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 alpha 11 [28 jan 2021]
26 * Deprecated the obsolete X9.31 RSA key generation related functions
27 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
28 BN_X931_generate_prime_ex().
32 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
33 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
34 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
35 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
36 were used to collect all necessary data to form a HTTP request, and to
37 perform the HTTP transfer with that request. With OpenSSL 3.0, the
38 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
39 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
40 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
41 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
42 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
43 OSSL_HTTP_REQ_CTX_set_max_response_length().
45 *Rich Salz and Richard Levitte*
47 * Validation of SM2 keys has been separated from the validation of regular EC
48 keys, allowing to improve the SM2 validation process to reject loaded private
49 keys that are not conforming to the SM2 ISO standard.
50 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
55 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
56 switches: a validation failure triggers an early exit, returning a failure
57 exit status to the parent process.
61 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
62 to ignore unknown ciphers.
66 * The `-cipher-commands` and `-digest-commands` options
67 of the command line utility `list` have been deprecated.
68 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
72 * All of the low level EC_KEY functions have been deprecated including:
74 EC_KEY_OpenSSL, EC_KEY_get_default_method, EC_KEY_set_default_method,
75 EC_KEY_get_method, EC_KEY_set_method, EC_KEY_new_method
76 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
77 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
78 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
79 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
80 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign,
81 EC_KEY_METHOD_get_verify,
82 EC_KEY_new_ex, EC_KEY_new, EC_KEY_get_flags, EC_KEY_set_flags,
83 EC_KEY_clear_flags, EC_KEY_decoded_from_explicit_params,
84 EC_KEY_new_by_curve_name_ex, EC_KEY_new_by_curve_name, EC_KEY_free,
85 EC_KEY_copy, EC_KEY_dup, EC_KEY_up_ref, EC_KEY_get0_engine,
86 EC_KEY_get0_group, EC_KEY_set_group, EC_KEY_get0_private_key,
87 EC_KEY_set_private_key, EC_KEY_get0_public_key, EC_KEY_set_public_key,
88 EC_KEY_get_enc_flags, EC_KEY_set_enc_flags, EC_KEY_get_conv_form,
89 EC_KEY_set_conv_form, EC_KEY_set_ex_data, EC_KEY_get_ex_data,
90 EC_KEY_set_asn1_flag, EC_KEY_generate_key, EC_KEY_check_key, EC_KEY_can_sign,
91 EC_KEY_set_public_key_affine_coordinates, EC_KEY_key2buf, EC_KEY_oct2key,
92 EC_KEY_oct2priv, EC_KEY_priv2oct and EC_KEY_priv2buf.
93 Applications that need to implement an EC_KEY_METHOD need to consider
94 implementation of the functionality in a special provider.
95 For replacement of the functions manipulating the EC_KEY objects
96 see the EVP_PKEY-EC(7) manual page.
98 Additionally functions that read and write EC_KEY objects such as
99 o2i_ECPublicKey, i2o_ECPublicKey, ECParameters_print_fp, EC_KEY_print_fp,
100 d2i_ECPKParameters, d2i_ECParameters, d2i_ECPrivateKey, d2i_ECPrivateKey_bio,
101 d2i_ECPrivateKey_fp, d2i_EC_PUBKEY, d2i_EC_PUBKEY_bio, d2i_EC_PUBKEY_fp,
102 i2d_ECPKParameters, i2d_ECParameters, i2d_ECPrivateKey, i2d_ECPrivateKey_bio,
103 i2d_ECPrivateKey_fp, i2d_EC_PUBKEY, i2d_EC_PUBKEY_bio and i2d_EC_PUBKEY_fp
104 have also been deprecated. Applications should instead use the
105 OSSL_DECODER and OSSL_ENCODER APIs to read and write EC files.
107 Finally functions that assign or obtain EC_KEY objects from an EVP_PKEY such as
108 EVP_PKEY_assign_EC_KEY, EVP_PKEY_get0_EC_KEY, EVP_PKEY_get1_EC_KEY and
109 EVP_PKEY_set1_EC_KEY are also deprecated. Applications should instead either
110 read or write an EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER
111 APIs. Or load an EVP_PKEY directly from EC data using EVP_PKEY_fromdata().
113 *Shane Lontis, Paul Dale, Richard Levitte, and Tomas Mraz*
115 * Deprecated all the libcrypto and libssl error string loading
116 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
117 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
118 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
119 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
120 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
121 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
122 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
123 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
124 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
125 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
127 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
128 now loads error strings automatically.
132 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
133 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
134 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
135 are to be used by servers requiring ephemeral DH keys. Instead applications
136 should consider using the built-in DH parameters that are available by
137 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
138 necessary then applications can use the alternative functions
139 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
140 replacement for the "callback" functions. The callback was originally useful
141 in order to have different parameters for export and non-export ciphersuites.
142 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
143 functions should be replaced by one of the other methods described above.
147 * The `-crypt` option to the `passwd` command line tool has been removed.
151 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
156 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
158 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
159 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
160 The inverse ciphers use AES decryption for wrapping, and
161 AES encryption for unwrapping.
165 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
166 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
167 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
168 OpenSSL 3.0 these are replaced by the more generic functions
169 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
170 The old versions have been converted to deprecated macros that just call the
175 * The security callback, which can be customised by application code, supports
176 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
177 in the "other" parameter. In most places this is what is passed. All these
178 places occur server side. However there was one client side call of this
179 security operation and it passed a DH object instead. This is incorrect
180 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
181 of the other locations. Therefore this client side call has been changed to
182 pass an EVP_PKEY instead.
186 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
187 interface. Their functionality remains unchanged.
191 * Added new option for 'openssl list', '-providers', which will display the
192 list of loaded providers, their names, version and status. It optionally
193 displays their gettable parameters.
197 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
198 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
199 type is internally recognised so the workaround is no longer needed.
201 Functionality is still retained as it is, but will only work with
202 EVP_PKEYs with a legacy internal key.
206 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
207 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
211 * Changed all "STACK" functions to be macros instead of inline functions. Macro
212 parameters are still checked for type safety at compile time via helper
217 * Remove the RAND_DRBG API
219 The RAND_DRBG API did not fit well into the new provider concept as
220 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
221 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
222 and some of its API calls are rather low-level. This holds in particular
223 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
225 Adding a compatibility layer to continue supporting the RAND_DRBG API as
226 a legacy API for a regular deprecation period turned out to come at the
227 price of complicating the new provider API unnecessarily. Since the
228 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
231 *Paul Dale and Matthias St. Pierre*
233 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
234 as well as actual hostnames.
238 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
239 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
240 conversely, silently ignore DTLS protocol version bounds when configuring
241 TLS-based contexts. The commands can be repeated to set bounds of both
242 types. The same applies with the corresponding "min_protocol" and
243 "max_protocol" command-line switches, in case some application uses both TLS
246 SSL_CTX instances that are created for a fixed protocol version (e.g.
247 `TLSv1_server_method()`) also silently ignore version bounds. Previously
248 attempts to apply bounds to these protocol versions would result in an
249 error. Now only the "version-flexible" SSL_CTX instances are subject to
250 limits in configuration files in command-line options.
254 * Deprecated the `ENGINE` API. Engines should be replaced with providers
259 * Reworked the recorded ERR codes to make better space for system errors.
260 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
261 given code is a system error (true) or an OpenSSL error (false).
265 * Reworked the test perl framework to better allow parallel testing.
267 *Nicola Tuveri and David von Oheimb*
269 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
270 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
274 * 'Configure' has been changed to figure out the configuration target if
275 none is given on the command line. Consequently, the 'config' script is
276 now only a mere wrapper. All documentation is changed to only mention
279 *Rich Salz and Richard Levitte*
281 * Added a library context `OSSL_LIB_CTX` that applications as well as
282 other libraries can use to form a separate context within which
283 libcrypto operations are performed.
285 There are two ways this can be used:
287 - Directly, by passing a library context to functions that take
288 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
290 - Indirectly, by creating a new library context and then assigning
291 it as the new default, with `OSSL_LIB_CTX_set0_default`.
293 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
294 apart from the functions directly related to `OSSL_LIB_CTX`, accept
295 NULL to indicate that the default library context should be used.
297 Library code that changes the default library context using
298 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
299 second call before returning to the caller.
301 _(Note: the library context was initially called `OPENSSL_CTX` and
302 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
306 * Handshake now fails if Extended Master Secret extension is dropped
311 * Dropped interactive mode from the `openssl` program. From now on,
312 running it without arguments is equivalent to `openssl help`.
316 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
317 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
318 While the old function names have been retained for backward compatibility
319 they should not be used in new developments
320 because their return values are confusing: Unlike other `_cmp()` functions
321 they do not return 0 in case their arguments are equal.
325 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
326 `EC_GROUP_get_field_type()`.
330 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
331 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
332 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
333 Applications should rely on the library automatically assigning a suitable
334 EC_METHOD internally upon EC_GROUP construction.
338 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
339 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
340 assigned internally without application intervention.
341 Users of EC_GROUP_new() should switch to a different suitable constructor.
345 * Add CAdES-BES signature verification support, mostly derived
346 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
348 *Filipe Raimundo da Silva*
350 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
354 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
355 functions are not widely used and now OpenSSL automatically perform this
356 conversion when needed.
360 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
361 EC_KEY_precompute_mult(). These functions are not widely used and
362 applications should instead switch to named curves which OpenSSL has
363 hardcoded lookup tables for.
367 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
368 should instead use the L<EC_POINT_mul(3)> function.
372 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
373 that are not applicable to the new provider model. Applications should
374 instead use EVP_default_properties_is_fips_enabled() and
375 EVP_default_properties_enable_fips().
379 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
380 is set, an unexpected EOF is ignored, it pretends a close notify was received
381 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
385 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
386 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
387 used and applications should instead use the
388 L<EC_POINT_set_affine_coordinates(3)> and
389 L<EC_POINT_get_affine_coordinates(3)> functions.
393 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
394 arrays to be more easily constructed via a series of utility functions.
395 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
396 the various push functions and finally convert to a passable OSSL_PARAM
397 array using OSSL_PARAM_BLD_to_param().
401 * The security strength of SHA1 and MD5 based signatures in TLS has been
402 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
403 working at the default security level of 1 and instead requires security
404 level 0. The security level can be changed either using the cipher string
405 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
409 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
410 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
411 internal keys, if they correspond to one of those built in types.
415 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
416 contain a provider side internal key.
420 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
421 They are old functions that we don't use, and that you could disable with
422 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
426 * Project text documents not yet having a proper file name extension
427 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
428 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
432 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
433 have been converted to Markdown with the goal to produce documents
434 which not only look pretty when viewed online in the browser, but
435 remain well readable inside a plain text editor.
437 To achieve this goal, a 'minimalistic' Markdown style has been applied
438 which avoids formatting elements that interfere too much with the
439 reading flow in the text file. For example, it
441 * avoids [ATX headings][] and uses [setext headings][] instead
442 (which works for `<h1>` and `<h2>` headings only).
443 * avoids [inline links][] and uses [reference links][] instead.
444 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
446 [ATX headings]: https://github.github.com/gfm/#atx-headings
447 [setext headings]: https://github.github.com/gfm/#setext-headings
448 [inline links]: https://github.github.com/gfm/#inline-link
449 [reference links]: https://github.github.com/gfm/#reference-link
450 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
451 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
453 *Matthias St. Pierre*
455 * The test suite is changed to preserve results of each test recipe.
456 A new directory test-runs/ with subdirectories named like the
457 test recipes are created in the build tree for this purpose.
461 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
462 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
463 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
465 *David von Oheimb, Martin Peylo*
467 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
468 The legacy OCSP-focused and only partly documented API is retained for
469 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
473 * Added `util/check-format.pl`, a tool for checking adherence to the
474 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
475 The checks performed are incomplete and yield some false positives.
476 Still the tool should be useful for detecting most typical glitches.
480 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
481 If domain name resolution yields multiple IP addresses all of them are tried
482 after `connect()` failures.
486 * All of the low level RSA functions have been deprecated including:
488 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
489 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
490 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
491 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
492 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
493 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
494 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
495 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
496 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
497 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
498 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
499 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
500 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
501 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
502 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
503 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
504 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
505 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
506 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
507 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
508 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
509 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
510 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
511 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
512 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
513 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
514 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
515 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
517 Use of these low level functions has been informally discouraged for a long
518 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
519 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
520 L<EVP_PKEY_decrypt(3)>.
524 * X509 certificates signed using SHA1 are no longer allowed at security
526 In TLS/SSL the default security level is 1. It can be set either
527 using the cipher string with `@SECLEVEL`, or calling
528 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
529 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
531 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
532 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
533 options of the commands.
537 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
538 modified to use PKEY APIs. These commands are now in maintenance mode
539 and no new features will be added to them.
543 * The command line utility rsautl has been deprecated.
544 Instead use the pkeyutl program.
548 * The command line utilities genrsa and rsa have been modified to use PKEY
549 APIs. They now write PKCS#8 keys by default. These commands are now in
550 maintenance mode and no new features will be added to them.
554 * All of the low level DH functions have been deprecated including:
556 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
557 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
558 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
559 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
560 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
561 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
562 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
563 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
564 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
565 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
566 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
567 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
568 DH_meth_get_generate_params and DH_meth_set_generate_params.
570 Use of these low level functions has been informally discouraged for a long
571 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
572 and L<EVP_PKEY_derive(3)>.
574 Additionally functions that read and write DH objects such as d2i_DHparams,
575 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
576 functions have also been deprecated. Applications should instead use the
577 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
579 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
580 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
581 `EVP_PKEY_set1_DH()` are also deprecated.
582 Applications should instead either read or write an
583 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
584 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
586 *Paul Dale and Matt Caswell*
588 * All of the low level DSA functions have been deprecated including:
590 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
591 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
592 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
593 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
594 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
595 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
596 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
597 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
598 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
599 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
600 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
601 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
602 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
603 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
604 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
605 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
606 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
608 Use of these low level functions has been informally discouraged for a long
609 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
610 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
614 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
615 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
616 This means that applications don't have to look at the curve NID and
617 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
618 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
619 a no-op when the EVP_PKEY is already of the given type.
621 Parameter and key generation is also reworked to make it possible
622 to generate EVP_PKEY_SM2 parameters and keys without having to go
623 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
624 However, code that does the latter will still work as before.
628 * Deprecated low level ECDH and ECDSA functions. These include:
630 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
631 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
634 Use of these low level functions has been informally discouraged for a long
635 time. Instead applications should use the EVP_PKEY_derive(3),
636 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
640 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
641 and EVP_PKEY_decrypt() instead.
642 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
643 and EVP_PKEY_encrypt() instead.
647 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
648 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
649 a new formulation to include all the things it can be used for,
650 as well as words of caution.
654 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
655 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
659 * All of the low level HMAC functions have been deprecated including:
661 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
662 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
665 Use of these low level functions has been informally discouraged for a long
666 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
667 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
668 and L<EVP_MAC_final(3)>.
672 * Over two thousand fixes were made to the documentation, including:
673 - Common options (such as -rand/-writerand, TLS version control, etc)
674 were refactored and point to newly-enhanced descriptions in openssl.pod.
675 - Added style conformance for all options (with help from Richard Levitte),
676 documented all reported missing options, added a CI build to check
677 that all options are documented and that no unimplemented options
679 - Documented some internals, such as all use of environment variables.
680 - Addressed all internal broken L<> references.
684 * All of the low level CMAC functions have been deprecated including:
686 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
687 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
689 Use of these low level functions has been informally discouraged for a long
690 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
691 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
692 and L<EVP_MAC_final(3)>.
696 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
697 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
700 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
701 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
702 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
703 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
704 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
705 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
706 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
707 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
708 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
709 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
711 Use of these low level functions has been informally discouraged
712 for a long time. Applications should use the EVP_DigestInit_ex(3),
713 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
717 * Corrected the documentation of the return values from the `EVP_DigestSign*`
718 set of functions. The documentation mentioned negative values for some
719 errors, but this was never the case, so the mention of negative values
722 Code that followed the documentation and thereby check with something
723 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
727 * All of the low level cipher functions have been deprecated including:
729 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
730 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
731 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
732 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
733 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
734 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
735 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
736 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
737 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
738 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
739 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
740 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
741 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
742 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
743 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
744 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
745 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
746 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
747 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
748 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
749 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
750 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
751 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
752 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
753 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
754 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
755 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
756 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
757 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
759 Use of these low level functions has been informally discouraged for
760 a long time. Applications should use the high level EVP APIs, e.g.
761 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
762 equivalently named decrypt functions instead.
764 *Matt Caswell and Paul Dale*
766 * Removed include/openssl/opensslconf.h.in and replaced it with
767 include/openssl/configuration.h.in, which differs in not including
768 <openssl/macros.h>. A short header include/openssl/opensslconf.h
769 was added to include both.
771 This allows internal hacks where one might need to modify the set
772 of configured macros, for example this if deprecated symbols are
773 still supposed to be available internally:
775 #include <openssl/configuration.h>
777 #undef OPENSSL_NO_DEPRECATED
778 #define OPENSSL_SUPPRESS_DEPRECATED
780 #include <openssl/macros.h>
782 This should not be used by applications that use the exported
783 symbols, as that will lead to linking errors.
787 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
788 used in exponentiation with 512-bit moduli. No EC algorithms are
789 affected. Analysis suggests that attacks against 2-prime RSA1024,
790 3-prime RSA1536, and DSA1024 as a result of this defect would be very
791 difficult to perform and are not believed likely. Attacks against DH512
792 are considered just feasible. However, for an attack the target would
793 have to re-use the DH512 private key, which is not recommended anyway.
794 Also applications directly using the low level API BN_mod_exp may be
795 affected if they use BN_FLG_CONSTTIME.
800 * Most memory-debug features have been deprecated, and the functionality
801 replaced with no-ops.
805 * Added documentation for the STACK API.
809 * Introduced a new method type and API, OSSL_ENCODER, to
810 represent generic encoders. An implementation is expected to
811 be able to encode an object associated with a given name (such
812 as an algorithm name for an asymmetric key) into forms given by
813 implementation properties.
815 Encoders are primarily used from inside libcrypto, through
816 calls to functions like EVP_PKEY_print_private(),
817 PEM_write_bio_PrivateKey() and similar.
819 Encoders are specified in such a way that they can be made to
820 directly handle the provider side portion of an object, if this
821 provider side part comes from the same provider as the encoder
822 itself, but can also be made to handle objects in parametrized
823 form (as an OSSL_PARAM array of data). This allows a provider to
824 offer generic encoders as a service for any other provider.
828 * Added a .pragma directive to the syntax of configuration files, to
829 allow varying behavior in a supported and predictable manner.
830 Currently added pragma:
834 This allows dollar signs to be a keyword character unless it's
835 followed by a opening brace or parenthesis. This is useful for
836 platforms where dollar signs are commonly used in names, such as
837 volume names and system directory names on VMS.
841 * Added functionality to create an EVP_PKEY from user data. This
842 is effectively the same as creating a RSA, DH or DSA object and
843 then assigning them to an EVP_PKEY, but directly using algorithm
844 agnostic EVP functions. A benefit is that this should be future
845 proof for public key algorithms to come.
849 * Change the interpretation of the '--api' configuration option to
850 mean that this is a desired API compatibility level with no
851 further meaning. The previous interpretation, that this would
852 also mean to remove all deprecated symbols up to and including
853 the given version, no requires that 'no-deprecated' is also used
854 in the configuration.
856 When building applications, the desired API compatibility level
857 can be set with the OPENSSL_API_COMPAT macro like before. For
858 API compatibility version below 3.0, the old style numerical
859 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
860 For version 3.0 and on, the value is expected to be the decimal
861 value calculated from the major and minor version like this:
863 MAJOR * 10000 + MINOR * 100
867 -DOPENSSL_API_COMPAT=30000 For 3.0
868 -DOPENSSL_API_COMPAT=30200 For 3.2
870 To hide declarations that are deprecated up to and including the
871 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
872 given when building the application as well.
876 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
877 access to certificate and CRL stores via URIs and OSSL_STORE
880 This adds the following functions:
882 - X509_LOOKUP_store()
883 - X509_STORE_load_file()
884 - X509_STORE_load_path()
885 - X509_STORE_load_store()
886 - SSL_add_store_cert_subjects_to_stack()
887 - SSL_CTX_set_default_verify_store()
888 - SSL_CTX_load_verify_file()
889 - SSL_CTX_load_verify_dir()
890 - SSL_CTX_load_verify_store()
894 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
895 The presence of this system service is determined at run-time.
899 * Added functionality to create an EVP_PKEY context based on data
900 for methods from providers. This takes an algorithm name and a
901 property query string and simply stores them, with the intent
902 that any operation that uses this context will use those strings
903 to fetch the needed methods implicitly, thereby making the port
904 of application written for pre-3.0 OpenSSL easier.
908 * The undocumented function NCONF_WIN32() has been deprecated; for
909 conversion details see the HISTORY section of doc/man5/config.pod
913 * Introduced the new functions EVP_DigestSignInit_ex() and
914 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
915 EVP_DigestVerifyUpdate() have been converted to functions. See the man
916 pages for further details.
920 * Over two thousand fixes were made to the documentation, including:
921 adding missing command flags, better style conformance, documentation
924 *Rich Salz, Richard Levitte*
926 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
927 X25519, X448, Ed25519 and Ed448.
931 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
936 * Deprecated the public definition of `ERR_STATE` as well as the function
937 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
942 * Added ERR functionality to give callers access to the stored function
943 names that have replaced the older function code based functions.
945 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
946 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
947 ERR_peek_error_all() and ERR_peek_last_error_all().
949 These functions have become deprecated: ERR_get_error_line(),
950 ERR_get_error_line_data(), ERR_peek_error_line_data(),
951 ERR_peek_last_error_line_data() and ERR_func_error_string().
953 Users are recommended to use ERR_get_error_all(), or to pick information
954 with ERR_peek functions and finish off with getting the error code by using
959 * Extended testing to be verbose for failing tests only. The make variables
960 VERBOSE_FAILURE or VF can be used to enable this:
962 $ make VF=1 test # Unix
963 $ mms /macro=(VF=1) test ! OpenVMS
964 $ nmake VF=1 test # Windows
968 * Added the `-copy_extensions` option to the `x509` command for use with
969 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
970 all extensions in the request are copied to the certificate or vice versa.
972 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
974 * Added the `-copy_extensions` option to the `req` command for use with
975 `-x509`. When given with the `copy` or `copyall` argument,
976 all extensions in the certification request are copied to the certificate.
980 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
981 they generate are by default RFC 5280 compliant in the following sense:
982 There is a subjectKeyIdentifier extension with a hash value of the public key
983 and for not self-signed certs there is an authorityKeyIdentifier extension
984 with a keyIdentifier field or issuer information identifying the signing key.
985 This is done unless some configuration overrides the new default behavior,
986 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
990 * Added several checks to `X509_verify_cert()` according to requirements in
991 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
992 (which may be done by using the CLI option `-x509_strict`):
993 * The basicConstraints of CA certificates must be marked critical.
994 * CA certificates must explicitly include the keyUsage extension.
995 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
996 * The issuer name of any certificate must not be empty.
997 * The subject name of CA certs, certs with keyUsage crlSign,
998 and certs without subjectAlternativeName must not be empty.
999 * If a subjectAlternativeName extension is given it must not be empty.
1000 * The signatureAlgorithm field and the cert signature must be consistent.
1001 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
1002 must not be marked critical.
1003 * The authorityKeyIdentifier must be given for X.509v3 certs
1004 unless they are self-signed.
1005 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
1009 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
1010 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
1014 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1015 used even when parsing explicit parameters, when loading a encoded key
1016 or calling `EC_GROUP_new_from_ecpkparameters()`/
1017 `EC_GROUP_new_from_ecparameters()`.
1018 This prevents bypass of security hardening and performance gains,
1019 especially for curves with specialized EC_METHODs.
1020 By default, if a key encoded with explicit parameters is loaded and later
1021 encoded, the output is still encoded with explicit parameters, even if
1022 internally a "named" EC_GROUP is used for computation.
1026 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1027 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1028 NULL. After this change, only the cofactor parameter can be NULL. It also
1029 does some minimal sanity checks on the passed order.
1034 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1035 An attack is simple, if the first CMS_recipientInfo is valid but the
1036 second CMS_recipientInfo is chosen ciphertext. If the second
1037 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1038 encryption key will be replaced by garbage, and the message cannot be
1039 decoded, but if the RSA decryption fails, the correct encryption key is
1040 used and the recipient will not notice the attack.
1041 As a work around for this potential attack the length of the decrypted
1042 key must be equal to the cipher default key length, in case the
1043 certifiate is not given and all recipientInfo are tried out.
1044 The old behaviour can be re-enabled in the CMS code by setting the
1045 CMS_DEBUG_DECRYPT flag.
1049 * Early start up entropy quality from the DEVRANDOM seed source has been
1050 improved for older Linux systems. The RAND subsystem will wait for
1051 /dev/random to be producing output before seeding from /dev/urandom.
1052 The seeded state is stored for future library initialisations using
1053 a system global shared memory segment. The shared memory identifier
1054 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1055 the desired value. The default identifier is 114.
1059 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1060 when primes for RSA keys are computed.
