072 the implementation prior to it being automatically set the first time
12073 any other ERR function takes place. Ie. an application can call "get",
12074 pass the return value to a module it has just loaded, and that module
12075 can call its own "set" function using that value. This means the
12076 module's "ERR" operations will use (and modify) the error state in the
12077 application and not in its own statically linked copy of OpenSSL code.
12081 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12082 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12083 the operation, and provides a more encapsulated way for external code
12084 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12085 to use these functions rather than manually incrementing the counts.
12087 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12091 * Add EVP test program.
12095 * Add symmetric cipher support to ENGINE. Expect the API to change!
12099 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12100 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12101 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12102 These allow a CRL to be built without having to access X509_CRL fields
12103 directly. Modify 'ca' application to use new functions.
12107 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12108 bug workarounds. Rollback attack detection is a security feature.
12109 The problem will only arise on OpenSSL servers when TLSv1 is not
12110 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12111 Software authors not wanting to support TLSv1 will have special reasons
12112 for their choice and can explicitly enable this option.
12114 *Bodo Moeller, Lutz Jaenicke*
12116 * Rationalise EVP so it can be extended: don't include a union of
12117 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12118 (similar to those existing for EVP_CIPHER_CTX).
12123 EVP_MD_CTX_init(&md); /* new function call */
12124 EVP_DigestInit(&md, EVP_sha1());
12125 EVP_DigestUpdate(&md, in, len);
12126 EVP_DigestFinal(&md, out, NULL);
12127 EVP_MD_CTX_cleanup(&md); /* new function call */
12131 * Make DES key schedule conform to the usual scheme, as well as
12132 correcting its structure. This means that calls to DES functions
12133 now have to pass a pointer to a des_key_schedule instead of a
12134 plain des_key_schedule (which was actually always a pointer
12137 des_key_schedule ks;
12139 des_set_key_checked(..., &ks);
12140 des_ncbc_encrypt(..., &ks, ...);
12142 (Note that a later change renames 'des_...' into 'DES_...'.)
12146 * Initial reduction of linker bloat: the use of some functions, such as
12147 PEM causes large amounts of unused functions to be linked in due to
12148 poor organisation. For example pem_all.c contains every PEM function
12149 which has a knock on effect of linking in large amounts of (unused)
12150 ASN1 code. Grouping together similar functions and splitting unrelated
12151 functions prevents this.
12155 * Cleanup of EVP macros.
12159 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12160 correct `_ecb suffix`.
12164 * Add initial OCSP responder support to ocsp application. The
12165 revocation information is handled using the text based index
12166 use by the ca application. The responder can either handle
12167 requests generated internally, supplied in files (for example
12168 via a CGI script) or using an internal minimal server.
12172 * Add configuration choices to get zlib compression for TLS.
12176 * Changes to Kerberos SSL for RFC 2712 compliance:
12177 1. Implemented real KerberosWrapper, instead of just using
12178 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12179 2. Implemented optional authenticator field of KerberosWrapper.
12181 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12182 and authenticator structs; see crypto/krb5/.
12184 Generalized Kerberos calls to support multiple Kerberos libraries.
12185 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12186 via Richard Levitte*
12188 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12189 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12190 values for each of the key sizes rather than having just
12191 parameters (and 'speed' generating keys each time).
12195 * Speed up EVP routines.
12198 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12199 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12200 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12201 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12203 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12204 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12205 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12208 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12210 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12214 * Added the OS2-EMX target.
12216 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12218 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12219 to support NCONF routines in extension code. New function CONF_set_nconf()
12220 to allow functions which take an NCONF to also handle the old LHASH
12221 structure: this means that the old CONF compatible routines can be
12222 retained (in particular wrt extensions) without having to duplicate the
12223 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12227 * Enhance the general user interface with mechanisms for inner control
12228 and with possibilities to have yes/no kind of prompts.
12232 * Change all calls to low level digest routines in the library and
12233 applications to use EVP. Add missing calls to HMAC_cleanup() and
12234 don't assume HMAC_CTX can be copied using memcpy().
12236 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12238 * Add the possibility to control engines through control names but with
12239 arbitrary arguments instead of just a string.
12240 Change the key loaders to take a UI_METHOD instead of a callback
12241 function pointer. NOTE: this breaks binary compatibility with earlier
12242 versions of OpenSSL [engine].
12243 Adapt the nCipher code for these new conditions and add a card insertion
12248 * Enhance the general user interface with mechanisms to better support
12249 dialog box interfaces, application-defined prompts, the possibility
12250 to use defaults (for example default passwords from somewhere else)
12251 and interrupts/cancellations.
12255 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12256 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12260 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12261 tidy up some unnecessarily weird code in 'sk_new()').
12263 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12265 * Change the key loading routines for ENGINEs to use the same kind
12266 callback (pem_password_cb) as all other routines that need this
12271 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12272 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12273 than this minimum value is recommended.
12277 * New random seeder for OpenVMS, using the system process statistics
12278 that are easily reachable.
12282 * Windows apparently can't transparently handle global
12283 variables defined in DLLs. Initialisations such as:
12285 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12287 won't compile. This is used by the any applications that need to
12288 declare their own ASN1 modules. This was fixed by adding the option
12289 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12290 needed for static libraries under Win32.
12294 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12295 setting of purpose and trust fields. New X509_STORE trust and
12296 purpose functions and tidy up setting in other SSL functions.
12300 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12301 structure. These are inherited by X509_STORE_CTX when it is
12302 initialised. This allows various defaults to be set in the
12303 X509_STORE structure (such as flags for CRL checking and custom
12304 purpose or trust settings) for functions which only use X509_STORE_CTX
12305 internally such as S/MIME.
12307 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12308 trust settings if they are not set in X509_STORE. This allows X509_STORE
12309 purposes and trust (in S/MIME for example) to override any set by default.
12311 Add command line options for CRL checking to smime, s_client and s_server
12316 * Initial CRL based revocation checking. If the CRL checking flag(s)
12317 are set then the CRL is looked up in the X509_STORE structure and
12318 its validity and signature checked, then if the certificate is found
12319 in the CRL the verify fails with a revoked error.
12321 Various new CRL related callbacks added to X509_STORE_CTX structure.
12323 Command line options added to 'verify' application to support this.
12325 This needs some additional work, such as being able to handle multiple
12326 CRLs with different times, extension based lookup (rather than just
12327 by subject name) and ultimately more complete V2 CRL extension
12332 * Add a general user interface API (crypto/ui/). This is designed
12333 to replace things like des_read_password and friends (backward
12334 compatibility functions using this new API are provided).
12335 The purpose is to remove prompting functions from the DES code
12336 section as well as provide for prompting through dialog boxes in
12337 a window system and the like.
12341 * Add "ex_data" support to ENGINE so implementations can add state at a
12342 per-structure level rather than having to store it globally.
12346 * Make it possible for ENGINE structures to be copied when retrieved by
12347 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12348 This causes the "original" ENGINE structure to act like a template,
12349 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12350 operational state can be localised to each ENGINE structure, despite the
12351 fact they all share the same "methods". New ENGINE structures returned in
12352 this case have no functional references and the return value is the single
12353 structural reference. This matches the single structural reference returned
12354 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12359 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12360 needs to match any other type at all we need to manually clear the
12365 * Changes to the "openssl engine" utility to include;
12366 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12367 about an ENGINE's available control commands.
12368 - executing control commands from command line arguments using the
12369 '-pre' and '-post' switches. '-post' is only used if '-t' is
12370 specified and the ENGINE is successfully initialised. The syntax for
12371 the individual commands are colon-separated, for example;
12372 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12376 * New dynamic control command support for ENGINEs. ENGINEs can now
12377 declare their own commands (numbers), names (strings), descriptions,
12378 and input types for run-time discovery by calling applications. A
12379 subset of these commands are implicitly classed as "executable"
12380 depending on their input type, and only these can be invoked through
12381 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12382 can be based on user input, config files, etc). The distinction is
12383 that "executable" commands cannot return anything other than a boolean
12384 result and can only support numeric or string input, whereas some
12385 discoverable commands may only be for direct use through
12386 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12387 pointers, or other custom uses. The "executable" commands are to
12388 support parameterisations of ENGINE behaviour that can be
12389 unambiguously defined by ENGINEs and used consistently across any
12390 OpenSSL-based application. Commands have been added to all the
12391 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12392 control over shared-library paths without source code alterations.
12396 * Changed all ENGINE implementations to dynamically allocate their
12397 ENGINEs rather than declaring them statically. Apart from this being
12398 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12399 this also allows the implementations to compile without using the
12400 internal engine_int.h header.
12404 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12405 'const' value. Any code that should be able to modify a RAND_METHOD
12406 should already have non-const pointers to it (ie. they should only
12407 modify their own ones).
12411 * Made a variety of little tweaks to the ENGINE code.
12412 - "atalla" and "ubsec" string definitions were moved from header files
12413 to C code. "nuron" string definitions were placed in variables
12414 rather than hard-coded - allowing parameterisation of these values
12415 later on via ctrl() commands.
12416 - Removed unused "#if 0"'d code.
12417 - Fixed engine list iteration code so it uses ENGINE_free() to release
12418 structural references.
12419 - Constified the RAND_METHOD element of ENGINE structures.
12420 - Constified various get/set functions as appropriate and added
12421 missing functions (including a catch-all ENGINE_cpy that duplicates
12422 all ENGINE values onto a new ENGINE except reference counts/state).
12423 - Removed NULL parameter checks in get/set functions. Setting a method
12424 or function to NULL is a way of cancelling out a previously set
12425 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12426 and doesn't justify the extra error symbols and code.
12427 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12428 flags from engine_int.h to engine.h.
12429 - Changed prototypes for ENGINE handler functions (init(), finish(),
12430 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12434 * Implement binary inversion algorithm for BN_mod_inverse in addition
12435 to the algorithm using long division. The binary algorithm can be
12436 used only if the modulus is odd. On 32-bit systems, it is faster
12437 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12438 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12439 up to 450 bits. In 64-bit environments, the binary algorithm
12440 appears to be advantageous for much longer moduli; here we use it
12441 for moduli up to 2048 bits.
12445 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12446 could not support the combine flag in choice fields.
12450 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12451 extensions from a certificate request to the certificate.
12455 * Allow multiple 'certopt' and 'nameopt' options to be separated
12456 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12457 file: this allows the display of the certificate about to be
12458 signed to be customised, to allow certain fields to be included
12459 or excluded and extension details. The old system didn't display
12460 multicharacter strings properly, omitted fields not in the policy
12461 and couldn't display additional details such as extensions.
12465 * Function EC_POINTs_mul for multiple scalar multiplication
12466 of an arbitrary number of elliptic curve points
12467 \sum scalars[i]*points[i],
12468 optionally including the generator defined for the EC_GROUP:
12469 scalar*generator + \sum scalars[i]*points[i].
12471 EC_POINT_mul is a simple wrapper function for the typical case
12472 that the point list has just one item (besides the optional
12477 * First EC_METHODs for curves over GF(p):
12479 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12480 operations and provides various method functions that can also
12481 operate with faster implementations of modular arithmetic.
12483 EC_GFp_mont_method() reuses most functions that are part of
12484 EC_GFp_simple_method, but uses Montgomery arithmetic.
12486 *Bodo Moeller; point addition and point doubling
12487 implementation directly derived from source code provided by
12488 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12490 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12491 crypto/ec/ec_lib.c):
12493 Curves are EC_GROUP objects (with an optional group generator)
12494 based on EC_METHODs that are built into the library.
12496 Points are EC_POINT objects based on EC_GROUP objects.
12498 Most of the framework would be able to handle curves over arbitrary
12499 finite fields, but as there are no obvious types for fields other
12500 than GF(p), some functions are limited to that for now.
12504 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12505 that the file contains a complete HTTP response.
12509 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12510 change the def and num file printf format specifier from "%-40sXXX"
12511 to "%-39s XXX". The latter will always guarantee a space after the
12512 field while the former will cause them to run together if the field
12513 is 40 of more characters long.
12517 * Constify the cipher and digest 'method' functions and structures
12518 and modify related functions to take constant EVP_MD and EVP_CIPHER
12523 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12524 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12528 * Modify `EVP_Digest*()` routines so they now return values. Although the
12529 internal software routines can never fail additional hardware versions
12534 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12536 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12537 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12540 ERR_R_NESTED_ASN1_ERROR
12542 ERR_R_MISSING_ASN1_EOS
12543 were 4 .. 9, conflicting with
12544 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12546 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12547 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12549 Add new error code 'ERR_R_INTERNAL_ERROR'.
12553 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12558 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12559 sets the subject name for a new request or supersedes the
12560 subject name in a given request. Formats that can be parsed are
12561 'CN=Some Name, OU=myOU, C=IT'
12563 'CN=Some Name/OU=myOU/C=IT'.
12565 Add options '-batch' and '-verbose' to 'openssl req'.
12567 *Massimiliano Pala <madwolf@hackmasters.net>*
12569 * Introduce the possibility to access global variables through
12570 functions on platform were that's the best way to handle exporting
12571 global variables in shared libraries. To enable this functionality,
12572 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12573 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12574 is normally done by Configure or something similar).
12576 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12577 in the source file (foo.c) like this:
12579 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12580 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12582 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12583 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12585 OPENSSL_DECLARE_GLOBAL(int,foo);
12586 #define foo OPENSSL_GLOBAL_REF(foo)
12587 OPENSSL_DECLARE_GLOBAL(double,bar);
12588 #define bar OPENSSL_GLOBAL_REF(bar)
12590 The #defines are very important, and therefore so is including the
12591 header file everywhere where the defined globals are used.
12593 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12594 of ASN.1 items, but that structure is a bit different.
12596 The largest change is in util/mkdef.pl which has been enhanced with
12597 better and easier to understand logic to choose which symbols should
12598 go into the Windows .def files as well as a number of fixes and code
12599 cleanup (among others, algorithm keywords are now sorted
12600 lexicographically to avoid constant rewrites).
12604 * In BN_div() keep a copy of the sign of 'num' before writing the
12605 result to 'rm' because if rm==num the value will be overwritten
12606 and produce the wrong result if 'num' is negative: this caused
12607 problems with BN_mod() and BN_nnmod().
12611 * Function OCSP_request_verify(). This checks the signature on an
12612 OCSP request and verifies the signer certificate. The signer
12613 certificate is just checked for a generic purpose and OCSP request
12618 * Add OCSP_check_validity() function to check the validity of OCSP
12619 responses. OCSP responses are prepared in real time and may only
12620 be a few seconds old. Simply checking that the current time lies
12621 between thisUpdate and nextUpdate max reject otherwise valid responses
12622 caused by either OCSP responder or client clock inaccuracy. Instead
12623 we allow thisUpdate and nextUpdate to fall within a certain period of
12624 the current time. The age of the response can also optionally be
12625 checked. Two new options -validity_period and -status_age added to
12630 * If signature or public key algorithm is unrecognized print out its
12631 OID rather that just UNKNOWN.
12635 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12636 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12637 ID to be generated from the issuer certificate alone which can then be
12638 passed to OCSP_id_issuer_cmp().
12642 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12643 ASN1 modules to export functions returning ASN1_ITEM pointers
12644 instead of the ASN1_ITEM structures themselves. This adds several
12645 new macros which allow the underlying ASN1 function/structure to
12646 be accessed transparently. As a result code should not use ASN1_ITEM
12647 references directly (such as &X509_it) but instead use the relevant
12648 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12649 use of the new ASN1 code on platforms where exporting structures
12650 is problematical (for example in shared libraries) but exporting
12651 functions returning pointers to structures is not.
12655 * Add support for overriding the generation of SSL/TLS session IDs.
12656 These callbacks can be registered either in an SSL_CTX or per SSL.
12657 The purpose of this is to allow applications to control, if they wish,
12658 the arbitrary values chosen for use as session IDs, particularly as it
12659 can be useful for session caching in multiple-server environments. A
12660 command-line switch for testing this (and any client code that wishes
12661 to use such a feature) has been added to "s_server".
12663 *Geoff Thorpe, Lutz Jaenicke*
12665 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12666 of the form `#if defined(...) || defined(...) || ...` and
12667 `#if !defined(...) && !defined(...) && ...`. This also avoids
12668 the growing number of special cases it was previously handling.
12672 * Make all configuration macros available for application by making
12673 sure they are available in opensslconf.h, by giving them names starting
12674 with `OPENSSL_` to avoid conflicts with other packages and by making
12675 sure e_os2.h will cover all platform-specific cases together with
12677 Additionally, it is now possible to define configuration/platform-
12678 specific names (called "system identities"). In the C code, these
12679 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12680 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12681 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12686 * New option -set_serial to 'req' and 'x509' this allows the serial
12687 number to use to be specified on the command line. Previously self
12688 signed certificates were hard coded with serial number 0 and the
12689 CA options of 'x509' had to use a serial number in a file which was
12694 * New options to 'ca' utility to support V2 CRL entry extensions.
12695 Currently CRL reason, invalidity date and hold instruction are
12696 supported. Add new CRL extensions to V3 code and some new objects.
12700 * New function EVP_CIPHER_CTX_set_padding() this is used to
12701 disable standard block padding (aka PKCS#5 padding) in the EVP
12702 API, which was previously mandatory. This means that the data is
12703 not padded in any way and so the total length much be a multiple
12704 of the block size, otherwise an error occurs.
12708 * Initial (incomplete) OCSP SSL support.
12712 * New function OCSP_parse_url(). This splits up a URL into its host,
12713 port and path components: primarily to parse OCSP URLs. New -url
12714 option to ocsp utility.
12718 * New nonce behavior. The return value of OCSP_check_nonce() now
12719 reflects the various checks performed. Applications can decide
12720 whether to tolerate certain situations such as an absent nonce
12721 in a response when one was present in a request: the ocsp application
12722 just prints out a warning. New function OCSP_add1_basic_nonce()
12723 this is to allow responders to include a nonce in a response even if
12724 the request is nonce-less.
12728 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12729 skipped when using openssl x509 multiple times on a single input file,
12730 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12734 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12735 set string type: to handle setting ASN1_TIME structures. Fix ca
12736 utility to correctly initialize revocation date of CRLs.
12740 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12741 the clients preferred ciphersuites and rather use its own preferences.
12742 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12743 Internet Explorer by ensuring unchanged hash method during stepup.
12744 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12748 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12749 to aes and add a new 'exist' option to print out symbols that don't
12754 * Additional options to ocsp utility to allow flags to be set and
12755 additional certificates supplied.
12759 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12760 OCSP client a number of certificate to only verify the response
12765 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12766 handle the new API. Currently only ECB, CBC modes supported. Add new
12769 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12770 Encryption Standard (AES) Ciphersuites for Transport Layer
12771 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12772 not enabled by default and were not part of the "ALL" ciphersuite
12773 alias because they were not yet official; they could be
12774 explicitly requested by specifying the "AESdraft" ciphersuite
12775 group alias. In the final release of OpenSSL 0.9.7, the group
12776 alias is called "AES" and is part of "ALL".)
12778 *Ben Laurie, Steve Henson, Bodo Moeller*
12780 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12781 request to response.
12785 * Functions for OCSP responders. OCSP_request_onereq_count(),
12786 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12787 extract information from a certificate request. OCSP_response_create()
12788 creates a response and optionally adds a basic response structure.
12789 OCSP_basic_add1_status() adds a complete single response to a basic
12790 response and returns the OCSP_SINGLERESP structure just added (to allow
12791 extensions to be included for example). OCSP_basic_add1_cert() adds a
12792 certificate to a basic response and OCSP_basic_sign() signs a basic
12793 response with various flags. New helper functions ASN1_TIME_check()
12794 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12795 (converts ASN1_TIME to GeneralizedTime).
12799 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12800 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12801 structure from a certificate. X509_pubkey_digest() digests the public_key
12802 contents: this is used in various key identifiers.
12806 * Make sk_sort() tolerate a NULL argument.
12808 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12810 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12811 passed by the function are trusted implicitly. If any of them signed the
12812 response then it is assumed to be valid and is not verified.
12816 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12817 to data. This was previously part of the PKCS7 ASN1 code. This
12818 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12819 *Steve Henson, reported by Kenneth R. Robinette
12820 <support@securenetterm.com>*
12822 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12823 routines: without these tracing memory leaks is very painful.
12824 Fix leaks in PKCS12 and PKCS7 routines.
12828 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12829 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12830 effectively meant GeneralizedTime would never be used. Now it
12831 is initialised to -1 but X509_time_adj() now has to check the value
12832 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12833 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12834 *Steve Henson, reported by Kenneth R. Robinette
12835 <support@securenetterm.com>*
12837 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12838 result in a zero length in the ASN1_INTEGER structure which was
12839 not consistent with the structure when d2i_ASN1_INTEGER() was used
12840 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12841 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12842 where it did not print out a minus for negative ASN1_INTEGER.