1061 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1062 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1063 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1064 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1065 This avoids possible fingerprinting of newly generated RSA modules.
1069 * Correct the extended master secret constant on EBCDIC systems. Without this
1070 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1071 negotiate EMS will fail. Unfortunately this also means that TLS connections
1072 between EBCDIC systems with this fix, and EBCDIC systems without this
1073 fix will fail if they negotiate EMS.
1077 * Changed the library initialisation so that the config file is now loaded
1078 by default. This was already the case for libssl. It now occurs for both
1079 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1080 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1084 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1085 where the former acts as a replacement for `ERR_put_error()`, and the
1086 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1087 `ERR_raise_data()` adds more flexibility by taking a format string and
1088 an arbitrary number of arguments following it, to be processed with
1093 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1094 to check if a named provider is loaded and available. When called, it
1095 will also activate all fallback providers if such are still present.
1099 * Enforce a minimum DH modulus size of 512 bits.
1103 * Changed DH parameters to generate the order q subgroup instead of 2q.
1104 Previously generated DH parameters are still accepted by DH_check
1105 but DH_generate_key works around that by clearing bit 0 of the
1106 private key for those. This avoids leaking bit 0 of the private key.
1110 * Significantly reduce secure memory usage by the randomness pools.
1114 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1119 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1120 algorithms. An implementation of a key exchange algorithm can be obtained
1121 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1122 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1123 the older EVP_PKEY_derive_init() function. See the man pages for the new
1124 functions for further details.
1128 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1132 * Removed the function names from error messages and deprecated the
1135 * Removed NextStep support and the macro OPENSSL_UNISTD
1139 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1140 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1141 Also removed "export var as function" capability; we do not export
1142 variables, only functions.
1146 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1147 an error and 1 indicating success. In previous versions of OpenSSL this
1148 was a void type. If a key was set longer than the maximum possible this
1153 * Support SM2 signing and verification schemes with X509 certificate.
1157 * Use SHA256 as the default digest for TS query in the `ts` app.
1161 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1162 This checks that the salt length is at least 128 bits, the derived key
1163 length is at least 112 bits, and that the iteration count is at least 1000.
1164 For backwards compatibility these checks are disabled by default in the
1165 default provider, but are enabled by default in the fips provider.
1166 To enable or disable these checks use the control
1167 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1171 * Default cipher lists/suites are now available via a function, the
1172 #defines are deprecated.
1176 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1177 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1178 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1182 * Join the directories crypto/x509 and crypto/x509v3
1186 * Added command 'openssl kdf' that uses the EVP_KDF API.
1190 * Added command 'openssl mac' that uses the EVP_MAC API.
1194 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1195 as default directories. Also added the command 'openssl info'
1196 for scripting purposes.
1200 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1201 deprecated. These undocumented functions were never integrated into the EVP
1202 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1203 Bi-directional IGE mode. These modes were never formally standardised and
1204 usage of these functions is believed to be very small. In particular
1205 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1206 is ever used. The security implications are believed to be minimal, but
1207 this issue was never fixed for backwards compatibility reasons. New code
1208 should not use these modes.
1212 * Add prediction resistance to the DRBG reseeding process.
1216 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1217 mandated by IEEE Std 1619-2018.
1221 * Added newline escaping functionality to a filename when using openssl dgst.
1222 This output format is to replicate the output format found in the `*sum`
1223 checksum programs. This aims to preserve backward compatibility.
1225 *Matt Eaton, Richard Levitte, and Paul Dale*
1227 * Removed the heartbeat message in DTLS feature, as it has very
1228 little usage and doesn't seem to fulfill a valuable purpose.
1229 The configuration option is now deprecated.
1233 * Changed the output of 'openssl {digestname} < file' to display the
1234 digest name in its output.
1238 * Added a new generic trace API which provides support for enabling
1239 instrumentation through trace output. This feature is mainly intended
1240 as an aid for developers and is disabled by default. To utilize it,
1241 OpenSSL needs to be configured with the `enable-trace` option.
1243 If the tracing API is enabled, the application can activate trace output
1244 by registering BIOs as trace channels for a number of tracing and debugging
1247 The `openssl` program has been expanded to enable any of the types
1248 available via environment variables defined by the user, and serves as
1249 one possible example on how to use this functionality.
1251 *Richard Levitte & Matthias St. Pierre*
1253 * Added build tests for C++. These are generated files that only do one
1254 thing, to include one public OpenSSL head file each. This tests that
1255 the public header files can be usefully included in a C++ application.
1257 This test isn't enabled by default. It can be enabled with the option
1258 'enable-buildtest-c++'.
1262 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1266 * Add KMAC to EVP_MAC.
1270 * Added property based algorithm implementation selection framework to
1275 * Added SCA hardening for modular field inversion in EC_GROUP through
1276 a new dedicated field_inv() pointer in EC_METHOD.
1277 This also addresses a leakage affecting conversions from projective
1278 to affine coordinates.
1280 *Billy Bob Brumley, Nicola Tuveri*
1282 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1283 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1284 those algorithms that were already supported through the EVP_PKEY API
1285 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1286 and scrypt are now wrappers that call EVP_KDF.
1290 * Build devcrypto engine as a dynamic engine.
1292 *Eneas U de Queiroz*
1294 * Add keyed BLAKE2 to EVP_MAC.
1298 * Fix a bug in the computation of the endpoint-pair shared secret used
1299 by DTLS over SCTP. This breaks interoperability with older versions
1300 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1301 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1302 interoperability with such broken implementations. However, enabling
1303 this switch breaks interoperability with correct implementations.
1305 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1306 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1310 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1314 * Change the license to the Apache License v2.0.
1318 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1320 - Major releases (indicated by incrementing the MAJOR release number)
1321 may introduce incompatible API/ABI changes.
1322 - Minor releases (indicated by incrementing the MINOR release number)
1323 may introduce new features but retain API/ABI compatibility.
1324 - Patch releases (indicated by incrementing the PATCH number)
1325 are intended for bug fixes and other improvements of existing
1326 features only (like improving performance or adding documentation)
1327 and retain API/ABI compatibility.
1331 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1335 * Remove the 'dist' target and add a tarball building script. The
1336 'dist' target has fallen out of use, and it shouldn't be
1337 necessary to configure just to create a source distribution.
1341 * Recreate the OS390-Unix config target. It no longer relies on a
1342 special script like it did for OpenSSL pre-1.1.0.
1346 * Instead of having the source directories listed in Configure, add
1347 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1352 * Add GMAC to EVP_MAC.
1356 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1360 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1361 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1362 to facilitate the continued use of MACs through raw private keys in
1363 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1367 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1368 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1372 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1373 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1374 are retained for backwards compatibility.
1378 * AES-XTS mode now enforces that its two keys are different to mitigate
1379 the attacked described in "Efficient Instantiations of Tweakable
1380 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1381 Details of this attack can be obtained from:
1382 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1386 * Rename the object files, i.e. give them other names than in previous
1387 versions. Their names now include the name of the final product, as
1388 well as its type mnemonic (bin, lib, shlib).
1392 * Added new option for 'openssl list', '-objects', which will display the
1393 list of built in objects, i.e. OIDs with names.
1397 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1398 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1403 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1404 improves application performance by removing data copies and providing
1405 applications with zero-copy system calls such as sendfile and splice.
1409 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1410 option is set, openssl cleanses (zeroize) plaintext bytes from
1411 internal buffers after delivering them to the application. Note,
1412 the application is still responsible for cleansing other copies
1413 (e.g.: data received by SSL_read(3)).
1417 * `PKCS12_parse` now maintains the order of the parsed certificates
1418 when outputting them via `*ca` (rather than reversing it).
1422 * Deprecated pthread fork support methods. These were unused so no
1423 replacement is required.
1425 - OPENSSL_fork_prepare()
1426 - OPENSSL_fork_parent()
1427 - OPENSSL_fork_child()
1434 ### Changes between 1.1.1i and 1.1.1j [xx XXX xxxx]
1436 * Fixed SRP_Calc_client_key so that it uses constant time. The previous
1437 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1438 could be exploited in a side channel attack to recover the password. Since
1439 the attack is local host only this is outside of the current OpenSSL
1440 threat model and therefore no CVE is assigned.
1442 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1447 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1449 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1450 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1451 If an attacker can control both items being compared then this could lead
1452 to a possible denial of service attack. OpenSSL itself uses the
1453 GENERAL_NAME_cmp function for two purposes:
1454 1) Comparing CRL distribution point names between an available CRL and a
1455 CRL distribution point embedded in an X509 certificate
1456 2) When verifying that a timestamp response token signer matches the
1457 timestamp authority name (exposed via the API functions
1458 TS_RESP_verify_response and TS_RESP_verify_token)
1463 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1465 * Certificates with explicit curve parameters are now disallowed in
1466 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1470 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1471 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1472 conversely, silently ignore DTLS protocol version bounds when configuring
1473 TLS-based contexts. The commands can be repeated to set bounds of both
1474 types. The same applies with the corresponding "min_protocol" and
1475 "max_protocol" command-line switches, in case some application uses both TLS
1478 SSL_CTX instances that are created for a fixed protocol version (e.g.
1479 TLSv1_server_method()) also silently ignore version bounds. Previously
1480 attempts to apply bounds to these protocol versions would result in an
1481 error. Now only the "version-flexible" SSL_CTX instances are subject to
1482 limits in configuration files in command-line options.
1486 * Handshake now fails if Extended Master Secret extension is dropped
1491 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1493 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1495 * Fixed segmentation fault in SSL_check_chain()
1496 Server or client applications that call the SSL_check_chain() function
1497 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1498 dereference as a result of incorrect handling of the
1499 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1500 or unrecognised signature algorithm is received from the peer. This could
1501 be exploited by a malicious peer in a Denial of Service attack.
1506 * Added AES consttime code for no-asm configurations
1507 an optional constant time support for AES was added
1508 when building openssl for no-asm.
1509 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1510 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1511 At this time this feature is by default disabled.
1512 It will be enabled by default in 3.0.
1516 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1518 * Revert the change of EOF detection while reading in libssl to avoid
1519 regressions in applications depending on the current way of reporting
1520 the EOF. As the existing method is not fully accurate the change to
1521 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1522 branch and will be present in the 3.0 release.
1526 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1527 when primes for RSA keys are computed.
1528 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1529 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1530 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1531 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1532 This avoids possible fingerprinting of newly generated RSA modules.
1536 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1538 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1539 while reading in libssl then we would report an error back to the
1540 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1541 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1542 therefore give a hint as to what went wrong.
1546 * Check that ed25519 and ed448 are allowed by the security level. Previously
1547 signature algorithms not using an MD were not being checked that they were
1548 allowed by the security level.
1552 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1553 was not quite right. The behaviour was not consistent between resumption
1554 and normal handshakes, and also not quite consistent with historical
1555 behaviour. The behaviour in various scenarios has been clarified and
1556 it has been updated to make it match historical behaviour as closely as
1561 * *[VMS only]* The header files that the VMS compilers include automatically,
1562 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1563 that the C++ compiler doesn't understand. This is a shortcoming in the
1564 compiler, but can be worked around with `__cplusplus` guards.
1566 C++ applications that use OpenSSL libraries must be compiled using the
1567 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1568 functions. Otherwise, only functions with symbols of less than 31
1569 characters can be used, as the linker will not be able to successfully
1570 resolve symbols with longer names.
1574 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1575 The presence of this system service is determined at run-time.
1579 * Added newline escaping functionality to a filename when using openssl dgst.
1580 This output format is to replicate the output format found in the `*sum`
1581 checksum programs. This aims to preserve backward compatibility.
1583 *Matt Eaton, Richard Levitte, and Paul Dale*
1585 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1590 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1592 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1593 number generator (RNG). This was intended to include protection in the
1594 event of a fork() system call in order to ensure that the parent and child
1595 processes did not share the same RNG state. However this protection was not
1596 being used in the default case.
1598 A partial mitigation for this issue is that the output from a high
1599 precision timer is mixed into the RNG state so the likelihood of a parent
1600 and child process sharing state is significantly reduced.
1602 If an application already calls OPENSSL_init_crypto() explicitly using
1603 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1606 *Matthias St. Pierre*
1608 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1609 used even when parsing explicit parameters, when loading a encoded key
1610 or calling `EC_GROUP_new_from_ecpkparameters()`/
1611 `EC_GROUP_new_from_ecparameters()`.
1612 This prevents bypass of security hardening and performance gains,
1613 especially for curves with specialized EC_METHODs.
1614 By default, if a key encoded with explicit parameters is loaded and later
1615 encoded, the output is still encoded with explicit parameters, even if
1616 internally a "named" EC_GROUP is used for computation.
1620 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1621 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1622 NULL. After this change, only the cofactor parameter can be NULL. It also
1623 does some minimal sanity checks on the passed order.
1628 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1629 An attack is simple, if the first CMS_recipientInfo is valid but the
1630 second CMS_recipientInfo is chosen ciphertext. If the second
1631 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1632 encryption key will be replaced by garbage, and the message cannot be
1633 decoded, but if the RSA decryption fails, the correct encryption key is
1634 used and the recipient will not notice the attack.
1635 As a work around for this potential attack the length of the decrypted
1636 key must be equal to the cipher default key length, in case the
1637 certifiate is not given and all recipientInfo are tried out.
1638 The old behaviour can be re-enabled in the CMS code by setting the
1639 CMS_DEBUG_DECRYPT flag.
1644 * Early start up entropy quality from the DEVRANDOM seed source has been
1645 improved for older Linux systems. The RAND subsystem will wait for
1646 /dev/random to be producing output before seeding from /dev/urandom.
1647 The seeded state is stored for future library initialisations using
1648 a system global shared memory segment. The shared memory identifier
1649 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1650 the desired value. The default identifier is 114.
1654 * Correct the extended master secret constant on EBCDIC systems. Without this
1655 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1656 negotiate EMS will fail. Unfortunately this also means that TLS connections
1657 between EBCDIC systems with this fix, and EBCDIC systems without this
1658 fix will fail if they negotiate EMS.
1662 * Use Windows installation paths in the mingw builds
1664 Mingw isn't a POSIX environment per se, which means that Windows
1665 paths should be used for installation.
1670 * Changed DH_check to accept parameters with order q and 2q subgroups.
1671 With order 2q subgroups the bit 0 of the private key is not secret
1672 but DH_generate_key works around that by clearing bit 0 of the
1673 private key for those. This avoids leaking bit 0 of the private key.
1677 * Significantly reduce secure memory usage by the randomness pools.
1681 * Revert the DEVRANDOM_WAIT feature for Linux systems
1683 The DEVRANDOM_WAIT feature added a select() call to wait for the
1684 /dev/random device to become readable before reading from the
1685 /dev/urandom device.
1687 It turned out that this change had negative side effects on
1688 performance which were not acceptable. After some discussion it
1689 was decided to revert this feature and leave it up to the OS
1690 resp. the platform maintainer to ensure a proper initialization
1691 during early boot time.
1693 *Matthias St. Pierre*
1695 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1697 * Add build tests for C++. These are generated files that only do one
1698 thing, to include one public OpenSSL head file each. This tests that
1699 the public header files can be usefully included in a C++ application.
1701 This test isn't enabled by default. It can be enabled with the option
1702 'enable-buildtest-c++'.
1706 * Enable SHA3 pre-hashing for ECDSA and DSA.
1710 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1711 This changes the size when using the `genpkey` command when no size is given.
1712 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1713 generation commands to use 2048 bits by default.
1717 * Reorganize the manual pages to consistently have RETURN VALUES,
1718 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1719 util/fix-doc-nits accordingly.
1721 *Paul Yang, Joshua Lock*
1723 * Add the missing accessor EVP_PKEY_get0_engine()
1727 * Have commands like `s_client` and `s_server` output the signature scheme
1728 along with other cipher suite parameters when debugging.
1732 * Make OPENSSL_config() error agnostic again.
1736 * Do the error handling in RSA decryption constant time.
1740 * Prevent over long nonces in ChaCha20-Poly1305.
1742 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1743 for every encryption operation. RFC 7539 specifies that the nonce value
1744 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1745 and front pads the nonce with 0 bytes if it is less than 12
1746 bytes. However it also incorrectly allows a nonce to be set of up to 16
1747 bytes. In this case only the last 12 bytes are significant and any
1748 additional leading bytes are ignored.
1750 It is a requirement of using this cipher that nonce values are
1751 unique. Messages encrypted using a reused nonce value are susceptible to
1752 serious confidentiality and integrity attacks. If an application changes
1753 the default nonce length to be longer than 12 bytes and then makes a
1754 change to the leading bytes of the nonce expecting the new value to be a
1755 new unique nonce then such an application could inadvertently encrypt
1756 messages with a reused nonce.
1758 Additionally the ignored bytes in a long nonce are not covered by the
1759 integrity guarantee of this cipher. Any application that relies on the
1760 integrity of these ignored leading bytes of a long nonce may be further
1761 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1762 is safe because no such use sets such a long nonce value. However user
1763 applications that use this cipher directly and set a non-default nonce
1764 length to be longer than 12 bytes may be vulnerable.
1766 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1772 * Add DEVRANDOM_WAIT feature for Linux systems
1774 On older Linux systems where the getrandom() system call is not available,
1775 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1776 Contrary to getrandom(), the /dev/urandom device will not block during
1777 early boot when the kernel CSPRNG has not been seeded yet.
1779 To mitigate this known weakness, use select() to wait for /dev/random to
1780 become readable before reading from /dev/urandom.
1782 * Ensure that SM2 only uses SM3 as digest algorithm
1786 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1788 * Change the info callback signals for the start and end of a post-handshake
1789 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1790 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1791 confused by this and assume that a TLSv1.2 renegotiation has started. This
1792 can break KeyUpdate handling. Instead we no longer signal the start and end
1793 of a post handshake message exchange (although the messages themselves are
1794 still signalled). This could break some applications that were expecting
1795 the old signals. However without this KeyUpdate is not usable for many
1800 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1802 * Timing vulnerability in DSA signature generation
1804 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1805 timing side channel attack. An attacker could use variations in the signing
1806 algorithm to recover the private key.
1808 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1813 * Timing vulnerability in ECDSA signature generation
1815 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1816 timing side channel attack. An attacker could use variations in the signing
1817 algorithm to recover the private key.
1819 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1824 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1825 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1826 of two gigabytes and the error handling improved.
1828 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1829 categorized as a normal bug, not a security issue, because the DRBG reseeds
1830 automatically and is fully functional even without additional randomness
1831 provided by the application.
1833 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1835 * Add a new ClientHello callback. Provides a callback interface that gives
1836 the application the ability to adjust the nascent SSL object at the
1837 earliest stage of ClientHello processing, immediately after extensions have
1838 been collected but before they have been processed. In particular, this
1839 callback can adjust the supported TLS versions in response to the contents
1844 * Add SM2 base algorithm support.
1848 * s390x assembly pack: add (improved) hardware-support for the following
1849 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1850 aes-cfb/cfb8, aes-ecb.
1854 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1855 parameter is no longer accepted, as it leads to a corrupt table. NULL
1856 pem_str is reserved for alias entries only.
1860 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1861 step for prime curves. The new implementation is based on formulae from
1862 differential addition-and-doubling in homogeneous projective coordinates
1863 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1864 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1865 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1866 to work in projective coordinates.
1868 *Billy Bob Brumley, Nicola Tuveri*
1870 * Change generating and checking of primes so that the error rate of not
1871 being prime depends on the intended use based on the size of the input.
1872 For larger primes this will result in more rounds of Miller-Rabin.
1873 The maximal error rate for primes with more than 1080 bits is lowered
1876 *Kurt Roeckx, Annie Yousar*
1878 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1882 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1883 moving between systems, and to avoid confusion when a Windows build is
1884 done with mingw vs with MSVC. For POSIX installs, there's still a
1885 symlink or copy named 'tsget' to avoid that confusion as well.
1889 * Revert blinding in ECDSA sign and instead make problematic addition
1890 length-invariant. Switch even to fixed-length Montgomery multiplication.
1894 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1895 step for binary curves. The new implementation is based on formulae from
1896 differential addition-and-doubling in mixed Lopez-Dahab projective
1897 coordinates, modified to independently blind the operands.
1899 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1901 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1902 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1903 EC_METHODs to implement their own specialized "ladder step", to take
1904 advantage of more favorable coordinate systems or more efficient
1905 differential addition-and-doubling algorithms.
1907 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1909 * Modified the random device based seed sources to keep the relevant
1910 file descriptors open rather than reopening them on each access.
1911 This allows such sources to operate in a chroot() jail without
1912 the associated device nodes being available. This behaviour can be
1913 controlled using RAND_keep_random_devices_open().
1917 * Numerous side-channel attack mitigations have been applied. This may have
1918 performance impacts for some algorithms for the benefit of improved
1919 security. Specific changes are noted in this change log by their respective
1924 * AIX shared library support overhaul. Switch to AIX "natural" way of
1925 handling shared libraries, which means collecting shared objects of
1926 different versions and bitnesses in one common archive. This allows to
1927 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1928 doesn't affect the way 3rd party applications are linked, only how
1929 multi-version installation is managed.
1933 * Make ec_group_do_inverse_ord() more robust and available to other
1934 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1935 mitigations are applied to the fallback BN_mod_inverse().
1936 When using this function rather than BN_mod_inverse() directly, new
1937 EC cryptosystem implementations are then safer-by-default.
1941 * Add coordinate blinding for EC_POINT and implement projective
1942 coordinate blinding for generic prime curves as a countermeasure to
1943 chosen point SCA attacks.
1945 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1947 * Add blinding to ECDSA and DSA signatures to protect against side channel
1948 attacks discovered by Keegan Ryan (NCC Group).
1952 * Enforce checking in the `pkeyutl` command to ensure that the input
1953 length does not exceed the maximum supported digest length when performing
1954 a sign, verify or verifyrecover operation.
1958 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1959 I/O in combination with something like select() or poll() will hang. This
1960 can be turned off again using SSL_CTX_clear_mode().
1961 Many applications do not properly handle non-application data records, and
1962 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1963 around the problems in those applications, but can also break some.
1964 It's recommended to read the manpages about SSL_read(), SSL_write(),
1965 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1966 SSL_CTX_set_read_ahead() again.
1970 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1971 now allow empty (zero character) pass phrases.
1975 * Apply blinding to binary field modular inversion and remove patent
1976 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1980 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1981 binary and prime elliptic curves.
1985 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1986 constant time fixed point multiplication.
1990 * Revise elliptic curve scalar multiplication with timing attack
1991 defenses: ec_wNAF_mul redirects to a constant time implementation
1992 when computing fixed point and variable point multiplication (which
1993 in OpenSSL are mostly used with secret scalars in keygen, sign,
1994 ECDH derive operations).
1995 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1998 * Updated CONTRIBUTING
2002 * Updated DRBG / RAND to request nonce and additional low entropy
2003 randomness from the system.
2005 *Matthias St. Pierre*
2007 * Updated 'openssl rehash' to use OpenSSL consistent default.
2011 * Moved the load of the ssl_conf module to libcrypto, which helps
2012 loading engines that libssl uses before libssl is initialised.
2016 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
2020 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
2022 *Ingo Schwarze, Rich Salz*
2024 * Added output of accepting IP address and port for 'openssl s_server'
2028 * Added a new API for TLSv1.3 ciphersuites:
2029 SSL_CTX_set_ciphersuites()
2030 SSL_set_ciphersuites()
2034 * Memory allocation failures consistently add an error to the error
2039 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2040 in libcrypto when run as setuid/setgid.
2044 * Load any config file by default when libssl is used.
2048 * Added new public header file <openssl/rand_drbg.h> and documentation
2049 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2051 *Matthias St. Pierre*
2053 * QNX support removed (cannot find contributors to get their approval
2054 for the license change).
2058 * TLSv1.3 replay protection for early data has been implemented. See the
2059 SSL_read_early_data() man page for further details.
2063 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2064 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2065 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2066 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2067 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2068 configuration has been separated out. See the ciphers man page or the
2069 SSL_CTX_set_ciphersuites() man page for more information.
2073 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2074 in responder mode now supports the new "-multi" option, which
2075 spawns the specified number of child processes to handle OCSP
2076 requests. The "-timeout" option now also limits the OCSP
2077 responder's patience to wait to receive the full client request
2078 on a newly accepted connection. Child processes are respawned
2079 as needed, and the CA index file is automatically reloaded
2080 when changed. This makes it possible to run the "ocsp" responder
2081 as a long-running service, making the OpenSSL CA somewhat more
2082 feature-complete. In this mode, most diagnostic messages logged
2083 after entering the event loop are logged via syslog(3) rather than
2088 * Added support for X448 and Ed448. Heavily based on original work by
2093 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2094 objects loaded. This adds the functions OSSL_STORE_expect() and
2095 OSSL_STORE_find() as well as needed tools to construct searches and
2096 get the search data out of them.
2100 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2101 version of OpenSSL should review their configuration settings to ensure
2102 that they are still appropriate for TLSv1.3. For further information see:
2103 <https://wiki.openssl.org/index.php/TLS1.3>
2107 * Grand redesign of the OpenSSL random generator
2109 The default RAND method now utilizes an AES-CTR DRBG according to
2110 NIST standard SP 800-90Ar1. The new random generator is essentially
2111 a port of the default random generator from the OpenSSL FIPS 2.0
2112 object module. It is a hybrid deterministic random bit generator
2113 using an AES-CTR bit stream and which seeds and reseeds itself
2114 automatically using trusted system entropy sources.
2116 Some of its new features are:
2117 - Support for multiple DRBG instances with seed chaining.
2118 - The default RAND method makes use of a DRBG.
2119 - There is a public and private DRBG instance.
2120 - The DRBG instances are fork-safe.
2121 - Keep all global DRBG instances on the secure heap if it is enabled.
2122 - The public and private DRBG instance are per thread for lock free
2125 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2127 * Changed Configure so it only says what it does and doesn't dump
2128 so much data. Instead, ./configdata.pm should be used as a script
2129 to display all sorts of configuration data.
2133 * Added processing of "make variables" to Configure.
2137 * Added SHA512/224 and SHA512/256 algorithm support.
2141 * The last traces of Netware support, first removed in 1.1.0, have
2146 * Get rid of Makefile.shared, and in the process, make the processing
2147 of certain files (rc.obj, or the .def/.map/.opt files produced from
2148 the ordinal files) more visible and hopefully easier to trace and
2149 debug (or make silent).
2153 * Make it possible to have environment variable assignments as
2154 arguments to config / Configure.
2158 * Add multi-prime RSA (RFC 8017) support.
2162 * Add SM3 implemented according to GB/T 32905-2016
2163 *Jack Lloyd <jack.lloyd@ribose.com>,*
2164 *Ronald Tse <ronald.tse@ribose.com>,*
2165 *Erick Borsboom <erick.borsboom@ribose.com>*
2167 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2168 as documented in RFC6066.
2169 Based on a patch from Tomasz Moń
2171 *Filipe Raimundo da Silva*
2173 * Add SM4 implemented according to GB/T 32907-2016.
2174 *Jack Lloyd <jack.lloyd@ribose.com>,*
2175 *Ronald Tse <ronald.tse@ribose.com>,*
2176 *Erick Borsboom <erick.borsboom@ribose.com>*
2178 * Reimplement -newreq-nodes and ERR_error_string_n; the
2179 original author does not agree with the license change.
2183 * Add ARIA AEAD TLS support.
2187 * Some macro definitions to support VS6 have been removed. Visual
2188 Studio 6 has not worked since 1.1.0
2192 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2193 without clearing the errors.
2197 * Add "atfork" functions. If building on a system that without
2198 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2199 requirements. The RAND facility now uses/requires this.
2207 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2208 not possible to disable entirely. However, it's still possible to
2209 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2212 To disable, configure with 'no-ui-console'. 'no-ui' is still
2213 possible to use as an alias. Check at compile time with the
2214 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2215 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2219 * Add a STORE module, which implements a uniform and URI based reader of
2220 stores that can contain keys, certificates, CRLs and numerous other
2221 objects. The main API is loosely based on a few stdio functions,
2222 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2223 OSSL_STORE_error and OSSL_STORE_close.
2224 The implementation uses backends called "loaders" to implement arbitrary
2225 URI schemes. There is one built in "loader" for the 'file' scheme.
2229 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2230 then adjusted to work on FreeBSD 8.4 as well.
2231 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2232 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2236 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2237 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2238 error code calls like this:
2240 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2242 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2243 that can be encoded in C. For the foreseeable future, this will only
2246 *Richard Levitte and Tim Hudson*
2248 * Removed BSD cryptodev engine.
2252 * Add a build target 'build_all_generated', to build all generated files
2253 and only that. This can be used to prepare everything that requires
2254 things like perl for a system that lacks perl and then move everything
2255 to that system and do the rest of the build there.
2259 * In the UI interface, make it possible to duplicate the user data. This
2260 can be used by engines that need to retain the data for a longer time
2261 than just the call where this user data is passed.
2265 * Ignore the '-named_curve auto' value for compatibility of applications
2268 *Tomas Mraz <tmraz@fedoraproject.org>*
2270 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2271 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2272 alerts across multiple records (some of which could be empty). In practice
2273 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2274 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2275 support this at all. Supporting it adds significant complexity to the
2276 record layer, and its removal is unlikely to cause interoperability
2281 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2282 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2283 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2288 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2289 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2291 *Richard Levitte, Andy Polyakov*
2293 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2298 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2299 platform rather than 'mingw'.
2303 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2304 success if they are asked to add an object which already exists
2305 in the store. This change cascades to other functions which load
2306 certificates and CRLs.
2310 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2311 facilitate stack unwinding even from assembly subroutines.
2315 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2316 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2320 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2321 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2322 which is the minimum version we support.
2326 * Certificate time validation (X509_cmp_time) enforces stricter
2327 compliance with RFC 5280. Fractional seconds and timezone offsets
2328 are no longer allowed.
2332 * Add support for ARIA
2336 * s_client will now send the Server Name Indication (SNI) extension by
2337 default unless the new "-noservername" option is used. The server name is
2338 based on the host provided to the "-connect" option unless overridden by
2339 using "-servername".
2343 * Add support for SipHash
2347 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2348 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2349 prevent issues where no progress is being made and the peer continually
2350 sends unrecognised record types, using up resources processing them.
2354 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2355 using the algorithm defined in
2356 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2360 * Heartbeat support has been removed; the ABI is changed for now.
2362 *Richard Levitte, Rich Salz*
2364 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2368 * The RSA "null" method, which was partially supported to avoid patent
2369 issues, has been replaced to always returns NULL.
2376 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2378 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2379 used even when parsing explicit parameters, when loading a encoded key
2380 or calling `EC_GROUP_new_from_ecpkparameters()`/
2381 `EC_GROUP_new_from_ecparameters()`.
2382 This prevents bypass of security hardening and performance gains,
2383 especially for curves with specialized EC_METHODs.
2384 By default, if a key encoded with explicit parameters is loaded and later
2385 encoded, the output is still encoded with explicit parameters, even if
2386 internally a "named" EC_GROUP is used for computation.
2390 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2391 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2392 NULL. After this change, only the cofactor parameter can be NULL. It also
2393 does some minimal sanity checks on the passed order.
2398 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2399 An attack is simple, if the first CMS_recipientInfo is valid but the
2400 second CMS_recipientInfo is chosen ciphertext. If the second
2401 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2402 encryption key will be replaced by garbage, and the message cannot be
2403 decoded, but if the RSA decryption fails, the correct encryption key is
2404 used and the recipient will not notice the attack.
2405 As a work around for this potential attack the length of the decrypted
2406 key must be equal to the cipher default key length, in case the
2407 certifiate is not given and all recipientInfo are tried out.
2408 The old behaviour can be re-enabled in the CMS code by setting the
2409 CMS_DEBUG_DECRYPT flag.
2414 * Use Windows installation paths in the mingw builds
2416 Mingw isn't a POSIX environment per se, which means that Windows
2417 paths should be used for installation.
2422 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2424 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2425 This changes the size when using the `genpkey` command when no size is given.
2426 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2427 generation commands to use 2048 bits by default.
2431 * Prevent over long nonces in ChaCha20-Poly1305.
2433 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2434 for every encryption operation. RFC 7539 specifies that the nonce value
2435 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2436 and front pads the nonce with 0 bytes if it is less than 12
2437 bytes. However it also incorrectly allows a nonce to be set of up to 16
2438 bytes. In this case only the last 12 bytes are significant and any
2439 additional leading bytes are ignored.
2441 It is a requirement of using this cipher that nonce values are
2442 unique. Messages encrypted using a reused nonce value are susceptible to
2443 serious confidentiality and integrity attacks. If an application changes
2444 the default nonce length to be longer than 12 bytes and then makes a
2445 change to the leading bytes of the nonce expecting the new value to be a
2446 new unique nonce then such an application could inadvertently encrypt
2447 messages with a reused nonce.
2449 Additionally the ignored bytes in a long nonce are not covered by the
2450 integrity guarantee of this cipher. Any application that relies on the
2451 integrity of these ignored leading bytes of a long nonce may be further
2452 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2453 is safe because no such use sets such a long nonce value. However user
2454 applications that use this cipher directly and set a non-default nonce
2455 length to be longer than 12 bytes may be vulnerable.
2457 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2463 * Added SCA hardening for modular field inversion in EC_GROUP through
2464 a new dedicated field_inv() pointer in EC_METHOD.
2465 This also addresses a leakage affecting conversions from projective
2466 to affine coordinates.
2468 *Billy Bob Brumley, Nicola Tuveri*
2470 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2471 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2475 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2479 * Remove the 'dist' target and add a tarball building script. The
2480 'dist' target has fallen out of use, and it shouldn't be
2481 necessary to configure just to create a source distribution.
2485 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2487 * Timing vulnerability in DSA signature generation
2489 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2490 timing side channel attack. An attacker could use variations in the signing
2491 algorithm to recover the private key.
2493 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2498 * Timing vulnerability in ECDSA signature generation
2500 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2501 timing side channel attack. An attacker could use variations in the signing
2502 algorithm to recover the private key.
2504 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2509 * Add coordinate blinding for EC_POINT and implement projective
2510 coordinate blinding for generic prime curves as a countermeasure to
2511 chosen point SCA attacks.
2513 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2515 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2517 * Client DoS due to large DH parameter
2519 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2520 malicious server can send a very large prime value to the client. This will
2521 cause the client to spend an unreasonably long period of time generating a
2522 key for this prime resulting in a hang until the client has finished. This
2523 could be exploited in a Denial Of Service attack.
2525 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2530 * Cache timing vulnerability in RSA Key Generation
2532 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2533 a cache timing side channel attack. An attacker with sufficient access to
2534 mount cache timing attacks during the RSA key generation process could
2535 recover the private key.
2537 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2538 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2543 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2544 parameter is no longer accepted, as it leads to a corrupt table. NULL
2545 pem_str is reserved for alias entries only.
2549 * Revert blinding in ECDSA sign and instead make problematic addition
2550 length-invariant. Switch even to fixed-length Montgomery multiplication.
2554 * Change generating and checking of primes so that the error rate of not
2555 being prime depends on the intended use based on the size of the input.
2556 For larger primes this will result in more rounds of Miller-Rabin.
2557 The maximal error rate for primes with more than 1080 bits is lowered
2560 *Kurt Roeckx, Annie Yousar*
2562 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2566 * Add blinding to ECDSA and DSA signatures to protect against side channel
2567 attacks discovered by Keegan Ryan (NCC Group).
2571 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2572 now allow empty (zero character) pass phrases.
2576 * Certificate time validation (X509_cmp_time) enforces stricter
2577 compliance with RFC 5280. Fractional seconds and timezone offsets
2578 are no longer allowed.
2582 * Fixed a text canonicalisation bug in CMS
2584 Where a CMS detached signature is used with text content the text goes
2585 through a canonicalisation process first prior to signing or verifying a
2586 signature. This process strips trailing space at the end of lines, converts
2587 line terminators to CRLF and removes additional trailing line terminators
2588 at the end of a file. A bug in the canonicalisation process meant that
2589 some characters, such as form-feed, were incorrectly treated as whitespace
2590 and removed. This is contrary to the specification (RFC5485). This fix
2591 could mean that detached text data signed with an earlier version of
2592 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2593 signed with a fixed OpenSSL may fail to verify with an earlier version of
2594 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2595 and use the "-binary" flag (for the "cms" command line application) or set
2596 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2600 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2602 * Constructed ASN.1 types with a recursive definition could exceed the stack
2604 Constructed ASN.1 types with a recursive definition (such as can be found
2605 in PKCS7) could eventually exceed the stack given malicious input with
2606 excessive recursion. This could result in a Denial Of Service attack. There
2607 are no such structures used within SSL/TLS that come from untrusted sources
2608 so this is considered safe.
2610 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2616 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2618 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2619 effectively reduced to only comparing the least significant bit of each
2620 byte. This allows an attacker to forge messages that would be considered as
2621 authenticated in an amount of tries lower than that guaranteed by the
2622 security claims of the scheme. The module can only be compiled by the
2623 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2625 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2631 * Add a build target 'build_all_generated', to build all generated files
2632 and only that. This can be used to prepare everything that requires
2633 things like perl for a system that lacks perl and then move everything
2634 to that system and do the rest of the build there.
2638 * Backport SSL_OP_NO_RENGOTIATION
2640 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2641 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2642 changes this is no longer possible in 1.1.0. Therefore the new
2643 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2644 1.1.0 to provide equivalent functionality.
2646 Note that if an application built against 1.1.0h headers (or above) is run
2647 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2648 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2652 * Removed the OS390-Unix config target. It relied on a script that doesn't
2657 * rsaz_1024_mul_avx2 overflow bug on x86_64
2659 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2660 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2661 Analysis suggests that attacks against RSA and DSA as a result of this
2662 defect would be very difficult to perform and are not believed likely.
2663 Attacks against DH1024 are considered just feasible, because most of the
2664 work necessary to deduce information about a private key may be performed
2665 offline. The amount of resources required for such an attack would be
2666 significant. However, for an attack on TLS to be meaningful, the server
2667 would have to share the DH1024 private key among multiple clients, which is
2668 no longer an option since CVE-2016-0701.
2670 This only affects processors that support the AVX2 but not ADX extensions
2671 like Intel Haswell (4th generation).
2673 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2674 was originally found via the OSS-Fuzz project.
2679 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2681 * bn_sqrx8x_internal carry bug on x86_64
2683 There is a carry propagating bug in the x86_64 Montgomery squaring
2684 procedure. No EC algorithms are affected. Analysis suggests that attacks
2685 against RSA and DSA as a result of this defect would be very difficult to
2686 perform and are not believed likely. Attacks against DH are considered just
2687 feasible (although very difficult) because most of the work necessary to
2688 deduce information about a private key may be performed offline. The amount
2689 of resources required for such an attack would be very significant and
2690 likely only accessible to a limited number of attackers. An attacker would
2691 additionally need online access to an unpatched system using the target
2692 private key in a scenario with persistent DH parameters and a private
2693 key that is shared between multiple clients.
2695 This only affects processors that support the BMI1, BMI2 and ADX extensions
2696 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2698 This issue was reported to OpenSSL by the OSS-Fuzz project.
2703 * Malformed X.509 IPAddressFamily could cause OOB read
2705 If an X.509 certificate has a malformed IPAddressFamily extension,
2706 OpenSSL could do a one-byte buffer overread. The most likely result
2707 would be an erroneous display of the certificate in text format.
2709 This issue was reported to OpenSSL by the OSS-Fuzz project.
2714 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2716 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2717 platform rather than 'mingw'.
2721 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2722 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2723 which is the minimum version we support.
2727 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2729 * Encrypt-Then-Mac renegotiation crash
2731 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2732 negotiated where it was not in the original handshake (or vice-versa) then
2733 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2734 and servers are affected.
2736 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2741 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2743 * Truncated packet could crash via OOB read
2745 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2746 cipher is being used, then a truncated packet can cause that host to
2747 perform an out-of-bounds read, usually resulting in a crash.
2749 This issue was reported to OpenSSL by Robert Święcki of Google.
2754 * Bad (EC)DHE parameters cause a client crash
2756 If a malicious server supplies bad parameters for a DHE or ECDHE key
2757 exchange then this can result in the client attempting to dereference a
2758 NULL pointer leading to a client crash. This could be exploited in a Denial
2761 This issue was reported to OpenSSL by Guido Vranken.
2766 * BN_mod_exp may produce incorrect results on x86_64
2768 There is a carry propagating bug in the x86_64 Montgomery squaring
2769 procedure. No EC algorithms are affected. Analysis suggests that attacks
2770 against RSA and DSA as a result of this defect would be very difficult to
2771 perform and are not believed likely. Attacks against DH are considered just
2772 feasible (although very difficult) because most of the work necessary to
2773 deduce information about a private key may be performed offline. The amount
2774 of resources required for such an attack would be very significant and
2775 likely only accessible to a limited number of attackers. An attacker would
2776 additionally need online access to an unpatched system using the target
2777 private key in a scenario with persistent DH parameters and a private
2778 key that is shared between multiple clients. For example this can occur by
2779 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2780 similar to CVE-2015-3193 but must be treated as a separate problem.
2782 This issue was reported to OpenSSL by the OSS-Fuzz project.
2787 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2789 * ChaCha20/Poly1305 heap-buffer-overflow
2791 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2792 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2793 crash. This issue is not considered to be exploitable beyond a DoS.
2795 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2800 * CMS Null dereference
2802 Applications parsing invalid CMS structures can crash with a NULL pointer
2803 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2804 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2805 structure callback if an attempt is made to free certain invalid encodings.
2806 Only CHOICE structures using a callback which do not handle NULL value are
2809 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2814 * Montgomery multiplication may produce incorrect results
2816 There is a carry propagating bug in the Broadwell-specific Montgomery
2817 multiplication procedure that handles input lengths divisible by, but
2818 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2819 and DH private keys are impossible. This is because the subroutine in
2820 question is not used in operations with the private key itself and an input
2821 of the attacker's direct choice. Otherwise the bug can manifest itself as
2822 transient authentication and key negotiation failures or reproducible
2823 erroneous outcome of public-key operations with specially crafted input.
2824 Among EC algorithms only Brainpool P-512 curves are affected and one
2825 presumably can attack ECDH key negotiation. Impact was not analyzed in
2826 detail, because pre-requisites for attack are considered unlikely. Namely
2827 multiple clients have to choose the curve in question and the server has to
2828 share the private key among them, neither of which is default behaviour.
2829 Even then only clients that chose the curve will be affected.
2831 This issue was publicly reported as transient failures and was not
2832 initially recognized as a security issue. Thanks to Richard Morgan for
2833 providing reproducible case.
2838 * Removed automatic addition of RPATH in shared libraries and executables,
2839 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2843 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2845 * Fix Use After Free for large message sizes
2847 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2848 message larger than approx 16k is received then the underlying buffer to
2849 store the incoming message is reallocated and moved. Unfortunately a
2850 dangling pointer to the old location is left which results in an attempt to
2851 write to the previously freed location. This is likely to result in a
2852 crash, however it could potentially lead to execution of arbitrary code.
2854 This issue only affects OpenSSL 1.1.0a.
2856 This issue was reported to OpenSSL by Robert Święcki.
2861 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2863 * OCSP Status Request extension unbounded memory growth
2865 A malicious client can send an excessively large OCSP Status Request
2866 extension. If that client continually requests renegotiation, sending a
2867 large OCSP Status Request extension each time, then there will be unbounded
2868 memory growth on the server. This will eventually lead to a Denial Of
2869 Service attack through memory exhaustion. Servers with a default
2870 configuration are vulnerable even if they do not support OCSP. Builds using
2871 the "no-ocsp" build time option are not affected.
2873 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2878 * SSL_peek() hang on empty record
2880 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2881 sends an empty record. This could be exploited by a malicious peer in a
2882 Denial Of Service attack.
2884 This issue was reported to OpenSSL by Alex Gaynor.
2889 * Excessive allocation of memory in tls_get_message_header() and
2890 dtls1_preprocess_fragment()
2892 A (D)TLS message includes 3 bytes for its length in the header for the
2893 message. This would allow for messages up to 16Mb in length. Messages of
2894 this length are excessive and OpenSSL includes a check to ensure that a
2895 peer is sending reasonably sized messages in order to avoid too much memory
2896 being consumed to service a connection. A flaw in the logic of version
2897 1.1.0 means that memory for the message is allocated too early, prior to
2898 the excessive message length check. Due to way memory is allocated in
2899 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2900 to service a connection. This could lead to a Denial of Service through
2901 memory exhaustion. However, the excessive message length check still takes
2902 place, and this would cause the connection to immediately fail. Assuming
2903 that the application calls SSL_free() on the failed connection in a timely
2904 manner then the 21Mb of allocated memory will then be immediately freed
2905 again. Therefore the excessive memory allocation will be transitory in
2906 nature. This then means that there is only a security impact if:
2908 1) The application does not call SSL_free() in a timely manner in the event
2909 that the connection fails
2911 2) The application is working in a constrained environment where there is
2912 very little free memory
2914 3) The attacker initiates multiple connection attempts such that there are
2915 multiple connections in a state where memory has been allocated for the
2916 connection; SSL_free() has not yet been called; and there is insufficient
2917 memory to service the multiple requests.
2919 Except in the instance of (1) above any Denial Of Service is likely to be
2920 transitory because as soon as the connection fails the memory is
2921 subsequently freed again in the SSL_free() call. However there is an
2922 increased risk during this period of application crashes due to the lack of
2923 memory - which would then mean a more serious Denial of Service.
2925 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2926 (CVE-2016-6307 and CVE-2016-6308)
2930 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2931 had to be removed. Primary reason is that vendor assembler can't
2932 assemble our modules with -KPIC flag. As result it, assembly
2933 support, was not even available as option. But its lack means
2934 lack of side-channel resistant code, which is incompatible with
2935 security by todays standards. Fortunately gcc is readily available
2936 prepackaged option, which we firmly point at...
2940 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2942 * Windows command-line tool supports UTF-8 opt-in option for arguments
2943 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2944 (to any value) allows Windows user to access PKCS#12 file generated
2945 with Windows CryptoAPI and protected with non-ASCII password, as well
2946 as files generated under UTF-8 locale on Linux also protected with
2951 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2952 have been disabled by default and removed from DEFAULT, just like RC4.
2953 See the RC4 item below to re-enable both.
2957 * The method for finding the storage location for the Windows RAND seed file
2958 has changed. First we check %RANDFILE%. If that is not set then we check
2959 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2960 all else fails we fall back to C:\.
2964 * The EVP_EncryptUpdate() function has had its return type changed from void
2965 to int. A return of 0 indicates and error while a return of 1 indicates
2970 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2971 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2972 off the constant time implementation for RSA, DSA and DH have been made
2973 no-ops and deprecated.
2977 * Windows RAND implementation was simplified to only get entropy by
2978 calling CryptGenRandom(). Various other RAND-related tickets
2981 *Joseph Wylie Yandle, Rich Salz*
2983 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2984 and `OPENSSL_LH_`, respectively. The old names are available
2985 with API compatibility. They new names are now completely documented.
2989 * Unify TYPE_up_ref(obj) methods signature.
2990 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2991 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2992 int (instead of void) like all others TYPE_up_ref() methods.
2993 So now these methods also check the return value of CRYPTO_atomic_add(),
2994 and the validity of object reference counter.
2996 *fdasilvayy@gmail.com*
2998 * With Windows Visual Studio builds, the .pdb files are installed
2999 alongside the installed libraries and executables. For a static
3000 library installation, ossl_static.pdb is the associate compiler
3001 generated .pdb file to be used when linking programs.
3005 * Remove openssl.spec. Packaging files belong with the packagers.
3009 * Automatic Darwin/OSX configuration has had a refresh, it will now
3010 recognise x86_64 architectures automatically. You can still decide
3011 to build for a different bitness with the environment variable
3012 KERNEL_BITS (can be 32 or 64), for example:
3014 KERNEL_BITS=32 ./config
3018 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
3019 256 bit AES and HMAC with SHA256.
3023 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
3027 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3031 * To enable users to have their own config files and build file templates,
3032 Configure looks in the directory indicated by the environment variable
3033 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3034 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3035 name and is used as is.
3039 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3040 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3041 X509_CERT_FILE_CTX was removed.
3045 * "shared" builds are now the default. To create only static libraries use
3046 the "no-shared" Configure option.
3050 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3051 All of these option have not worked for some while and are fundamental
3056 * Make various cleanup routines no-ops and mark them as deprecated. Most
3057 global cleanup functions are no longer required because they are handled
3058 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3059 Explicitly de-initing can cause problems (e.g. where a library that uses
3060 OpenSSL de-inits, but an application is still using it). The affected
3061 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3062 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3063 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3064 COMP_zlib_cleanup().
3068 * --strict-warnings no longer enables runtime debugging options
3069 such as REF_DEBUG. Instead, debug options are automatically
3070 enabled with '--debug' builds.
3072 *Andy Polyakov, Emilia Käsper*
3074 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3075 have been moved out of the public header files. New functions for managing
3076 these have been added.
3080 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3081 objects have been moved out of the public header files. New
3082 functions for managing these have been added.
3086 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3087 have been moved out of the public header files. New functions for managing
3088 these have been added.
3092 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3093 moved out of the public header files. New functions for managing these
3098 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3102 * Removed the mk1mf build scripts.
3106 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3107 it is always safe to #include a header now.
3111 * Removed the aged BC-32 config and all its supporting scripts
3115 * Removed support for Ultrix, Netware, and OS/2.
3119 * Add support for HKDF.
3121 *Alessandro Ghedini*
3123 * Add support for blake2b and blake2s
3127 * Added support for "pipelining". Ciphers that have the
3128 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3129 encryptions/decryptions simultaneously. There are currently no built-in
3130 ciphers with this property but the expectation is that engines will be able
3131 to offer it to significantly improve throughput. Support has been extended
3132 into libssl so that multiple records for a single connection can be
3133 processed in one go (for >=TLS 1.1).
3137 * Added the AFALG engine. This is an async capable engine which is able to
3138 offload work to the Linux kernel. In this initial version it only supports
3139 AES128-CBC. The kernel must be version 4.1.0 or greater.
3143 * OpenSSL now uses a new threading API. It is no longer necessary to
3144 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3145 are two supported threading models: pthreads and windows threads. It is
3146 also possible to configure OpenSSL at compile time for "no-threads". The
3147 old threading API should no longer be used. The functions have been
3148 replaced with "no-op" compatibility macros.