12846 * Add summary printout to ocsp utility. The various functions which
12847 convert status values to strings have been renamed to:
12848 OCSP_response_status_str(), OCSP_cert_status_str() and
12849 OCSP_crl_reason_str() and are no longer static. New options
12850 to verify nonce values and to disable verification. OCSP response
12851 printout format cleaned up.
12855 * Add additional OCSP certificate checks. These are those specified
12856 in RFC2560. This consists of two separate checks: the CA of the
12857 certificate being checked must either be the OCSP signer certificate
12858 or the issuer of the OCSP signer certificate. In the latter case the
12859 OCSP signer certificate must contain the OCSP signing extended key
12860 usage. This check is performed by attempting to match the OCSP
12861 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12862 in the OCSP_CERTID structures of the response.
12866 * Initial OCSP certificate verification added to OCSP_basic_verify()
12867 and related routines. This uses the standard OpenSSL certificate
12868 verify routines to perform initial checks (just CA validity) and
12869 to obtain the certificate chain. Then additional checks will be
12870 performed on the chain. Currently the root CA is checked to see
12871 if it is explicitly trusted for OCSP signing. This is used to set
12872 a root CA as a global signing root: that is any certificate that
12873 chains to that CA is an acceptable OCSP signing certificate.
12877 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12878 extensions from a separate configuration file.
12879 As when reading extensions from the main configuration file,
12880 the '-extensions ...' option may be used for specifying the
12883 *Massimiliano Pala <madwolf@comune.modena.it>*
12885 * New OCSP utility. Allows OCSP requests to be generated or
12886 read. The request can be sent to a responder and the output
12887 parsed, outputted or printed in text form. Not complete yet:
12888 still needs to check the OCSP response validity.
12892 * New subcommands for 'openssl ca':
12893 `openssl ca -status <serial>` prints the status of the cert with
12894 the given serial number (according to the index file).
12895 `openssl ca -updatedb` updates the expiry status of certificates
12898 *Massimiliano Pala <madwolf@comune.modena.it>*
12900 * New '-newreq-nodes' command option to CA.pl. This is like
12901 '-newreq', but calls 'openssl req' with the '-nodes' option
12902 so that the resulting key is not encrypted.
12904 *Damien Miller <djm@mindrot.org>*
12906 * New configuration for the GNU Hurd.
12908 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12910 * Initial code to implement OCSP basic response verify. This
12911 is currently incomplete. Currently just finds the signer's
12912 certificate and verifies the signature on the response.
12916 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12917 value of OPENSSLDIR. This is available via the new '-d' option
12918 to 'openssl version', and is also included in 'openssl version -a'.
12922 * Allowing defining memory allocation callbacks that will be given
12923 file name and line number information in additional arguments
12924 (a `const char*` and an int). The basic functionality remains, as
12925 well as the original possibility to just replace malloc(),
12926 realloc() and free() by functions that do not know about these
12927 additional arguments. To register and find out the current
12928 settings for extended allocation functions, the following
12929 functions are provided:
12931 CRYPTO_set_mem_ex_functions
12932 CRYPTO_set_locked_mem_ex_functions
12933 CRYPTO_get_mem_ex_functions
12934 CRYPTO_get_locked_mem_ex_functions
12936 These work the same way as CRYPTO_set_mem_functions and friends.
12937 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12938 extended allocation function is enabled.
12939 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12940 a conventional allocation function is enabled.
12942 *Richard Levitte, Bodo Moeller*
12944 * Finish off removing the remaining LHASH function pointer casts.
12945 There should no longer be any prototype-casting required when using
12946 the LHASH abstraction, and any casts that remain are "bugs". See
12947 the callback types and macros at the head of lhash.h for details
12948 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12952 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12953 If /dev/[u]random devices are not available or do not return enough
12954 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12956 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12957 /etc/entropy will be queried once each in this sequence, querying stops
12958 when enough entropy was collected without querying more sockets.
12962 * Change the Unix RAND_poll() variant to be able to poll several
12963 random devices, as specified by DEVRANDOM, until a sufficient amount
12964 of data has been collected. We spend at most 10 ms on each file
12965 (select timeout) and read in non-blocking mode. DEVRANDOM now
12966 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12967 (previously it was just the string "/dev/urandom"), so on typical
12968 platforms the 10 ms delay will never occur.
12969 Also separate out the Unix variant to its own file, rand_unix.c.
12970 For VMS, there's a currently-empty rand_vms.c.
12974 * Move OCSP client related routines to ocsp_cl.c. These
12975 provide utility functions which an application needing
12976 to issue a request to an OCSP responder and analyse the
12977 response will typically need: as opposed to those which an
12978 OCSP responder itself would need which will be added later.
12980 OCSP_request_sign() signs an OCSP request with an API similar
12981 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12982 response. OCSP_response_get1_basic() extracts basic response
12983 from response. OCSP_resp_find_status(): finds and extracts status
12984 information from an OCSP_CERTID structure (which will be created
12985 when the request structure is built). These are built from lower
12986 level functions which work on OCSP_SINGLERESP structures but
12987 won't normally be used unless the application wishes to examine
12988 extensions in the OCSP response for example.
12990 Replace nonce routines with a pair of functions.
12991 OCSP_request_add1_nonce() adds a nonce value and optionally
12992 generates a random value. OCSP_check_nonce() checks the
12993 validity of the nonce in an OCSP response.
12997 * Change function OCSP_request_add() to OCSP_request_add0_id().
12998 This doesn't copy the supplied OCSP_CERTID and avoids the
12999 need to free up the newly created id. Change return type
13000 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13001 This can then be used to add extensions to the request.
13002 Deleted OCSP_request_new(), since most of its functionality
13003 is now in OCSP_REQUEST_new() (and the case insensitive name
13004 clash) apart from the ability to set the request name which
13005 will be added elsewhere.
13009 * Update OCSP API. Remove obsolete extensions argument from
13010 various functions. Extensions are now handled using the new
13011 OCSP extension code. New simple OCSP HTTP function which
13012 can be used to send requests and parse the response.
13016 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13017 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13018 uses the special reorder version of SET OF to sort the attributes
13019 and reorder them to match the encoded order. This resolves a long
13020 standing problem: a verify on a PKCS7 structure just after signing
13021 it used to fail because the attribute order did not match the
13022 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13023 it uses the received order. This is necessary to tolerate some broken
13024 software that does not order SET OF. This is handled by encoding
13025 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13026 to produce the required SET OF.
13030 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13031 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13032 files to get correct declarations of the ASN.1 item variables.
13036 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13037 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13038 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13039 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13040 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13041 ASN1_ITEM and no wrapper functions.
13045 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13046 replace the old function pointer based I/O routines. Change most of
13047 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13051 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13052 lines, recognize more "algorithms" that can be deselected, and make
13053 it complain about algorithm deselection that isn't recognised.
13057 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13058 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13059 to use new functions. Add NO_ASN1_OLD which can be set to remove
13060 some old style ASN1 functions: this can be used to determine if old
13061 code will still work when these eventually go away.
13065 * New extension functions for OCSP structures, these follow the
13066 same conventions as certificates and CRLs.
13070 * New function X509V3_add1_i2d(). This automatically encodes and
13071 adds an extension. Its behaviour can be customised with various
13072 flags to append, replace or delete. Various wrappers added for
13073 certificates and CRLs.
13077 * Fix to avoid calling the underlying ASN1 print routine when
13078 an extension cannot be parsed. Correct a typo in the
13079 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13083 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13084 entries for variables.
13088 * Add functionality to apps/openssl.c for detecting locking
13089 problems: As the program is single-threaded, all we have
13090 to do is register a locking callback using an array for
13091 storing which locks are currently held by the program.
13095 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13096 SSL_get_ex_data_X509_STORE_idx(), which is used in
13097 ssl_verify_cert_chain() and thus can be called at any time
13098 during TLS/SSL handshakes so that thread-safety is essential.
13099 Unfortunately, the ex_data design is not at all suited
13100 for multi-threaded use, so it probably should be abolished.
13104 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13106 *Broadcom, tweaked and integrated by Geoff Thorpe*
13108 * Move common extension printing code to new function
13109 X509V3_print_extensions(). Reorganise OCSP print routines and
13110 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13114 * New function X509_signature_print() to remove duplication in some
13119 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13120 set (this was treated exactly the same as SET OF previously). This
13121 is used to reorder the STACK representing the structure to match the
13122 encoding. This will be used to get round a problem where a PKCS7
13123 structure which was signed could not be verified because the STACK
13124 order did not reflect the encoded order.
13128 * Reimplement the OCSP ASN1 module using the new code.
13132 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13133 for its ASN1 operations. The old style function pointers still exist
13134 for now but they will eventually go away.
13138 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13139 completely replaces the old ASN1 functionality with a table driven
13140 encoder and decoder which interprets an ASN1_ITEM structure describing
13141 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13142 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13143 has also been converted to the new form.
13147 * Change BN_mod_exp_recp so that negative moduli are tolerated
13148 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13149 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13150 for negative moduli.
13154 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13155 of not touching the result's sign bit.
13159 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13164 * Changed the LHASH code to use prototypes for callbacks, and created
13165 macros to declare and implement thin (optionally static) functions
13166 that provide type-safety and avoid function pointer casting for the
13167 type-specific callbacks.
13171 * Added Kerberos Cipher Suites to be used with TLS, as written in
13173 *Veers Staats <staatsvr@asc.hpc.mil>,
13174 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13176 * Reformat the FAQ so the different questions and answers can be divided
13177 in sections depending on the subject.
13181 * Have the zlib compression code load ZLIB.DLL dynamically under
13186 * New function BN_mod_sqrt for computing square roots modulo a prime
13187 (using the probabilistic Tonelli-Shanks algorithm unless
13188 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13189 be handled deterministically).
13191 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13193 * Make BN_mod_inverse faster by explicitly handling small quotients
13194 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13195 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13199 * New function BN_kronecker.
13203 * Fix BN_gcd so that it works on negative inputs; the result is
13204 positive unless both parameters are zero.
13205 Previously something reasonably close to an infinite loop was
13206 possible because numbers could be growing instead of shrinking
13207 in the implementation of Euclid's algorithm.
13211 * Fix BN_is_word() and BN_is_one() macros to take into account the
13212 sign of the number in question.
13214 Fix BN_is_word(a,w) to work correctly for w == 0.
13216 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13217 because its test if the absolute value of 'a' equals 'w'.
13218 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13219 it exists mostly for use in the implementations of BN_is_zero(),
13220 BN_is_one(), and BN_is_word().
13224 * New function BN_swap.
13228 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13229 the exponentiation functions are more likely to produce reasonable
13230 results on negative inputs.
13234 * Change BN_mod_mul so that the result is always non-negative.
13235 Previously, it could be negative if one of the factors was negative;
13236 I don't think anyone really wanted that behaviour.
13240 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13241 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13242 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13243 and add new functions:
13252 BN_mod_lshift1_quick
13254 BN_mod_lshift_quick
13256 These functions always generate non-negative results.
13258 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13259 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13261 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13262 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13263 be reduced modulo `m`.
13265 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13268 The following entry accidentally appeared in the CHANGES file
13269 distributed with OpenSSL 0.9.7. The modifications described in
13270 it do *not* apply to OpenSSL 0.9.7.
13272 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13273 was actually never needed) and in BN_mul(). The removal in BN_mul()
13274 required a small change in bn_mul_part_recursive() and the addition
13275 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13276 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13277 bn_sub_words() and bn_add_words() except they take arrays with
13283 * In 'openssl passwd', verify passwords read from the terminal
13284 unless the '-salt' option is used (which usually means that
13285 verification would just waste user's time since the resulting
13286 hash is going to be compared with some given password hash)
13287 or the new '-noverify' option is used.
13289 This is an incompatible change, but it does not affect
13290 non-interactive use of 'openssl passwd' (passwords on the command
13291 line, '-stdin' option, '-in ...' option) and thus should not
13292 cause any problems.
13296 * Remove all references to RSAref, since there's no more need for it.
13300 * Make DSO load along a path given through an environment variable
13301 (SHLIB_PATH) with shl_load().
13305 * Constify the ENGINE code as a result of BIGNUM constification.
13306 Also constify the RSA code and most things related to it. In a
13307 few places, most notable in the depth of the ASN.1 code, ugly
13308 casts back to non-const were required (to be solved at a later
13313 * Make it so the openssl application has all engines loaded by default.
13317 * Constify the BIGNUM routines a little more.
13321 * Add the following functions:
13323 ENGINE_load_cswift()
13325 ENGINE_load_atalla()
13326 ENGINE_load_nuron()
13327 ENGINE_load_builtin_engines()
13329 That way, an application can itself choose if external engines that
13330 are built-in in OpenSSL shall ever be used or not. The benefit is
13331 that applications won't have to be linked with libdl or other dso
13332 libraries unless it's really needed.
13334 Changed 'openssl engine' to load all engines on demand.
13335 Changed the engine header files to avoid the duplication of some
13336 declarations (they differed!).
13340 * 'openssl engine' can now list capabilities.
13344 * Better error reporting in 'openssl engine'.
13348 * Never call load_dh_param(NULL) in s_server.
13352 * Add engine application. It can currently list engines by name and
13353 identity, and test if they are actually available.
13357 * Improve RPM specification file by forcing symbolic linking and making
13358 sure the installed documentation is also owned by root.root.
13360 *Damien Miller <djm@mindrot.org>*
13362 * Give the OpenSSL applications more possibilities to make use of
13363 keys (public as well as private) handled by engines.
13367 * Add OCSP code that comes from CertCo.
13371 * Add VMS support for the Rijndael code.
13375 * Added untested support for Nuron crypto accelerator.
13379 * Add support for external cryptographic devices. This code was
13380 previously distributed separately as the "engine" branch.
13382 *Geoff Thorpe, Richard Levitte*
13384 * Rework the filename-translation in the DSO code. It is now possible to
13385 have far greater control over how a "name" is turned into a filename
13386 depending on the operating environment and any oddities about the
13387 different shared library filenames on each system.
13391 * Support threads on FreeBSD-elf in Configure.
13395 * Fix for SHA1 assembly problem with MASM: it produces
13396 warnings about corrupt line number information when assembling
13397 with debugging information. This is caused by the overlapping
13400 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13403 NCONF_get_number() has no error checking at all. As a replacement,
13404 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13405 promoted strongly. The old NCONF_get_number is kept around for
13406 binary backward compatibility.
13407 Make it possible for methods to load from something other than a BIO,
13408 by providing a function pointer that is given a name instead of a BIO.
13409 For example, this could be used to load configuration data from an
13414 * Fix for non blocking accept BIOs. Added new I/O special reason
13415 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13416 with non blocking I/O was not possible because no retry code was
13417 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13422 * Added the beginnings of Rijndael support.
13426 * Fix for bug in DirectoryString mask setting. Add support for
13427 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13428 to allow certificate printing to more controllable, additional
13429 'certopt' option to 'x509' to allow new printing options to be
13434 * Clean old EAY MD5 hack from e_os.h.
13438 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13440 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13441 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13443 *Joe Orton, Steve Henson*
13445 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13447 * Fix additional bug revealed by the NISCC test suite:
13449 Stop bug triggering large recursion when presented with
13450 certain ASN.1 tags ([CVE-2003-0851])
13454 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13456 * Fix various bugs revealed by running the NISCC test suite:
13458 Stop out of bounds reads in the ASN1 code when presented with
13459 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13461 If verify callback ignores invalid public key errors don't try to check
13462 certificate signature with the NULL public key.
13466 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13467 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13472 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13473 extra data after the compression methods not only for TLS 1.0
13474 but also for SSL 3.0 (as required by the specification).
13476 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13478 * Change X509_certificate_type() to mark the key as exported/exportable
13479 when it's 512 *bits* long, not 512 bytes.
13483 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13485 * Countermeasure against the Klima-Pokorny-Rosa extension of
13486 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13487 a protocol version number mismatch like a decryption error
13488 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13492 * Turn on RSA blinding by default in the default implementation
13493 to avoid a timing attack. Applications that don't want it can call
13494 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13495 They would be ill-advised to do so in most cases.
13497 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13499 * Change RSA blinding code so that it works when the PRNG is not
13500 seeded (in this case, the secret RSA exponent is abused as
13501 an unpredictable seed -- if it is not unpredictable, there
13502 is no point in blinding anyway). Make RSA blinding thread-safe
13503 by remembering the creator's thread ID in rsa->blinding and
13504 having all other threads use local one-time blinding factors
13505 (this requires more computation than sharing rsa->blinding, but
13506 avoids excessive locking; and if an RSA object is not shared
13507 between threads, blinding will still be very fast).
13511 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13513 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13514 via timing by performing a MAC computation even if incorrect
13515 block cipher padding has been found. This is a countermeasure
13516 against active attacks where the attacker has to distinguish
13517 between bad padding and a MAC verification error. ([CVE-2003-0078])
13519 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13520 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13521 Martin Vuagnoux (EPFL, Ilion)*
13523 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13525 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13526 memory from its contents. This is done with a counter that will
13527 place alternating values in each byte. This can be used to solve
13528 two issues: 1) the removal of calls to memset() by highly optimizing
13529 compilers, and 2) cleansing with other values than 0, since those can
13530 be read through on certain media, for example a swap space on disk.
13534 * Bugfix: client side session caching did not work with external caching,
13535 because the session->cipher setting was not restored when reloading
13536 from the external cache. This problem was masked, when
13537 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13538 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13542 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13543 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13545 *Zeev Lieber <zeev-l@yahoo.com>*
13547 * Undo an undocumented change introduced in 0.9.6e which caused
13548 repeated calls to OpenSSL_add_all_ciphers() and
13549 OpenSSL_add_all_digests() to be ignored, even after calling
13554 * Change the default configuration reader to deal with last line not
13555 being properly terminated.
13559 * Change X509_NAME_cmp() so it applies the special rules on handling
13560 DN values that are of type PrintableString, as well as RDNs of type
13561 emailAddress where the value has the type ia5String.
13563 *stefank@valicert.com via Richard Levitte*
13565 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13566 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13567 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13568 the bitwise-OR of the two for use by the majority of applications
13569 wanting this behaviour, and update the docs. The documented
13570 behaviour and actual behaviour were inconsistent and had been
13571 changing anyway, so this is more a bug-fix than a behavioural
13574 *Geoff Thorpe, diagnosed by Nadav Har'El*
13576 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13577 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13581 * Fix initialization code race conditions in
13582 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13583 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13584 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13585 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13586 ssl2_get_cipher_by_char(),
13587 ssl3_get_cipher_by_char().
13589 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13591 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13592 the cached sessions are flushed, as the remove_cb() might use ex_data
13593 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13594 (see [openssl.org #212]).
13596 *Geoff Thorpe, Lutz Jaenicke*
13598 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13599 length, instead of the encoding length to d2i_ASN1_OBJECT.
13603 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13605 * [In 0.9.6g-engine release:]
13606 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13608 *Lynn Gazis <lgazis@rainbow.com>*
13610 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13612 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13613 and get fix the header length calculation.
13614 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13615 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13617 * Use proper error handling instead of 'assertions' in buffer
13618 overflow checks added in 0.9.6e. This prevents DoS (the
13619 assertions could call abort()).
13621 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13623 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13625 * Add various sanity checks to asn1_get_length() to reject
13626 the ASN1 length bytes if they exceed sizeof(long), will appear
13627 negative or the content length exceeds the length of the
13630 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13632 * Fix cipher selection routines: ciphers without encryption had no flags
13633 for the cipher strength set and where therefore not handled correctly
13634 by the selection routines (PR #130).
13638 * Fix EVP_dsa_sha macro.
13643 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13644 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13645 that was added in OpenSSL 0.9.6d.
13647 As the countermeasure turned out to be incompatible with some
13648 broken SSL implementations, the new option is part of SSL_OP_ALL.
13649 SSL_OP_ALL is usually employed when compatibility with weird SSL
13650 implementations is desired (e.g. '-bugs' option to 's_client' and
13651 's_server'), so the new option is automatically set in many
13656 * Changes in security patch:
13658 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13659 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13660 Air Force Materiel Command, USAF, under agreement number
13663 * Add various sanity checks to asn1_get_length() to reject
13664 the ASN1 length bytes if they exceed sizeof(long), will appear
13665 negative or the content length exceeds the length of the
13666 supplied buffer. ([CVE-2002-0659])
13668 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13670 * Assertions for various potential buffer overflows, not known to
13671 happen in practice.