3150 *Alessandro Ghedini, Matt Caswell*
3152 * Modify behavior of ALPN to invoke callback after SNI/servername
3153 callback, such that updates to the SSL_CTX affect ALPN.
3157 * Add SSL_CIPHER queries for authentication and key-exchange.
3161 * Changes to the DEFAULT cipherlist:
3162 - Prefer (EC)DHE handshakes over plain RSA.
3163 - Prefer AEAD ciphers over legacy ciphers.
3164 - Prefer ECDSA over RSA when both certificates are available.
3165 - Prefer TLSv1.2 ciphers/PRF.
3166 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3171 * Change the ECC default curve list to be this, in order: x25519,
3172 secp256r1, secp521r1, secp384r1.
3176 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3177 disabled by default. They can be re-enabled using the
3178 enable-weak-ssl-ciphers option to Configure.
3182 * If the server has ALPN configured, but supports no protocols that the
3183 client advertises, send a fatal "no_application_protocol" alert.
3184 This behaviour is SHALL in RFC 7301, though it isn't universally
3185 implemented by other servers.
3189 * Add X25519 support.
3190 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3191 for public and private key encoding using the format documented in
3192 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3193 key generation and key derivation.
3195 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3200 * Deprecate SRP_VBASE_get_by_user.
3201 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3202 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3203 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3204 seed, even if the seed is configured.
3206 Users should use SRP_VBASE_get1_by_user instead. Note that in
3207 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3208 also that even though configuring the SRP seed attempts to hide
3209 invalid usernames by continuing the handshake with fake
3210 credentials, this behaviour is not constant time and no strong
3211 guarantees are made that the handshake is indistinguishable from
3212 that of a valid user.
3216 * Configuration change; it's now possible to build dynamic engines
3217 without having to build shared libraries and vice versa. This
3218 only applies to the engines in `engines/`, those in `crypto/engine/`
3219 will always be built into libcrypto (i.e. "static").
3221 Building dynamic engines is enabled by default; to disable, use
3222 the configuration option "disable-dynamic-engine".
3224 The only requirements for building dynamic engines are the
3225 presence of the DSO module and building with position independent
3226 code, so they will also automatically be disabled if configuring
3227 with "disable-dso" or "disable-pic".
3229 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3230 are also taken away from openssl/opensslconf.h, as they are
3235 * Configuration change; if there is a known flag to compile
3236 position independent code, it will always be applied on the
3237 libcrypto and libssl object files, and never on the application
3238 object files. This means other libraries that use routines from
3239 libcrypto / libssl can be made into shared libraries regardless
3240 of how OpenSSL was configured.
3242 If this isn't desirable, the configuration options "disable-pic"
3243 or "no-pic" can be used to disable the use of PIC. This will
3244 also disable building shared libraries and dynamic engines.
3248 * Removed JPAKE code. It was experimental and has no wide use.
3252 * The INSTALL_PREFIX Makefile variable has been renamed to
3253 DESTDIR. That makes for less confusion on what this variable
3254 is for. Also, the configuration option --install_prefix is
3259 * Heartbeat for TLS has been removed and is disabled by default
3260 for DTLS; configure with enable-heartbeats. Code that uses the
3261 old #define's might need to be updated.
3263 *Emilia Käsper, Rich Salz*
3265 * Rename REF_CHECK to REF_DEBUG.
3269 * New "unified" build system
3271 The "unified" build system is aimed to be a common system for all
3272 platforms we support. With it comes new support for VMS.
3274 This system builds supports building in a different directory tree
3275 than the source tree. It produces one Makefile (for unix family
3276 or lookalikes), or one descrip.mms (for VMS).
3278 The source of information to make the Makefile / descrip.mms is
3279 small files called 'build.info', holding the necessary
3280 information for each directory with source to compile, and a
3281 template in Configurations, like unix-Makefile.tmpl or
3284 With this change, the library names were also renamed on Windows
3285 and on VMS. They now have names that are closer to the standard
3286 on Unix, and include the major version number, and in certain
3287 cases, the architecture they are built for. See "Notes on shared
3288 libraries" in INSTALL.
3290 We rely heavily on the perl module Text::Template.
3294 * Added support for auto-initialisation and de-initialisation of the library.
3295 OpenSSL no longer requires explicit init or deinit routines to be called,
3296 except in certain circumstances. See the OPENSSL_init_crypto() and
3297 OPENSSL_init_ssl() man pages for further information.
3301 * The arguments to the DTLSv1_listen function have changed. Specifically the
3302 "peer" argument is now expected to be a BIO_ADDR object.
3304 * Rewrite of BIO networking library. The BIO library lacked consistent
3305 support of IPv6, and adding it required some more extensive
3306 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3307 which hold all types of addresses and chains of address information.
3308 It also introduces a new API, with functions like BIO_socket,
3309 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3310 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3311 have been adapted accordingly.
3315 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3320 * CRIME protection: disable compression by default, even if OpenSSL is
3321 compiled with zlib enabled. Applications can still enable compression
3322 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3323 using the SSL_CONF library to configure compression.
3327 * The signature of the session callback configured with
3328 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3329 was explicitly marked as `const unsigned char*` instead of
3334 * Always DPURIFY. Remove the use of uninitialized memory in the
3335 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3339 * Removed many obsolete configuration items, including
3340 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3341 MD2_CHAR, MD2_INT, MD2_LONG
3343 IDEA_SHORT, IDEA_LONG
3344 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3346 *Rich Salz, with advice from Andy Polyakov*
3348 * Many BN internals have been moved to an internal header file.
3350 *Rich Salz with help from Andy Polyakov*
3352 * Configuration and writing out the results from it has changed.
3353 Files such as Makefile include/openssl/opensslconf.h and are now
3354 produced through general templates, such as Makefile.in and
3355 crypto/opensslconf.h.in and some help from the perl module
3358 Also, the center of configuration information is no longer
3359 Makefile. Instead, Configure produces a perl module in
3360 configdata.pm which holds most of the config data (in the hash
3361 table %config), the target data that comes from the target
3362 configuration in one of the `Configurations/*.conf` files (in
3367 * To clarify their intended purposes, the Configure options
3368 --prefix and --openssldir change their semantics, and become more
3369 straightforward and less interdependent.
3371 --prefix shall be used exclusively to give the location INSTALLTOP
3372 where programs, scripts, libraries, include files and manuals are
3373 going to be installed. The default is now /usr/local.
3375 --openssldir shall be used exclusively to give the default
3376 location OPENSSLDIR where certificates, private keys, CRLs are
3377 managed. This is also where the default openssl.cnf gets
3379 If the directory given with this option is a relative path, the
3380 values of both the --prefix value and the --openssldir value will
3381 be combined to become OPENSSLDIR.
3382 The default for --openssldir is INSTALLTOP/ssl.
3384 Anyone who uses --openssldir to specify where OpenSSL is to be
3385 installed MUST change to use --prefix instead.
3389 * The GOST engine was out of date and therefore it has been removed. An up
3390 to date GOST engine is now being maintained in an external repository.
3391 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3392 support for GOST ciphersuites (these are only activated if a GOST engine
3397 * EGD is no longer supported by default; use enable-egd when
3400 *Ben Kaduk and Rich Salz*
3402 * The distribution now has Makefile.in files, which are used to
3403 create Makefile's when Configure is run. *Configure must be run
3404 before trying to build now.*
3408 * The return value for SSL_CIPHER_description() for error conditions
3413 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3415 Obtaining and performing DNSSEC validation of TLSA records is
3416 the application's responsibility. The application provides
3417 the TLSA records of its choice to OpenSSL, and these are then
3418 used to authenticate the peer.
3420 The TLSA records need not even come from DNS. They can, for
3421 example, be used to implement local end-entity certificate or
3422 trust-anchor "pinning", where the "pin" data takes the form
3423 of TLSA records, which can augment or replace verification
3424 based on the usual WebPKI public certification authorities.
3428 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3429 continues to support deprecated interfaces in default builds.
3430 However, applications are strongly advised to compile their
3431 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3432 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3433 or the 1.1.0 releases.
3435 In environments in which all applications have been ported to
3436 not use any deprecated interfaces OpenSSL's Configure script
3437 should be used with the --api=1.1.0 option to entirely remove
3438 support for the deprecated features from the library and
3439 unconditionally disable them in the installed headers.
3440 Essentially the same effect can be achieved with the "no-deprecated"
3441 argument to Configure, except that this will always restrict
3442 the build to just the latest API, rather than a fixed API
3445 As applications are ported to future revisions of the API,
3446 they should update their compile-time OPENSSL_API_COMPAT define
3447 accordingly, but in most cases should be able to continue to
3448 compile with later releases.
3450 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3451 0x10000000L and 0x00908000L, respectively. However those
3452 versions did not support the OPENSSL_API_COMPAT feature, and
3453 so applications are not typically tested for explicit support
3454 of just the undeprecated features of either release.
3458 * Add support for setting the minimum and maximum supported protocol.
3459 It can bet set via the SSL_set_min_proto_version() and
3460 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3461 MaxProtocol. It's recommended to use the new APIs to disable
3462 protocols instead of disabling individual protocols using
3463 SSL_set_options() or SSL_CONF's Protocol. This change also
3464 removes support for disabling TLS 1.2 in the OpenSSL TLS
3465 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3469 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3473 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3474 and integrates ECDSA and ECDH functionality into EC. Implementations can
3475 now redirect key generation and no longer need to convert to or from
3478 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3479 include the ec.h header file instead.
3483 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3484 ciphers who are no longer supported and drops support the ephemeral RSA key
3485 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3489 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3490 opaque. For HMAC_CTX, the following constructors and destructors
3493 HMAC_CTX *HMAC_CTX_new(void);
3494 void HMAC_CTX_free(HMAC_CTX *ctx);
3496 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3497 destroy such methods has been added. See EVP_MD_meth_new(3) and
3498 EVP_CIPHER_meth_new(3) for documentation.
3501 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3502 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3503 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3504 an already created structure.
3505 2) For consistency with the majority of our object creators and
3506 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3507 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3508 for deprecated builds.
3512 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3513 cryptographic operations to be performed asynchronously as long as an
3514 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3515 further details. Libssl has also had this capability integrated with the
3516 introduction of the new mode SSL_MODE_ASYNC and associated error
3517 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3518 pages. This work was developed in partnership with Intel Corp.
3522 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3523 always enabled now. If you want to disable the support you should
3524 exclude it using the list of supported ciphers. This also means that the
3525 "-no_ecdhe" option has been removed from s_server.
3529 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3530 SSL_{CTX_}set1_curves() which can set a list.
3534 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3535 curve you want to support using SSL_{CTX_}set1_curves().
3539 * State machine rewrite. The state machine code has been significantly
3540 refactored in order to remove much duplication of code and solve issues
3541 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3542 further details). This change does have some associated API changes.
3543 Notably the SSL_state() function has been removed and replaced by
3544 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3545 SSL_set_state() has been removed altogether. The previous handshake states
3546 defined in ssl.h and ssl3.h have also been removed.
3550 * All instances of the string "ssleay" in the public API were replaced
3551 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3552 Some error codes related to internal RSA_eay API's were renamed.
3556 * The demo files in crypto/threads were moved to demo/threads.
3560 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3563 *Matt Caswell, Rich Salz*
3565 * New ASN.1 embed macro.
3567 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3568 structure is not allocated: it is part of the parent. That is instead of
3576 This reduces memory fragmentation and make it impossible to accidentally
3577 set a mandatory field to NULL.
3579 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3580 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3581 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3586 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3590 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3591 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3592 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3593 DES and RC4 ciphersuites.
3597 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3598 This changes the decoding behaviour for some invalid messages,
3599 though the change is mostly in the more lenient direction, and
3600 legacy behaviour is preserved as much as possible.
3604 * Fix no-stdio build.
3605 *David Woodhouse <David.Woodhouse@intel.com> and also*
3606 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3608 * New testing framework
3609 The testing framework has been largely rewritten and is now using
3610 perl and the perl modules Test::Harness and an extended variant of
3611 Test::More called OpenSSL::Test to do its work. All test scripts in
3612 test/ have been rewritten into test recipes, and all direct calls to
3613 executables in test/Makefile have become individual recipes using the
3614 simplified testing OpenSSL::Test::Simple.
3616 For documentation on our testing modules, do:
3618 perldoc test/testlib/OpenSSL/Test/Simple.pm
3619 perldoc test/testlib/OpenSSL/Test.pm
3623 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3624 are used; the latter aborts on memory leaks (usually checked on exit).
3625 Some undocumented "set malloc, etc., hooks" functions were removed
3626 and others were changed. All are now documented.
3630 * In DSA_generate_parameters_ex, if the provided seed is too short,
3633 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3635 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3636 from RFC4279, RFC4785, RFC5487, RFC5489.
3638 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3639 original RSA_PSK patch.
3643 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3644 era flag was never set throughout the codebase (only read). Also removed
3645 SSL3_FLAGS_POP_BUFFER which was only used if
3646 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3650 * Changed the default name options in the "ca", "crl", "req" and "x509"
3651 to be "oneline" instead of "compat".
3655 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3656 not aware of clients that still exhibit this bug, and the workaround
3657 hasn't been working properly for a while.
3661 * The return type of BIO_number_read() and BIO_number_written() as well as
3662 the corresponding num_read and num_write members in the BIO structure has
3663 changed from unsigned long to uint64_t. On platforms where an unsigned
3664 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3669 * Given the pervasive nature of TLS extensions it is inadvisable to run
3670 OpenSSL without support for them. It also means that maintaining
3671 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3672 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3676 * Removed support for the two export grade static DH ciphersuites
3677 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3678 were newly added (along with a number of other static DH ciphersuites) to
3679 1.0.2. However the two export ones have *never* worked since they were
3680 introduced. It seems strange in any case to be adding new export
3681 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3685 * Version negotiation has been rewritten. In particular SSLv23_method(),
3686 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3687 and turned into macros which simply call the new preferred function names
3688 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3689 should use the new names instead. Also as part of this change the ssl23.h
3690 header file has been removed.
3694 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3695 code and the associated standard is no longer considered fit-for-purpose.
3699 * RT2547 was closed. When generating a private key, try to make the
3700 output file readable only by the owner. This behavior change might
3701 be noticeable when interacting with other software.
3703 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3708 * Added HTTP GET support to the ocsp command.
3712 * Changed default digest for the dgst and enc commands from MD5 to
3717 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3721 * Added support for TLS extended master secret from
3722 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3723 initial patch which was a great help during development.
3727 * All libssl internal structures have been removed from the public header
3728 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3729 now redundant). Users should not attempt to access internal structures
3730 directly. Instead they should use the provided API functions.
3734 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3735 Access to deprecated functions can be re-enabled by running config with
3736 "enable-deprecated". In addition applications wishing to use deprecated
3737 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3738 will, by default, disable some transitive includes that previously existed
3739 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3743 * Added support for OCB mode. OpenSSL has been granted a patent license
3744 compatible with the OpenSSL license for use of OCB. Details are available
3745 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3746 for OCB can be removed by calling config with no-ocb.
3750 * SSLv2 support has been removed. It still supports receiving a SSLv2
3751 compatible client hello.
3755 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3756 done while fixing the error code for the key-too-small case.
3758 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3760 * CA.sh has been removed; use CA.pl instead.
3764 * Removed old DES API.
3768 * Remove various unsupported platforms:
3774 Sinix/ReliantUNIX RM400
3779 16-bit platforms such as WIN16
3783 * Clean up OPENSSL_NO_xxx #define's
3784 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3785 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3786 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3787 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3788 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3789 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3790 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3791 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3792 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3793 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3797 * Cleaned up dead code
3798 Remove all but one '#ifdef undef' which is to be looked at.
3802 * Clean up calling of xxx_free routines.
3803 Just like free(), fix most of the xxx_free routines to accept
3804 NULL. Remove the non-null checks from callers. Save much code.
3808 * Add secure heap for storage of private keys (when possible).
3809 Add BIO_s_secmem(), CBIGNUM, etc.
3810 Contributed by Akamai Technologies under our Corporate CLA.
3814 * Experimental support for a new, fast, unbiased prime candidate generator,
3815 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3817 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3819 * New output format NSS in the sess_id command line tool. This allows
3820 exporting the session id and the master key in NSS keylog format.
3822 *Martin Kaiser <martin@kaiser.cx>*
3824 * Harmonize version and its documentation. -f flag is used to display
3827 *mancha <mancha1@zoho.com>*
3829 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3830 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3832 *mancha <mancha1@zoho.com>*
3834 * Fix some double frees. These are not thought to be exploitable.
3836 *mancha <mancha1@zoho.com>*
3838 * A missing bounds check in the handling of the TLS heartbeat extension
3839 can be used to reveal up to 64k of memory to a connected client or
3842 Thanks for Neel Mehta of Google Security for discovering this bug and to
3843 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3844 preparing the fix ([CVE-2014-0160])
3846 *Adam Langley, Bodo Moeller*
3848 * Fix for the attack described in the paper "Recovering OpenSSL
3849 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3850 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3851 <http://eprint.iacr.org/2014/140>
3853 Thanks to Yuval Yarom and Naomi Benger for discovering this
3854 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3856 *Yuval Yarom and Naomi Benger*
3858 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3859 this fixes a limitation in previous versions of OpenSSL.
3863 * Experimental encrypt-then-mac support.
3865 Experimental support for encrypt then mac from
3866 draft-gutmann-tls-encrypt-then-mac-02.txt
3868 To enable it set the appropriate extension number (0x42 for the test
3869 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3871 For non-compliant peers (i.e. just about everything) this should have no
3874 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3878 * Add EVP support for key wrapping algorithms, to avoid problems with
3879 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3880 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3881 algorithms and include tests cases.
3885 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3890 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3891 MGF1 digest and OAEP label.
3895 * Make openssl verify return errors.
3897 *Chris Palmer <palmer@google.com> and Ben Laurie*
3899 * New function ASN1_TIME_diff to calculate the difference between two
3900 ASN1_TIME structures or one structure and the current time.
3904 * Update fips_test_suite to support multiple command line options. New
3905 test to induce all self test errors in sequence and check expected
3910 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3911 sign or verify all in one operation.
3915 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3916 test programs and fips_test_suite. Includes functionality to parse
3917 the minimal script output of fipsalgest.pl directly.
3921 * Add authorisation parameter to FIPS_module_mode_set().
3925 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3929 * Use separate DRBG fields for internal and external flags. New function
3930 FIPS_drbg_health_check() to perform on demand health checking. Add
3931 generation tests to fips_test_suite with reduced health check interval to
3932 demonstrate periodic health checking. Add "nodh" option to
3933 fips_test_suite to skip very slow DH test.
3937 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3942 * More extensive health check for DRBG checking many more failure modes.
3943 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3944 combination: call this in fips_test_suite.
3948 * Add support for canonical generation of DSA parameter 'g'. See
3951 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3952 POST to handle HMAC cases.
3956 * Add functions FIPS_module_version() and FIPS_module_version_text()
3957 to return numerical and string versions of the FIPS module number.
3961 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3962 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3963 outside the validated module in the FIPS capable OpenSSL.
3967 * Minor change to DRBG entropy callback semantics. In some cases
3968 there is no multiple of the block length between min_len and
3969 max_len. Allow the callback to return more than max_len bytes
3970 of entropy but discard any extra: it is the callback's responsibility
3971 to ensure that the extra data discarded does not impact the
3972 requested amount of entropy.
3976 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3977 information in FIPS186-3, SP800-57 and SP800-131A.
3981 * CCM support via EVP. Interface is very similar to GCM case except we
3982 must supply all data in one chunk (i.e. no update, final) and the
3983 message length must be supplied if AAD is used. Add algorithm test
3988 * Initial version of POST overhaul. Add POST callback to allow the status
3989 of POST to be monitored and/or failures induced. Modify fips_test_suite
3990 to use callback. Always run all selftests even if one fails.
3994 * XTS support including algorithm test driver in the fips_gcmtest program.
3995 Note: this does increase the maximum key length from 32 to 64 bytes but
3996 there should be no binary compatibility issues as existing applications
3997 will never use XTS mode.
4001 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
4002 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
4003 performs algorithm blocking for unapproved PRNG types. Also do not
4004 set PRNG type in FIPS_mode_set(): leave this to the application.
4005 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
4006 the standard OpenSSL PRNG: set additional data to a date time vector.
4010 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
4011 This shouldn't present any incompatibility problems because applications
4012 shouldn't be using these directly and any that are will need to rethink
4013 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
4017 * Extensive self tests and health checking required by SP800-90 DRBG.
4018 Remove strength parameter from FIPS_drbg_instantiate and always
4019 instantiate at maximum supported strength.
4023 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
4027 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4031 * New function DH_compute_key_padded() to compute a DH key and pad with
4032 leading zeroes if needed: this complies with SP800-56A et al.
4036 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4037 anything, incomplete, subject to change and largely untested at present.
4041 * Modify fipscanisteronly build option to only build the necessary object
4042 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4046 * Add experimental option FIPSSYMS to give all symbols in
4047 fipscanister.o and FIPS or fips prefix. This will avoid
4048 conflicts with future versions of OpenSSL. Add perl script
4049 util/fipsas.pl to preprocess assembly language source files
4050 and rename any affected symbols.
4054 * Add selftest checks and algorithm block of non-fips algorithms in
4055 FIPS mode. Remove DES2 from selftests.
4059 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4060 return internal method without any ENGINE dependencies. Add new
4061 tiny fips sign and verify functions.
4065 * New build option no-ec2m to disable characteristic 2 code.
4069 * New build option "fipscanisteronly". This only builds fipscanister.o
4070 and (currently) associated fips utilities. Uses the file Makefile.fips
4071 instead of Makefile.org as the prototype.
4075 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4076 Update fips_gcmtest to use IV generator.
4080 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4081 setting output buffer to NULL. The `*Final` function must be
4082 called although it will not retrieve any additional data. The tag
4083 can be set or retrieved with a ctrl. The IV length is by default 12
4084 bytes (96 bits) but can be set to an alternative value. If the IV
4085 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4090 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4091 underlying do_cipher function handles all cipher semantics itself
4092 including padding and finalisation. This is useful if (for example)
4093 an ENGINE cipher handles block padding itself. The behaviour of
4094 do_cipher is subtly changed if this flag is set: the return value
4095 is the number of characters written to the output buffer (zero is
4096 no longer an error code) or a negative error code. Also if the
4097 input buffer is NULL and length 0 finalisation should be performed.
4101 * If a candidate issuer certificate is already part of the constructed
4102 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4106 * Improve forward-security support: add functions
4108 void SSL_CTX_set_not_resumable_session_callback(
4109 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4110 void SSL_set_not_resumable_session_callback(
4111 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4113 for use by SSL/TLS servers; the callback function will be called whenever a
4114 new session is created, and gets to decide whether the session may be
4115 cached to make it resumable (return 0) or not (return 1). (As by the
4116 SSL/TLS protocol specifications, the session_id sent by the server will be
4117 empty to indicate that the session is not resumable; also, the server will
4118 not generate RFC 4507 (RFC 5077) session tickets.)
4120 A simple reasonable callback implementation is to return is_forward_secure.
4121 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4122 by the SSL/TLS server library, indicating whether it can provide forward
4125 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4127 * New -verify_name option in command line utilities to set verification
4132 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4133 Add CMAC pkey methods.
4137 * Experimental renegotiation in s_server -www mode. If the client
4138 browses /reneg connection is renegotiated. If /renegcert it is
4139 renegotiated requesting a certificate.
4143 * Add an "external" session cache for debugging purposes to s_server. This
4144 should help trace issues which normally are only apparent in deployed
4145 multi-process servers.
4149 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4150 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4151 BIO_set_cipher() and some obscure PEM functions were changed so they
4152 can now return an error. The RAND changes required a change to the
4153 RAND_METHOD structure.
4157 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4158 a gcc attribute to warn if the result of a function is ignored. This
4159 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4160 whose return value is often ignored.
4164 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4165 These allow SCTs (signed certificate timestamps) to be requested and
4166 validated when establishing a connection.
4168 *Rob Percival <robpercival@google.com>*
4173 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4175 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4176 used even when parsing explicit parameters, when loading a encoded key
4177 or calling `EC_GROUP_new_from_ecpkparameters()`/
4178 `EC_GROUP_new_from_ecparameters()`.
4179 This prevents bypass of security hardening and performance gains,
4180 especially for curves with specialized EC_METHODs.
4181 By default, if a key encoded with explicit parameters is loaded and later
4182 encoded, the output is still encoded with explicit parameters, even if
4183 internally a "named" EC_GROUP is used for computation.
4187 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4188 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4189 NULL. After this change, only the cofactor parameter can be NULL. It also
4190 does some minimal sanity checks on the passed order.
4195 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4196 An attack is simple, if the first CMS_recipientInfo is valid but the
4197 second CMS_recipientInfo is chosen ciphertext. If the second
4198 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4199 encryption key will be replaced by garbage, and the message cannot be
4200 decoded, but if the RSA decryption fails, the correct encryption key is
4201 used and the recipient will not notice the attack.
4202 As a work around for this potential attack the length of the decrypted
4203 key must be equal to the cipher default key length, in case the
4204 certifiate is not given and all recipientInfo are tried out.