13673 *Ben Laurie (CHATS)*
13675 * Various temporary buffers to hold ASCII versions of integers were
13676 too small for 64 bit platforms. ([CVE-2002-0655])
13677 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13679 * Remote buffer overflow in SSL3 protocol - an attacker could
13680 supply an oversized session ID to a client. ([CVE-2002-0656])
13682 *Ben Laurie (CHATS)*
13684 * Remote buffer overflow in SSL2 protocol - an attacker could
13685 supply an oversized client master key. ([CVE-2002-0656])
13687 *Ben Laurie (CHATS)*
13689 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13691 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13692 encoded as NULL) with id-dsa-with-sha1.
13694 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13696 * Check various `X509_...()` return values in apps/req.c.
13698 *Nils Larsch <nla@trustcenter.de>*
13700 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13701 an end-of-file condition would erroneously be flagged, when the CRLF
13702 was just at the end of a processed block. The bug was discovered when
13703 processing data through a buffering memory BIO handing the data to a
13704 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13705 <ptsekov@syntrex.com> and Nedelcho Stanev.
13709 * Implement a countermeasure against a vulnerability recently found
13710 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13711 before application data chunks to avoid the use of known IVs
13712 with data potentially chosen by the attacker.
13716 * Fix length checks in ssl3_get_client_hello().
13720 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13721 to prevent ssl3_read_internal() from incorrectly assuming that
13722 ssl3_read_bytes() found application data while handshake
13723 processing was enabled when in fact s->s3->in_read_app_data was
13724 merely automatically cleared during the initial handshake.
13726 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13728 * Fix object definitions for Private and Enterprise: they were not
13729 recognized in their shortname (=lowercase) representation. Extend
13730 obj_dat.pl to issue an error when using undefined keywords instead
13731 of silently ignoring the problem (Svenning Sorensen
13732 <sss@sss.dnsalias.net>).
13736 * Fix DH_generate_parameters() so that it works for 'non-standard'
13737 generators, i.e. generators other than 2 and 5. (Previously, the
13738 code did not properly initialise the 'add' and 'rem' values to
13739 BN_generate_prime().)
13741 In the new general case, we do not insist that 'generator' is
13742 actually a primitive root: This requirement is rather pointless;
13743 a generator of the order-q subgroup is just as good, if not
13748 * Map new X509 verification errors to alerts. Discovered and submitted by
13749 Tom Wu <tom@arcot.com>.
13753 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13754 returning non-zero before the data has been completely received
13755 when using non-blocking I/O.
13757 *Bodo Moeller; problem pointed out by John Hughes*
13759 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13761 *Ben Laurie, Lutz Jaenicke*
13763 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13764 Yoram Zahavi <YoramZ@gilian.com>).
13768 * Add information about CygWin 1.3 and on, and preserve proper
13769 configuration for the versions before that.
13771 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13773 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13774 check whether we deal with a copy of a session and do not delete from
13775 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13776 <izhar@checkpoint.com>.
13780 * Do not store session data into the internal session cache, if it
13781 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13782 flag is set). Proposed by Aslam <aslam@funk.com>.
13786 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13791 * [In 0.9.6d-engine release:]
13792 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13794 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13796 * Add the configuration target linux-s390x.
13798 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13800 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13801 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13802 variable as an indication that a ClientHello message has been
13803 received. As the flag value will be lost between multiple
13804 invocations of ssl3_accept when using non-blocking I/O, the
13805 function may not be aware that a handshake has actually taken
13806 place, thus preventing a new session from being added to the
13809 To avoid this problem, we now set s->new_session to 2 instead of
13810 using a local variable.
13812 *Lutz Jaenicke, Bodo Moeller*
13814 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13815 if the SSL_R_LENGTH_MISMATCH error is detected.
13817 *Geoff Thorpe, Bodo Moeller*
13819 * New 'shared_ldflag' column in Configure platform table.
13823 * Fix EVP_CIPHER_mode macro.
13825 *"Dan S. Camper" <dan@bti.net>*
13827 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13828 type, we must throw them away by setting rr->length to 0.
13830 *D P Chang <dpc@qualys.com>*
13832 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13834 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13835 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13836 worked incorrectly for those cases where range = `10..._2` and
13837 `3*range` is two bits longer than range.)
13841 * Only add signing time to PKCS7 structures if it is not already
13846 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13847 OBJ_ld_ce should be OBJ_id_ce.
13848 Also some ip-pda OIDs in crypto/objects/objects.txt were
13849 incorrect (cf. RFC 3039).
13851 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13853 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13854 returns early because it has nothing to do.
13856 *Andy Schneider <andy.schneider@bjss.co.uk>*
13858 * [In 0.9.6c-engine release:]
13859 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13861 *Andy Schneider <andy.schneider@bjss.co.uk>*
13863 * [In 0.9.6c-engine release:]
13864 Add support for Cryptographic Appliance's keyserver technology.
13865 (Use engine 'keyclient')
13867 *Cryptographic Appliances and Geoff Thorpe*
13869 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13870 is called via tools/c89.sh because arguments have to be
13871 rearranged (all '-L' options must appear before the first object
13874 *Richard Shapiro <rshapiro@abinitio.com>*
13876 * [In 0.9.6c-engine release:]
13877 Add support for Broadcom crypto accelerator cards, backported
13880 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13882 * [In 0.9.6c-engine release:]
13883 Add support for SureWare crypto accelerator cards from
13884 Baltimore Technologies. (Use engine 'sureware')
13886 *Baltimore Technologies and Mark Cox*
13888 * [In 0.9.6c-engine release:]
13889 Add support for crypto accelerator cards from Accelerated
13890 Encryption Processing, www.aep.ie. (Use engine 'aep')
13892 *AEP Inc. and Mark Cox*
13894 * Add a configuration entry for gcc on UnixWare.
13896 *Gary Benson <gbenson@redhat.com>*
13898 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13899 messages are stored in a single piece (fixed-length part and
13900 variable-length part combined) and fix various bugs found on the way.
13904 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13905 instead. BIO_gethostbyname() does not know what timeouts are
13906 appropriate, so entries would stay in cache even when they have
13908 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13910 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13911 faced with a pathologically small ClientHello fragment that does
13912 not contain client_version: Instead of aborting with an error,
13913 simply choose the highest available protocol version (i.e.,
13914 TLS 1.0 unless it is disabled). In practice, ClientHello
13915 messages are never sent like this, but this change gives us
13916 strictly correct behaviour at least for TLS.
13920 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13921 never resets s->method to s->ctx->method when called from within
13922 one of the SSL handshake functions.
13924 *Bodo Moeller; problem pointed out by Niko Baric*
13926 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13927 (sent using the client's version number) if client_version is
13928 smaller than the protocol version in use. Also change
13929 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13930 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13931 the client will at least see that alert.
13935 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13940 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13941 client receives HelloRequest while in a handshake.
13943 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13945 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13946 should end in 'break', not 'goto end' which circumvents various
13947 cleanups done in state SSL_ST_OK. But session related stuff
13948 must be disabled for SSL_ST_OK in the case that we just sent a
13951 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13952 before just sending a HelloRequest.
13954 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13956 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13957 reveal whether illegal block cipher padding was found or a MAC
13958 verification error occurred. (Neither SSLerr() codes nor alerts
13959 are directly visible to potential attackers, but the information
13960 may leak via logfiles.)
13962 Similar changes are not required for the SSL 2.0 implementation
13963 because the number of padding bytes is sent in clear for SSL 2.0,
13964 and the extra bytes are just ignored. However ssl/s2_pkt.c
13965 failed to verify that the purported number of padding bytes is in
13970 * Add OpenUNIX-8 support including shared libraries
13971 (Boyd Lynn Gerber <gerberb@zenez.com>).
13975 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13976 'wristwatch attack' using huge encoding parameters (cf.
13977 James H. Manger's CRYPTO 2001 paper). Note that the
13978 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13979 encoding parameters and hence was not vulnerable.
13983 * BN_sqr() bug fix.
13985 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
13987 * Rabin-Miller test analyses assume uniformly distributed witnesses,
13988 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
13989 followed by modular reduction.
13991 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
13993 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
13994 equivalent based on BN_pseudo_rand() instead of BN_rand().
13998 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
13999 This function was broken, as the check for a new client hello message
14000 to handle SGC did not allow these large messages.
14001 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14005 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14009 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14010 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14014 * Rework the configuration and shared library support for Tru64 Unix.
14015 The configuration part makes use of modern compiler features and
14016 still retains old compiler behavior for those that run older versions
14017 of the OS. The shared library support part includes a variant that
14018 uses the RPATH feature, and is available through the special
14019 configuration target "alpha-cc-rpath", which will never be selected
14022 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14024 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14025 with the same message size as in ssl3_get_certificate_request().
14026 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14027 messages might inadvertently be reject as too long.
14029 *Petr Lampa <lampa@fee.vutbr.cz>*
14031 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14035 * Modified SSL library such that the verify_callback that has been set
14036 specifically for an SSL object with SSL_set_verify() is actually being
14037 used. Before the change, a verify_callback set with this function was
14038 ignored and the verify_callback() set in the SSL_CTX at the time of
14039 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14040 to allow the necessary settings.
14044 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14045 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14046 done automatically (in contradiction to the requirements of the C
14047 standard). This made problems when used from OpenSSH.
14051 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14052 dh->length and always used
14054 BN_rand_range(priv_key, dh->p).
14056 BN_rand_range() is not necessary for Diffie-Hellman, and this
14057 specific range makes Diffie-Hellman unnecessarily inefficient if
14058 dh->length (recommended exponent length) is much smaller than the
14059 length of dh->p. We could use BN_rand_range() if the order of
14060 the subgroup was stored in the DH structure, but we only have
14065 BN_rand(priv_key, l, ...)
14067 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14074 RSA_eay_public_encrypt
14075 RSA_eay_private_decrypt
14076 RSA_eay_private_encrypt (signing)
14077 RSA_eay_public_decrypt (signature verification)
14079 (default implementations for RSA_public_encrypt,
14080 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14081 always reject numbers >= n.
14085 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14086 to synchronize access to 'locking_thread'. This is necessary on
14087 systems where access to 'locking_thread' (an 'unsigned long'
14088 variable) is not atomic.
14092 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14093 *before* setting the 'crypto_lock_rand' flag. The previous code had
14094 a race condition if 0 is a valid thread ID.
14096 *Travis Vitek <vitek@roguewave.com>*
14098 * Add support for shared libraries under Irix.
14100 *Albert Chin-A-Young <china@thewrittenword.com>*
14102 * Add configuration option to build on Linux on both big-endian and
14103 little-endian MIPS.
14105 *Ralf Baechle <ralf@uni-koblenz.de>*
14107 * Add the possibility to create shared libraries on HP-UX.
14111 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14113 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14114 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14115 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14116 PRNG state recovery was possible based on the output of
14117 one PRNG request appropriately sized to gain knowledge on
14118 'md' followed by enough consecutive 1-byte PRNG requests
14119 to traverse all of 'state'.
14121 1. When updating 'md_local' (the current thread's copy of 'md')
14122 during PRNG output generation, hash all of the previous
14123 'md_local' value, not just the half used for PRNG output.
14125 2. Make the number of bytes from 'state' included into the hash
14126 independent from the number of PRNG bytes requested.
14128 The first measure alone would be sufficient to avoid
14129 Markku-Juhani's attack. (Actually it had never occurred
14130 to me that the half of 'md_local' used for chaining was the
14131 half from which PRNG output bytes were taken -- I had always
14132 assumed that the secret half would be used.) The second
14133 measure makes sure that additional data from 'state' is never
14134 mixed into 'md_local' in small portions; this heuristically
14135 further strengthens the PRNG.
14139 * Fix crypto/bn/asm/mips3.s.
14143 * When only the key is given to "enc", the IV is undefined. Print out
14144 an error message in this case.
14148 * Handle special case when X509_NAME is empty in X509 printing routines.
14152 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14153 positive and less than q.
14157 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14158 used: it isn't thread safe and the add_lock_callback should handle
14161 *Paul Rose <Paul.Rose@bridge.com>*
14163 * Verify that incoming data obeys the block size in
14164 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14170 *Ulf Möller, Bodo Möller*
14172 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14173 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14174 when fixing the server behaviour for backwards-compatible 'client
14175 hello' messages. (Note that the attack is impractical against
14176 SSL 3.0 and TLS 1.0 anyway because length and version checking
14177 means that the probability of guessing a valid ciphertext is
14178 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14181 Before 0.9.5, the countermeasure (hide the error by generating a
14182 random 'decryption result') did not work properly because
14183 ERR_clear_error() was missing, meaning that SSL_get_error() would
14184 detect the supposedly ignored error.
14186 Both problems are now fixed.
14190 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14191 (previously it was 1024).
14195 * Fix for compatibility mode trust settings: ignore trust settings
14196 unless some valid trust or reject settings are present.
14200 * Fix for blowfish EVP: its a variable length cipher.
14204 * Fix various bugs related to DSA S/MIME verification. Handle missing
14205 parameters in DSA public key structures and return an error in the
14206 DSA routines if parameters are absent.
14210 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14211 in the current directory if neither $RANDFILE nor $HOME was set.
14212 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14213 caused some confusion to Windows users who haven't defined $HOME.
14214 Thus RAND_file_name() is changed again: e_os.h can define a
14215 DEFAULT_HOME, which will be used if $HOME is not set.
14216 For Windows, we use "C:"; on other platforms, we still require
14217 environment variables.
14219 * Move 'if (!initialized) RAND_poll()' into regions protected by
14220 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14221 having multiple threads call RAND_poll() concurrently.
14225 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14226 combination of a flag and a thread ID variable.
14227 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14228 flag), *other* threads can enter ssleay_add_bytes without obeying
14229 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14230 that they do not hold after the first thread unsets add_do_not_lock).
14234 * Change bctest again: '-x' expressions are not available in all
14235 versions of 'test'.
14239 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14241 * Fix a couple of memory leaks in PKCS7_dataDecode()
14243 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14245 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14246 the default extension for executables, if any. Also, make the perl
14247 scripts that use symlink() to test if it really exists and use "cp"
14248 if it doesn't. All this made OpenSSL compilable and installable in
14253 * Fix for asn1_GetSequence() for indefinite length constructed data.
14254 If SEQUENCE is length is indefinite just set c->slen to the total
14255 amount of data available.
14257 *Steve Henson, reported by shige@FreeBSD.org*
14259 *This change does not apply to 0.9.7.*
14261 * Change bctest to avoid here-documents inside command substitution
14262 (workaround for FreeBSD /bin/sh bug).
14263 For compatibility with Ultrix, avoid shell functions (introduced
14264 in the bctest version that searches along $PATH).
14268 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14269 with des_encrypt() defined on some operating systems, like Solaris
14274 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14275 On the Importance of Eliminating Errors in Cryptographic
14276 Computations, J. Cryptology 14 (2001) 2, 101-119,
14277 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14281 * MIPS assembler BIGNUM division bug fix.
14285 * Disabled incorrect Alpha assembler code.
14289 * Fix PKCS#7 decode routines so they correctly update the length
14290 after reading an EOC for the EXPLICIT tag.
14294 *This change does not apply to 0.9.7.*
14296 * Fix bug in PKCS#12 key generation routines. This was triggered
14297 if a 3DES key was generated with a 0 initial byte. Include
14298 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14299 (but broken) behaviour.
14303 * Enhance bctest to search for a working bc along $PATH and print
14306 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14308 * Fix memory leaks in err.c: free err_data string if necessary;
14309 don't write to the wrong index in ERR_set_error_data.
14313 * Implement ssl23_peek (analogous to ssl23_read), which previously
14318 * Replace rdtsc with `_emit` statements for VC++ version 5.
14320 *Jeremy Cooper <jeremy@baymoo.org>*
14322 * Make it possible to reuse SSLv2 sessions.
14326 * In copy_email() check for >= 0 as a return value for
14327 X509_NAME_get_index_by_NID() since 0 is a valid index.
14329 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14331 * Avoid coredump with unsupported or invalid public keys by checking if
14332 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14333 PKCS7_verify() fails with non detached data.
14337 * Don't use getenv in library functions when run as setuid/setgid.
14338 New function OPENSSL_issetugid().
14342 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14343 due to incorrect handling of multi-threading:
14345 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14347 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14349 3. Count how many times MemCheck_off() has been called so that
14350 nested use can be treated correctly. This also avoids
14351 inband-signalling in the previous code (which relied on the
14352 assumption that thread ID 0 is impossible).
14356 * Add "-rand" option also to s_client and s_server.
14360 * Fix CPU detection on Irix 6.x.
14361 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14362 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14364 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14369 *This change does not apply to 0.9.7.*
14371 * Use the cached encoding of an X509_NAME structure rather than
14372 copying it. This is apparently the reason for the libsafe "errors"
14373 but the code is actually correct.
14377 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14378 Bleichenbacher's DSA attack.
14379 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14380 to be set and top=0 forces the highest bit to be set; top=-1 is new
14381 and leaves the highest bit random.
14383 *Ulf Moeller, Bodo Moeller*
14385 * In the `NCONF_...`-based implementations for `CONF_...` queries
14386 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14387 a temporary CONF structure with the data component set to NULL
14388 (which gives segmentation faults in lh_retrieve).
14389 Instead, use NULL for the CONF pointer in CONF_get_string and
14390 CONF_get_number (which may use environment variables) and directly
14391 return NULL from CONF_get_section.
14395 * Fix potential buffer overrun for EBCDIC.
14399 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14400 keyUsage if basicConstraints absent for a CA.
14404 * Make SMIME_write_PKCS7() write mail header values with a format that
14405 is more generally accepted (no spaces before the semicolon), since
14406 some programs can't parse those values properly otherwise. Also make
14407 sure BIO's that break lines after each write do not create invalid
14412 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14413 macros previously used would not encode an empty SEQUENCE OF
14414 and break the signature.
14418 *This change does not apply to 0.9.7.*
14420 * Zero the premaster secret after deriving the master secret in
14425 * Add some EVP_add_digest_alias registrations (as found in
14426 OpenSSL_add_all_digests()) to SSL_library_init()
14427 aka OpenSSL_add_ssl_algorithms(). This provides improved
14428 compatibility with peers using X.509 certificates
14429 with unconventional AlgorithmIdentifier OIDs.
14433 * Fix for Irix with NO_ASM.
14435 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14437 * ./config script fixes.
14439 *Ulf Moeller, Richard Levitte*
14441 * Fix 'openssl passwd -1'.
14445 * Change PKCS12_key_gen_asc() so it can cope with non null
14446 terminated strings whose length is passed in the passlen
14447 parameter, for example from PEM callbacks. This was done
14448 by adding an extra length parameter to asc2uni().
14450 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14452 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14453 call failed, free the DSA structure.
14457 * Fix to uni2asc() to cope with zero length Unicode strings.
14458 These are present in some PKCS#12 files.
14462 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14463 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14464 when writing a 32767 byte record.
14466 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14468 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14469 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14471 (RSA objects have a reference count access to which is protected
14472 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14473 so they are meant to be shared between threads.)
14474 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14475 "Reddie, Steven" <Steven.Reddie@ca.com>*
14477 * Fix a deadlock in CRYPTO_mem_leaks().
14481 * Use better test patterns in bntest.
14485 * rand_win.c fix for Borland C.
14489 * BN_rshift bugfix for n == 0.
14493 * Add a 'bctest' script that checks for some known 'bc' bugs
14494 so that 'make test' does not abort just because 'bc' is broken.
14498 * Store verify_result within SSL_SESSION also for client side to
14499 avoid potential security hole. (Re-used sessions on the client side
14500 always resulted in verify_result==X509_V_OK, not using the original
14501 result of the server certificate verification.)
14505 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14506 SSL3_RT_APPLICATION_DATA, return 0.
14507 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14512 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14513 releases, have been re-implemented by renaming the previous
14514 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14515 and ssl3_read_internal, respectively, and adding 'peek' parameters
14516 to them. The new ssl[23]_{read,peek} functions are calls to
14517 ssl[23]_read_internal with the 'peek' flag set appropriately.
14518 A 'peek' parameter has also been added to ssl3_read_bytes, which
14519 does the actual work for ssl3_read_internal.
14523 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14524 the method-specific "init()" handler. Also clean up ex_data after
14525 calling the method-specific "finish()" handler. Previously, this was
14526 happening the other way round.
14530 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14531 The previous value, 12, was not always sufficient for BN_mod_exp().
14535 * Make sure that shared libraries get the internal name engine with
14536 the full version number and not just 0. This should mark the
14537 shared libraries as not backward compatible. Of course, this should
14538 be changed again when we can guarantee backward binary compatibility.
14542 * Fix typo in get_cert_by_subject() in by_dir.c
14544 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14546 * Rework the system to generate shared libraries:
14548 - Make note of the expected extension for the shared libraries and
14549 if there is a need for symbolic links from for example libcrypto.so.0
14550 to libcrypto.so.0.9.7. There is extended info in Configure for
14553 - Make as few rebuilds of the shared libraries as possible.
14555 - Still avoid linking the OpenSSL programs with the shared libraries.
14557 - When installing, install the shared libraries separately from the
14562 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14564 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14565 and not in SSL_clear because the latter is also used by the
14566 accept/connect functions; previously, the settings made by
14567 SSL_set_read_ahead would be lost during the handshake.
14569 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14571 * Correct util/mkdef.pl to be selective about disabled algorithms.
14572 Previously, it would create entries for disabled algorithms no
14577 * Added several new manual pages for SSL_* function.
14581 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14583 * In ssl23_get_client_hello, generate an error message when faced
14584 with an initial SSL 3.0/TLS record that is too small to contain the
14585 first two bytes of the ClientHello message, i.e. client_version.
14586 (Note that this is a pathologic case that probably has never happened
14587 in real life.) The previous approach was to use the version number
14588 from the record header as a substitute; but our protocol choice
14589 should not depend on that one because it is not authenticated
14590 by the Finished messages.
14594 * More robust randomness gathering functions for Windows.
14596 *Jeffrey Altman <jaltman@columbia.edu>*
14598 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14599 not set then we don't setup the error code for issuer check errors
14600 to avoid possibly overwriting other errors which the callback does
14601 handle. If an application does set the flag then we assume it knows
14602 what it is doing and can handle the new informational codes
14607 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14608 a general "ANY" type, as such it should be able to decode anything
14609 including tagged types. However it didn't check the class so it would
14610 wrongly interpret tagged types in the same way as their universal
14611 counterpart and unknown types were just rejected. Changed so that the
14612 tagged and unknown types are handled in the same way as a SEQUENCE:
14613 that is the encoding is stored intact. There is also a new type
14614 "V_ASN1_OTHER" which is used when the class is not universal, in this
14615 case we have no idea what the actual type is so we just lump them all
14620 * On VMS, stdout may very well lead to a file that is written to
14621 in a record-oriented fashion. That means that every write() will
14622 write a separate record, which will be read separately by the
14623 programs trying to read from it. This can be very confusing.
14625 The solution is to put a BIO filter in the way that will buffer
14626 text until a linefeed is reached, and then write everything a
14627 line at a time, so every record written will be an actual line,
14628 not chunks of lines and not (usually doesn't happen, but I've
14629 seen it once) several lines in one record. BIO_f_linebuffer() is
14632 Currently, it's a VMS-only method, because that's where it has
14633 been tested well enough.
14637 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14638 it can return incorrect results.
14639 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14640 but it was in 0.9.6-beta[12].)
14644 * Disable the check for content being present when verifying detached
14645 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14646 include zero length content when signing messages.
14650 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14651 BIO_ctrl (for BIO pairs).
14655 * Add DSO method for VMS.
14659 * Bug fix: Montgomery multiplication could produce results with the
14664 * Add RPM specification openssl.spec and modify it to build three
14665 packages. The default package contains applications, application
14666 documentation and run-time libraries. The devel package contains
14667 include files, static libraries and function documentation. The
14668 doc package contains the contents of the doc directory. The original
14669 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14673 * Add a large number of documentation files for many SSL routines.
14675 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14677 * Add a configuration entry for Sony News 4.
14679 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14681 * Don't set the two most significant bits to one when generating a
14682 random number < q in the DSA library.
14686 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14687 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14688 the underlying transport is blocking) if a handshake took place.
14689 (The default behaviour is needed by applications such as s_client
14690 and s_server that use select() to determine when to use SSL_read;
14691 but for applications that know in advance when to expect data, it
14692 just makes things more complicated.)
14696 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14701 * Add a few more EBCDIC conditionals that make `req` and `x509`
14702 work better on such systems.
14704 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14706 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14707 Update PKCS12_parse() so it copies the friendlyName and the
14708 keyid to the certificates aux info.
14712 * Fix bug in PKCS7_verify() which caused an infinite loop
14713 if there was more than one signature.
14715 *Sven Uszpelkat <su@celocom.de>*
14717 * Major change in util/mkdef.pl to include extra information
14718 about each symbol, as well as presenting variables as well
14719 as functions. This change means that there's n more need
14720 to rebuild the .num files when some algorithms are excluded.
14724 * Allow the verify time to be set by an application,
14725 rather than always using the current time.
14729 * Phase 2 verify code reorganisation. The certificate
14730 verify code now looks up an issuer certificate by a
14731 number of criteria: subject name, authority key id
14732 and key usage. It also verifies self signed certificates
14733 by the same criteria. The main comparison function is
14734 X509_check_issued() which performs these checks.
14736 Lot of changes were necessary in order to support this
14737 without completely rewriting the lookup code.
14739 Authority and subject key identifier are now cached.
14741 The LHASH 'certs' is X509_STORE has now been replaced
14742 by a STACK_OF(X509_OBJECT). This is mainly because an
14743 LHASH can't store or retrieve multiple objects with
14744 the same hash value.
14746 As a result various functions (which were all internal
14747 use only) have changed to handle the new X509_STORE
14748 structure. This will break anything that messed round
14749 with X509_STORE internally.
14751 The functions X509_STORE_add_cert() now checks for an
14752 exact match, rather than just subject name.
14754 The X509_STORE API doesn't directly support the retrieval
14755 of multiple certificates matching a given criteria, however
14756 this can be worked round by performing a lookup first
14757 (which will fill the cache with candidate certificates)
14758 and then examining the cache for matches. This is probably
14759 the best we can do without throwing out X509_LOOKUP
14760 entirely (maybe later...).
14762 The X509_VERIFY_CTX structure has been enhanced considerably.
14764 All certificate lookup operations now go via a get_issuer()
14765 callback. Although this currently uses an X509_STORE it
14766 can be replaced by custom lookups. This is a simple way
14767 to bypass the X509_STORE hackery necessary to make this
14768 work and makes it possible to use more efficient techniques
14769 in future. A very simple version which uses a simple
14770 STACK for its trusted certificate store is also provided
14771 using X509_STORE_CTX_trusted_stack().
14773 The verify_cb() and verify() callbacks now have equivalents
14774 in the X509_STORE_CTX structure.
14776 X509_STORE_CTX also has a 'flags' field which can be used
14777 to customise the verify behaviour.
14781 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14782 excludes S/MIME capabilities.
14786 * When a certificate request is read in keep a copy of the
14787 original encoding of the signed data and use it when outputting
14788 again. Signatures then use the original encoding rather than
14789 a decoded, encoded version which may cause problems if the
14790 request is improperly encoded.
14794 * For consistency with other BIO_puts implementations, call
14795 buffer_write(b, ...) directly in buffer_puts instead of calling
14798 In BIO_puts, increment b->num_write as in BIO_write.
14800 *Peter.Sylvester@EdelWeb.fr*
14802 * Fix BN_mul_word for the case where the word is 0. (We have to use
14803 BN_zero, we may not return a BIGNUM with an array consisting of
14804 words set to zero.)
14808 * Avoid calling abort() from within the library when problems are
14809 detected, except if preprocessor symbols have been defined
14810 (such as REF_CHECK, BN_DEBUG etc.).
14814 * New openssl application 'rsautl'. This utility can be
14815 used for low level RSA operations. DER public key
14816 BIO/fp routines also added.
14820 * New Configure entry and patches for compiling on QNX 4.
14822 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14824 * A demo state-machine implementation was sponsored by
14825 Nuron (<http://www.nuron.com/>) and is now available in
14826 demos/state_machine.
14830 * New options added to the 'dgst' utility for signature
14831 generation and verification.
14835 * Unrecognized PKCS#7 content types are now handled via a
14836 catch all ASN1_TYPE structure. This allows unsupported
14837 types to be stored as a "blob" and an application can
14838 encode and decode it manually.
14842 * Fix various signed/unsigned issues to make a_strex.c
14843 compile under VC++.
14845 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14847 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14848 length if passed a buffer. ASN1_INTEGER_to_BN failed
14849 if passed a NULL BN and its argument was negative.
14851 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14853 * Modification to PKCS#7 encoding routines to output definite
14854 length encoding. Since currently the whole structures are in
14855 memory there's not real point in using indefinite length
14856 constructed encoding. However if OpenSSL is compiled with
14857 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14861 * Added BIO_vprintf() and BIO_vsnprintf().
14865 * Added more prefixes to parse for in the strings written
14866 through a logging bio, to cover all the levels that are available
14867 through syslog. The prefixes are now:
14869 PANIC, EMERG, EMR => LOG_EMERG
14870 ALERT, ALR => LOG_ALERT
14871 CRIT, CRI => LOG_CRIT
14872 ERROR, ERR => LOG_ERR
14873 WARNING, WARN, WAR => LOG_WARNING
14874 NOTICE, NOTE, NOT => LOG_NOTICE
14875 INFO, INF => LOG_INFO
14876 DEBUG, DBG => LOG_DEBUG
14878 and as before, if none of those prefixes are present at the
14879 beginning of the string, LOG_ERR is chosen.
14881 On Win32, the `LOG_*` levels are mapped according to this:
14883 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14884 LOG_WARNING => EVENTLOG_WARNING_TYPE
14885 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14889 * Made it possible to reconfigure with just the configuration
14890 argument "reconf" or "reconfigure". The command line arguments
14891 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14892 and are retrieved from there when reconfiguring.
14898 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14900 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14904 * The obj_dat.pl script was messing up the sorting of object
14905 names. The reason was that it compared the quoted version
14906 of strings as a result "OCSP" > "OCSP Signing" because
14907 " > SPACE. Changed script to store unquoted versions of
14908 names and add quotes on output. It was also omitting some
14909 names from the lookup table if they were given a default
14910 value (that is if SN is missing it is given the same
14911 value as LN and vice versa), these are now added on the
14912 grounds that if an object has a name we should be able to
14913 look it up. Finally added warning output when duplicate
14914 short or long names are found.
14918 * Changes needed for Tandem NSK.
14920 *Scott Uroff <scott@xypro.com>*
14922 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14923 RSA_padding_check_SSLv23(), special padding was never detected
14924 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14925 version rollback attacks was not effective.
14927 In s23_clnt.c, don't use special rollback-attack detection padding
14928 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14929 client; similarly, in s23_srvr.c, don't do the rollback check if
14930 SSL 2.0 is the only protocol enabled in the server.
14934 * Make it possible to get hexdumps of unprintable data with 'openssl
14935 asn1parse'. By implication, the functions ASN1_parse_dump() and
14936 BIO_dump_indent() are added.
14940 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14941 these print out strings and name structures based on various
14942 flags including RFC2253 support and proper handling of
14943 multibyte characters. Added options to the 'x509' utility
14944 to allow the various flags to be set.
14948 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14949 Also change the functions X509_cmp_current_time() and
14950 X509_gmtime_adj() work with an ASN1_TIME structure,
14951 this will enable certificates using GeneralizedTime in validity
14952 dates to be checked.
14956 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14957 negative public key encodings) on by default,
14958 NO_NEG_PUBKEY_BUG can be set to disable it.
14962 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14963 content octets. An i2c_ASN1_OBJECT is unnecessary because
14964 the encoding can be trivially obtained from the structure.
14968 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14969 not read locks (`CRYPTO_r_[un]lock`).
14973 * A first attempt at creating official support for shared
14974 libraries through configuration. I've kept it so the
14975 default is static libraries only, and the OpenSSL programs
14976 are always statically linked for now, but there are
14977 preparations for dynamic linking in place.
14978 This has been tested on Linux and Tru64.
14982 * Randomness polling function for Win9x, as described in:
14983 Peter Gutmann, Software Generation of Practically Strong
14988 * Fix so PRNG is seeded in req if using an already existing
14993 * New options to smime application. -inform and -outform
14994 allow alternative formats for the S/MIME message including
14995 PEM and DER. The -content option allows the content to be
14996 specified separately. This should allow things like Netscape
14997 form signing output easier to verify.
15001 * Fix the ASN1 encoding of tags using the 'long form'.
15005 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15006 STRING types. These convert content octets to and from the
15007 underlying type. The actual tag and length octets are
15008 already assumed to have been read in and checked. These
15009 are needed because all other string types have virtually
15010 identical handling apart from the tag. By having versions
15011 of the ASN1 functions that just operate on content octets
15012 IMPLICIT tagging can be handled properly. It also allows
15013 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15014 and ASN1_INTEGER are identical apart from the tag.
15018 * Change the handling of OID objects as follows:
15020 - New object identifiers are inserted in objects.txt, following
15021 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15022 - objects.pl is used to process obj_mac.num and create a new
15024 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15027 This is currently kind of a hack, and the perl code in objects.pl
15028 isn't very elegant, but it works as I intended. The simplest way
15029 to check that it worked correctly is to look in obj_dat.h and
15030 check the array nid_objs and make sure the objects haven't moved
15031 around (this is important!). Additions are OK, as well as
15032 consistent name changes.
15036 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15040 * Addition of the command line parameter '-rand file' to 'openssl req'.
15041 The given file adds to whatever has already been seeded into the
15042 random pool through the RANDFILE configuration file option or
15043 environment variable, or the default random state file.
15047 * mkstack.pl now sorts each macro group into lexical order.
15048 Previously the output order depended on the order the files
15049 appeared in the directory, resulting in needless rewriting
15054 * Patches to make OpenSSL compile under Win32 again. Mostly
15055 work arounds for the VC++ problem that it treats func() as
15056 func(void). Also stripped out the parts of mkdef.pl that
15057 added extra typesafe functions: these no longer exist.
15061 * Reorganisation of the stack code. The macros are now all
15062 collected in safestack.h . Each macro is defined in terms of
15063 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15064 DEBUG_SAFESTACK is now handled in terms of function casts,
15065 this has the advantage of retaining type safety without the
15066 use of additional functions. If DEBUG_SAFESTACK is not defined
15067 then the non typesafe macros are used instead. Also modified the
15068 mkstack.pl script to handle the new form. Needs testing to see
15069 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15070 the default if no major problems. Similar behaviour for ASN1_SET_OF
15071 and PKCS12_STACK_OF.
15075 * When some versions of IIS use the 'NET' form of private key the
15076 key derivation algorithm is different. Normally MD5(password) is
15077 used as a 128 bit RC4 key. In the modified case
15078 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15079 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15080 as the old Netscape_RSA functions except they have an additional
15081 'sgckey' parameter which uses the modified algorithm. Also added
15082 an -sgckey command line option to the rsa utility. Thanks to
15083 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15084 algorithm to openssl-dev.
15088 * The evp_local.h macros were using 'c.##kname' which resulted in
15089 invalid expansion on some systems (SCO 5.0.5 for example).
15090 Corrected to 'c.kname'.
15092 *Phillip Porch <root@theporch.com>*
15094 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15095 a STACK of email addresses from a certificate or request, these look
15096 in the subject name and the subject alternative name extensions and
15097 omit any duplicate addresses.
15101 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15102 This makes DSA verification about 2 % faster.
15106 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15107 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15108 plus overhead for 1024 bit moduli).
15109 This makes exponentiations about 0.5 % faster for 1024 bit
15110 exponents (as measured by "openssl speed rsa2048").
15114 * Rename memory handling macros to avoid conflicts with other
15116 Malloc => OPENSSL_malloc
15117 Malloc_locked => OPENSSL_malloc_locked
15118 Realloc => OPENSSL_realloc
15119 Free => OPENSSL_free
15123 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15124 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15128 * CygWin32 support.
15130 *John Jarvie <jjarvie@newsguy.com>*
15132 * The type-safe stack code has been rejigged. It is now only compiled
15133 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15134 by default all type-specific stack functions are "#define"d back to
15135 standard stack functions. This results in more streamlined output
15136 but retains the type-safety checking possibilities of the original
15141 * The STACK code has been cleaned up, and certain type declarations
15142 that didn't make a lot of sense have been brought in line. This has
15143 also involved a cleanup of sorts in safestack.h to more correctly
15144 map type-safe stack functions onto their plain stack counterparts.
15145 This work has also resulted in a variety of "const"ifications of
15146 lots of the code, especially `_cmp` operations which should normally
15147 be prototyped with "const" parameters anyway.
15151 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15152 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15153 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15154 where all of 'md' is used each time the PRNG is used, but 'state'
15155 is used only indexed by a cyclic counter. As entropy may not be
15156 well distributed from the beginning, 'md' is important as a
15157 chaining variable. However, the output function chains only half
15158 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15159 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15160 in all of 'state' being rewritten, with the new values depending
15161 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15165 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15166 the handshake is continued after ssl_verify_cert_chain();
15167 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15168 can lead to 'unexplainable' connection aborts later.
15170 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15172 * Major EVP API cipher revision.
15173 Add hooks for extra EVP features. This allows various cipher
15174 parameters to be set in the EVP interface. Support added for variable
15175 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15176 setting of RC2 and RC5 parameters.
15178 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15181 Remove lots of duplicated code from the EVP library. For example *every*
15182 cipher init() function handles the 'iv' in the same way according to the
15183 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15184 for CFB and OFB modes they zero ctx->num.
15186 New functionality allows removal of S/MIME code RC2 hack.
15188 Most of the routines have the same form and so can be declared in terms
15191 By shifting this to the top level EVP_CipherInit() it can be removed from
15192 all individual ciphers. If the cipher wants to handle IVs or keys
15193 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15196 Change lots of functions like EVP_EncryptUpdate() to now return a
15197 value: although software versions of the algorithms cannot fail
15198 any installed hardware versions can.
15202 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15203 this option is set, tolerate broken clients that send the negotiated
15204 protocol version number instead of the requested protocol version
15209 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15210 i.e. non-zero for export ciphersuites, zero otherwise.
15211 Previous versions had this flag inverted, inconsistent with
15212 rsa_tmp_cb (..._TMP_RSA_CB).
15214 *Bodo Moeller; problem reported by Amit Chopra*
15216 * Add missing DSA library text string. Work around for some IIS
15217 key files with invalid SEQUENCE encoding.
15221 * Add a document (doc/standards.txt) that list all kinds of standards
15222 and so on that are implemented in OpenSSL.
15226 * Enhance c_rehash script. Old version would mishandle certificates
15227 with the same subject name hash and wouldn't handle CRLs at all.
15228 Added -fingerprint option to crl utility, to support new c_rehash
15233 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15237 * Fix for SSL server purpose checking. Server checking was
15238 rejecting certificates which had extended key usage present
15239 but no ssl client purpose.
15241 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15243 * Make PKCS#12 code work with no password. The PKCS#12 spec
15244 is a little unclear about how a blank password is handled.
15245 Since the password in encoded as a BMPString with terminating
15246 double NULL a zero length password would end up as just the
15247 double NULL. However no password at all is different and is
15248 handled differently in the PKCS#12 key generation code. NS
15249 treats a blank password as zero length. MSIE treats it as no
15250 password on export: but it will try both on import. We now do
15251 the same: PKCS12_parse() tries zero length and no password if
15252 the password is set to "" or NULL (NULL is now a valid password:
15253 it wasn't before) as does the pkcs12 application.
15257 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15258 perror when PEM_read_bio_X509_REQ fails, the error message must
15259 be obtained from the error queue.
15263 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15264 it in ERR_remove_state if appropriate, and change ERR_get_state
15265 accordingly to avoid race conditions (this is necessary because
15266 thread_hash is no longer constant once set).
15270 * Bugfix for linux-elf makefile.one.
15274 * RSA_get_default_method() will now cause a default
15275 RSA_METHOD to be chosen if one doesn't exist already.
15276 Previously this was only set during a call to RSA_new()
15277 or RSA_new_method(NULL) meaning it was possible for
15278 RSA_get_default_method() to return NULL.
15282 * Added native name translation to the existing DSO code
15283 that will convert (if the flag to do so is set) filenames
15284 that are sufficiently small and have no path information
15285 into a canonical native form. Eg. "blah" converted to
15286 "libblah.so" or "blah.dll" etc.