4205 The old behaviour can be re-enabled in the CMS code by setting the
4206 CMS_DEBUG_DECRYPT flag.
4211 * Document issue with installation paths in diverse Windows builds
4213 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4214 binaries and run-time config file.
4219 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4221 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4222 This changes the size when using the `genpkey` command when no size is given.
4223 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4224 generation commands to use 2048 bits by default.
4228 * Add FIPS support for Android Arm 64-bit
4230 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4231 Module in Version 2.0.10. For some reason, the corresponding target
4232 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4233 built with FIPS support on Android Arm 64-bit. This omission has been
4236 *Matthias St. Pierre*
4238 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4240 * 0-byte record padding oracle
4242 If an application encounters a fatal protocol error and then calls
4243 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4244 then OpenSSL can respond differently to the calling application if a 0 byte
4245 record is received with invalid padding compared to if a 0 byte record is
4246 received with an invalid MAC. If the application then behaves differently
4247 based on that in a way that is detectable to the remote peer, then this
4248 amounts to a padding oracle that could be used to decrypt data.
4250 In order for this to be exploitable "non-stitched" ciphersuites must be in
4251 use. Stitched ciphersuites are optimised implementations of certain
4252 commonly used ciphersuites. Also the application must call SSL_shutdown()
4253 twice even if a protocol error has occurred (applications should not do
4254 this but some do anyway).
4256 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4257 Aviram, with additional investigation by Steven Collison and Andrew
4258 Hourselt. It was reported to OpenSSL on 10th December 2018.
4263 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4267 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4269 * Microarchitecture timing vulnerability in ECC scalar multiplication
4271 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4272 shown to be vulnerable to a microarchitecture timing side channel attack.
4273 An attacker with sufficient access to mount local timing attacks during
4274 ECDSA signature generation could recover the private key.
4276 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4277 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4283 * Timing vulnerability in DSA signature generation
4285 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4286 timing side channel attack. An attacker could use variations in the signing
4287 algorithm to recover the private key.
4289 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4294 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4295 Module, accidentally introduced while backporting security fixes from the
4296 development branch and hindering the use of ECC in FIPS mode.
4300 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4302 * Client DoS due to large DH parameter
4304 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4305 malicious server can send a very large prime value to the client. This will
4306 cause the client to spend an unreasonably long period of time generating a
4307 key for this prime resulting in a hang until the client has finished. This
4308 could be exploited in a Denial Of Service attack.
4310 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4315 * Cache timing vulnerability in RSA Key Generation
4317 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4318 a cache timing side channel attack. An attacker with sufficient access to
4319 mount cache timing attacks during the RSA key generation process could
4320 recover the private key.
4322 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4323 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4328 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4329 parameter is no longer accepted, as it leads to a corrupt table. NULL
4330 pem_str is reserved for alias entries only.
4334 * Revert blinding in ECDSA sign and instead make problematic addition
4335 length-invariant. Switch even to fixed-length Montgomery multiplication.
4339 * Change generating and checking of primes so that the error rate of not
4340 being prime depends on the intended use based on the size of the input.
4341 For larger primes this will result in more rounds of Miller-Rabin.
4342 The maximal error rate for primes with more than 1080 bits is lowered
4345 *Kurt Roeckx, Annie Yousar*
4347 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4351 * Add blinding to ECDSA and DSA signatures to protect against side channel
4352 attacks discovered by Keegan Ryan (NCC Group).
4356 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4357 now allow empty (zero character) pass phrases.
4361 * Certificate time validation (X509_cmp_time) enforces stricter
4362 compliance with RFC 5280. Fractional seconds and timezone offsets
4363 are no longer allowed.
4367 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4369 * Constructed ASN.1 types with a recursive definition could exceed the stack
4371 Constructed ASN.1 types with a recursive definition (such as can be found
4372 in PKCS7) could eventually exceed the stack given malicious input with
4373 excessive recursion. This could result in a Denial Of Service attack. There
4374 are no such structures used within SSL/TLS that come from untrusted sources
4375 so this is considered safe.
4377 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4383 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4385 * Read/write after SSL object in error state
4387 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4388 mechanism. The intent was that if a fatal error occurred during a handshake
4389 then OpenSSL would move into the error state and would immediately fail if
4390 you attempted to continue the handshake. This works as designed for the
4391 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4392 SSL_connect()), however due to a bug it does not work correctly if
4393 SSL_read() or SSL_write() is called directly. In that scenario, if the
4394 handshake fails then a fatal error will be returned in the initial function
4395 call. If SSL_read()/SSL_write() is subsequently called by the application
4396 for the same SSL object then it will succeed and the data is passed without
4397 being decrypted/encrypted directly from the SSL/TLS record layer.
4399 In order to exploit this issue an application bug would have to be present
4400 that resulted in a call to SSL_read()/SSL_write() being issued after having
4401 already received a fatal error.
4403 This issue was reported to OpenSSL by David Benjamin (Google).
4408 * rsaz_1024_mul_avx2 overflow bug on x86_64
4410 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4411 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4412 Analysis suggests that attacks against RSA and DSA as a result of this
4413 defect would be very difficult to perform and are not believed likely.
4414 Attacks against DH1024 are considered just feasible, because most of the
4415 work necessary to deduce information about a private key may be performed
4416 offline. The amount of resources required for such an attack would be
4417 significant. However, for an attack on TLS to be meaningful, the server
4418 would have to share the DH1024 private key among multiple clients, which is
4419 no longer an option since CVE-2016-0701.
4421 This only affects processors that support the AVX2 but not ADX extensions
4422 like Intel Haswell (4th generation).
4424 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4425 was originally found via the OSS-Fuzz project.
4430 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4432 * bn_sqrx8x_internal carry bug on x86_64
4434 There is a carry propagating bug in the x86_64 Montgomery squaring
4435 procedure. No EC algorithms are affected. Analysis suggests that attacks
4436 against RSA and DSA as a result of this defect would be very difficult to
4437 perform and are not believed likely. Attacks against DH are considered just
4438 feasible (although very difficult) because most of the work necessary to
4439 deduce information about a private key may be performed offline. The amount
4440 of resources required for such an attack would be very significant and
4441 likely only accessible to a limited number of attackers. An attacker would
4442 additionally need online access to an unpatched system using the target
4443 private key in a scenario with persistent DH parameters and a private
4444 key that is shared between multiple clients.
4446 This only affects processors that support the BMI1, BMI2 and ADX extensions
4447 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4449 This issue was reported to OpenSSL by the OSS-Fuzz project.
4454 * Malformed X.509 IPAddressFamily could cause OOB read
4456 If an X.509 certificate has a malformed IPAddressFamily extension,
4457 OpenSSL could do a one-byte buffer overread. The most likely result
4458 would be an erroneous display of the certificate in text format.
4460 This issue was reported to OpenSSL by the OSS-Fuzz project.
4465 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4467 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4468 platform rather than 'mingw'.
4472 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4474 * Truncated packet could crash via OOB read
4476 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4477 cipher is being used, then a truncated packet can cause that host to
4478 perform an out-of-bounds read, usually resulting in a crash.
4480 This issue was reported to OpenSSL by Robert Święcki of Google.
4485 * BN_mod_exp may produce incorrect results on x86_64
4487 There is a carry propagating bug in the x86_64 Montgomery squaring
4488 procedure. No EC algorithms are affected. Analysis suggests that attacks
4489 against RSA and DSA as a result of this defect would be very difficult to
4490 perform and are not believed likely. Attacks against DH are considered just
4491 feasible (although very difficult) because most of the work necessary to
4492 deduce information about a private key may be performed offline. The amount
4493 of resources required for such an attack would be very significant and
4494 likely only accessible to a limited number of attackers. An attacker would
4495 additionally need online access to an unpatched system using the target
4496 private key in a scenario with persistent DH parameters and a private
4497 key that is shared between multiple clients. For example this can occur by
4498 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4499 similar to CVE-2015-3193 but must be treated as a separate problem.
4501 This issue was reported to OpenSSL by the OSS-Fuzz project.
4506 * Montgomery multiplication may produce incorrect results
4508 There is a carry propagating bug in the Broadwell-specific Montgomery
4509 multiplication procedure that handles input lengths divisible by, but
4510 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4511 and DH private keys are impossible. This is because the subroutine in
4512 question is not used in operations with the private key itself and an input
4513 of the attacker's direct choice. Otherwise the bug can manifest itself as
4514 transient authentication and key negotiation failures or reproducible
4515 erroneous outcome of public-key operations with specially crafted input.
4516 Among EC algorithms only Brainpool P-512 curves are affected and one
4517 presumably can attack ECDH key negotiation. Impact was not analyzed in
4518 detail, because pre-requisites for attack are considered unlikely. Namely
4519 multiple clients have to choose the curve in question and the server has to
4520 share the private key among them, neither of which is default behaviour.
4521 Even then only clients that chose the curve will be affected.
4523 This issue was publicly reported as transient failures and was not
4524 initially recognized as a security issue. Thanks to Richard Morgan for
4525 providing reproducible case.
4530 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4531 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4532 prevent issues where no progress is being made and the peer continually
4533 sends unrecognised record types, using up resources processing them.
4537 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4539 * Missing CRL sanity check
4541 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4542 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4543 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4545 This issue only affects the OpenSSL 1.0.2i
4550 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4552 * OCSP Status Request extension unbounded memory growth
4554 A malicious client can send an excessively large OCSP Status Request
4555 extension. If that client continually requests renegotiation, sending a
4556 large OCSP Status Request extension each time, then there will be unbounded
4557 memory growth on the server. This will eventually lead to a Denial Of
4558 Service attack through memory exhaustion. Servers with a default
4559 configuration are vulnerable even if they do not support OCSP. Builds using
4560 the "no-ocsp" build time option are not affected.
4562 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4567 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4570 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4576 * OOB write in MDC2_Update()
4578 An overflow can occur in MDC2_Update() either if called directly or
4579 through the EVP_DigestUpdate() function using MDC2. If an attacker
4580 is able to supply very large amounts of input data after a previous
4581 call to EVP_EncryptUpdate() with a partial block then a length check
4582 can overflow resulting in a heap corruption.
4584 The amount of data needed is comparable to SIZE_MAX which is impractical
4587 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4592 * Malformed SHA512 ticket DoS
4594 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4595 DoS attack where a malformed ticket will result in an OOB read which will
4598 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4599 a custom server callback and ticket lookup mechanism.
4601 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4606 * OOB write in BN_bn2dec()
4608 The function BN_bn2dec() does not check the return value of BN_div_word().
4609 This can cause an OOB write if an application uses this function with an
4610 overly large BIGNUM. This could be a problem if an overly large certificate
4611 or CRL is printed out from an untrusted source. TLS is not affected because
4612 record limits will reject an oversized certificate before it is parsed.
4614 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4619 * OOB read in TS_OBJ_print_bio()
4621 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4622 the total length the OID text representation would use and not the amount
4623 of data written. This will result in OOB reads when large OIDs are
4626 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4631 * Pointer arithmetic undefined behaviour
4633 Avoid some undefined pointer arithmetic
4635 A common idiom in the codebase is to check limits in the following manner:
4638 Where "p" points to some malloc'd data of SIZE bytes and
4641 "len" here could be from some externally supplied data (e.g. from a TLS
4644 The rules of C pointer arithmetic are such that "p + len" is only well
4645 defined where len <= SIZE. Therefore the above idiom is actually
4646 undefined behaviour.
4648 For example this could cause problems if some malloc implementation
4649 provides an address for "p" such that "p + len" actually overflows for
4650 values of len that are too big and therefore p + len < limit.
4652 This issue was reported to OpenSSL by Guido Vranken
4657 * Constant time flag not preserved in DSA signing
4659 Operations in the DSA signing algorithm should run in constant time in
4660 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4661 implementation means that a non-constant time codepath is followed for
4662 certain operations. This has been demonstrated through a cache-timing
4663 attack to be sufficient for an attacker to recover the private DSA key.
4665 This issue was reported by César Pereida (Aalto University), Billy Brumley
4666 (Tampere University of Technology), and Yuval Yarom (The University of
4667 Adelaide and NICTA).
4672 * DTLS buffered message DoS
4674 In a DTLS connection where handshake messages are delivered out-of-order
4675 those messages that OpenSSL is not yet ready to process will be buffered
4676 for later use. Under certain circumstances, a flaw in the logic means that
4677 those messages do not get removed from the buffer even though the handshake
4678 has been completed. An attacker could force up to approx. 15 messages to
4679 remain in the buffer when they are no longer required. These messages will
4680 be cleared when the DTLS connection is closed. The default maximum size for
4681 a message is 100k. Therefore the attacker could force an additional 1500k
4682 to be consumed per connection. By opening many simulataneous connections an
4683 attacker could cause a DoS attack through memory exhaustion.
4685 This issue was reported to OpenSSL by Quan Luo.
4690 * DTLS replay protection DoS
4692 A flaw in the DTLS replay attack protection mechanism means that records
4693 that arrive for future epochs update the replay protection "window" before
4694 the MAC for the record has been validated. This could be exploited by an
4695 attacker by sending a record for the next epoch (which does not have to
4696 decrypt or have a valid MAC), with a very large sequence number. This means
4697 that all subsequent legitimate packets are dropped causing a denial of
4698 service for a specific DTLS connection.
4700 This issue was reported to OpenSSL by the OCAP audit team.
4705 * Certificate message OOB reads
4707 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4708 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4709 theoretical DoS risk but this has not been observed in practice on common
4712 The messages affected are client certificate, client certificate request
4713 and server certificate. As a result the attack can only be performed
4714 against a client or a server which enables client authentication.
4716 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4721 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4723 * Prevent padding oracle in AES-NI CBC MAC check
4725 A MITM attacker can use a padding oracle attack to decrypt traffic
4726 when the connection uses an AES CBC cipher and the server support
4729 This issue was introduced as part of the fix for Lucky 13 padding
4730 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4731 constant time by making sure that always the same bytes are read and
4732 compared against either the MAC or padding bytes. But it no longer
4733 checked that there was enough data to have both the MAC and padding
4736 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4741 * Fix EVP_EncodeUpdate overflow
4743 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4744 Base64 encoding of binary data. If an attacker is able to supply very large
4745 amounts of input data then a length check can overflow resulting in a heap
4748 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4749 the `PEM_write_bio*` family of functions. These are mainly used within the
4750 OpenSSL command line applications, so any application which processes data
4751 from an untrusted source and outputs it as a PEM file should be considered
4752 vulnerable to this issue. User applications that call these APIs directly
4753 with large amounts of untrusted data may also be vulnerable.
4755 This issue was reported by Guido Vranken.
4760 * Fix EVP_EncryptUpdate overflow
4762 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4763 is able to supply very large amounts of input data after a previous call to
4764 EVP_EncryptUpdate() with a partial block then a length check can overflow
4765 resulting in a heap corruption. Following an analysis of all OpenSSL
4766 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4767 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4768 the first called function after an EVP_EncryptInit(), and therefore that
4769 specific call must be safe. The second form is where the length passed to
4770 EVP_EncryptUpdate() can be seen from the code to be some small value and
4771 therefore there is no possibility of an overflow. Since all instances are
4772 one of these two forms, it is believed that there can be no overflows in
4773 internal code due to this problem. It should be noted that
4774 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4775 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4776 of these calls have also been analysed too and it is believed there are no
4777 instances in internal usage where an overflow could occur.
4779 This issue was reported by Guido Vranken.
4784 * Prevent ASN.1 BIO excessive memory allocation
4786 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4787 a short invalid encoding can cause allocation of large amounts of memory
4788 potentially consuming excessive resources or exhausting memory.
4790 Any application parsing untrusted data through d2i BIO functions is
4791 affected. The memory based functions such as d2i_X509() are *not* affected.
4792 Since the memory based functions are used by the TLS library, TLS
4793 applications are not affected.
4795 This issue was reported by Brian Carpenter.
4802 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4803 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4804 in arbitrary stack data being returned in the buffer.
4806 This issue was reported by Guido Vranken.
4811 * Modify behavior of ALPN to invoke callback after SNI/servername
4812 callback, such that updates to the SSL_CTX affect ALPN.
4816 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4821 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4822 methods are enabled and ssl2 is disabled the methods return NULL.
4826 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4828 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4829 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4830 provide any "EXPORT" or "LOW" strength ciphers.
4834 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4835 is by default disabled at build-time. Builds that are not configured with
4836 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4837 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4838 will need to explicitly call either of:
4840 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4842 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4844 as appropriate. Even if either of those is used, or the application
4845 explicitly uses the version-specific SSLv2_method() or its client and
4846 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4847 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4848 ciphers, and SSLv2 56-bit DES are no longer available.
4853 * Fix a double-free in DSA code
4855 A double free bug was discovered when OpenSSL parses malformed DSA private
4856 keys and could lead to a DoS attack or memory corruption for applications
4857 that receive DSA private keys from untrusted sources. This scenario is
4860 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4866 * Disable SRP fake user seed to address a server memory leak.
4868 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4870 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4871 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4872 was changed to ignore the "fake user" SRP seed, even if the seed
4875 Users should use SRP_VBASE_get1_by_user instead. Note that in
4876 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4877 also that even though configuring the SRP seed attempts to hide
4878 invalid usernames by continuing the handshake with fake
4879 credentials, this behaviour is not constant time and no strong
4880 guarantees are made that the handshake is indistinguishable from
4881 that of a valid user.
4886 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4888 In the BN_hex2bn function the number of hex digits is calculated using an
4889 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4890 large values of `i` this can result in `bn_expand` not allocating any
4891 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4892 field as NULL leading to a subsequent NULL ptr deref. For very large values
4893 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4894 In this case memory is allocated to the internal BIGNUM data field, but it
4895 is insufficiently sized leading to heap corruption. A similar issue exists
4896 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4897 is ever called by user applications with very large untrusted hex/dec data.
4898 This is anticipated to be a rare occurrence.
4900 All OpenSSL internal usage of these functions use data that is not expected
4901 to be untrusted, e.g. config file data or application command line
4902 arguments. If user developed applications generate config file data based
4903 on untrusted data then it is possible that this could also lead to security
4904 consequences. This is also anticipated to be rare.
4906 This issue was reported to OpenSSL by Guido Vranken.
4911 * Fix memory issues in `BIO_*printf` functions
4913 The internal `fmtstr` function used in processing a "%s" format string in
4914 the `BIO_*printf` functions could overflow while calculating the length of a
4915 string and cause an OOB read when printing very long strings.
4917 Additionally the internal `doapr_outch` function can attempt to write to an
4918 OOB memory location (at an offset from the NULL pointer) in the event of a
4919 memory allocation failure. In 1.0.2 and below this could be caused where
4920 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4921 could be in processing a very long "%s" format string. Memory leaks can
4924 The first issue may mask the second issue dependent on compiler behaviour.
4925 These problems could enable attacks where large amounts of untrusted data
4926 is passed to the `BIO_*printf` functions. If applications use these functions
4927 in this way then they could be vulnerable. OpenSSL itself uses these
4928 functions when printing out human-readable dumps of ASN.1 data. Therefore
4929 applications that print this data could be vulnerable if the data is from
4930 untrusted sources. OpenSSL command line applications could also be
4931 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4932 as command line arguments.
4934 Libssl is not considered directly vulnerable. Additionally certificates etc
4935 received via remote connections via libssl are also unlikely to be able to
4936 trigger these issues because of message size limits enforced within libssl.
4938 This issue was reported to OpenSSL Guido Vranken.
4943 * Side channel attack on modular exponentiation
4945 A side-channel attack was found which makes use of cache-bank conflicts on
4946 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4947 of RSA keys. The ability to exploit this issue is limited as it relies on
4948 an attacker who has control of code in a thread running on the same
4949 hyper-threaded core as the victim thread which is performing decryptions.
4951 This issue was reported to OpenSSL by Yuval Yarom, The University of
4952 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4953 Nadia Heninger, University of Pennsylvania with more information at
4954 <http://cachebleed.info>.
4959 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4960 if no keysize is specified with default_bits. This fixes an
4961 omission in an earlier change that changed all RSA/DSA key generation
4962 commands to use 2048 bits by default.
4966 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4968 * DH small subgroups
4970 Historically OpenSSL only ever generated DH parameters based on "safe"
4971 primes. More recently (in version 1.0.2) support was provided for
4972 generating X9.42 style parameter files such as those required for RFC 5114
4973 support. The primes used in such files may not be "safe". Where an
4974 application is using DH configured with parameters based on primes that are
4975 not "safe" then an attacker could use this fact to find a peer's private
4976 DH exponent. This attack requires that the attacker complete multiple
4977 handshakes in which the peer uses the same private DH exponent. For example
4978 this could be used to discover a TLS server's private DH exponent if it's
4979 reusing the private DH exponent or it's using a static DH ciphersuite.
4981 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4982 TLS. It is not on by default. If the option is not set then the server
4983 reuses the same private DH exponent for the life of the server process and
4984 would be vulnerable to this attack. It is believed that many popular
4985 applications do set this option and would therefore not be at risk.
4987 The fix for this issue adds an additional check where a "q" parameter is
4988 available (as is the case in X9.42 based parameters). This detects the
4989 only known attack, and is the only possible defense for static DH
4990 ciphersuites. This could have some performance impact.
4992 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4993 default and cannot be disabled. This could have some performance impact.
4995 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
5000 * SSLv2 doesn't block disabled ciphers
5002 A malicious client can negotiate SSLv2 ciphers that have been disabled on
5003 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
5004 been disabled, provided that the SSLv2 protocol was not also disabled via
5007 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
5008 and Sebastian Schinzel.
5013 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
5015 * BN_mod_exp may produce incorrect results on x86_64
5017 There is a carry propagating bug in the x86_64 Montgomery squaring
5018 procedure. No EC algorithms are affected. Analysis suggests that attacks
5019 against RSA and DSA as a result of this defect would be very difficult to
5020 perform and are not believed likely. Attacks against DH are considered just
5021 feasible (although very difficult) because most of the work necessary to
5022 deduce information about a private key may be performed offline. The amount
5023 of resources required for such an attack would be very significant and
5024 likely only accessible to a limited number of attackers. An attacker would
5025 additionally need online access to an unpatched system using the target
5026 private key in a scenario with persistent DH parameters and a private
5027 key that is shared between multiple clients. For example this can occur by
5028 default in OpenSSL DHE based SSL/TLS ciphersuites.
5030 This issue was reported to OpenSSL by Hanno Böck.
5035 * Certificate verify crash with missing PSS parameter
5037 The signature verification routines will crash with a NULL pointer
5038 dereference if presented with an ASN.1 signature using the RSA PSS
5039 algorithm and absent mask generation function parameter. Since these
5040 routines are used to verify certificate signature algorithms this can be
5041 used to crash any certificate verification operation and exploited in a
5042 DoS attack. Any application which performs certificate verification is
5043 vulnerable including OpenSSL clients and servers which enable client
5046 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5051 * X509_ATTRIBUTE memory leak
5053 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5054 memory. This structure is used by the PKCS#7 and CMS routines so any
5055 application which reads PKCS#7 or CMS data from untrusted sources is
5056 affected. SSL/TLS is not affected.
5058 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5064 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5065 This changes the decoding behaviour for some invalid messages,
5066 though the change is mostly in the more lenient direction, and
5067 legacy behaviour is preserved as much as possible.
5071 * In DSA_generate_parameters_ex, if the provided seed is too short,
5074 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5076 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5078 * Alternate chains certificate forgery
5080 During certificate verification, OpenSSL will attempt to find an
5081 alternative certificate chain if the first attempt to build such a chain
5082 fails. An error in the implementation of this logic can mean that an
5083 attacker could cause certain checks on untrusted certificates to be
5084 bypassed, such as the CA flag, enabling them to use a valid leaf
5085 certificate to act as a CA and "issue" an invalid certificate.
5087 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5092 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5094 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5095 incompatibility in the handling of HMAC. The previous ABI has now been
5100 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5102 * Malformed ECParameters causes infinite loop
5104 When processing an ECParameters structure OpenSSL enters an infinite loop
5105 if the curve specified is over a specially malformed binary polynomial
5108 This can be used to perform denial of service against any
5109 system which processes public keys, certificate requests or
5110 certificates. This includes TLS clients and TLS servers with
5111 client authentication enabled.
5113 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5118 * Exploitable out-of-bounds read in X509_cmp_time
5120 X509_cmp_time does not properly check the length of the ASN1_TIME
5121 string and can read a few bytes out of bounds. In addition,
5122 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5125 An attacker can use this to craft malformed certificates and CRLs of
5126 various sizes and potentially cause a segmentation fault, resulting in
5127 a DoS on applications that verify certificates or CRLs. TLS clients
5128 that verify CRLs are affected. TLS clients and servers with client
5129 authentication enabled may be affected if they use custom verification
5132 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5133 independently by Hanno Böck.
5138 * PKCS7 crash with missing EnvelopedContent
5140 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5141 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5142 with missing content and trigger a NULL pointer dereference on parsing.
5144 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5145 structures from untrusted sources are affected. OpenSSL clients and
5146 servers are not affected.