15290 * New function ERR_error_string_n(e, buf, len) which is like
15291 ERR_error_string(e, buf), but writes at most 'len' bytes
15292 including the 0 terminator. For ERR_error_string_n, 'buf'
15295 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15297 * CONF library reworked to become more general. A new CONF
15298 configuration file reader "class" is implemented as well as a
15299 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15300 old `CONF_*` functions are still there, but are reimplemented to
15301 work in terms of the new functions. Also, a set of functions
15302 to handle the internal storage of the configuration data is
15303 provided to make it easier to write new configuration file
15304 reader "classes" (I can definitely see something reading a
15305 configuration file in XML format, for example), called `_CONF_*`,
15306 or "the configuration storage API"...
15308 The new configuration file reading functions are:
15310 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15311 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15313 NCONF_default, NCONF_WIN32
15315 NCONF_dump_fp, NCONF_dump_bio
15317 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15318 NCONF_new creates a new CONF object. This works in the same way
15319 as other interfaces in OpenSSL, like the BIO interface.
15320 `NCONF_dump_*` dump the internal storage of the configuration file,
15321 which is useful for debugging. All other functions take the same
15322 arguments as the old `CONF_*` functions with the exception of the
15323 first that must be a `CONF *` instead of a `LHASH *`.
15325 To make it easier to use the new classes with the old `CONF_*` functions,
15326 the function CONF_set_default_method is provided.
15330 * Add '-tls1' option to 'openssl ciphers', which was already
15331 mentioned in the documentation but had not been implemented.
15332 (This option is not yet really useful because even the additional
15333 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15337 * Initial DSO code added into libcrypto for letting OpenSSL (and
15338 OpenSSL-based applications) load shared libraries and bind to
15339 them in a portable way.
15341 *Geoff Thorpe, with contributions from Richard Levitte*
15343 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15345 * Make sure _lrotl and _lrotr are only used with MSVC.
15347 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15348 (the default implementation of RAND_status).
15350 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15351 to '-clrext' (= clear extensions), as intended and documented.
15352 *Bodo Moeller; inconsistency pointed out by Michael Attili
15353 <attili@amaxo.com>*
15355 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15356 was larger than the MD block size.
15358 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15360 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15361 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15362 using the passed key: if the passed key was a private key the result
15363 of X509_print(), for example, would be to print out all the private key
15368 * des_quad_cksum() byte order bug fix.
15369 *Ulf Möller, using the problem description in krb4-0.9.7, where
15370 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15372 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15375 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15377 * For easily testing in shell scripts whether some command
15378 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15379 returns with exit code 0 iff no command of the given name is available.
15380 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15381 the output goes to stdout and nothing is printed to stderr.
15382 Additional arguments are always ignored.
15384 Since for each cipher there is a command of the same name,
15385 the 'no-cipher' compilation switches can be tested this way.
15387 ('openssl no-XXX' is not able to detect pseudo-commands such
15388 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15392 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15396 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15397 is set; it will be thrown away anyway because each handshake creates
15399 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15400 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15401 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15402 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15406 * New s_client option -ign_eof: EOF at stdin is ignored, and
15407 'Q' and 'R' lose their special meanings (quit/renegotiate).
15408 This is part of what -quiet does; unlike -quiet, -ign_eof
15409 does not suppress any output.
15413 * Add compatibility options to the purpose and trust code. The
15414 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15415 accepts a certificate or CA, this was the previous behaviour,
15416 with all the associated security issues.
15418 X509_TRUST_COMPAT is the old trust behaviour: only and
15419 automatically trust self signed roots in certificate store. A
15420 new trust setting X509_TRUST_DEFAULT is used to specify that
15421 a purpose has no associated trust setting and it should instead
15422 use the value in the default purpose.
15426 * Fix the PKCS#8 DSA private key code so it decodes keys again
15427 and fix a memory leak.
15431 * In util/mkerr.pl (which implements 'make errors'), preserve
15432 reason strings from the previous version of the .c file, as
15433 the default to have only downcase letters (and digits) in
15434 automatically generated reasons codes is not always appropriate.
15438 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15439 using strerror. Previously, ERR_reason_error_string() returned
15440 library names as reason strings for SYSerr; but SYSerr is a special
15441 case where small numbers are errno values, not library numbers.
15445 * Add '-dsaparam' option to 'openssl dhparam' application. This
15446 converts DSA parameters into DH parameters. (When creating parameters,
15447 DSA_generate_parameters is used.)
15451 * Include 'length' (recommended exponent length) in C code generated
15452 by 'openssl dhparam -C'.
15456 * The second argument to set_label in perlasm was already being used
15457 so couldn't be used as a "file scope" flag. Moved to third argument
15462 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15463 instead of RAND_bytes for encryption IVs and salts.
15467 * Include RAND_status() into RAND_METHOD instead of implementing
15468 it only for md_rand.c Otherwise replacing the PRNG by calling
15469 RAND_set_rand_method would be impossible.
15473 * Don't let DSA_generate_key() enter an infinite loop if the random
15474 number generation fails.
15478 * New 'rand' application for creating pseudo-random output.
15482 * Added configuration support for Linux/IA64
15484 *Rolf Haberrecker <rolf@suse.de>*
15486 * Assembler module support for Mingw32.
15490 * Shared library support for HPUX (in shlib/).
15492 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15494 * Shared library support for Solaris gcc.
15496 *Lutz Behnke <behnke@trustcenter.de>*
15498 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15500 * PKCS7_encrypt() was adding text MIME headers twice because they
15501 were added manually and by SMIME_crlf_copy().
15505 * In bntest.c don't call BN_rand with zero bits argument.
15507 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15509 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15510 case was implemented. This caused BN_div_recp() to fail occasionally.
15514 * Add an optional second argument to the set_label() in the perl
15515 assembly language builder. If this argument exists and is set
15516 to 1 it signals that the assembler should use a symbol whose
15517 scope is the entire file, not just the current function. This
15518 is needed with MASM which uses the format label:: for this scope.
15520 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15522 * Change the ASN1 types so they are typedefs by default. Before
15523 almost all types were #define'd to ASN1_STRING which was causing
15524 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15529 * Change names of new functions to the new get1/get0 naming
15530 convention: After 'get1', the caller owns a reference count
15531 and has to call `..._free`; 'get0' returns a pointer to some
15532 data structure without incrementing reference counters.
15533 (Some of the existing 'get' functions increment a reference
15534 counter, some don't.)
15535 Similarly, 'set1' and 'add1' functions increase reference
15536 counters or duplicate objects.
15540 * Allow for the possibility of temp RSA key generation failure:
15541 the code used to assume it always worked and crashed on failure.
15545 * Fix potential buffer overrun problem in BIO_printf().
15546 *Ulf Möller, using public domain code by Patrick Powell; problem
15547 pointed out by David Sacerdote <das33@cornell.edu>*
15549 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15550 RAND_egd() and RAND_status(). In the command line application,
15551 the EGD socket can be specified like a seed file using RANDFILE
15556 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15557 Some CAs (e.g. Verisign) distribute certificates in this form.
15561 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15562 list to exclude them. This means that no special compilation option
15563 is needed to use anonymous DH: it just needs to be included in the
15568 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15569 EVP_MD_type. The old functionality is available in a new macro called
15570 EVP_MD_md(). Change code that uses it and update docs.
15574 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15575 where the `void *` argument is replaced by a function pointer argument.
15576 Previously `void *` was abused to point to functions, which works on
15577 many platforms, but is not correct. As these functions are usually
15578 called by macros defined in OpenSSL header files, most source code
15579 should work without changes.
15583 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15584 sections with information on -D... compiler switches used for
15585 compiling the library so that applications can see them. To enable
15586 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15587 must be defined. E.g.,
15588 #define OPENSSL_ALGORITHM_DEFINES
15589 #include <openssl/opensslconf.h>
15590 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15592 *Richard Levitte, Ulf and Bodo Möller*
15594 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15599 * Change the 'other' type in certificate aux info to a STACK_OF
15600 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15601 the required ASN1 format: arbitrary types determined by an OID.
15605 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15606 argument to 'req'. This is not because the function is newer or
15607 better than others it just uses the work 'NEW' in the certificate
15608 request header lines. Some software needs this.
15612 * Reorganise password command line arguments: now passwords can be
15613 obtained from various sources. Delete the PEM_cb function and make
15614 it the default behaviour: i.e. if the callback is NULL and the
15615 usrdata argument is not NULL interpret it as a null terminated pass
15616 phrase. If usrdata and the callback are NULL then the pass phrase
15617 is prompted for as usual.
15621 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15622 the support is automatically enabled. The resulting binaries will
15623 autodetect the card and use it if present.
15625 *Ben Laurie and Compaq Inc.*
15627 * Work around for Netscape hang bug. This sends certificate request
15628 and server done in one record. Since this is perfectly legal in the
15629 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15630 the bugs/SSLv3 entry for more info.
15634 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15638 * Add -rand argument to smime and pkcs12 applications and read/write
15643 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15647 * Add command line password options to the remaining applications.
15651 * Bug fix for BN_div_recp() for numerators with an even number of
15656 * More tests in bntest.c, and changed test_bn output.
15660 * ./config recognizes MacOS X now.
15664 * Bug fix for BN_div() when the first words of num and divisor are
15665 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15669 * Add support for various broken PKCS#8 formats, and command line
15670 options to produce them.
15674 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15675 get temporary BIGNUMs from a BN_CTX.
15679 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15684 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15685 include a #define from the old name to the new. The original intent
15686 was that statically linked binaries could for example just call
15687 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15688 link with digests. This never worked because SSLeay_add_all_digests()
15689 and SSLeay_add_all_ciphers() were in the same source file so calling
15690 one would link with the other. They are now in separate source files.
15694 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15698 * Use a less unusual form of the Miller-Rabin primality test (it used
15699 a binary algorithm for exponentiation integrated into the Miller-Rabin
15700 loop, our standard modexp algorithms are faster).
15704 * Support for the EBCDIC character set completed.
15706 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15708 * Source code cleanups: use const where appropriate, eliminate casts,
15709 use `void *` instead of `char *` in lhash.
15713 * Bugfix: ssl3_send_server_key_exchange was not restartable
15714 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15715 this the server could overwrite ephemeral keys that the client
15720 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15721 using 50 iterations of the Rabin-Miller test.
15723 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15724 iterations of the Rabin-Miller test as required by the appendix
15725 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15726 As BN_is_prime_fasttest includes trial division, DSA parameter
15727 generation becomes much faster.
15729 This implies a change for the callback functions in DSA_is_prime
15730 and DSA_generate_parameters: The callback function is called once
15731 for each positive witness in the Rabin-Miller test, not just
15732 occasionally in the inner loop; and the parameters to the
15733 callback function now provide an iteration count for the outer
15734 loop rather than for the current invocation of the inner loop.
15735 DSA_generate_parameters additionally can call the callback
15736 function with an 'iteration count' of -1, meaning that a
15737 candidate has passed the trial division test (when q is generated
15738 from an application-provided seed, trial division is skipped).
15742 * New function BN_is_prime_fasttest that optionally does trial
15743 division before starting the Rabin-Miller test and has
15744 an additional BN_CTX * argument (whereas BN_is_prime always
15745 has to allocate at least one BN_CTX).
15746 'callback(1, -1, cb_arg)' is called when a number has passed the
15747 trial division stage.
15751 * Fix for bug in CRL encoding. The validity dates weren't being handled
15756 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15760 * New function BN_pseudo_rand().
15764 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15765 bignum version of BN_from_montgomery() with the working code from
15766 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15771 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15772 made it impossible to use the same SSL_SESSION data structure in
15773 SSL2 clients in multiple threads.
15777 * The return value of RAND_load_file() no longer counts bytes obtained
15778 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15779 to seed the PRNG (previously an explicit byte count was required).
15781 *Ulf Möller, Bodo Möller*
15783 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15784 used `char *` instead of `void *` and had casts all over the place.
15788 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15792 * Retain source code compatibility for BN_prime_checks macro:
15793 BN_is_prime(..., BN_prime_checks, ...) now uses
15794 BN_prime_checks_for_size to determine the appropriate number of
15795 Rabin-Miller iterations.
15799 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15800 DH_CHECK_P_NOT_SAFE_PRIME.
15801 (Check if this is true? OpenPGP calls them "strong".)
15805 * Merge the functionality of "dh" and "gendh" programs into a new program
15806 "dhparam". The old programs are retained for now but will handle DH keys
15807 (instead of parameters) in future.
15811 * Make the ciphers, s_server and s_client programs check the return values
15812 when a new cipher list is set.
15816 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15817 ciphers. Before when the 56bit ciphers were enabled the sorting was
15820 The syntax for the cipher sorting has been extended to support sorting by
15821 cipher-strength (using the strength_bits hard coded in the tables).
15822 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15824 Fix a bug in the cipher-command parser: when supplying a cipher command
15825 string with an "undefined" symbol (neither command nor alphanumeric
15826 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15827 an error is flagged.
15829 Due to the strength-sorting extension, the code of the
15830 ssl_create_cipher_list() function was completely rearranged. I hope that
15831 the readability was also increased :-)
15833 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15835 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15836 for the first serial number and places 2 in the serial number file. This
15837 avoids problems when the root CA is created with serial number zero and
15838 the first user certificate has the same issuer name and serial number
15843 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15844 the new code. Add documentation for this stuff.
15848 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15849 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15850 structures and behave in an analogous way to the X509v3 functions:
15851 they shouldn't be called directly but wrapper functions should be used
15854 So we also now have some wrapper functions that call the X509at functions
15855 when passed certificate requests. (TO DO: similar things can be done with
15856 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15857 things. Some of these need some d2i or i2d and print functionality
15858 because they handle more complex structures.)
15862 * Add missing #ifndefs that caused missing symbols when building libssl
15863 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15864 NO_RSA in `ssl/s2*.c`.
15866 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15868 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15869 has a return value which indicates the quality of the random data
15870 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15871 error queue. New function RAND_pseudo_bytes() generates output that is
15872 guaranteed to be unique but not unpredictable. RAND_add is like
15873 RAND_seed, but takes an extra argument for an entropy estimate
15874 (RAND_seed always assumes full entropy).
15878 * Do more iterations of Rabin-Miller probable prime test (specifically,
15879 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15880 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15881 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15882 false-positive rate of at most 2^-80 for random input.
15886 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15890 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15891 in the 0.9.5 release), this returns the chain
15892 from an X509_CTX structure with a dup of the stack and all
15893 the X509 reference counts upped: so the stack will exist
15894 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15897 Also make SSL_SESSION_print() print out the verify return
15902 * Add manpage for the pkcs12 command. Also change the default
15903 behaviour so MAC iteration counts are used unless the new
15904 -nomaciter option is used. This improves file security and
15905 only older versions of MSIE (4.0 for example) need it.
15909 * Honor the no-xxx Configure options when creating .DEF files.
15913 * Add PKCS#10 attributes to field table: challengePassword,
15914 unstructuredName and unstructuredAddress. These are taken from
15915 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15916 international characters are used.
15918 More changes to X509_ATTRIBUTE code: allow the setting of types
15919 based on strings. Remove the 'loc' parameter when adding
15920 attributes because these will be a SET OF encoding which is sorted
15925 * Initial changes to the 'req' utility to allow request generation
15926 automation. This will allow an application to just generate a template
15927 file containing all the field values and have req construct the
15930 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15931 used all over the place including certificate requests and PKCS#7
15932 structures. They are currently handled manually where necessary with
15933 some primitive wrappers for PKCS#7. The new functions behave in a
15934 manner analogous to the X509 extension functions: they allow
15935 attributes to be looked up by NID and added.
15937 Later something similar to the X509V3 code would be desirable to
15938 automatically handle the encoding, decoding and printing of the
15939 more complex types. The string types like challengePassword can
15940 be handled by the string table functions.
15942 Also modified the multi byte string table handling. Now there is
15943 a 'global mask' which masks out certain types. The table itself
15944 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15945 is useful when for example there is only one permissible type
15946 (as in countryName) and using the mask might result in no valid
15951 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15952 SSL_get_peer_finished to allow applications to obtain the latest
15953 Finished messages sent to the peer or expected from the peer,
15954 respectively. (SSL_get_peer_finished is usually the Finished message
15955 actually received from the peer, otherwise the protocol will be aborted.)
15957 As the Finished message are message digests of the complete handshake
15958 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15959 be used for external authentication procedures when the authentication
15960 provided by SSL/TLS is not desired or is not enough.
15964 * Enhanced support for Alpha Linux is added. Now ./config checks if
15965 the host supports BWX extension and if Compaq C is present on the
15966 $PATH. Just exploiting of the BWX extension results in 20-30%
15967 performance kick for some algorithms, e.g. DES and RC4 to mention
15968 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15973 * Add support for MS "fast SGC". This is arguably a violation of the
15974 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15975 weak crypto and after checking the certificate is SGC a second one
15976 with strong crypto. MS SGC stops the first handshake after receiving
15977 the server certificate message and sends a second client hello. Since
15978 a server will typically do all the time consuming operations before
15979 expecting any further messages from the client (server key exchange
15980 is the most expensive) there is little difference between the two.
15982 To get OpenSSL to support MS SGC we have to permit a second client
15983 hello message after we have sent server done. In addition we have to
15984 reset the MAC if we do get this second client hello.
15988 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
15989 if a DER encoded private key is RSA or DSA traditional format. Changed
15990 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
15991 format DER encoded private key. Newer code should use PKCS#8 format which
15992 has the key type encoded in the ASN1 structure. Added DER private key
15993 support to pkcs8 application.
15997 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
15998 ciphersuites has been selected (as required by the SSL 3/TLS 1
15999 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16000 is set, we interpret this as a request to violate the specification
16001 (the worst that can happen is a handshake failure, and 'correct'
16002 behaviour would result in a handshake failure anyway).
16006 * In SSL_CTX_add_session, take into account that there might be multiple
16007 SSL_SESSION structures with the same session ID (e.g. when two threads
16008 concurrently obtain them from an external cache).
16009 The internal cache can handle only one SSL_SESSION with a given ID,
16010 so if there's a conflict, we now throw out the old one to achieve
16015 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16016 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16017 some routines that use cipher OIDs: some ciphers do not have OIDs
16018 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16023 * Simplify the trust setting structure and code. Now we just have
16024 two sequences of OIDs for trusted and rejected settings. These will
16025 typically have values the same as the extended key usage extension
16026 and any application specific purposes.
16028 The trust checking code now has a default behaviour: it will just
16029 check for an object with the same NID as the passed id. Functions can
16030 be provided to override either the default behaviour or the behaviour
16031 for a given id. SSL client, server and email already have functions
16032 in place for compatibility: they check the NID and also return "trusted"
16033 if the certificate is self signed.
16037 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16038 traditional format into an EVP_PKEY structure.
16042 * Add a password callback function PEM_cb() which either prompts for
16043 a password if usr_data is NULL or otherwise assumes it is a null
16044 terminated password. Allow passwords to be passed on command line
16045 environment or config files in a few more utilities.
16049 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16050 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16051 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16052 Update documentation.
16056 * Support for ASN1 "NULL" type. This could be handled before by using
16057 ASN1_TYPE but there wasn't any function that would try to read a NULL
16058 and produce an error if it couldn't. For compatibility we also have
16059 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16060 don't allocate anything because they don't need to.
16064 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16067 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16069 * Rebuild of the memory allocation routines used by OpenSSL code and
16070 possibly others as well. The purpose is to make an interface that
16071 provide hooks so anyone can build a separate set of allocation and
16072 deallocation routines to be used by OpenSSL, for example memory
16073 pool implementations, or something else, which was previously hard
16074 since Malloc(), Realloc() and Free() were defined as macros having
16075 the values malloc, realloc and free, respectively (except for Win32
16076 compilations). The same is provided for memory debugging code.
16077 OpenSSL already comes with functionality to find memory leaks, but
16078 this gives people a chance to debug other memory problems.
16080 With these changes, a new set of functions and macros have appeared:
16082 CRYPTO_set_mem_debug_functions() [F]
16083 CRYPTO_get_mem_debug_functions() [F]
16084 CRYPTO_dbg_set_options() [F]
16085 CRYPTO_dbg_get_options() [F]
16086 CRYPTO_malloc_debug_init() [M]
16088 The memory debug functions are NULL by default, unless the library
16089 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16090 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16091 gives the standard debugging functions that come with OpenSSL) or
16092 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16093 provided by the library user) must be used. When the standard
16094 debugging functions are used, CRYPTO_dbg_set_options can be used to
16095 request additional information:
16096 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16097 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16099 Also, things like CRYPTO_set_mem_functions will always give the
16100 expected result (the new set of functions is used for allocation
16101 and deallocation) at all times, regardless of platform and compiler
16104 To finish it up, some functions that were never use in any other
16105 way than through macros have a new API and new semantic:
16107 CRYPTO_dbg_malloc()
16108 CRYPTO_dbg_realloc()
16111 All macros of value have retained their old syntax.
16113 *Richard Levitte and Bodo Moeller*
16115 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16116 ordering of SMIMECapabilities wasn't in "strength order" and there
16117 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16122 * Some ASN1 types with illegal zero length encoding (INTEGER,
16123 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16125 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16127 * Merge in my S/MIME library for OpenSSL. This provides a simple
16128 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16129 functionality to handle multipart/signed properly) and a utility
16130 called 'smime' to call all this stuff. This is based on code I
16131 originally wrote for Celo who have kindly allowed it to be
16132 included in OpenSSL.
16136 * Add variants des_set_key_checked and des_set_key_unchecked of
16137 des_set_key (aka des_key_sched). Global variable des_check_key
16138 decides which of these is called by des_set_key; this way
16139 des_check_key behaves as it always did, but applications and
16140 the library itself, which was buggy for des_check_key == 1,
16141 have a cleaner way to pick the version they need.
16145 * New function PKCS12_newpass() which changes the password of a
16150 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16151 dynamic mix. In both cases the ids can be used as an index into the
16152 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16153 functions so they accept a list of the field values and the
16154 application doesn't need to directly manipulate the X509_TRUST
16159 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16164 * Modify the way the V3 extension code looks up extensions. This now
16165 works in a similar way to the object code: we have some "standard"
16166 extensions in a static table which is searched with OBJ_bsearch()
16167 and the application can add dynamic ones if needed. The file
16168 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16169 updated whenever a new extension is added to the core code and kept
16170 in ext_nid order. There is a simple program 'tabtest.c' which checks
16171 this. New extensions are not added too often so this file can readily
16172 be maintained manually.
16174 There are two big advantages in doing things this way. The extensions
16175 can be looked up immediately and no longer need to be "added" using
16176 X509V3_add_standard_extensions(): this function now does nothing.
16177 Side note: I get *lots* of email saying the extension code doesn't
16178 work because people forget to call this function.
16179 Also no dynamic allocation is done unless new extensions are added:
16180 so if we don't add custom extensions there is no need to call
16181 X509V3_EXT_cleanup().
16185 * Modify enc utility's salting as follows: make salting the default. Add a
16186 magic header, so unsalted files fail gracefully instead of just decrypting
16187 to garbage. This is because not salting is a big security hole, so people
16188 should be discouraged from doing it.
16192 * Fixes and enhancements to the 'x509' utility. It allowed a message
16193 digest to be passed on the command line but it only used this
16194 parameter when signing a certificate. Modified so all relevant
16195 operations are affected by the digest parameter including the
16196 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16197 DSA key was used because it didn't fix the digest.
16201 * Initial certificate chain verify code. Currently tests the untrusted
16202 certificates for consistency with the verify purpose (which is set
16203 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16205 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16206 this is because it will reject chains with invalid extensions whereas
16207 every previous version of OpenSSL and SSLeay made no checks at all.
16209 Trust code: checks the root CA for the relevant trust settings. Trust
16210 settings have an initial value consistent with the verify purpose: e.g.
16211 if the verify purpose is for SSL client use it expects the CA to be
16212 trusted for SSL client use. However the default value can be changed to
16213 permit custom trust settings: one example of this would be to only trust
16214 certificates from a specific "secure" set of CAs.
16216 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16217 which should be used for version portability: especially since the
16218 verify structure is likely to change more often now.
16220 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16221 to set them. If not set then assume SSL clients will verify SSL servers
16224 Two new options to the verify program: -untrusted allows a set of
16225 untrusted certificates to be passed in and -purpose which sets the
16226 intended purpose of the certificate. If a purpose is set then the
16227 new chain verify code is used to check extension consistency.
16231 * Support for the authority information access extension.
16235 * Modify RSA and DSA PEM read routines to transparently handle
16236 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16237 public keys in a format compatible with certificate
16238 SubjectPublicKeyInfo structures. Unfortunately there were already
16239 functions called *_PublicKey_* which used various odd formats so
16240 these are retained for compatibility: however the DSA variants were
16241 never in a public release so they have been deleted. Changed dsa/rsa
16242 utilities to handle the new format: note no releases ever handled public
16243 keys so we should be OK.
16245 The primary motivation for this change is to avoid the same fiasco
16246 that dogs private keys: there are several incompatible private key
16247 formats some of which are standard and some OpenSSL specific and
16248 require various evil hacks to allow partial transparent handling and
16249 even then it doesn't work with DER formats. Given the option anything
16250 other than PKCS#8 should be dumped: but the other formats have to
16251 stay in the name of compatibility.
16253 With public keys and the benefit of hindsight one standard format
16254 is used which works with EVP_PKEY, RSA or DSA structures: though
16255 it clearly returns an error if you try to read the wrong kind of key.
16257 Added a -pubkey option to the 'x509' utility to output the public key.
16258 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16259 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16260 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16261 that do the same as the `EVP_PKEY_assign_*()` except they up the
16262 reference count of the added key (they don't "swallow" the
16267 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16268 CRLs would fail if the file contained no certificates or no CRLs:
16269 added a new function to read in both types and return the number
16270 read: this means that if none are read it will be an error. The
16271 DER versions of the certificate and CRL reader would always fail
16272 because it isn't possible to mix certificates and CRLs in DER format
16273 without choking one or the other routine. Changed this to just read
16274 a certificate: this is the best we can do. Also modified the code
16275 in apps/verify.c to take notice of return codes: it was previously
16276 attempting to read in certificates from NULL pointers and ignoring
16277 any errors: this is one reason why the cert and CRL reader seemed
16278 to work. It doesn't check return codes from the default certificate
16279 routines: these may well fail if the certificates aren't installed.
16283 * Code to support otherName option in GeneralName.
16287 * First update to verify code. Change the verify utility
16288 so it warns if it is passed a self signed certificate:
16289 for consistency with the normal behaviour. X509_verify
16290 has been modified to it will now verify a self signed
16291 certificate if *exactly* the same certificate appears
16292 in the store: it was previously impossible to trust a
16293 single self signed certificate. This means that:
16294 openssl verify ss.pem
16295 now gives a warning about a self signed certificate but
16296 openssl verify -CAfile ss.pem ss.pem
16301 * For servers, store verify_result in SSL_SESSION data structure
16302 (and add it to external session representation).
16303 This is needed when client certificate verifications fails,
16304 but an application-provided verification callback (set by
16305 SSL_CTX_set_cert_verify_callback) allows accepting the session
16306 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16307 but returns 1): When the session is reused, we have to set
16308 ssl->verify_result to the appropriate error code to avoid
16311 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16313 * Fix a bug in the new PKCS#7 code: it didn't consider the
16314 case in PKCS7_dataInit() where the signed PKCS7 structure
16315 didn't contain any existing data because it was being created.
16317 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16319 * Add a salt to the key derivation routines in enc.c. This
16320 forms the first 8 bytes of the encrypted file. Also add a
16321 -S option to allow a salt to be input on the command line.
16325 * New function X509_cmp(). Oddly enough there wasn't a function
16326 to compare two certificates. We do this by working out the SHA1
16327 hash and comparing that. X509_cmp() will be needed by the trust
16332 * SSL_get1_session() is like SSL_get_session(), but increments
16333 the reference count in the SSL_SESSION returned.
16335 *Geoff Thorpe <geoff@eu.c2.net>*
16337 * Fix for 'req': it was adding a null to request attributes.
16338 Also change the X509_LOOKUP and X509_INFO code to handle
16339 certificate auxiliary information.
16343 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16348 * Add the possibility to add extra information to the memory leak
16349 detecting output, to form tracebacks, showing from where each
16350 allocation was originated: CRYPTO_push_info("constant string") adds
16351 the string plus current file name and line number to a per-thread
16352 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16353 is like calling CYRPTO_pop_info() until the stack is empty.
16354 Also updated memory leak detection code to be multi-thread-safe.
16358 * Add options -text and -noout to pkcs7 utility and delete the
16359 encryption options which never did anything. Update docs.
16363 * Add options to some of the utilities to allow the pass phrase
16364 to be included on either the command line (not recommended on
16365 OSes like Unix) or read from the environment. Update the
16366 manpages and fix a few bugs.
16370 * Add a few manpages for some of the openssl commands.
16374 * Fix the -revoke option in ca. It was freeing up memory twice,
16375 leaking and not finding already revoked certificates.
16379 * Extensive changes to support certificate auxiliary information.
16380 This involves the use of X509_CERT_AUX structure and X509_AUX
16381 functions. An X509_AUX function such as PEM_read_X509_AUX()
16382 can still read in a certificate file in the usual way but it
16383 will also read in any additional "auxiliary information". By
16384 doing things this way a fair degree of compatibility can be
16385 retained: existing certificates can have this information added
16386 using the new 'x509' options.
16388 Current auxiliary information includes an "alias" and some trust
16389 settings. The trust settings will ultimately be used in enhanced
16390 certificate chain verification routines: currently a certificate
16391 can only be trusted if it is self signed and then it is trusted
16396 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16397 The problem was that one of the replacement routines had not been working
16398 since SSLeay releases. For now the offending routine has been replaced
16399 with non-optimised assembler. Even so, this now gives around 95%
16400 performance improvement for 1024 bit RSA signs.
16404 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16405 handling. Most clients have the effective key size in bits equal to
16406 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16407 A few however don't do this and instead use the size of the decrypted key
16408 to determine the RC2 key length and the AlgorithmIdentifier to determine
16409 the effective key length. In this case the effective key length can still
16410 be 40 bits but the key length can be 168 bits for example. This is fixed
16411 by manually forcing an RC2 key into the EVP_PKEY structure because the
16412 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16413 the key length and effective key length are equal.
16417 * Add a bunch of functions that should simplify the creation of
16418 X509_NAME structures. Now you should be able to do:
16419 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16420 and have it automatically work out the correct field type and fill in
16421 the structures. The more adventurous can try:
16422 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16423 and it will (hopefully) work out the correct multibyte encoding.
16427 * Change the 'req' utility to use the new field handling and multibyte
16428 copy routines. Before the DN field creation was handled in an ad hoc
16429 way in req, ca, and x509 which was rather broken and didn't support
16430 BMPStrings or UTF8Strings. Since some software doesn't implement
16431 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16432 using the dirstring_type option. See the new comment in the default
16433 openssl.cnf for more info.
16437 * Make crypto/rand/md_rand.c more robust:
16438 - Assure unique random numbers after fork().
16439 - Make sure that concurrent threads access the global counter and
16440 md serializably so that we never lose entropy in them
16441 or use exactly the same state in multiple threads.
16442 Access to the large state is not always serializable because
16443 the additional locking could be a performance killer, and
16444 md should be large enough anyway.
16448 * New file apps/app_rand.c with commonly needed functionality
16449 for handling the random seed file.
16451 Use the random seed file in some applications that previously did not:
16453 dsaparam -genkey (which also ignored its '-rand' option),
16456 x509 (when signing).
16457 Except on systems with /dev/urandom, it is crucial to have a random
16458 seed file at least for key creation, DSA signing, and for DH exchanges;
16459 for RSA signatures we could do without one.
16461 gendh and gendsa (unlike genrsa) used to read only the first byte
16462 of each file listed in the '-rand' option. The function as previously
16463 found in genrsa is now in app_rand.c and is used by all programs
16464 that support '-rand'.
16468 * In RAND_write_file, use mode 0600 for creating files;
16469 don't just chmod when it may be too late.
16473 * Report an error from X509_STORE_load_locations
16474 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16478 * New function ASN1_mbstring_copy() this copies a string in either
16479 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16480 into an ASN1_STRING type. A mask of permissible types is passed
16481 and it chooses the "minimal" type to use or an error if not type
16486 * Add function equivalents to the various macros in asn1.h. The old
16487 macros are retained with an `M_` prefix. Code inside the library can
16488 use the `M_` macros. External code (including the openssl utility)
16489 should *NOT* in order to be "shared library friendly".
16493 * Add various functions that can check a certificate's extensions
16494 to see if it usable for various purposes such as SSL client,
16495 server or S/MIME and CAs of these types. This is currently
16496 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16497 verification. Also added a -purpose flag to x509 utility to
16498 print out all the purposes.
16502 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16507 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16508 for, obtain and decode and extension and obtain its critical flag.
16509 This allows all the necessary extension code to be handled in a
16510 single function call.
16514 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16515 platforms. See crypto/rc4/rc4_enc.c for further details.
16519 * New -noout option to asn1parse. This causes no output to be produced
16520 its main use is when combined with -strparse and -out to extract data
16521 from a file (which may not be in ASN.1 format).
16525 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16526 when producing the local key id.
16528 *Richard Levitte <levitte@stacken.kth.se>*
16530 * New option -dhparam in s_server. This allows a DH parameter file to be
16531 stated explicitly. If it is not stated then it tries the first server
16532 certificate file. The previous behaviour hard coded the filename
16537 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16538 a public key to be input or output. For example:
16539 openssl rsa -in key.pem -pubout -out pubkey.pem
16540 Also added necessary DSA public key functions to handle this.
16544 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16545 in the message. This was handled by allowing
16546 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16548 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16550 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16551 to the end of the strings whereas this didn't. This would cause problems
16552 if strings read with d2i_ASN1_bytes() were later modified.
16554 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16556 * Fix for base64 decode bug. When a base64 bio reads only one line of
16557 data and it contains EOF it will end up returning an error. This is
16558 caused by input 46 bytes long. The cause is due to the way base64
16559 BIOs find the start of base64 encoded data. They do this by trying a
16560 trial decode on each line until they find one that works. When they
16561 do a flag is set and it starts again knowing it can pass all the
16562 data directly through the decoder. Unfortunately it doesn't reset
16563 the context it uses. This means that if EOF is reached an attempt
16564 is made to pass two EOFs through the context and this causes the
16565 resulting error. This can also cause other problems as well. As is
16566 usual with these problems it takes *ages* to find and the fix is
16567 trivial: move one line.
16569 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16571 * Ugly workaround to get s_client and s_server working under Windows. The
16572 old code wouldn't work because it needed to select() on sockets and the
16573 tty (for keypresses and to see if data could be written). Win32 only
16574 supports select() on sockets so we select() with a 1s timeout on the
16575 sockets and then see if any characters are waiting to be read, if none
16576 are present then we retry, we also assume we can always write data to
16577 the tty. This isn't nice because the code then blocks until we've
16578 received a complete line of data and it is effectively polling the
16579 keyboard at 1s intervals: however it's quite a bit better than not
16580 working at all :-) A dedicated Windows application might handle this
16581 with an event loop for example.
16585 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16586 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16587 will be called when RSA_sign() and RSA_verify() are used. This is useful
16588 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16589 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16590 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16591 This necessitated the support of an extra signature type NID_md5_sha1
16592 for SSL signatures and modifications to the SSL library to use it instead
16593 of calling RSA_public_decrypt() and RSA_private_encrypt().
16597 * Add new -verify -CAfile and -CApath options to the crl program, these
16598 will lookup a CRL issuers certificate and verify the signature in a
16599 similar way to the verify program. Tidy up the crl program so it
16600 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16601 less strict. It will now permit CRL extensions even if it is not
16602 a V2 CRL: this will allow it to tolerate some broken CRLs.
16606 * Initialize all non-automatic variables each time one of the openssl
16607 sub-programs is started (this is necessary as they may be started
16608 multiple times from the "OpenSSL>" prompt).
16610 *Lennart Bang, Bodo Moeller*
16612 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16613 removing all other RSA functionality (this is what NO_RSA does). This
16614 is so (for example) those in the US can disable those operations covered
16615 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16620 * Non-copying interface to BIO pairs.
16621 (still largely untested)
16625 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16626 ASCII string. This was handled independently in various places before.
16630 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16631 UTF8 strings a character at a time.
16635 * Use client_version from client hello to select the protocol
16636 (s23_srvr.c) and for RSA client key exchange verification
16637 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16641 * Add various utility functions to handle SPKACs, these were previously
16642 handled by poking round in the structure internals. Added new function
16643 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16644 print, verify and generate SPKACs. Based on an original idea from
16645 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16649 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16653 * Allow the config file extension section to be overwritten on the
16654 command line. Based on an original idea from Massimiliano Pala
16655 <madwolf@comune.modena.it>. The new option is called -extensions
16656 and can be applied to ca, req and x509. Also -reqexts to override
16657 the request extensions in req and -crlexts to override the crl extensions
16662 * Add new feature to the SPKAC handling in ca. Now you can include
16663 the same field multiple times by preceding it by "XXXX." for example:
16666 this is the same syntax as used in the req config file.
16670 * Allow certificate extensions to be added to certificate requests. These
16671 are specified in a 'req_extensions' option of the req section of the
16672 config file. They can be printed out with the -text option to req but
16673 are otherwise ignored at present.
16677 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16678 data read consists of only the final block it would not decrypted because
16679 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16680 A misplaced 'break' also meant the decrypted final block might not be
16681 copied until the next read.
16685 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16686 a few extra parameters to the DH structure: these will be useful if
16687 for example we want the value of 'q' or implement X9.42 DH.
16691 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16692 provides hooks that allow the default DSA functions or functions on a
16693 "per key" basis to be replaced. This allows hardware acceleration and
16694 hardware key storage to be handled without major modification to the
16695 library. Also added low level modexp hooks and CRYPTO_EX structure and
16696 associated functions.
16700 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16701 as "read only": it can't be written to and the buffer it points to will
16702 not be freed. Reading from a read only BIO is much more efficient than
16703 a normal memory BIO. This was added because there are several times when
16704 an area of memory needs to be read from a BIO. The previous method was
16705 to create a memory BIO and write the data to it, this results in two
16706 copies of the data and an O(n^2) reading algorithm. There is a new
16707 function BIO_new_mem_buf() which creates a read only memory BIO from
16708 an area of memory. Also modified the PKCS#7 routines to use read only
16713 * Bugfix: ssl23_get_client_hello did not work properly when called in
16714 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16715 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16716 but a retry condition occurred while trying to read the rest.
16720 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16721 NID_pkcs7_encrypted by default: this was wrong since this should almost
16722 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16723 the encrypted data type: this is a more sensible place to put it and it
16724 allows the PKCS#12 code to be tidied up that duplicated this
16729 * Changed obj_dat.pl script so it takes its input and output files on
16730 the command line. This should avoid shell escape redirection problems
16735 * Initial support for certificate extension requests, these are included
16736 in things like Xenroll certificate requests. Included functions to allow
16737 extensions to be obtained and added.
16741 * -crlf option to s_client and s_server for sending newlines as
16742 CRLF (as required by many protocols).
16746 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16748 * Install libRSAglue.a when OpenSSL is built with RSAref.
16750 *Ralf S. Engelschall*
16752 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16754 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16756 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16761 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16762 DH parameters/keys (q is lost during that conversion, but the resulting
16763 DH parameters contain its length).
16765 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16766 much faster than DH_generate_parameters (which creates parameters
16767 where `p = 2*q + 1`), and also the smaller q makes DH computations
16768 much more efficient (160-bit exponentiation instead of 1024-bit
16769 exponentiation); so this provides a convenient way to support DHE
16770 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16771 utter importance to use
16772 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16774 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16775 when such DH parameters are used, because otherwise small subgroup
16776 attacks may become possible!
16780 * Avoid memory leak in i2d_DHparams.
16784 * Allow the -k option to be used more than once in the enc program:
16785 this allows the same encrypted message to be read by multiple recipients.
16789 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16790 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16791 it will always use the numerical form of the OID, even if it has a short
16796 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16797 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16798 otherwise bn_mod_exp was called. In the case of hardware keys for example
16799 no private key components need be present and it might store extra data
16800 in the RSA structure, which cannot be accessed from bn_mod_exp.
16801 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16802 private key operations.
16806 * Added support for SPARC Linux.
16810 * pem_password_cb function type incompatibly changed from
16811 typedef int pem_password_cb(char *buf, int size, int rwflag);
16813 ....(char *buf, int size, int rwflag, void *userdata);
16814 so that applications can pass data to their callbacks:
16815 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16816 additional void * argument, which is just handed through whenever
16817 the password callback is called.
16819 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16821 New function SSL_CTX_set_default_passwd_cb_userdata.
16823 Compatibility note: As many C implementations push function arguments
16824 onto the stack in reverse order, the new library version is likely to
16825 interoperate with programs that have been compiled with the old
16826 pem_password_cb definition (PEM_whatever takes some data that
16827 happens to be on the stack as its last argument, and the callback
16828 just ignores this garbage); but there is no guarantee whatsoever that
16831 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16832 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16833 problems not only on Windows, but also on some Unix platforms.