5148 This issue was reported to OpenSSL by Michal Zalewski (Google).
5153 * CMS verify infinite loop with unknown hash function
5155 When verifying a signedData message the CMS code can enter an infinite loop
5156 if presented with an unknown hash function OID. This can be used to perform
5157 denial of service against any system which verifies signedData messages using
5159 This issue was reported to OpenSSL by Johannes Bauer.
5164 * Race condition handling NewSessionTicket
5166 If a NewSessionTicket is received by a multi-threaded client when attempting to
5167 reuse a previous ticket then a race condition can occur potentially leading to
5168 a double free of the ticket data.
5173 * Only support 256-bit or stronger elliptic curves with the
5174 'ecdh_auto' setting (server) or by default (client). Of supported
5175 curves, prefer P-256 (both).
5179 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5181 * ClientHello sigalgs DoS fix
5183 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5184 invalid signature algorithms extension a NULL pointer dereference will
5185 occur. This can be exploited in a DoS attack against the server.
5187 This issue was was reported to OpenSSL by David Ramos of Stanford
5191 *Stephen Henson and Matt Caswell*
5193 * Multiblock corrupted pointer fix
5195 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5196 feature only applies on 64 bit x86 architecture platforms that support AES
5197 NI instructions. A defect in the implementation of "multiblock" can cause
5198 OpenSSL's internal write buffer to become incorrectly set to NULL when
5199 using non-blocking IO. Typically, when the user application is using a
5200 socket BIO for writing, this will only result in a failed connection.
5201 However if some other BIO is used then it is likely that a segmentation
5202 fault will be triggered, thus enabling a potential DoS attack.
5204 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5209 * Segmentation fault in DTLSv1_listen fix
5211 The DTLSv1_listen function is intended to be stateless and processes the
5212 initial ClientHello from many peers. It is common for user code to loop
5213 over the call to DTLSv1_listen until a valid ClientHello is received with
5214 an associated cookie. A defect in the implementation of DTLSv1_listen means
5215 that state is preserved in the SSL object from one invocation to the next
5216 that can lead to a segmentation fault. Errors processing the initial
5217 ClientHello can trigger this scenario. An example of such an error could be
5218 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5221 This issue was reported to OpenSSL by Per Allansson.
5226 * Segmentation fault in ASN1_TYPE_cmp fix
5228 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5229 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5230 certificate signature algorithm consistency this can be used to crash any
5231 certificate verification operation and exploited in a DoS attack. Any
5232 application which performs certificate verification is vulnerable including
5233 OpenSSL clients and servers which enable client authentication.
5238 * Segmentation fault for invalid PSS parameters fix
5240 The signature verification routines will crash with a NULL pointer
5241 dereference if presented with an ASN.1 signature using the RSA PSS
5242 algorithm and invalid parameters. Since these routines are used to verify
5243 certificate signature algorithms this can be used to crash any
5244 certificate verification operation and exploited in a DoS attack. Any
5245 application which performs certificate verification is vulnerable including
5246 OpenSSL clients and servers which enable client authentication.
5248 This issue was was reported to OpenSSL by Brian Carpenter.
5253 * ASN.1 structure reuse memory corruption fix
5255 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5256 memory corruption via an invalid write. Such reuse is and has been
5257 strongly discouraged and is believed to be rare.
5259 Applications that parse structures containing CHOICE or ANY DEFINED BY
5260 components may be affected. Certificate parsing (d2i_X509 and related
5261 functions) are however not affected. OpenSSL clients and servers are
5267 * PKCS7 NULL pointer dereferences fix
5269 The PKCS#7 parsing code does not handle missing outer ContentInfo
5270 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5271 missing content and trigger a NULL pointer dereference on parsing.
5273 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5274 otherwise parse PKCS#7 structures from untrusted sources are
5275 affected. OpenSSL clients and servers are not affected.
5277 This issue was reported to OpenSSL by Michal Zalewski (Google).
5282 * DoS via reachable assert in SSLv2 servers fix
5284 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5285 servers that both support SSLv2 and enable export cipher suites by sending
5286 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5288 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5289 (OpenSSL development team).
5294 * Empty CKE with client auth and DHE fix
5296 If client auth is used then a server can seg fault in the event of a DHE
5297 ciphersuite being selected and a zero length ClientKeyExchange message
5298 being sent by the client. This could be exploited in a DoS attack.
5303 * Handshake with unseeded PRNG fix
5305 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5306 with an unseeded PRNG. The conditions are:
5307 - The client is on a platform where the PRNG has not been seeded
5308 automatically, and the user has not seeded manually
5309 - A protocol specific client method version has been used (i.e. not
5310 SSL_client_methodv23)
5311 - A ciphersuite is used that does not require additional random data from
5312 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5314 If the handshake succeeds then the client random that has been used will
5315 have been generated from a PRNG with insufficient entropy and therefore the
5316 output may be predictable.
5318 For example using the following command with an unseeded openssl will
5319 succeed on an unpatched platform:
5321 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5326 * Use After Free following d2i_ECPrivatekey error fix
5328 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5329 could cause a use after free condition. This, in turn, could cause a double
5330 free in several private key parsing functions (such as d2i_PrivateKey
5331 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5332 for applications that receive EC private keys from untrusted
5333 sources. This scenario is considered rare.
5335 This issue was discovered by the BoringSSL project and fixed in their
5341 * X509_to_X509_REQ NULL pointer deref fix
5343 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5344 the certificate key is invalid. This function is rarely used in practice.
5346 This issue was discovered by Brian Carpenter.
5351 * Removed the export ciphers from the DEFAULT ciphers
5355 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5357 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5358 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5359 So far those who have to target multiple platforms would compromise
5360 and argue that binary targeting say ARMv5 would still execute on
5361 ARMv8. "Universal" build resolves this compromise by providing
5362 near-optimal performance even on newer platforms.
5366 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5367 (other platforms pending).
5369 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5371 * Add support for the SignedCertificateTimestampList certificate and
5372 OCSP response extensions from RFC6962.
5376 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5377 for corner cases. (Certain input points at infinity could lead to
5378 bogus results, with non-infinity inputs mapped to infinity too.)
5382 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5383 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5384 common cases are optimized and there still is room for further
5385 improvements. Vector Permutation AES for Altivec is also added.
5389 * Add support for little-endian ppc64 Linux target.
5391 *Marcelo Cerri (IBM)*
5393 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5394 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5395 are optimized and there still is room for further improvements.
5396 Both 32- and 64-bit modes are supported.
5398 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5400 * Improved ARMv7 NEON support.
5404 * Support for SPARC Architecture 2011 crypto extensions, first
5405 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5406 SHA256/512, MD5, GHASH and modular exponentiation.
5408 *Andy Polyakov, David Miller*
5410 * Accelerated modular exponentiation for Intel processors, a.k.a.
5413 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5415 * Support for new and upcoming Intel processors, including AVX2,
5416 BMI and SHA ISA extensions. This includes additional "stitched"
5417 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5420 This work was sponsored by Intel Corp.
5424 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5425 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5426 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5430 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5431 this fixes a limitation in previous versions of OpenSSL.
5435 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5436 MGF1 digest and OAEP label.
5440 * Add EVP support for key wrapping algorithms, to avoid problems with
5441 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5442 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5443 algorithms and include tests cases.
5447 * Add functions to allocate and set the fields of an ECDSA_METHOD
5450 *Douglas E. Engert, Steve Henson*
5452 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5453 difference in days and seconds between two tm or ASN1_TIME structures.
5457 * Add -rev test option to s_server to just reverse order of characters
5458 received by client and send back to server. Also prints an abbreviated
5459 summary of the connection parameters.
5463 * New option -brief for s_client and s_server to print out a brief summary
5464 of connection parameters.
5468 * Add callbacks for arbitrary TLS extensions.
5470 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5472 * New option -crl_download in several openssl utilities to download CRLs
5473 from CRLDP extension in certificates.
5477 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5481 * New function X509_CRL_diff to generate a delta CRL from the difference
5482 of two full CRLs. Add support to "crl" utility.
5486 * New functions to set lookup_crls function and to retrieve
5487 X509_STORE from X509_STORE_CTX.
5491 * Print out deprecated issuer and subject unique ID fields in
5496 * Extend OCSP I/O functions so they can be used for simple general purpose
5497 HTTP as well as OCSP. New wrapper function which can be used to download
5498 CRLs using the OCSP API.
5502 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5506 * `SSL_CONF*` functions. These provide a common framework for application
5507 configuration using configuration files or command lines.
5511 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5512 message callback and prints the results. Needs compile time option
5513 "enable-ssl-trace". New options to s_client and s_server to enable
5518 * New ctrl and macro to retrieve supported points extensions.
5519 Print out extension in s_server and s_client.
5523 * New functions to retrieve certificate signature and signature
5528 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5533 * New Suite B modes for TLS code. These use and enforce the requirements
5534 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5535 only use Suite B curves. The Suite B modes can be set by using the
5536 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5540 * New chain verification flags for Suite B levels of security. Check
5541 algorithms are acceptable when flags are set in X509_verify_cert.
5545 * Make tls1_check_chain return a set of flags indicating checks passed
5546 by a certificate chain. Add additional tests to handle client
5547 certificates: checks for matching certificate type and issuer name
5552 * If an attempt is made to use a signature algorithm not in the peer
5553 preference list abort the handshake. If client has no suitable
5554 signature algorithms in response to a certificate request do not
5555 use the certificate.
5559 * If server EC tmp key is not in client preference list abort handshake.
5563 * Add support for certificate stores in CERT structure. This makes it
5564 possible to have different stores per SSL structure or one store in
5565 the parent SSL_CTX. Include distinct stores for certificate chain
5566 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5567 to build and store a certificate chain in CERT structure: returning
5568 an error if the chain cannot be built: this will allow applications
5569 to test if a chain is correctly configured.
5571 Note: if the CERT based stores are not set then the parent SSL_CTX
5572 store is used to retain compatibility with existing behaviour.
5576 * New function ssl_set_client_disabled to set a ciphersuite disabled
5577 mask based on the current session, check mask when sending client
5578 hello and checking the requested ciphersuite.
5582 * New ctrls to retrieve and set certificate types in a certificate
5583 request message. Print out received values in s_client. If certificate
5584 types is not set with custom values set sensible values based on
5585 supported signature algorithms.
5589 * Support for distinct client and server supported signature algorithms.
5593 * Add certificate callback. If set this is called whenever a certificate
5594 is required by client or server. An application can decide which
5595 certificate chain to present based on arbitrary criteria: for example
5596 supported signature algorithms. Add very simple example to s_server.
5597 This fixes many of the problems and restrictions of the existing client
5598 certificate callback: for example you can now clear an existing
5599 certificate and specify the whole chain.
5603 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5604 the certificate can be used for (if anything). Set valid_flags field
5605 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5606 to have similar checks in it.
5608 Add new "cert_flags" field to CERT structure and include a "strict mode".
5609 This enforces some TLS certificate requirements (such as only permitting
5610 certificate signature algorithms contained in the supported algorithms
5611 extension) which some implementations ignore: this option should be used
5612 with caution as it could cause interoperability issues.
5616 * Update and tidy signature algorithm extension processing. Work out
5617 shared signature algorithms based on preferences and peer algorithms
5618 and print them out in s_client and s_server. Abort handshake if no
5619 shared signature algorithms.
5623 * Add new functions to allow customised supported signature algorithms
5624 for SSL and SSL_CTX structures. Add options to s_client and s_server
5629 * New function SSL_certs_clear() to delete all references to certificates
5630 from an SSL structure. Before this once a certificate had been added
5631 it couldn't be removed.
5635 * Integrate hostname, email address and IP address checking with certificate
5636 verification. New verify options supporting checking in openssl utility.
5640 * Fixes and wildcard matching support to hostname and email checking
5641 functions. Add manual page.
5643 *Florian Weimer (Red Hat Product Security Team)*
5645 * New functions to check a hostname email or IP address against a
5646 certificate. Add options x509 utility to print results of checks against
5651 * Fix OCSP checking.
5653 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5655 * Initial experimental support for explicitly trusted non-root CAs.
5656 OpenSSL still tries to build a complete chain to a root but if an
5657 intermediate CA has a trust setting included that is used. The first
5658 setting is used: whether to trust (e.g., -addtrust option to the x509
5663 * Add -trusted_first option which attempts to find certificates in the
5664 trusted store even if an untrusted chain is also supplied.
5668 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5669 platform support for Linux and Android.
5673 * Support for linux-x32, ILP32 environment in x86_64 framework.
5677 * Experimental multi-implementation support for FIPS capable OpenSSL.
5678 When in FIPS mode the approved implementations are used as normal,
5679 when not in FIPS mode the internal unapproved versions are used instead.
5680 This means that the FIPS capable OpenSSL isn't forced to use the
5681 (often lower performance) FIPS implementations outside FIPS mode.
5685 * Transparently support X9.42 DH parameters when calling
5686 PEM_read_bio_DHparameters. This means existing applications can handle
5687 the new parameter format automatically.
5691 * Initial experimental support for X9.42 DH parameter format: mainly
5692 to support use of 'q' parameter for RFC5114 parameters.
5696 * Add DH parameters from RFC5114 including test data to dhtest.
5700 * Support for automatic EC temporary key parameter selection. If enabled
5701 the most preferred EC parameters are automatically used instead of
5702 hardcoded fixed parameters. Now a server just has to call:
5703 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5704 support ECDH and use the most appropriate parameters.
5708 * Enhance and tidy EC curve and point format TLS extension code. Use
5709 static structures instead of allocation if default values are used.
5710 New ctrls to set curves we wish to support and to retrieve shared curves.
5711 Print out shared curves in s_server. New options to s_server and s_client
5712 to set list of supported curves.
5716 * New ctrls to retrieve supported signature algorithms and
5717 supported curve values as an array of NIDs. Extend openssl utility
5718 to print out received values.
5722 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5723 between NIDs and the more common NIST names such as "P-256". Enhance
5724 ecparam utility and ECC method to recognise the NIST names for curves.
5728 * Enhance SSL/TLS certificate chain handling to support different
5729 chains for each certificate instead of one chain in the parent SSL_CTX.
5733 * Support for fixed DH ciphersuite client authentication: where both
5734 server and client use DH certificates with common parameters.
5738 * Support for fixed DH ciphersuites: those requiring DH server
5743 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5745 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5746 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5747 X509_CINF_get_signature were reverted post internal team review.
5752 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5754 * OCSP Status Request extension unbounded memory growth
5756 A malicious client can send an excessively large OCSP Status Request
5757 extension. If that client continually requests renegotiation, sending a
5758 large OCSP Status Request extension each time, then there will be unbounded
5759 memory growth on the server. This will eventually lead to a Denial Of
5760 Service attack through memory exhaustion. Servers with a default
5761 configuration are vulnerable even if they do not support OCSP. Builds using
5762 the "no-ocsp" build time option are not affected.
5764 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5769 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5772 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5778 * OOB write in MDC2_Update()
5780 An overflow can occur in MDC2_Update() either if called directly or
5781 through the EVP_DigestUpdate() function using MDC2. If an attacker
5782 is able to supply very large amounts of input data after a previous
5783 call to EVP_EncryptUpdate() with a partial block then a length check
5784 can overflow resulting in a heap corruption.
5786 The amount of data needed is comparable to SIZE_MAX which is impractical
5789 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5794 * Malformed SHA512 ticket DoS
5796 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5797 DoS attack where a malformed ticket will result in an OOB read which will
5800 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5801 a custom server callback and ticket lookup mechanism.
5803 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5808 * OOB write in BN_bn2dec()
5810 The function BN_bn2dec() does not check the return value of BN_div_word().
5811 This can cause an OOB write if an application uses this function with an
5812 overly large BIGNUM. This could be a problem if an overly large certificate
5813 or CRL is printed out from an untrusted source. TLS is not affected because
5814 record limits will reject an oversized certificate before it is parsed.
5816 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5821 * OOB read in TS_OBJ_print_bio()
5823 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5824 the total length the OID text representation would use and not the amount
5825 of data written. This will result in OOB reads when large OIDs are
5828 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5833 * Pointer arithmetic undefined behaviour
5835 Avoid some undefined pointer arithmetic
5837 A common idiom in the codebase is to check limits in the following manner:
5840 Where "p" points to some malloc'd data of SIZE bytes and
5843 "len" here could be from some externally supplied data (e.g. from a TLS
5846 The rules of C pointer arithmetic are such that "p + len" is only well
5847 defined where len <= SIZE. Therefore the above idiom is actually
5848 undefined behaviour.
5850 For example this could cause problems if some malloc implementation
5851 provides an address for "p" such that "p + len" actually overflows for
5852 values of len that are too big and therefore p + len < limit.
5854 This issue was reported to OpenSSL by Guido Vranken
5859 * Constant time flag not preserved in DSA signing
5861 Operations in the DSA signing algorithm should run in constant time in
5862 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5863 implementation means that a non-constant time codepath is followed for
5864 certain operations. This has been demonstrated through a cache-timing
5865 attack to be sufficient for an attacker to recover the private DSA key.
5867 This issue was reported by César Pereida (Aalto University), Billy Brumley
5868 (Tampere University of Technology), and Yuval Yarom (The University of
5869 Adelaide and NICTA).
5874 * DTLS buffered message DoS
5876 In a DTLS connection where handshake messages are delivered out-of-order
5877 those messages that OpenSSL is not yet ready to process will be buffered
5878 for later use. Under certain circumstances, a flaw in the logic means that
5879 those messages do not get removed from the buffer even though the handshake
5880 has been completed. An attacker could force up to approx. 15 messages to
5881 remain in the buffer when they are no longer required. These messages will
5882 be cleared when the DTLS connection is closed. The default maximum size for
5883 a message is 100k. Therefore the attacker could force an additional 1500k
5884 to be consumed per connection. By opening many simulataneous connections an
5885 attacker could cause a DoS attack through memory exhaustion.
5887 This issue was reported to OpenSSL by Quan Luo.
5892 * DTLS replay protection DoS
5894 A flaw in the DTLS replay attack protection mechanism means that records
5895 that arrive for future epochs update the replay protection "window" before
5896 the MAC for the record has been validated. This could be exploited by an
5897 attacker by sending a record for the next epoch (which does not have to
5898 decrypt or have a valid MAC), with a very large sequence number. This means
5899 that all subsequent legitimate packets are dropped causing a denial of
5900 service for a specific DTLS connection.
5902 This issue was reported to OpenSSL by the OCAP audit team.
5907 * Certificate message OOB reads
5909 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5910 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5911 theoretical DoS risk but this has not been observed in practice on common
5914 The messages affected are client certificate, client certificate request
5915 and server certificate. As a result the attack can only be performed
5916 against a client or a server which enables client authentication.
5918 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5923 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5925 * Prevent padding oracle in AES-NI CBC MAC check
5927 A MITM attacker can use a padding oracle attack to decrypt traffic
5928 when the connection uses an AES CBC cipher and the server support
5931 This issue was introduced as part of the fix for Lucky 13 padding
5932 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5933 constant time by making sure that always the same bytes are read and
5934 compared against either the MAC or padding bytes. But it no longer
5935 checked that there was enough data to have both the MAC and padding
5938 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5943 * Fix EVP_EncodeUpdate overflow
5945 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5946 Base64 encoding of binary data. If an attacker is able to supply very large
5947 amounts of input data then a length check can overflow resulting in a heap
5950 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5951 the `PEM_write_bio*` family of functions. These are mainly used within the
5952 OpenSSL command line applications, so any application which processes data
5953 from an untrusted source and outputs it as a PEM file should be considered
5954 vulnerable to this issue. User applications that call these APIs directly
5955 with large amounts of untrusted data may also be vulnerable.
5957 This issue was reported by Guido Vranken.
5962 * Fix EVP_EncryptUpdate overflow
5964 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5965 is able to supply very large amounts of input data after a previous call to
5966 EVP_EncryptUpdate() with a partial block then a length check can overflow
5967 resulting in a heap corruption. Following an analysis of all OpenSSL
5968 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5969 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5970 the first called function after an EVP_EncryptInit(), and therefore that
5971 specific call must be safe. The second form is where the length passed to
5972 EVP_EncryptUpdate() can be seen from the code to be some small value and
5973 therefore there is no possibility of an overflow. Since all instances are
5974 one of these two forms, it is believed that there can be no overflows in
5975 internal code due to this problem. It should be noted that
5976 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5977 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5978 of these calls have also been analysed too and it is believed there are no
5979 instances in internal usage where an overflow could occur.
5981 This issue was reported by Guido Vranken.
5986 * Prevent ASN.1 BIO excessive memory allocation
5988 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5989 a short invalid encoding can casuse allocation of large amounts of memory
5990 potentially consuming excessive resources or exhausting memory.
5992 Any application parsing untrusted data through d2i BIO functions is
5993 affected. The memory based functions such as d2i_X509() are *not* affected.
5994 Since the memory based functions are used by the TLS library, TLS
5995 applications are not affected.
5997 This issue was reported by Brian Carpenter.
6004 ASN1 Strings that are over 1024 bytes can cause an overread in applications
6005 using the X509_NAME_oneline() function on EBCDIC systems. This could result
6006 in arbitrary stack data being returned in the buffer.
6008 This issue was reported by Guido Vranken.
6013 * Modify behavior of ALPN to invoke callback after SNI/servername
6014 callback, such that updates to the SSL_CTX affect ALPN.
6018 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
6023 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
6024 methods are enabled and ssl2 is disabled the methods return NULL.
6028 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6030 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6031 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6032 provide any "EXPORT" or "LOW" strength ciphers.
6036 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6037 is by default disabled at build-time. Builds that are not configured with
6038 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6039 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6040 will need to explicitly call either of:
6042 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6044 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6046 as appropriate. Even if either of those is used, or the application
6047 explicitly uses the version-specific SSLv2_method() or its client and
6048 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6049 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6050 ciphers, and SSLv2 56-bit DES are no longer available.
6055 * Fix a double-free in DSA code
6057 A double free bug was discovered when OpenSSL parses malformed DSA private
6058 keys and could lead to a DoS attack or memory corruption for applications
6059 that receive DSA private keys from untrusted sources. This scenario is
6062 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6068 * Disable SRP fake user seed to address a server memory leak.
6070 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6072 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6073 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6074 was changed to ignore the "fake user" SRP seed, even if the seed
6077 Users should use SRP_VBASE_get1_by_user instead. Note that in
6078 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6079 also that even though configuring the SRP seed attempts to hide
6080 invalid usernames by continuing the handshake with fake
6081 credentials, this behaviour is not constant time and no strong
6082 guarantees are made that the handshake is indistinguishable from
6083 that of a valid user.
6088 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6090 In the BN_hex2bn function the number of hex digits is calculated using an
6091 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6092 large values of `i` this can result in `bn_expand` not allocating any
6093 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6094 field as NULL leading to a subsequent NULL ptr deref. For very large values
6095 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6096 In this case memory is allocated to the internal BIGNUM data field, but it
6097 is insufficiently sized leading to heap corruption. A similar issue exists
6098 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6099 is ever called by user applications with very large untrusted hex/dec data.
6100 This is anticipated to be a rare occurrence.
6102 All OpenSSL internal usage of these functions use data that is not expected
6103 to be untrusted, e.g. config file data or application command line
6104 arguments. If user developed applications generate config file data based
6105 on untrusted data then it is possible that this could also lead to security
6106 consequences. This is also anticipated to be rare.
6108 This issue was reported to OpenSSL by Guido Vranken.
6113 * Fix memory issues in `BIO_*printf` functions
6115 The internal `fmtstr` function used in processing a "%s" format string in
6116 the `BIO_*printf` functions could overflow while calculating the length of a
6117 string and cause an OOB read when printing very long strings.
6119 Additionally the internal `doapr_outch` function can attempt to write to an
6120 OOB memory location (at an offset from the NULL pointer) in the event of a
6121 memory allocation failure. In 1.0.2 and below this could be caused where
6122 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6123 could be in processing a very long "%s" format string. Memory leaks can
6126 The first issue may mask the second issue dependent on compiler behaviour.
6127 These problems could enable attacks where large amounts of untrusted data
6128 is passed to the `BIO_*printf` functions. If applications use these functions
6129 in this way then they could be vulnerable. OpenSSL itself uses these
6130 functions when printing out human-readable dumps of ASN.1 data. Therefore
6131 applications that print this data could be vulnerable if the data is from
6132 untrusted sources. OpenSSL command line applications could also be
6133 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6134 as command line arguments.
6136 Libssl is not considered directly vulnerable. Additionally certificates etc
6137 received via remote connections via libssl are also unlikely to be able to
6138 trigger these issues because of message size limits enforced within libssl.
6140 This issue was reported to OpenSSL Guido Vranken.
6145 * Side channel attack on modular exponentiation
6147 A side-channel attack was found which makes use of cache-bank conflicts on
6148 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6149 of RSA keys. The ability to exploit this issue is limited as it relies on
6150 an attacker who has control of code in a thread running on the same
6151 hyper-threaded core as the victim thread which is performing decryptions.
6153 This issue was reported to OpenSSL by Yuval Yarom, The University of
6154 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6155 Nadia Heninger, University of Pennsylvania with more information at
6156 <http://cachebleed.info>.
6161 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6162 if no keysize is specified with default_bits. This fixes an
6163 omission in an earlier change that changed all RSA/DSA key generation
6164 commands to use 2048 bits by default.