16834 To avoid problematic command lines, these definitions are now in an
16835 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16836 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16840 * MIPS III/IV assembler module is reimplemented.
16844 * More DES library cleanups: remove references to srand/rand and
16845 delete an unused file.
16849 * Add support for the free Netwide assembler (NASM) under Win32,
16850 since not many people have MASM (ml) and it can be hard to obtain.
16851 This is currently experimental but it seems to work OK and pass all
16852 the tests. Check out INSTALL.W32 for info.
16856 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16857 without temporary keys kept an extra copy of the server key,
16858 and connections with temporary keys did not free everything in case
16863 * New function RSA_check_key and new openssl rsa option -check
16864 for verifying the consistency of RSA keys.
16866 *Ulf Moeller, Bodo Moeller*
16868 * Various changes to make Win32 compile work:
16869 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16870 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16871 comparison" warnings.
16872 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16876 * Add a debugging option to PKCS#5 v2 key generation function: when
16877 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16878 derived keys are printed to stderr.
16882 * Copy the flags in ASN1_STRING_dup().
16884 *Roman E. Pavlov <pre@mo.msk.ru>*
16886 * The x509 application mishandled signing requests containing DSA
16887 keys when the signing key was also DSA and the parameters didn't match.
16889 It was supposed to omit the parameters when they matched the signing key:
16890 the verifying software was then supposed to automatically use the CA's
16891 parameters if they were absent from the end user certificate.
16893 Omitting parameters is no longer recommended. The test was also
16894 the wrong way round! This was probably due to unusual behaviour in
16895 EVP_cmp_parameters() which returns 1 if the parameters match.
16896 This meant that parameters were omitted when they *didn't* match and
16897 the certificate was useless. Certificates signed with 'ca' didn't have
16900 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16902 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16903 The interface is as follows:
16904 Applications can use
16905 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16906 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16907 "off" is now the default.
16908 The library internally uses
16909 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16910 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16911 to disable memory-checking temporarily.
16913 Some inconsistent states that previously were possible (and were
16914 even the default) are now avoided.
16916 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16917 with each memory chunk allocated; this is occasionally more helpful
16918 than just having a counter.
16920 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16922 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16927 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16928 which largely parallels "options", but is for changing API behaviour,
16929 whereas "options" are about protocol behaviour.
16930 Initial "mode" flags are:
16932 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16933 a single record has been written.
16934 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16935 retries use the same buffer location.
16936 (But all of the contents must be
16941 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16944 * Fix problems with no-hmac etc.
16946 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16948 * New functions RSA_get_default_method(), RSA_set_method() and
16949 RSA_get_method(). These allows replacement of RSA_METHODs without having
16950 to mess around with the internals of an RSA structure.
16954 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16955 Also really enable memory leak checks in openssl.c and in some
16958 *Chad C. Mulligan, Bodo Moeller*
16960 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16961 up the length of negative integers. This has now been simplified to just
16962 store the length when it is first determined and use it later, rather
16963 than trying to keep track of where data is copied and updating it to
16965 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16967 * Add a new function PKCS7_signatureVerify. This allows the verification
16968 of a PKCS#7 signature but with the signing certificate passed to the
16969 function itself. This contrasts with PKCS7_dataVerify which assumes the
16970 certificate is present in the PKCS#7 structure. This isn't always the
16971 case: certificates can be omitted from a PKCS#7 structure and be
16972 distributed by "out of band" means (such as a certificate database).
16976 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16977 function prototypes in pem.h, also change util/mkdef.pl to add the
16978 necessary function names.
16982 * mk1mf.pl (used by Windows builds) did not properly read the
16983 options set by Configure in the top level Makefile, and Configure
16984 was not even able to write more than one option correctly.
16985 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
16989 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
16990 file to be loaded from a BIO or FILE pointer. The BIO version will
16991 for example allow memory BIOs to contain config info.
16995 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
16996 Whoever hopes to achieve shared-library compatibility across versions
16997 must use this, not the compile-time macro.
16998 (Exercise 0.9.4: Which is the minimum library version required by
17000 Note: All this applies only to multi-threaded programs, others don't
17005 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17006 through a BIO pair triggered the default case, i.e.
17007 SSLerr(...,SSL_R_UNKNOWN_STATE).
17011 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17012 can use the SSL library even if none of the specific BIOs is
17017 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17018 for the encoded length.
17020 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17022 * Add initial documentation of the X509V3 functions.
17026 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17027 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17028 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17029 secure PKCS#8 private key format with a high iteration count.
17033 * Fix determination of Perl interpreter: A perl or perl5
17034 *directory* in $PATH was also accepted as the interpreter.
17036 *Ralf S. Engelschall*
17038 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17039 wrong with it but it was very old and did things like calling
17040 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17041 unusual formatting.
17045 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17046 to use the new extension code.
17050 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17051 with macros. This should make it easier to change their form, add extra
17052 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17057 * Add to configuration table a new entry that can specify an alternative
17058 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17059 according to Mark Crispin <MRC@Panda.COM>.
17063 * DES CBC did not update the IV. Weird.
17067 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17068 Changing the behaviour of the former might break existing programs --
17069 where IV updating is needed, des_ncbc_encrypt can be used.
17072 * When bntest is run from "make test" it drives bc to check its
17073 calculations, as well as internally checking them. If an internal check
17074 fails, it needs to cause bc to give a non-zero result or make test carries
17075 on without noticing the failure. Fixed.
17079 * DES library cleanups.
17083 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17084 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17085 ciphers. NOTE: although the key derivation function has been verified
17086 against some published test vectors it has not been extensively tested
17087 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17092 * Instead of "mkdir -p", which is not fully portable, use new
17093 Perl script "util/mkdir-p.pl".
17097 * Rewrite the way password based encryption (PBE) is handled. It used to
17098 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17099 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17100 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17101 the 'parameter' field of the AlgorithmIdentifier is passed to the
17102 underlying key generation function so it must do its own ASN1 parsing.
17103 This has also changed the EVP_PBE_CipherInit() function which now has a
17104 'parameter' argument instead of literal salt and iteration count values
17105 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17109 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17110 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17111 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17112 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17113 value was just used as a "magic string" and not used directly its
17114 value doesn't matter.
17118 * Introduce some semblance of const correctness to BN. Shame C doesn't
17123 * "linux-sparc64" configuration (ultrapenguin).
17125 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17126 "linux-sparc" configuration.
17128 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17130 * config now generates no-xxx options for missing ciphers.
17134 * Support the EBCDIC character set (work in progress).
17135 File ebcdic.c not yet included because it has a different license.
17137 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17139 * Support BS2000/OSD-POSIX.
17141 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17143 * Make callbacks for key generation use `void *` instead of `char *`.
17147 * Make S/MIME samples compile (not yet tested).
17151 * Additional typesafe stacks.
17155 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17159 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17161 * New configuration variant "sco5-gcc".
17163 * Updated some demos.
17165 *Sean O Riordain, Wade Scholine*
17167 * Add missing BIO_free at exit of pkcs12 application.
17171 * Fix memory leak in conf.c.
17175 * Updates for Win32 to assembler version of MD5.
17179 * Set #! path to perl in apps/der_chop to where we found it
17180 instead of using a fixed path.
17184 * SHA library changes for irix64-mips4-cc.
17188 * Improvements for VMS support.
17192 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17194 * Bignum library bug fix. IRIX 6 passes "make test" now!
17195 This also avoids the problems with SC4.2 and unpatched SC5.
17197 *Andy Polyakov <appro@fy.chalmers.se>*
17199 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17200 These are required because of the typesafe stack would otherwise break
17201 existing code. If old code used a structure member which used to be STACK
17202 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17203 sk_num or sk_value it would produce an error because the num, data members
17204 are not present in STACK_OF. Now it just produces a warning. sk_set
17205 replaces the old method of assigning a value to sk_value
17206 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17207 that does this will no longer work (and should use sk_set instead) but
17208 this could be regarded as a "questionable" behaviour anyway.
17212 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17213 correctly handle encrypted S/MIME data.
17217 * Change type of various DES function arguments from des_cblock
17218 (which means, in function argument declarations, pointer to char)
17219 to des_cblock * (meaning pointer to array with 8 char elements),
17220 which allows the compiler to do more typechecking; it was like
17221 that back in SSLeay, but with lots of ugly casts.
17223 Introduce new type const_des_cblock.
17227 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17228 problems: find RecipientInfo structure that matches recipient certificate
17229 and initialise the ASN1 structures properly based on passed cipher.
17233 * Belatedly make the BN tests actually check the results.
17237 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17238 to and from BNs: it was completely broken. New compilation option
17239 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17240 key elements as negative integers.
17244 * Reorganize and speed up MD5.
17246 *Andy Polyakov <appro@fy.chalmers.se>*
17250 *Richard Levitte <richard@levitte.org>*
17252 * New option -out to asn1parse to allow the parsed structure to be
17253 output to a file. This is most useful when combined with the -strparse
17254 option to examine the output of things like OCTET STRINGS.
17258 * Make SSL library a little more fool-proof by not requiring any longer
17259 that `SSL_set_{accept,connect}_state` be called before
17260 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17261 in many applications because usually everything *appeared* to work as
17262 intended anyway -- now it really works as intended).
17266 * Move openssl.cnf out of lib/.
17270 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17271 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17272 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17274 *Ralf S. Engelschall*
17276 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17277 handle PKCS#7 enveloped data properly.
17279 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17281 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17282 copying pointers. The cert_st handling is changed by this in
17283 various ways (and thus what used to be known as ctx->default_cert
17284 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17285 any longer when s->cert does not give us what we need).
17286 ssl_cert_instantiate becomes obsolete by this change.
17287 As soon as we've got the new code right (possibly it already is?),
17288 we have solved a couple of bugs of the earlier code where s->cert
17289 was used as if it could not have been shared with other SSL structures.
17291 Note that using the SSL API in certain dirty ways now will result
17292 in different behaviour than observed with earlier library versions:
17293 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17294 does not influence s as it used to.
17296 In order to clean up things more thoroughly, inside SSL_SESSION
17297 we don't use CERT any longer, but a new structure SESS_CERT
17298 that holds per-session data (if available); currently, this is
17299 the peer's certificate chain and, for clients, the server's certificate
17300 and temporary key. CERT holds only those values that can have
17301 meaningful defaults in an SSL_CTX.
17305 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17306 from the internal representation. Various PKCS#7 fixes: remove some
17307 evil casts and set the enc_dig_alg field properly based on the signing
17312 * Allow PKCS#12 password to be set from the command line or the
17313 environment. Let 'ca' get its config file name from the environment
17314 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17319 * Allow certificate policies extension to use an IA5STRING for the
17320 organization field. This is contrary to the PKIX definition but
17321 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17326 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17327 without disallowing inline assembler and the like for non-pedantic builds.
17331 * Support Borland C++ builder.
17333 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17339 * SHA-1 cleanups and performance enhancements.
17341 *Andy Polyakov <appro@fy.chalmers.se>*
17343 * Sparc v8plus assembler for the bignum library.
17345 *Andy Polyakov <appro@fy.chalmers.se>*
17347 * Accept any -xxx and +xxx compiler options in Configure.
17351 * Update HPUX configuration.
17355 * Add missing `sk_<type>_unshift()` function to safestack.h
17357 *Ralf S. Engelschall*
17359 * New function SSL_CTX_use_certificate_chain_file that sets the
17360 "extra_cert"s in addition to the certificate. (This makes sense
17361 only for "PEM" format files, as chains as a whole are not
17366 * Support verify_depth from the SSL API.
17367 x509_vfy.c had what can be considered an off-by-one-error:
17368 Its depth (which was not part of the external interface)
17369 was actually counting the number of certificates in a chain;
17370 now it really counts the depth.
17374 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17375 instead of X509err, which often resulted in confusing error
17376 messages since the error codes are not globally unique
17377 (e.g. an alleged error in ssl3_accept when a certificate
17378 didn't match the private key).
17380 * New function SSL_CTX_set_session_id_context that allows to set a default
17381 value (so that you don't need SSL_set_session_id_context for each
17382 connection using the SSL_CTX).
17386 * OAEP decoding bug fix.
17390 * Support INSTALL_PREFIX for package builders, as proposed by
17395 * New Configure options "threads" and "no-threads". For systems
17396 where the proper compiler options are known (currently Solaris
17397 and Linux), "threads" is the default.
17401 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17405 * Install various scripts to $(OPENSSLDIR)/misc, not to
17406 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17407 such as /usr/local/bin.
17411 * "make linux-shared" to build shared libraries.
17413 *Niels Poppe <niels@netbox.org>*
17415 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17419 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17420 extension adding in x509 utility.
17424 * Remove NOPROTO sections and error code comments.
17428 * Partial rewrite of the DEF file generator to now parse the ANSI
17433 * New Configure options --prefix=DIR and --openssldir=DIR.
17437 * Complete rewrite of the error code script(s). It is all now handled
17438 by one script at the top level which handles error code gathering,
17439 header rewriting and C source file generation. It should be much better
17440 than the old method: it now uses a modified version of Ulf's parser to
17441 read the ANSI prototypes in all header files (thus the old K&R definitions
17442 aren't needed for error creation any more) and do a better job of
17443 translating function codes into names. The old 'ASN1 error code embedded
17444 in a comment' is no longer necessary and it doesn't use .err files which
17445 have now been deleted. Also the error code call doesn't have to appear all
17446 on one line (which resulted in some large lines...).
17450 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17454 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17455 0 (which usually indicates a closed connection), but continue reading.
17459 * Fix some race conditions.
17463 * Add support for CRL distribution points extension. Add Certificate
17464 Policies and CRL distribution points documentation.
17468 * Move the autogenerated header file parts to crypto/opensslconf.h.
17472 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17473 8 of keying material. Merlin has also confirmed interop with this fix
17474 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17476 *Merlin Hughes <merlin@baltimore.ie>*
17478 * Fix lots of warnings.
17480 *Richard Levitte <levitte@stacken.kth.se>*
17482 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17483 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17485 *Richard Levitte <levitte@stacken.kth.se>*
17487 * Fix problems with sizeof(long) == 8.
17489 *Andy Polyakov <appro@fy.chalmers.se>*
17491 * Change functions to ANSI C.
17495 * Fix typos in error codes.
17497 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17499 * Remove defunct assembler files from Configure.
17503 * SPARC v8 assembler BIGNUM implementation.
17505 *Andy Polyakov <appro@fy.chalmers.se>*
17507 * Support for Certificate Policies extension: both print and set.
17508 Various additions to support the r2i method this uses.
17512 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17513 return a const string when you are expecting an allocated buffer.
17517 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17518 types DirectoryString and DisplayText.
17522 * Add code to allow r2i extensions to access the configuration database,
17523 add an LHASH database driver and add several ctx helper functions.
17527 * Fix an evil bug in bn_expand2() which caused various BN functions to
17528 fail when they extended the size of a BIGNUM.
17532 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17533 support typesafe stack.
17537 * Fix typo in SSL_[gs]et_options().
17539 *Nils Frostberg <nils@medcom.se>*
17541 * Delete various functions and files that belonged to the (now obsolete)
17542 old X509V3 handling code.
17546 * New Configure option "rsaref".
17550 * Don't auto-generate pem.h.
17554 * Introduce type-safe ASN.1 SETs.
17558 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17560 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17562 * Introduce type-safe STACKs. This will almost certainly break lots of code
17563 that links with OpenSSL (well at least cause lots of warnings), but fear
17564 not: the conversion is trivial, and it eliminates loads of evil casts. A
17565 few STACKed things have been converted already. Feel free to convert more.
17566 In the fullness of time, I'll do away with the STACK type altogether.
17570 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17571 specified in `<certfile>` by updating the entry in the index.txt file.
17572 This way one no longer has to edit the index.txt file manually for
17573 revoking a certificate. The -revoke option does the gory details now.
17575 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17577 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17578 `-text` option at all and this way the `-noout -text` combination was
17579 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17581 *Ralf S. Engelschall*
17583 * Make sure a corresponding plain text error message exists for the
17584 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17585 verify callback function determined that a certificate was revoked.
17587 *Ralf S. Engelschall*
17589 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17590 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17591 all available ciphers including rc5, which was forgotten until now.
17592 In order to let the testing shell script know which algorithms
17593 are available, a new (up to now undocumented) command
17594 `openssl list-cipher-commands` is used.
17598 * Bugfix: s_client occasionally would sleep in select() when
17599 it should have checked SSL_pending() first.
17603 * New functions DSA_do_sign and DSA_do_verify to provide access to
17604 the raw DSA values prior to ASN.1 encoding.
17608 * Tweaks to Configure
17610 *Niels Poppe <niels@netbox.org>*
17612 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17617 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17621 * New config option to avoid instructions that are illegal on the 80386.
17622 The default code is faster, but requires at least a 486.
17626 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17627 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17628 same as SSL2_VERSION anyway.
17632 * New "-showcerts" option for s_client.
17636 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17637 application. Various cleanups and fixes.
17641 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17642 modify error routines to work internally. Add error codes and PBE init
17643 to library startup routines.
17647 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17648 packing functions to asn1 and evp. Changed function names and error
17649 codes along the way.
17653 * PKCS12 integration: and so it begins... First of several patches to
17654 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17655 objects to objects.h
17659 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17660 and display support for Thawte strong extranet extension.
17664 * Add LinuxPPC support.
17666 *Jeff Dubrule <igor@pobox.org>*
17668 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17669 bn_div_words in alpha.s.
17671 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17673 * Make sure the RSA OAEP test is skipped under -DRSAref because
17674 OAEP isn't supported when OpenSSL is built with RSAref.
17676 *Ulf Moeller <ulf@fitug.de>*
17678 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17679 so they no longer are missing under -DNOPROTO.
17681 *Soren S. Jorvang <soren@t.dk>*
17683 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17685 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17686 doesn't work when the session is reused. Coming soon!
17690 * Fix a security hole, that allows sessions to be reused in the wrong
17691 context thus bypassing client cert protection! All software that uses
17692 client certs and session caches in multiple contexts NEEDS PATCHING to
17693 allow session reuse! A fuller solution is in the works.
17695 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17697 * Some more source tree cleanups (removed obsolete files
17698 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17699 permission on "config" script to be executable) and a fix for the INSTALL
17702 *Ulf Moeller <ulf@fitug.de>*
17704 * Remove some legacy and erroneous uses of malloc, free instead of
17707 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17709 * Make rsa_oaep_test return non-zero on error.
17711 *Ulf Moeller <ulf@fitug.de>*
17713 * Add support for native Solaris shared libraries. Configure
17714 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17715 if someone would make that last step automatic.
17717 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17719 * ctx_size was not built with the right compiler during "make links". Fixed.
17723 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17724 except NULL ciphers". This means the default cipher list will no longer
17725 enable NULL ciphers. They need to be specifically enabled e.g. with
17726 the string "DEFAULT:eNULL".
17730 * Fix to RSA private encryption routines: if p < q then it would
17731 occasionally produce an invalid result. This will only happen with
17732 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17736 * Be less restrictive and allow also `perl util/perlpath.pl
17737 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17738 because this way one can also use an interpreter named `perl5` (which is
17739 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17740 installed as `perl`).
17742 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17744 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17746 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17748 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17749 advapi32.lib to Win32 build and change the pem test comparison
17750 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17751 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17752 and crypto/des/ede_cbcm_enc.c.
17756 * DES quad checksum was broken on big-endian architectures. Fixed.
17760 * Comment out two functions in bio.h that aren't implemented. Fix up the
17761 Win32 test batch file so it (might) work again. The Win32 test batch file
17762 is horrible: I feel ill....
17766 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17767 in e_os.h. Audit of header files to check ANSI and non ANSI
17768 sections: 10 functions were absent from non ANSI section and not exported
17769 from Windows DLLs. Fixed up libeay.num for new functions.
17773 * Make `openssl version` output lines consistent.
17775 *Ralf S. Engelschall*
17777 * Fix Win32 symbol export lists for BIO functions: Added
17778 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17779 to ms/libeay{16,32}.def.
17781 *Ralf S. Engelschall*
17783 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17784 fine under Unix and passes some trivial tests I've now added. But the
17785 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17786 added to make sure no one expects that this stuff really works in the
17787 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17788 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17791 *Ralf S. Engelschall*
17793 * Fix the generation of two part addresses in perl.
17795 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17797 * Add config entry for Linux on MIPS.
17799 *John Tobey <jtobey@channel1.com>*
17801 * Make links whenever Configure is run, unless we are on Windoze.