6168 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6170 * Protection for DH small subgroup attacks
6172 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6173 switched on by default and cannot be disabled. This could have some
6178 * SSLv2 doesn't block disabled ciphers
6180 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6181 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6182 been disabled, provided that the SSLv2 protocol was not also disabled via
6185 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6186 and Sebastian Schinzel.
6191 * Reject DH handshakes with parameters shorter than 1024 bits.
6195 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6197 * Certificate verify crash with missing PSS parameter
6199 The signature verification routines will crash with a NULL pointer
6200 dereference if presented with an ASN.1 signature using the RSA PSS
6201 algorithm and absent mask generation function parameter. Since these
6202 routines are used to verify certificate signature algorithms this can be
6203 used to crash any certificate verification operation and exploited in a
6204 DoS attack. Any application which performs certificate verification is
6205 vulnerable including OpenSSL clients and servers which enable client
6208 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6213 * X509_ATTRIBUTE memory leak
6215 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6216 memory. This structure is used by the PKCS#7 and CMS routines so any
6217 application which reads PKCS#7 or CMS data from untrusted sources is
6218 affected. SSL/TLS is not affected.
6220 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6226 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6227 This changes the decoding behaviour for some invalid messages,
6228 though the change is mostly in the more lenient direction, and
6229 legacy behaviour is preserved as much as possible.
6233 * In DSA_generate_parameters_ex, if the provided seed is too short,
6234 use a random seed, as already documented.
6236 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6238 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6240 * Alternate chains certificate forgery
6242 During certificate verfification, OpenSSL will attempt to find an
6243 alternative certificate chain if the first attempt to build such a chain
6244 fails. An error in the implementation of this logic can mean that an
6245 attacker could cause certain checks on untrusted certificates to be
6246 bypassed, such as the CA flag, enabling them to use a valid leaf
6247 certificate to act as a CA and "issue" an invalid certificate.
6249 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6255 * Race condition handling PSK identify hint
6257 If PSK identity hints are received by a multi-threaded client then
6258 the values are wrongly updated in the parent SSL_CTX structure. This can
6259 result in a race condition potentially leading to a double free of the
6265 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6267 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6268 incompatibility in the handling of HMAC. The previous ABI has now been
6271 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6273 * Malformed ECParameters causes infinite loop
6275 When processing an ECParameters structure OpenSSL enters an infinite loop
6276 if the curve specified is over a specially malformed binary polynomial
6279 This can be used to perform denial of service against any
6280 system which processes public keys, certificate requests or
6281 certificates. This includes TLS clients and TLS servers with
6282 client authentication enabled.
6284 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6289 * Exploitable out-of-bounds read in X509_cmp_time
6291 X509_cmp_time does not properly check the length of the ASN1_TIME
6292 string and can read a few bytes out of bounds. In addition,
6293 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6296 An attacker can use this to craft malformed certificates and CRLs of
6297 various sizes and potentially cause a segmentation fault, resulting in
6298 a DoS on applications that verify certificates or CRLs. TLS clients
6299 that verify CRLs are affected. TLS clients and servers with client
6300 authentication enabled may be affected if they use custom verification
6303 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6304 independently by Hanno Böck.
6309 * PKCS7 crash with missing EnvelopedContent
6311 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6312 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6313 with missing content and trigger a NULL pointer dereference on parsing.
6315 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6316 structures from untrusted sources are affected. OpenSSL clients and
6317 servers are not affected.
6319 This issue was reported to OpenSSL by Michal Zalewski (Google).
6324 * CMS verify infinite loop with unknown hash function
6326 When verifying a signedData message the CMS code can enter an infinite loop
6327 if presented with an unknown hash function OID. This can be used to perform
6328 denial of service against any system which verifies signedData messages using
6330 This issue was reported to OpenSSL by Johannes Bauer.
6335 * Race condition handling NewSessionTicket
6337 If a NewSessionTicket is received by a multi-threaded client when attempting to
6338 reuse a previous ticket then a race condition can occur potentially leading to
6339 a double free of the ticket data.
6344 * Reject DH handshakes with parameters shorter than 768 bits.
6346 *Kurt Roeckx and Emilia Kasper*
6348 * dhparam: generate 2048-bit parameters by default.
6350 *Kurt Roeckx and Emilia Kasper*
6352 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6354 * Segmentation fault in ASN1_TYPE_cmp fix
6356 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6357 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6358 certificate signature algorithm consistency this can be used to crash any
6359 certificate verification operation and exploited in a DoS attack. Any
6360 application which performs certificate verification is vulnerable including
6361 OpenSSL clients and servers which enable client authentication.
6366 * ASN.1 structure reuse memory corruption fix
6368 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6369 memory corruption via an invalid write. Such reuse is and has been
6370 strongly discouraged and is believed to be rare.
6372 Applications that parse structures containing CHOICE or ANY DEFINED BY
6373 components may be affected. Certificate parsing (d2i_X509 and related
6374 functions) are however not affected. OpenSSL clients and servers are
6380 * PKCS7 NULL pointer dereferences fix
6382 The PKCS#7 parsing code does not handle missing outer ContentInfo
6383 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6384 missing content and trigger a NULL pointer dereference on parsing.
6386 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6387 otherwise parse PKCS#7 structures from untrusted sources are
6388 affected. OpenSSL clients and servers are not affected.
6390 This issue was reported to OpenSSL by Michal Zalewski (Google).
6395 * DoS via reachable assert in SSLv2 servers fix
6397 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6398 servers that both support SSLv2 and enable export cipher suites by sending
6399 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6401 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6402 (OpenSSL development team).
6407 * Use After Free following d2i_ECPrivatekey error fix
6409 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6410 could cause a use after free condition. This, in turn, could cause a double
6411 free in several private key parsing functions (such as d2i_PrivateKey
6412 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6413 for applications that receive EC private keys from untrusted
6414 sources. This scenario is considered rare.
6416 This issue was discovered by the BoringSSL project and fixed in their
6422 * X509_to_X509_REQ NULL pointer deref fix
6424 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6425 the certificate key is invalid. This function is rarely used in practice.
6427 This issue was discovered by Brian Carpenter.
6432 * Removed the export ciphers from the DEFAULT ciphers
6436 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6438 * Build fixes for the Windows and OpenVMS platforms
6440 *Matt Caswell and Richard Levitte*
6442 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6444 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6445 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6446 dereference. This could lead to a Denial Of Service attack. Thanks to
6447 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6452 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6453 dtls1_buffer_record function under certain conditions. In particular this
6454 could occur if an attacker sent repeated DTLS records with the same
6455 sequence number but for the next epoch. The memory leak could be exploited
6456 by an attacker in a Denial of Service attack through memory exhaustion.
6457 Thanks to Chris Mueller for reporting this issue.
6462 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6463 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6464 method would be set to NULL which could later result in a NULL pointer
6465 dereference. Thanks to Frank Schmirler for reporting this issue.
6470 * Abort handshake if server key exchange message is omitted for ephemeral
6473 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6474 reporting this issue.
6479 * Remove non-export ephemeral RSA code on client and server. This code
6480 violated the TLS standard by allowing the use of temporary RSA keys in
6481 non-export ciphersuites and could be used by a server to effectively
6482 downgrade the RSA key length used to a value smaller than the server
6483 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6484 INRIA or reporting this issue.
6489 * Fixed issue where DH client certificates are accepted without verification.
6490 An OpenSSL server will accept a DH certificate for client authentication
6491 without the certificate verify message. This effectively allows a client to
6492 authenticate without the use of a private key. This only affects servers
6493 which trust a client certificate authority which issues certificates
6494 containing DH keys: these are extremely rare and hardly ever encountered.
6495 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6501 * Ensure that the session ID context of an SSL is updated when its
6502 SSL_CTX is updated via SSL_set_SSL_CTX.
6504 The session ID context is typically set from the parent SSL_CTX,
6505 and can vary with the CTX.
6509 * Fix various certificate fingerprint issues.
6511 By using non-DER or invalid encodings outside the signed portion of a
6512 certificate the fingerprint can be changed without breaking the signature.
6513 Although no details of the signed portion of the certificate can be changed
6514 this can cause problems with some applications: e.g. those using the
6515 certificate fingerprint for blacklists.
6517 1. Reject signatures with non zero unused bits.
6519 If the BIT STRING containing the signature has non zero unused bits reject
6520 the signature. All current signature algorithms require zero unused bits.
6522 2. Check certificate algorithm consistency.
6524 Check the AlgorithmIdentifier inside TBS matches the one in the
6525 certificate signature. NB: this will result in signature failure
6526 errors for some broken certificates.
6528 Thanks to Konrad Kraszewski from Google for reporting this issue.
6530 3. Check DSA/ECDSA signatures use DER.
6532 Re-encode DSA/ECDSA signatures and compare with the original received
6533 signature. Return an error if there is a mismatch.
6535 This will reject various cases including garbage after signature
6536 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6537 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6538 (negative or with leading zeroes).
6540 Further analysis was conducted and fixes were developed by Stephen Henson
6541 of the OpenSSL core team.
6547 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6548 results on some platforms, including x86_64. This bug occurs at random
6549 with a very low probability, and is not known to be exploitable in any
6550 way, though its exact impact is difficult to determine. Thanks to Pieter
6551 Wuille (Blockstream) who reported this issue and also suggested an initial
6552 fix. Further analysis was conducted by the OpenSSL development team and
6553 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6554 the OpenSSL core team.
6559 * Do not resume sessions on the server if the negotiated protocol
6560 version does not match the session's version. Resuming with a different
6561 version, while not strictly forbidden by the RFC, is of questionable
6562 sanity and breaks all known clients.
6564 *David Benjamin, Emilia Käsper*
6566 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6567 early CCS messages during renegotiation. (Note that because
6568 renegotiation is encrypted, this early CCS was not exploitable.)
6572 * Tighten client-side session ticket handling during renegotiation:
6573 ensure that the client only accepts a session ticket if the server sends
6574 the extension anew in the ServerHello. Previously, a TLS client would
6575 reuse the old extension state and thus accept a session ticket if one was
6576 announced in the initial ServerHello.
6578 Similarly, ensure that the client requires a session ticket if one
6579 was advertised in the ServerHello. Previously, a TLS client would
6580 ignore a missing NewSessionTicket message.
6584 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6588 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6589 sends a carefully crafted handshake message, to cause OpenSSL to fail
6590 to free up to 64k of memory causing a memory leak. This could be
6591 exploited in a Denial Of Service attack. This issue affects OpenSSL
6592 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6593 whether SRTP is used or configured. Implementations of OpenSSL that
6594 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6596 The fix was developed by the OpenSSL team.
6601 * Session Ticket Memory Leak.
6603 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6604 integrity of that ticket is first verified. In the event of a session
6605 ticket integrity check failing, OpenSSL will fail to free memory
6606 causing a memory leak. By sending a large number of invalid session
6607 tickets an attacker could exploit this issue in a Denial Of Service
6613 * Build option no-ssl3 is incomplete.
6615 When OpenSSL is configured with "no-ssl3" as a build option, servers
6616 could accept and complete a SSL 3.0 handshake, and clients could be
6617 configured to send them.
6620 *Akamai and the OpenSSL team*
6622 * Add support for TLS_FALLBACK_SCSV.
6623 Client applications doing fallback retries should call
6624 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6627 *Adam Langley, Bodo Moeller*
6629 * Add additional DigestInfo checks.
6631 Re-encode DigestInto in DER and check against the original when
6632 verifying RSA signature: this will reject any improperly encoded
6633 DigestInfo structures.
6635 Note: this is a precautionary measure and no attacks are currently known.
6639 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6641 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6642 SRP code can be overrun an internal buffer. Add sanity check that
6643 g, A, B < N to SRP code.
6645 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6646 Group for discovering this issue.
6651 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6652 TLS 1.0 instead of higher protocol versions when the ClientHello message
6653 is badly fragmented. This allows a man-in-the-middle attacker to force a
6654 downgrade to TLS 1.0 even if both the server and the client support a
6655 higher protocol version, by modifying the client's TLS records.
6657 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6658 researching this issue.
6663 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6664 to a denial of service attack. A malicious server can crash the client
6665 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6666 ciphersuite and sending carefully crafted handshake messages.
6668 Thanks to Felix Gröbert (Google) for discovering and researching this
6674 * By sending carefully crafted DTLS packets an attacker could cause openssl
6675 to leak memory. This can be exploited through a Denial of Service attack.
6676 Thanks to Adam Langley for discovering and researching this issue.
6681 * An attacker can force openssl to consume large amounts of memory whilst
6682 processing DTLS handshake messages. This can be exploited through a
6683 Denial of Service attack.
6684 Thanks to Adam Langley for discovering and researching this issue.
6689 * An attacker can force an error condition which causes openssl to crash
6690 whilst processing DTLS packets due to memory being freed twice. This
6691 can be exploited through a Denial of Service attack.
6692 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6698 * If a multithreaded client connects to a malicious server using a resumed
6699 session and the server sends an ec point format extension it could write
6700 up to 255 bytes to freed memory.
6702 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6708 * A malicious server can crash an OpenSSL client with a null pointer
6709 dereference (read) by specifying an SRP ciphersuite even though it was not
6710 properly negotiated with the client. This can be exploited through a
6711 Denial of Service attack.
6713 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6714 discovering and researching this issue.
6719 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6720 X509_name_oneline, X509_name_print_ex et al. to leak some information
6721 from the stack. Applications may be affected if they echo pretty printing
6722 output to the attacker.
6724 Thanks to Ivan Fratric (Google) for discovering this issue.
6727 *Emilia Käsper, and Steve Henson*
6729 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6730 for corner cases. (Certain input points at infinity could lead to
6731 bogus results, with non-infinity inputs mapped to infinity too.)
6735 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6737 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6738 handshake can force the use of weak keying material in OpenSSL
6739 SSL/TLS clients and servers.
6741 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6742 researching this issue. ([CVE-2014-0224])
6744 *KIKUCHI Masashi, Steve Henson*
6746 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6747 OpenSSL DTLS client the code can be made to recurse eventually crashing
6750 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6753 *Imre Rad, Steve Henson*
6755 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6756 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6757 client or server. This is potentially exploitable to run arbitrary
6758 code on a vulnerable client or server.
6760 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6762 *Jüri Aedla, Steve Henson*
6764 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6765 are subject to a denial of service attack.
6767 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6768 this issue. ([CVE-2014-3470])
6770 *Felix Gröbert, Ivan Fratric, Steve Henson*
6772 * Harmonize version and its documentation. -f flag is used to display
6775 *mancha <mancha1@zoho.com>*
6777 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6778 in i2d_ECPrivateKey.
6780 *mancha <mancha1@zoho.com>*
6782 * Fix some double frees. These are not thought to be exploitable.
6784 *mancha <mancha1@zoho.com>*
6786 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6788 * A missing bounds check in the handling of the TLS heartbeat extension
6789 can be used to reveal up to 64k of memory to a connected client or
6792 Thanks for Neel Mehta of Google Security for discovering this bug and to
6793 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6794 preparing the fix ([CVE-2014-0160])
6796 *Adam Langley, Bodo Moeller*
6798 * Fix for the attack described in the paper "Recovering OpenSSL
6799 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6800 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6801 <http://eprint.iacr.org/2014/140>
6803 Thanks to Yuval Yarom and Naomi Benger for discovering this
6804 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6806 *Yuval Yarom and Naomi Benger*
6808 * TLS pad extension: draft-agl-tls-padding-03
6810 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6811 TLS client Hello record length value would otherwise be > 255 and
6812 less that 512 pad with a dummy extension containing zeroes so it
6813 is at least 512 bytes long.
6815 *Adam Langley, Steve Henson*
6817 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6819 * Fix for TLS record tampering bug. A carefully crafted invalid
6820 handshake could crash OpenSSL with a NULL pointer exception.
6821 Thanks to Anton Johansson for reporting this issues.
6824 * Keep original DTLS digest and encryption contexts in retransmission
6825 structures so we can use the previous session parameters if they need
6826 to be resent. ([CVE-2013-6450])
6830 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6831 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6832 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6833 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6834 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6835 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6837 *Rob Stradling, Adam Langley*
6839 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6841 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6842 supporting platforms or when small records were transferred.
6844 *Andy Polyakov, Steve Henson*
6846 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6848 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6850 This addresses the flaw in CBC record processing discovered by
6851 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6852 at: <http://www.isg.rhul.ac.uk/tls/>
6854 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6855 Security Group at Royal Holloway, University of London
6856 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6857 Emilia Käsper for the initial patch.
6860 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6862 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6863 ciphersuites which can be exploited in a denial of service attack.
6864 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6865 and detecting this bug and to Wolfgang Ettlinger
6866 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6871 * Return an error when checking OCSP signatures when key is NULL.
6872 This fixes a DoS attack. ([CVE-2013-0166])
6876 * Make openssl verify return errors.
6878 *Chris Palmer <palmer@google.com> and Ben Laurie*
6880 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6881 the right response is stapled. Also change SSL_get_certificate()
6882 so it returns the certificate actually sent.
6883 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6885 *Rob Stradling <rob.stradling@comodo.com>*
6887 * Fix possible deadlock when decoding public keys.
6891 * Don't use TLS 1.0 record version number in initial client hello
6896 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6898 * Sanity check record length before skipping explicit IV in TLS
6899 1.2, 1.1 and DTLS to fix DoS attack.
6901 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6902 fuzzing as a service testing platform.
6907 * Initialise tkeylen properly when encrypting CMS messages.
6908 Thanks to Solar Designer of Openwall for reporting this issue.
6912 * In FIPS mode don't try to use composite ciphers as they are not
6917 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6919 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6920 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6921 mean any application compiled against OpenSSL 1.0.0 headers setting
6922 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6923 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6924 0x10000000L Any application which was previously compiled against
6925 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6926 will need to be recompiled as a result. Letting be results in
6927 inability to disable specifically TLS 1.1 and in client context,
6928 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6932 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6933 disable just protocol X, but all protocols above X *if* there are
6934 protocols *below* X still enabled. In more practical terms it means
6935 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6936 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6937 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6942 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6944 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6945 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6946 in CRYPTO_realloc_clean.
6948 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6949 issue and to Adam Langley <agl@chromium.org> for fixing it.
6952 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6954 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6958 * Workarounds for some broken servers that "hang" if a client hello
6959 record length exceeds 255 bytes.
6961 1. Do not use record version number > TLS 1.0 in initial client
6962 hello: some (but not all) hanging servers will now work.
6963 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6964 the number of ciphers sent in the client hello. This should be
6965 set to an even number, such as 50, for example by passing:
6966 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6967 Most broken servers should now work.
6968 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6969 TLS 1.2 client support entirely.
6973 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6977 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6979 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6980 STRING form instead of a DigestInfo.
6984 * The format used for MDC2 RSA signatures is inconsistent between EVP
6985 and the RSA_sign/RSA_verify functions. This was made more apparent when
6986 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6987 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6988 the correct format in RSA_verify so both forms transparently work.
6992 * Some servers which support TLS 1.0 can choke if we initially indicate
6993 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6994 encrypted premaster secret. As a workaround use the maximum permitted
6995 client version in client hello, this should keep such servers happy
6996 and still work with previous versions of OpenSSL.
7000 * Add support for TLS/DTLS heartbeats.
7002 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7004 * Add support for SCTP.
7006 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7008 * Improved PRNG seeding for VOS.
7010 *Paul Green <Paul.Green@stratus.com>*
7012 * Extensive assembler packs updates, most notably:
7014 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
7015 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
7016 - x86_64: bit-sliced AES implementation;
7017 - ARM: NEON support, contemporary platforms optimizations;
7018 - s390x: z196 support;
7019 - `*`: GHASH and GF(2^m) multiplication implementations;
7023 * Make TLS-SRP code conformant with RFC 5054 API cleanup
7024 (removal of unnecessary code)
7026 *Peter Sylvester <peter.sylvester@edelweb.fr>*
7028 * Add TLS key material exporter from RFC 5705.
7032 * Add DTLS-SRTP negotiation from RFC 5764.
7036 * Add Next Protocol Negotiation,
7037 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7038 disabled with a no-npn flag to config or Configure. Code donated
7041 *Adam Langley <agl@google.com> and Ben Laurie*
7043 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7044 NIST-P256, NIST-P521, with constant-time single point multiplication on
7045 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7046 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7047 Code made available under Apache License version 2.0.
7049 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7050 line to include this in your build of OpenSSL, and run "make depend" (or
7051 "make update"). This enables the following EC_METHODs:
7053 EC_GFp_nistp224_method()
7054 EC_GFp_nistp256_method()
7055 EC_GFp_nistp521_method()
7057 EC_GROUP_new_by_curve_name() will automatically use these (while
7058 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7061 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7063 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7064 all platforms. Move ssize_t definition from e_os.h to the public
7065 header file e_os2.h as it now appears in public header file cms.h
7069 * New -sigopt option to the ca, req and x509 utilities. Additional
7070 signature parameters can be passed using this option and in
7075 * Add RSA PSS signing function. This will generate and set the
7076 appropriate AlgorithmIdentifiers for PSS based on those in the
7077 corresponding EVP_MD_CTX structure. No application support yet.
7081 * Support for companion algorithm specific ASN1 signing routines.
7082 New function ASN1_item_sign_ctx() signs a pre-initialised
7083 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7084 the appropriate parameters.
7088 * Add new algorithm specific ASN1 verification initialisation function
7089 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7090 handling will be the same no matter what EVP_PKEY_METHOD is used.
7091 Add a PSS handler to support verification of PSS signatures: checked
7092 against a number of sample certificates.
7096 * Add signature printing for PSS. Add PSS OIDs.
7098 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7100 * Add algorithm specific signature printing. An individual ASN1 method
7101 can now print out signatures instead of the standard hex dump.
7103 More complex signatures (e.g. PSS) can print out more meaningful
7104 information. Include DSA version that prints out the signature
7109 * Password based recipient info support for CMS library: implementing
7114 * Split password based encryption into PBES2 and PBKDF2 functions. This
7115 neatly separates the code into cipher and PBE sections and is required
7116 for some algorithms that split PBES2 into separate pieces (such as
7117 password based CMS).
7121 * Session-handling fixes:
7122 - Fix handling of connections that are resuming with a session ID,
7123 but also support Session Tickets.
7124 - Fix a bug that suppressed issuing of a new ticket if the client
7125 presented a ticket with an expired session.
7126 - Try to set the ticket lifetime hint to something reasonable.
7127 - Make tickets shorter by excluding irrelevant information.
7128 - On the client side, don't ignore renewed tickets.
7130 *Adam Langley, Bodo Moeller (Google)*
7132 * Fix PSK session representation.
7136 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7138 This work was sponsored by Intel.
7142 * Add GCM support to TLS library. Some custom code is needed to split
7143 the IV between the fixed (from PRF) and explicit (from TLS record)
7144 portions. This adds all GCM ciphersuites supported by RFC5288 and
7145 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7146 add a special AESGCM string for GCM only.
7150 * Expand range of ctrls for AES GCM. Permit setting invocation
7151 field on decrypt and retrieval of invocation field only on encrypt.
7155 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7156 As required by RFC5289 these ciphersuites cannot be used if for
7157 versions of TLS earlier than 1.2.
7161 * For FIPS capable OpenSSL interpret a NULL default public key method
7162 as unset and return the appropriate default but do *not* set the default.
7163 This means we can return the appropriate method in applications that
7164 switch between FIPS and non-FIPS modes.
7168 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7169 ENGINE is used then we cannot handle that in the FIPS module so we
7170 keep original code iff non-FIPS operations are allowed.
7174 * Add -attime option to openssl utilities.
7176 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7178 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7182 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7183 FIPS EC methods unconditionally for now.
7187 * New build option no-ec2m to disable characteristic 2 code.
7191 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7192 all cases can be covered as some introduce binary incompatibilities.
7196 * Redirect RSA operations to FIPS module including keygen,
7197 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7201 * Add similar low level API blocking to ciphers.
7205 * Low level digest APIs are not approved in FIPS mode: any attempt
7206 to use these will cause a fatal error. Applications that *really* want
7207 to use them can use the `private_*` version instead.
7211 * Redirect cipher operations to FIPS module for FIPS builds.
7215 * Redirect digest operations to FIPS module for FIPS builds.
7219 * Update build system to add "fips" flag which will link in fipscanister.o
7220 for static and shared library builds embedding a signature if needed.
7224 * Output TLS supported curves in preference order instead of numerical
7225 order. This is currently hardcoded for the highest order curves first.
7226 This should be configurable so applications can judge speed vs strength.
7230 * Add TLS v1.2 server support for client authentication.
7234 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7239 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7240 FIPS modules versions.
7244 * Add TLS v1.2 client side support for client authentication. Keep cache
7245 of handshake records longer as we don't know the hash algorithm to use
7246 until after the certificate request message is received.
7250 * Initial TLS v1.2 client support. Add a default signature algorithms
7251 extension including all the algorithms we support. Parse new signature
7252 format in client key exchange. Relax some ECC signing restrictions for
7253 TLS v1.2 as indicated in RFC5246.
7257 * Add server support for TLS v1.2 signature algorithms extension. Switch
7258 to new signature format when needed using client digest preference.
7259 All server ciphersuites should now work correctly in TLS v1.2. No client
7260 support yet and no support for client certificates.