17805 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17806 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17811 * Add a useful kludge to allow package maintainers to specify compiler and
17812 other platforms details on the command line without having to patch the
17813 Configure script every time: One now can use
17814 `perl Configure <id>:<details>`,
17815 i.e. platform ids are allowed to have details appended
17816 to them (separated by colons). This is treated as there would be a static
17817 pre-configured entry in Configure's %table under key `<id>` with value
17818 `<details>` and `perl Configure <id>` is called. So, when you want to
17819 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17820 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17821 now, which overrides the FreeBSD-elf entry on-the-fly.
17823 *Ralf S. Engelschall*
17825 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17829 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17830 on the `perl Configure ...` command line. This way one can compile
17831 OpenSSL libraries with Position Independent Code (PIC) which is needed
17832 for linking it into DSOs.
17834 *Ralf S. Engelschall*
17836 * Remarkably, export ciphers were totally broken and no-one had noticed!
17841 * Cleaned up the LICENSE document: The official contact for any license
17842 questions now is the OpenSSL core team under openssl-core@openssl.org.
17843 And add a paragraph about the dual-license situation to make sure people
17844 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17845 to the OpenSSL toolkit.
17847 *Ralf S. Engelschall*
17849 * General source tree makefile cleanups: Made `making xxx in yyy...`
17850 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17851 Additionally cleaned up the `make links` target: Remove unnecessary
17852 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17853 to speed processing and no longer clutter the display with confusing
17854 stuff. Instead only the actually done links are displayed.
17856 *Ralf S. Engelschall*
17858 * Permit null encryption ciphersuites, used for authentication only. It used
17859 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17860 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17865 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17866 signed attributes when verifying signatures (this would break them),
17867 the detached data encoding was wrong and public keys obtained using
17868 X509_get_pubkey() weren't freed.
17872 * Add text documentation for the BUFFER functions. Also added a work around
17873 to a Win95 console bug. This was triggered by the password read stuff: the
17874 last character typed gets carried over to the next fread(). If you were
17875 generating a new cert request using 'req' for example then the last
17876 character of the passphrase would be CR which would then enter the first
17881 * Added the new 'Includes OpenSSL Cryptography Software' button as
17882 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17883 button and can be used by applications based on OpenSSL to show the
17884 relationship to the OpenSSL project.
17886 *Ralf S. Engelschall*
17888 * Remove confusing variables in function signatures in files
17889 ssl/ssl_lib.c and ssl/ssl.h.
17891 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17893 * Don't install bss_file.c under PREFIX/include/
17895 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17897 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17898 functions that return function pointers and has support for NT specific
17899 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17900 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17901 unsigned to signed types: this was killing the Win32 compile.
17905 * Add new certificate file to stack functions,
17906 SSL_add_dir_cert_subjects_to_stack() and
17907 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17908 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17909 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17910 This means that Apache-SSL and similar packages don't have to mess around
17911 to add as many CAs as they want to the preferred list.
17915 * Experiment with doxygen documentation. Currently only partially applied to
17917 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17918 openssl.doxy as the configuration file.
17922 * Get rid of remaining C++-style comments which strict C compilers hate.
17924 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17926 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17927 compiled in by default: it has problems with large keys.
17931 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17932 DH private keys and/or callback functions which directly correspond to
17933 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17934 is needed for applications which have to configure certificates on a
17935 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17937 For the RSA certificate situation is makes no difference, but
17938 for the DSA certificate situation this fixes the "no shared cipher"
17939 problem where the OpenSSL cipher selection procedure failed because the
17940 temporary keys were not overtaken from the context and the API provided
17941 no way to reconfigure them.
17942 The new functions now let applications reconfigure the stuff and they
17943 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17944 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17945 non-public-API function ssl_cert_instantiate() is used as a helper
17946 function and also to reduce code redundancy inside ssl_rsa.c.
17948 *Ralf S. Engelschall*
17950 * Move s_server -dcert and -dkey options out of the undocumented feature
17951 area because they are useful for the DSA situation and should be
17952 recognized by the users.
17954 *Ralf S. Engelschall*
17956 * Fix the cipher decision scheme for export ciphers: the export bits are
17957 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17958 SSL_EXP_MASK. So, the original variable has to be used instead of the
17959 already masked variable.
17961 *Richard Levitte <levitte@stacken.kth.se>*
17963 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17965 *Richard Levitte <levitte@stacken.kth.se>*
17967 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17968 from `int` to `unsigned int` because it is a length and initialized by
17969 EVP_DigestFinal() which expects an `unsigned int *`.
17971 *Richard Levitte <levitte@stacken.kth.se>*
17973 * Don't hard-code path to Perl interpreter on shebang line of Configure
17974 script. Instead use the usual Shell->Perl transition trick.
17976 *Ralf S. Engelschall*
17978 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17979 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17980 -noout -modulus` as it's already the case for `openssl rsa -noout
17981 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17982 currently the public key is printed (a decision which was already done by
17983 `openssl dsa -modulus` in the past) which serves a similar purpose.
17984 Additionally the NO_RSA no longer completely removes the whole -modulus
17985 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
17988 *Ralf S. Engelschall*
17990 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
17991 BIO. See the source (crypto/evp/bio_ok.c) for more info.
17993 *Arne Ansper <arne@ats.cyber.ee>*
17995 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
17996 to be added. Now both 'req' and 'ca' can use new objects defined in the
18001 * Add cool BIO that does syslog (or event log on NT).
18003 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18005 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18006 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18007 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18008 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18012 * Add preliminary config info for new extension code.
18016 * Make RSA_NO_PADDING really use no padding.
18018 *Ulf Moeller <ulf@fitug.de>*
18020 * Generate errors when private/public key check is done.
18024 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18025 for some CRL extensions and new objects added.
18029 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18030 key usage extension and fuller support for authority key id.
18034 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18035 padding method for RSA, which is recommended for new applications in PKCS
18036 #1 v2.0 (RFC 2437, October 1998).
18037 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18038 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18039 against Bleichbacher's attack on RSA.
18040 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18043 * Updates to the new SSL compression code
18045 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18047 * Fix so that the version number in the master secret, when passed
18048 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18049 (because the server will not accept higher), that the version number
18050 is 0x03,0x01, not 0x03,0x00
18052 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18054 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
18055 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
18056 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
18060 * Support for RAW extensions where an arbitrary extension can be
18061 created by including its DER encoding. See apps/openssl.cnf for
18066 * Make sure latest Perl versions don't interpret some generated C array
18067 code as Perl array code in the crypto/err/err_genc.pl script.
18069 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18071 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18072 not many people have the assembler. Various Win32 compilation fixes and
18073 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18074 build instructions.
18078 * Modify configure script 'Configure' to automatically create crypto/date.h
18079 file under Win32 and also build pem.h from pem.org. New script
18080 util/mkfiles.pl to create the MINFO file on environments that can't do a
18081 'make files': perl util/mkfiles.pl >MINFO should work.
18085 * Major rework of DES function declarations, in the pursuit of correctness
18086 and purity. As a result, many evil casts evaporated, and some weirdness,
18087 too. You may find this causes warnings in your code. Zapping your evil
18088 casts will probably fix them. Mostly.
18092 * Fix for a typo in asn1.h. Bug fix to object creation script
18093 obj_dat.pl. It considered a zero in an object definition to mean
18094 "end of object": none of the objects in objects.h have any zeros
18095 so it wasn't spotted.
18097 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18099 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18100 Masking (CBCM). In the absence of test vectors, the best I have been able
18101 to do is check that the decrypt undoes the encrypt, so far. Send me test
18102 vectors if you have them.
18106 * Correct calculation of key length for export ciphers (too much space was
18107 allocated for null ciphers). This has not been tested!
18111 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18112 message is now correct (it understands "crypto" and "ssl" on its
18113 command line). There is also now an "update" option. This will update
18114 the util/ssleay.num and util/libeay.num files with any new functions.
18116 perl util/mkdef.pl crypto ssl update
18117 it will update them.
18121 * Overhauled the Perl interface:
18122 - ported BN stuff to OpenSSL's different BN library
18123 - made the perl/ source tree CVS-aware
18124 - renamed the package from SSLeay to OpenSSL (the files still contain
18125 their history because I've copied them in the repository)
18126 - removed obsolete files (the test scripts will be replaced
18127 by better Test::Harness variants in the future)
18129 *Ralf S. Engelschall*
18131 * First cut for a very conservative source tree cleanup:
18132 1. merge various obsolete readme texts into doc/ssleay.txt
18133 where we collect the old documents and readme texts.
18134 2. remove the first part of files where I'm already sure that we no
18135 longer need them because of three reasons: either they are just temporary
18136 files which were left by Eric or they are preserved original files where
18137 I've verified that the diff is also available in the CVS via "cvs diff
18138 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18139 the crypto/md/ stuff).
18141 *Ralf S. Engelschall*
18143 * More extension code. Incomplete support for subject and issuer alt
18144 name, issuer and authority key id. Change the i2v function parameters
18145 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18146 what that's for :-) Fix to ASN1 macro which messed up
18147 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18151 * Preliminary support for ENUMERATED type. This is largely copied from the
18156 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18158 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18160 * Make sure `make rehash` target really finds the `openssl` program.
18162 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18164 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18165 like to hear about it if this slows down other processors.
18169 * Add CygWin32 platform information to Configure script.
18171 *Alan Batie <batie@aahz.jf.intel.com>*
18173 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18175 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18177 * New program nseq to manipulate netscape certificate sequences
18181 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18186 * Fixes to BN code. Previously the default was to define BN_RECURSION
18187 but the BN code had some problems that would cause failures when
18188 doing certificate verification and some other functions.
18190 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18192 * Add ASN1 and PEM code to support netscape certificate sequences.
18196 * Add ASN1 and PEM code to support netscape certificate sequences.
18200 * Add several PKIX and private extended key usage OIDs.
18204 * Modify the 'ca' program to handle the new extension code. Modify
18205 openssl.cnf for new extension format, add comments.
18209 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18210 and add a sample to openssl.cnf so req -x509 now adds appropriate
18215 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18216 error code, add initial support to X509_print() and x509 application.
18220 * Takes a deep breath and start adding X509 V3 extension support code. Add
18221 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18222 stuff is currently isolated and isn't even compiled yet.
18226 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18227 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18228 Removed the versions check from X509 routines when loading extensions:
18229 this allows certain broken certificates that don't set the version
18230 properly to be processed.
18234 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18235 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18236 can still be regenerated with "make depend".
18240 * Spelling mistake in C version of CAST-128.
18242 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18244 * Changes to the error generation code. The perl script err-code.pl
18245 now reads in the old error codes and retains the old numbers, only
18246 adding new ones if necessary. It also only changes the .err files if new
18247 codes are added. The makefiles have been modified to only insert errors
18248 when needed (to avoid needlessly modifying header files). This is done
18249 by only inserting errors if the .err file is newer than the auto generated
18250 C file. To rebuild all the error codes from scratch (the old behaviour)
18251 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18252 or delete all the .err files.
18256 * CAST-128 was incorrectly implemented for short keys. The C version has
18257 been fixed, but is untested. The assembler versions are also fixed, but
18258 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18259 to regenerate it if needed.
18260 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18261 Hagino <itojun@kame.net>*
18263 * File was opened incorrectly in randfile.c.
18265 *Ulf Möller <ulf@fitug.de>*
18267 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18268 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18269 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18270 al: it's just almost always a UTCTime. Note this patch adds new error
18271 codes so do a "make errors" if there are problems.
18275 * Correct Linux 1 recognition in config.
18277 *Ulf Möller <ulf@fitug.de>*
18279 * Remove pointless MD5 hash when using DSA keys in ca.
18281 *Anonymous <nobody@replay.com>*
18283 * Generate an error if given an empty string as a cert directory. Also
18284 generate an error if handed NULL (previously returned 0 to indicate an
18285 error, but didn't set one).
18287 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18289 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18293 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18294 parameters. This was causing a warning which killed off the Win32 compile.
18298 * Remove C++ style comments from crypto/bn/bn_local.h.
18300 *Neil Costigan <neil.costigan@celocom.com>*
18302 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18303 based on a text string, looking up short and long names and finally
18304 "dot" format. The "dot" format stuff didn't work. Added new function
18305 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18306 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18307 OID is not part of the table.
18311 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18312 X509_LOOKUP_by_alias().
18316 * Sort openssl functions by name.
18320 * Get the gendsa program working (hopefully) and add it to app list. Remove
18321 encryption from sample DSA keys (in case anyone is interested the password
18326 * Make *all* `*_free` functions accept a NULL pointer.
18328 *Frans Heymans <fheymans@isaserver.be>*
18330 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18333 *Anonymous <nobody@replay.com>*
18335 * s_server should send the CAfile as acceptable CAs, not its own cert.
18337 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18339 * Don't blow it for numeric -newkey arguments to apps/req.
18341 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18343 * Temp key "for export" tests were wrong in s3_srvr.c.
18345 *Anonymous <nobody@replay.com>*
18347 * Add prototype for temp key callback functions
18348 SSL_CTX_set_tmp_{rsa,dh}_callback().
18352 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18353 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18357 * X509_name_add_entry() freed the wrong thing after an error.
18359 *Arne Ansper <arne@ats.cyber.ee>*
18361 * rsa_eay.c would attempt to free a NULL context.
18363 *Arne Ansper <arne@ats.cyber.ee>*
18365 * BIO_s_socket() had a broken should_retry() on Windoze.
18367 *Arne Ansper <arne@ats.cyber.ee>*
18369 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18371 *Arne Ansper <arne@ats.cyber.ee>*
18373 * Make sure the already existing X509_STORE->depth variable is initialized
18374 in X509_STORE_new(), but document the fact that this variable is still
18375 unused in the certificate verification process.
18377 *Ralf S. Engelschall*
18379 * Fix the various library and apps files to free up pkeys obtained from
18380 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18384 * Fix reference counting in X509_PUBKEY_get(). This makes
18385 demos/maurice/example2.c work, amongst others, probably.
18387 *Steve Henson and Ben Laurie*
18389 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18390 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18391 are no longer created. This way we have a single and consistent command
18392 line interface `openssl <command>`, similar to `cvs <command>`.
18394 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18396 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18397 BIT STRING wrapper always have zero unused bits.
18401 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18405 * Make the top-level INSTALL documentation easier to understand.
18409 * Makefiles updated to exit if an error occurs in a sub-directory
18410 make (including if user presses ^C) [Paul Sutton]
18412 * Make Montgomery context stuff explicit in RSA data structure.
18416 * Fix build order of pem and err to allow for generated pem.h.
18420 * Fix renumbering bug in X509_NAME_delete_entry().
18424 * Enhanced the err-ins.pl script so it makes the error library number
18425 global and can add a library name. This is needed for external ASN1 and
18426 other error libraries.
18430 * Fixed sk_insert which never worked properly.
18434 * Fix ASN1 macros so they can handle indefinite length constructed
18435 EXPLICIT tags. Some non standard certificates use these: they can now
18440 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18441 into a single doc/ssleay.txt bundle. This way the information is still
18442 preserved but no longer messes up this directory. Now it's new room for
18443 the new set of documentation files.
18445 *Ralf S. Engelschall*
18447 * SETs were incorrectly DER encoded. This was a major pain, because they
18448 shared code with SEQUENCEs, which aren't coded the same. This means that
18449 almost everything to do with SETs or SEQUENCEs has either changed name or
18450 number of arguments.
18452 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18454 * Fix test data to work with the above.
18458 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18459 was already fixed by Eric for 0.9.1 it seems.
18461 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18463 * Autodetect FreeBSD3.
18467 * Fix various bugs in Configure. This affects the following platforms:
18471 unixware-2.0-pentium
18476 * Eliminate generated files from CVS. Reorder tests to regenerate files
18477 before they are needed.
18481 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18485 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18487 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18488 changed SSLeay to OpenSSL in version strings.
18490 *Ralf S. Engelschall*
18492 * Some fixups to the top-level documents.
18496 * Fixed the nasty bug where rsaref.h was not found under compile-time
18497 because the symlink to include/ was missing.
18499 *Ralf S. Engelschall*
18501 * Incorporated the popular no-RSA/DSA-only patches
18502 which allow to compile a RSA-free SSLeay.
18504 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18506 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18507 when "ssleay" is still not found.
18509 *Ralf S. Engelschall*
18511 * Added more platforms to Configure: Cray T3E, HPUX 11,
18513 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18515 * Updated the README file.
18517 *Ralf S. Engelschall*
18519 * Added various .cvsignore files in the CVS repository subdirs
18520 to make a "cvs update" really silent.
18522 *Ralf S. Engelschall*
18524 * Recompiled the error-definition header files and added
18525 missing symbols to the Win32 linker tables.
18527 *Ralf S. Engelschall*
18529 * Cleaned up the top-level documents;
18530 o new files: CHANGES and LICENSE
18531 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18532 o merged COPYRIGHT into LICENSE
18533 o removed obsolete TODO file
18534 o renamed MICROSOFT to INSTALL.W32
18536 *Ralf S. Engelschall*
18538 * Removed dummy files from the 0.9.1b source tree:
18539 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18540 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18541 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18542 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18543 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18545 *Ralf S. Engelschall*
18547 * Added various platform portability fixes.
18551 * The Genesis of the OpenSSL rpject:
18552 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18553 Young and Tim J. Hudson created while they were working for C2Net until
18556 *The OpenSSL Project*
18558 ### Changes between 0.9.0b and 0.9.1b [not released]
18560 * Updated a few CA certificates under certs/
18564 * Changed some BIGNUM api stuff.
18568 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18569 DGUX x86, Linux Alpha, etc.
18573 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18574 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18579 * Add -strparse option to asn1pars program which parses nested
18582 *Dr Stephen Henson <shenson@bigfoot.com>*
18584 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18588 * DSA fix for "ca" program.
18592 * Added "-genkey" option to "dsaparam" program.
18596 * Added RIPE MD160 (rmd160) message digest.
18600 * Added -a (all) option to "ssleay version" command.
18604 * Added PLATFORM define which is the id given to Configure.
18608 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18612 * Extended the ASN.1 parser routines.
18616 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18620 * Added a BN_CTX to the BN library.
18624 * Fixed the weak key values in DES library
18628 * Changed API in EVP library for cipher aliases.
18632 * Added support for RC2/64bit cipher.
18636 * Converted the lhash library to the crypto/mem.c functions.
18640 * Added more recognized ASN.1 object ids.
18644 * Added more RSA padding checks for SSL/TLS.
18648 * Added BIO proxy/filter functionality.
18652 * Added extra_certs to SSL_CTX which can be used
18653 send extra CA certificates to the client in the CA cert chain sending
18654 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18658 * Now Fortezza is denied in the authentication phase because
18659 this is key exchange mechanism is not supported by SSLeay at all.
18663 * Additional PKCS1 checks.
18667 * Support the string "TLSv1" for all TLS v1 ciphers.
18671 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18672 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18676 * Fixed a few memory leaks.
18680 * Fixed various code and comment typos.
18684 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18685 bytes sent in the client random.
18687 *Edward Bishop <ebishop@spyglass.com>*
18691 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18692 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18693 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18694 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18695 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18696 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18697 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18698 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18699 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18700 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18701 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18702 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18703 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18704 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18705 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18706 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18707 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18708 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18709 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18710 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18711 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18712 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18713 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18714 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18715 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18716 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18717 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18718 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18719 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18720 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18721 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18722 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18723 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18724 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18725 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18726 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18727 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18728 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18729 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18730 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18731 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18732 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18733 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18734 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18735 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18736 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18737 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18738 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18739 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18740 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18741 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18742 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18743 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18744 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18745 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18746 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18747 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18748 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18749 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18750 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18751 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18752 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18753 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18754 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18755 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18756 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18757 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18758 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18759 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18760 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18761 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18762 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18763 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18764 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18765 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18766 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18767 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18768 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18769 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18770 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18771 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18772 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18773 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18774 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18775 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18776 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18777 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18778 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18779 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18780 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18781 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18782 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18783 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18784 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18785 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18786 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18787 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18788 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18789 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18790 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18791 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18792 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18793 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18794 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18795 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18796 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18797 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18798 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18799 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18800 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18801 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18802 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18803 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18804 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18805 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18806 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18807 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18808 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18809 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18810 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18811 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18812 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18813 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18814 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18815 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18816 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18817 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18818 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18819 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18820 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18821 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18822 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18823 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18824 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18825 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18826 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18827 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18828 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18829 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18830 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18831 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18832 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18833 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18834 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18835 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18836 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18837 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18838 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18839 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18840 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18841 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18842 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18843 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18844 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18845 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18846 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18847 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18848 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18849 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18850 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18851 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18852 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655