7264 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7265 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7266 ciphersuites. At present only RSA key exchange ciphersuites work with
7267 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7268 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7269 and version checking.
7273 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7274 with this defined it will not be affected by any changes to ssl internal
7275 structures. Add several utility functions to allow openssl application
7276 to work with OPENSSL_NO_SSL_INTERN defined.
7280 * A long standing patch to add support for SRP from EdelWeb (Peter
7281 Sylvester and Christophe Renou) was integrated.
7282 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7283 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7286 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7290 * Permit abbreviated handshakes when renegotiating using the function
7291 SSL_renegotiate_abbreviated().
7293 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7295 * Add call to ENGINE_register_all_complete() to
7296 ENGINE_load_builtin_engines(), so some implementations get used
7297 automatically instead of needing explicit application support.
7301 * Add support for TLS key exporter as described in RFC5705.
7303 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7305 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7306 a few changes are required:
7308 Add SSL_OP_NO_TLSv1_1 flag.
7309 Add TLSv1_1 methods.
7310 Update version checking logic to handle version 1.1.
7311 Add explicit IV handling (ported from DTLS code).
7312 Add command line options to s_client/s_server.
7319 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7321 * X509_ATTRIBUTE memory leak
7323 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7324 memory. This structure is used by the PKCS#7 and CMS routines so any
7325 application which reads PKCS#7 or CMS data from untrusted sources is
7326 affected. SSL/TLS is not affected.
7328 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7334 * Race condition handling PSK identify hint
7336 If PSK identity hints are received by a multi-threaded client then
7337 the values are wrongly updated in the parent SSL_CTX structure. This can
7338 result in a race condition potentially leading to a double free of the
7344 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7346 * Malformed ECParameters causes infinite loop
7348 When processing an ECParameters structure OpenSSL enters an infinite loop
7349 if the curve specified is over a specially malformed binary polynomial
7352 This can be used to perform denial of service against any
7353 system which processes public keys, certificate requests or
7354 certificates. This includes TLS clients and TLS servers with
7355 client authentication enabled.
7357 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7362 * Exploitable out-of-bounds read in X509_cmp_time
7364 X509_cmp_time does not properly check the length of the ASN1_TIME
7365 string and can read a few bytes out of bounds. In addition,
7366 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7369 An attacker can use this to craft malformed certificates and CRLs of
7370 various sizes and potentially cause a segmentation fault, resulting in
7371 a DoS on applications that verify certificates or CRLs. TLS clients
7372 that verify CRLs are affected. TLS clients and servers with client
7373 authentication enabled may be affected if they use custom verification
7376 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7377 independently by Hanno Böck.
7382 * PKCS7 crash with missing EnvelopedContent
7384 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7385 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7386 with missing content and trigger a NULL pointer dereference on parsing.
7388 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7389 structures from untrusted sources are affected. OpenSSL clients and
7390 servers are not affected.
7392 This issue was reported to OpenSSL by Michal Zalewski (Google).
7397 * CMS verify infinite loop with unknown hash function
7399 When verifying a signedData message the CMS code can enter an infinite loop
7400 if presented with an unknown hash function OID. This can be used to perform
7401 denial of service against any system which verifies signedData messages using
7403 This issue was reported to OpenSSL by Johannes Bauer.
7408 * Race condition handling NewSessionTicket
7410 If a NewSessionTicket is received by a multi-threaded client when attempting to
7411 reuse a previous ticket then a race condition can occur potentially leading to
7412 a double free of the ticket data.
7417 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7419 * Segmentation fault in ASN1_TYPE_cmp fix
7421 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7422 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7423 certificate signature algorithm consistency this can be used to crash any
7424 certificate verification operation and exploited in a DoS attack. Any
7425 application which performs certificate verification is vulnerable including
7426 OpenSSL clients and servers which enable client authentication.
7431 * ASN.1 structure reuse memory corruption fix
7433 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7434 memory corruption via an invalid write. Such reuse is and has been
7435 strongly discouraged and is believed to be rare.
7437 Applications that parse structures containing CHOICE or ANY DEFINED BY
7438 components may be affected. Certificate parsing (d2i_X509 and related
7439 functions) are however not affected. OpenSSL clients and servers are
7445 * PKCS7 NULL pointer dereferences fix
7447 The PKCS#7 parsing code does not handle missing outer ContentInfo
7448 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7449 missing content and trigger a NULL pointer dereference on parsing.
7451 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7452 otherwise parse PKCS#7 structures from untrusted sources are
7453 affected. OpenSSL clients and servers are not affected.
7455 This issue was reported to OpenSSL by Michal Zalewski (Google).
7460 * DoS via reachable assert in SSLv2 servers fix
7462 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7463 servers that both support SSLv2 and enable export cipher suites by sending
7464 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7466 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7467 (OpenSSL development team).
7472 * Use After Free following d2i_ECPrivatekey error fix
7474 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7475 could cause a use after free condition. This, in turn, could cause a double
7476 free in several private key parsing functions (such as d2i_PrivateKey
7477 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7478 for applications that receive EC private keys from untrusted
7479 sources. This scenario is considered rare.
7481 This issue was discovered by the BoringSSL project and fixed in their
7487 * X509_to_X509_REQ NULL pointer deref fix
7489 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7490 the certificate key is invalid. This function is rarely used in practice.
7492 This issue was discovered by Brian Carpenter.
7497 * Removed the export ciphers from the DEFAULT ciphers
7501 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7503 * Build fixes for the Windows and OpenVMS platforms
7505 *Matt Caswell and Richard Levitte*
7507 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7509 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7510 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7511 dereference. This could lead to a Denial Of Service attack. Thanks to
7512 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7517 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7518 dtls1_buffer_record function under certain conditions. In particular this
7519 could occur if an attacker sent repeated DTLS records with the same
7520 sequence number but for the next epoch. The memory leak could be exploited
7521 by an attacker in a Denial of Service attack through memory exhaustion.
7522 Thanks to Chris Mueller for reporting this issue.
7527 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7528 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7529 method would be set to NULL which could later result in a NULL pointer
7530 dereference. Thanks to Frank Schmirler for reporting this issue.
7535 * Abort handshake if server key exchange message is omitted for ephemeral
7538 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7539 reporting this issue.
7544 * Remove non-export ephemeral RSA code on client and server. This code
7545 violated the TLS standard by allowing the use of temporary RSA keys in
7546 non-export ciphersuites and could be used by a server to effectively
7547 downgrade the RSA key length used to a value smaller than the server
7548 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7549 INRIA or reporting this issue.
7554 * Fixed issue where DH client certificates are accepted without verification.
7555 An OpenSSL server will accept a DH certificate for client authentication
7556 without the certificate verify message. This effectively allows a client to
7557 authenticate without the use of a private key. This only affects servers
7558 which trust a client certificate authority which issues certificates
7559 containing DH keys: these are extremely rare and hardly ever encountered.
7560 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7566 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7567 results on some platforms, including x86_64. This bug occurs at random
7568 with a very low probability, and is not known to be exploitable in any
7569 way, though its exact impact is difficult to determine. Thanks to Pieter
7570 Wuille (Blockstream) who reported this issue and also suggested an initial
7571 fix. Further analysis was conducted by the OpenSSL development team and
7572 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7573 the OpenSSL core team.
7578 * Fix various certificate fingerprint issues.
7580 By using non-DER or invalid encodings outside the signed portion of a
7581 certificate the fingerprint can be changed without breaking the signature.
7582 Although no details of the signed portion of the certificate can be changed
7583 this can cause problems with some applications: e.g. those using the
7584 certificate fingerprint for blacklists.
7586 1. Reject signatures with non zero unused bits.
7588 If the BIT STRING containing the signature has non zero unused bits reject
7589 the signature. All current signature algorithms require zero unused bits.
7591 2. Check certificate algorithm consistency.
7593 Check the AlgorithmIdentifier inside TBS matches the one in the
7594 certificate signature. NB: this will result in signature failure
7595 errors for some broken certificates.
7597 Thanks to Konrad Kraszewski from Google for reporting this issue.
7599 3. Check DSA/ECDSA signatures use DER.
7601 Reencode DSA/ECDSA signatures and compare with the original received
7602 signature. Return an error if there is a mismatch.
7604 This will reject various cases including garbage after signature
7605 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7606 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7607 (negative or with leading zeroes).
7609 Further analysis was conducted and fixes were developed by Stephen Henson
7610 of the OpenSSL core team.
7616 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7618 * Session Ticket Memory Leak.
7620 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7621 integrity of that ticket is first verified. In the event of a session
7622 ticket integrity check failing, OpenSSL will fail to free memory
7623 causing a memory leak. By sending a large number of invalid session
7624 tickets an attacker could exploit this issue in a Denial Of Service
7630 * Build option no-ssl3 is incomplete.
7632 When OpenSSL is configured with "no-ssl3" as a build option, servers
7633 could accept and complete a SSL 3.0 handshake, and clients could be
7634 configured to send them.
7637 *Akamai and the OpenSSL team*
7639 * Add support for TLS_FALLBACK_SCSV.
7640 Client applications doing fallback retries should call
7641 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7644 *Adam Langley, Bodo Moeller*
7646 * Add additional DigestInfo checks.
7648 Reencode DigestInto in DER and check against the original when
7649 verifying RSA signature: this will reject any improperly encoded
7650 DigestInfo structures.
7652 Note: this is a precautionary measure and no attacks are currently known.
7656 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7658 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7659 to a denial of service attack. A malicious server can crash the client
7660 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7661 ciphersuite and sending carefully crafted handshake messages.
7663 Thanks to Felix Gröbert (Google) for discovering and researching this
7669 * By sending carefully crafted DTLS packets an attacker could cause openssl
7670 to leak memory. This can be exploited through a Denial of Service attack.
7671 Thanks to Adam Langley for discovering and researching this issue.
7676 * An attacker can force openssl to consume large amounts of memory whilst
7677 processing DTLS handshake messages. This can be exploited through a
7678 Denial of Service attack.
7679 Thanks to Adam Langley for discovering and researching this issue.
7684 * An attacker can force an error condition which causes openssl to crash
7685 whilst processing DTLS packets due to memory being freed twice. This
7686 can be exploited through a Denial of Service attack.
7687 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7693 * If a multithreaded client connects to a malicious server using a resumed
7694 session and the server sends an ec point format extension it could write
7695 up to 255 bytes to freed memory.
7697 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7703 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7704 X509_name_oneline, X509_name_print_ex et al. to leak some information
7705 from the stack. Applications may be affected if they echo pretty printing
7706 output to the attacker.
7708 Thanks to Ivan Fratric (Google) for discovering this issue.
7711 *Emilia Käsper, and Steve Henson*
7713 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7714 for corner cases. (Certain input points at infinity could lead to
7715 bogus results, with non-infinity inputs mapped to infinity too.)
7719 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7721 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7722 handshake can force the use of weak keying material in OpenSSL
7723 SSL/TLS clients and servers.
7725 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7726 researching this issue. ([CVE-2014-0224])
7728 *KIKUCHI Masashi, Steve Henson*
7730 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7731 OpenSSL DTLS client the code can be made to recurse eventually crashing
7734 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7737 *Imre Rad, Steve Henson*
7739 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7740 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7741 client or server. This is potentially exploitable to run arbitrary
7742 code on a vulnerable client or server.
7744 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7746 *Jüri Aedla, Steve Henson*
7748 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7749 are subject to a denial of service attack.
7751 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7752 this issue. ([CVE-2014-3470])
7754 *Felix Gröbert, Ivan Fratric, Steve Henson*
7756 * Harmonize version and its documentation. -f flag is used to display
7759 *mancha <mancha1@zoho.com>*
7761 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7762 in i2d_ECPrivateKey.
7764 *mancha <mancha1@zoho.com>*
7766 * Fix some double frees. These are not thought to be exploitable.
7768 *mancha <mancha1@zoho.com>*
7770 * Fix for the attack described in the paper "Recovering OpenSSL
7771 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7772 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7773 <http://eprint.iacr.org/2014/140>
7775 Thanks to Yuval Yarom and Naomi Benger for discovering this
7776 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7778 *Yuval Yarom and Naomi Benger*
7780 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7782 * Keep original DTLS digest and encryption contexts in retransmission
7783 structures so we can use the previous session parameters if they need
7784 to be resent. ([CVE-2013-6450])
7788 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7789 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7790 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7791 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7792 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7793 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7795 *Rob Stradling, Adam Langley*
7797 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7799 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7801 This addresses the flaw in CBC record processing discovered by
7802 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7803 at: <http://www.isg.rhul.ac.uk/tls/>
7805 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7806 Security Group at Royal Holloway, University of London
7807 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7808 Emilia Käsper for the initial patch.
7811 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7813 * Return an error when checking OCSP signatures when key is NULL.
7814 This fixes a DoS attack. ([CVE-2013-0166])
7818 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7819 the right response is stapled. Also change SSL_get_certificate()
7820 so it returns the certificate actually sent.
7821 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7822 (This is a backport)
7824 *Rob Stradling <rob.stradling@comodo.com>*
7826 * Fix possible deadlock when decoding public keys.
7830 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7832 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7835 * Sanity check record length before skipping explicit IV in DTLS
7838 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7839 fuzzing as a service testing platform.
7844 * Initialise tkeylen properly when encrypting CMS messages.
7845 Thanks to Solar Designer of Openwall for reporting this issue.
7849 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7851 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7852 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7853 in CRYPTO_realloc_clean.
7855 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7856 issue and to Adam Langley <agl@chromium.org> for fixing it.
7859 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7861 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7863 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7864 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7865 content decryption and always return the same error. Note: this attack
7866 needs on average 2^20 messages so it only affects automated senders. The
7867 old behaviour can be re-enabled in the CMS code by setting the
7868 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7869 an MMA defence is not necessary.
7870 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7871 this issue. ([CVE-2012-0884])
7875 * Fix CVE-2011-4619: make sure we really are receiving a
7876 client hello before rejecting multiple SGC restarts. Thanks to
7877 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7881 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7883 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7884 Thanks to Antonio Martin, Enterprise Secure Access Research and
7885 Development, Cisco Systems, Inc. for discovering this bug and
7886 preparing a fix. ([CVE-2012-0050])
7890 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7892 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7893 of the Vaudenay padding oracle attack on CBC mode encryption
7894 which enables an efficient plaintext recovery attack against
7895 the OpenSSL implementation of DTLS. Their attack exploits timing
7896 differences arising during decryption processing. A research
7897 paper describing this attack can be found at:
7898 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7899 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7900 Security Group at Royal Holloway, University of London
7901 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7902 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7903 for preparing the fix. ([CVE-2011-4108])
7905 *Robin Seggelmann, Michael Tuexen*
7907 * Clear bytes used for block padding of SSL 3.0 records.
7910 *Adam Langley (Google)*
7912 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7913 Kadianakis <desnacked@gmail.com> for discovering this issue and
7914 Adam Langley for preparing the fix. ([CVE-2011-4619])
7916 *Adam Langley (Google)*
7918 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7920 *Andrey Kulikov <amdeich@gmail.com>*
7922 * Prevent malformed RFC3779 data triggering an assertion failure.
7923 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7924 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7926 *Rob Austein <sra@hactrn.net>*
7928 * Improved PRNG seeding for VOS.
7930 *Paul Green <Paul.Green@stratus.com>*
7932 * Fix ssl_ciph.c set-up race.
7934 *Adam Langley (Google)*
7936 * Fix spurious failures in ecdsatest.c.
7938 *Emilia Käsper (Google)*
7940 * Fix the BIO_f_buffer() implementation (which was mixing different
7941 interpretations of the `..._len` fields).
7943 *Adam Langley (Google)*
7945 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7946 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7947 threads won't reuse the same blinding coefficients.
7949 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7950 lock to call BN_BLINDING_invert_ex, and avoids one use of
7951 BN_BLINDING_update for each BN_BLINDING structure (previously,
7952 the last update always remained unused).
7954 *Emilia Käsper (Google)*
7956 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7958 *Bob Buckholz (Google)*
7960 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7962 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7963 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7965 *Kaspar Brand <ossl@velox.ch>*
7967 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7968 for multi-threaded use of ECDH. ([CVE-2011-3210])
7970 *Adam Langley (Google)*
7972 * Fix x509_name_ex_d2i memory leak on bad inputs.
7976 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7977 signature public key algorithm by using OID xref utilities instead.
7978 Before this you could only use some ECC ciphersuites with SHA1 only.
7982 * Add protection against ECDSA timing attacks as mentioned in the paper
7983 by Billy Bob Brumley and Nicola Tuveri, see:
7984 <http://eprint.iacr.org/2011/232.pdf>
7986 *Billy Bob Brumley and Nicola Tuveri*
7988 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7990 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7992 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7994 * Fix bug in string printing code: if *any* escaping is enabled we must
7995 escape the escape character (backslash) or the resulting string is
8000 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
8002 * Disable code workaround for ancient and obsolete Netscape browsers
8003 and servers: an attacker can use it in a ciphersuite downgrade attack.
8004 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
8008 * Fixed J-PAKE implementation error, originally discovered by
8009 Sebastien Martini, further info and confirmation from Stefan
8010 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
8014 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
8016 * Fix extension code to avoid race conditions which can result in a buffer
8017 overrun vulnerability: resumed sessions must not be modified as they can
8018 be shared by multiple threads. CVE-2010-3864
8022 * Fix WIN32 build system to correctly link an ENGINE directory into
8027 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
8029 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8032 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8034 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8036 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8037 context. The operation can be customised via the ctrl mechanism in
8038 case ENGINEs want to include additional functionality.
8042 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8046 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8047 output hashes compatible with older versions of OpenSSL.
8049 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8051 * Fix compression algorithm handling: if resuming a session use the
8052 compression algorithm of the resumed session instead of determining
8053 it from client hello again. Don't allow server to change algorithm.
8057 * Add load_crls() function to commands tidying load_certs() too. Add option
8058 to verify utility to allow additional CRLs to be included.
8062 * Update OCSP request code to permit adding custom headers to the request:
8063 some responders need this.
8067 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8070 *Julia Lawall <julia@diku.dk>*
8072 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8073 needlessly dereferenced structures, used obsolete functions and
8074 didn't handle all updated verify codes correctly.
8078 * Disable MD2 in the default configuration.
8082 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8083 indicate the initial BIO being pushed or popped. This makes it possible
8084 to determine whether the BIO is the one explicitly called or as a result
8085 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8086 it handles reference counts correctly and doesn't zero out the I/O bio
8087 when it is not being explicitly popped. WARNING: applications which
8088 included workarounds for the old buggy behaviour will need to be modified
8089 or they could free up already freed BIOs.
8093 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8094 renaming to all platforms (within the 0.9.8 branch, this was
8095 done conditionally on Netware platforms to avoid a name clash).
8097 *Guenter <lists@gknw.net>*
8099 * Add ECDHE and PSK support to DTLS.
8101 *Michael Tuexen <tuexen@fh-muenster.de>*
8103 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8108 * Add "missing" function EVP_MD_flags() (without this the only way to
8109 retrieve a digest flags is by accessing the structure directly. Update
8110 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8111 or cipher is registered as in the "from" argument. Print out all
8112 registered digests in the dgst usage message instead of manually
8113 attempting to work them out.
8117 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8118 this allows the use of compression and extensions. Change default cipher
8119 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8120 by default unless an application cipher string requests it.
8124 * Alter match criteria in PKCS12_parse(). It used to try to use local
8125 key ids to find matching certificates and keys but some PKCS#12 files
8126 don't follow the (somewhat unwritten) rules and this strategy fails.
8127 Now just gather all certificates together and the first private key
8128 then look for the first certificate that matches the key.
8132 * Support use of registered digest and cipher names for dgst and cipher
8133 commands instead of having to add each one as a special case. So now
8140 openssl dgst -sha256 foo
8142 and this works for ENGINE based algorithms too.
8146 * Update Gost ENGINE to support parameter files.
8148 *Victor B. Wagner <vitus@cryptocom.ru>*
8150 * Support GeneralizedTime in ca utility.
8152 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8154 * Enhance the hash format used for certificate directory links. The new
8155 form uses the canonical encoding (meaning equivalent names will work
8156 even if they aren't identical) and uses SHA1 instead of MD5. This form
8157 is incompatible with the older format and as a result c_rehash should
8158 be used to rebuild symbolic links.
8162 * Make PKCS#8 the default write format for private keys, replacing the
8163 traditional format. This form is standardised, more secure and doesn't
8164 include an implicit MD5 dependency.
8168 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8169 committed to OpenSSL should pass this lot as a minimum.
8173 * Add session ticket override functionality for use by EAP-FAST.
8175 *Jouni Malinen <j@w1.fi>*
8177 * Modify HMAC functions to return a value. Since these can be implemented
8178 in an ENGINE errors can occur.
8182 * Type-checked OBJ_bsearch_ex.
8186 * Type-checked OBJ_bsearch. Also some constification necessitated
8187 by type-checking. Still to come: TXT_DB, bsearch(?),
8188 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8193 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8194 seconds to a tm structure directly, instead of going through OS
8195 specific date routines. This avoids any issues with OS routines such
8196 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8197 and X509_time_adj_ex() to cover the extended range. The existing
8198 X509_time_adj() is still usable and will no longer have any date issues.
8202 * Delta CRL support. New use deltas option which will attempt to locate
8203 and search any appropriate delta CRLs available.
8205 This work was sponsored by Google.
8209 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8210 code and add additional score elements. Validate alternate CRL paths
8211 as part of the CRL checking and indicate a new error "CRL path validation
8212 error" in this case. Applications wanting additional details can use
8213 the verify callback and check the new "parent" field. If this is not
8214 NULL CRL path validation is taking place. Existing applications won't
8215 see this because it requires extended CRL support which is off by
8218 This work was sponsored by Google.
8222 * Support for freshest CRL extension.
8224 This work was sponsored by Google.
8228 * Initial indirect CRL support. Currently only supported in the CRLs
8229 passed directly and not via lookup. Process certificate issuer
8230 CRL entry extension and lookup CRL entries by bother issuer name
8231 and serial number. Check and process CRL issuer entry in IDP extension.
8233 This work was sponsored by Google.
8237 * Add support for distinct certificate and CRL paths. The CRL issuer
8238 certificate is validated separately in this case. Only enabled if
8239 an extended CRL support flag is set: this flag will enable additional
8240 CRL functionality in future.
8242 This work was sponsored by Google.
8246 * Add support for policy mappings extension.
8248 This work was sponsored by Google.
8252 * Fixes to pathlength constraint, self issued certificate handling,
8253 policy processing to align with RFC3280 and PKITS tests.
8255 This work was sponsored by Google.
8259 * Support for name constraints certificate extension. DN, email, DNS
8260 and URI types are currently supported.
8262 This work was sponsored by Google.
8266 * To cater for systems that provide a pointer-based thread ID rather
8267 than numeric, deprecate the current numeric thread ID mechanism and
8268 replace it with a structure and associated callback type. This
8269 mechanism allows a numeric "hash" to be extracted from a thread ID in
8270 either case, and on platforms where pointers are larger than 'long',
8271 mixing is done to help ensure the numeric 'hash' is usable even if it
8272 can't be guaranteed unique. The default mechanism is to use "&errno"
8273 as a pointer-based thread ID to distinguish between threads.
8275 Applications that want to provide their own thread IDs should now use
8276 CRYPTO_THREADID_set_callback() to register a callback that will call
8277 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8279 Note that ERR_remove_state() is now deprecated, because it is tied
8280 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8281 to free the current thread's error state should be replaced by
8282 ERR_remove_thread_state(NULL).
8284 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8285 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8286 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8287 application was previously providing a numeric thread callback that
8288 was inappropriate for distinguishing threads, then uniqueness might
8289 have been obtained with &errno that happened immediately in the
8290 intermediate development versions of OpenSSL; this is no longer the
8291 case, the numeric thread callback will now override the automatic use
8294 *Geoff Thorpe, with help from Bodo Moeller*
8296 * Initial support for different CRL issuing certificates. This covers a
8297 simple case where the self issued certificates in the chain exist and
8298 the real CRL issuer is higher in the existing chain.
8300 This work was sponsored by Google.
8304 * Removed effectively defunct crypto/store from the build.
8308 * Revamp of STACK to provide stronger type-checking. Still to come:
8309 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8310 ASN1_STRING, CONF_VALUE.
8314 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8315 RAM on SSL connections. This option can save about 34k per idle SSL.
8319 * Revamp of LHASH to provide stronger type-checking. Still to come:
8320 STACK, TXT_DB, bsearch, qsort.
8324 * Initial support for Cryptographic Message Syntax (aka CMS) based
8325 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8326 support for data, signedData, compressedData, digestedData and
8327 encryptedData, envelopedData types included. Scripts to check against
8328 RFC4134 examples draft and interop and consistency checks of many
8329 content types and variants.
8333 * Add options to enc utility to support use of zlib compression BIO.
8337 * Extend mk1mf to support importing of options and assembly language
8338 files from Configure script, currently only included in VC-WIN32.
8339 The assembly language rules can now optionally generate the source
8340 files from the associated perl scripts.
8344 * Implement remaining functionality needed to support GOST ciphersuites.
8345 Interop testing has been performed using CryptoPro implementations.
8347 *Victor B. Wagner <vitus@cryptocom.ru>*
8349 * s390x assembler pack.
8353 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8358 * Implement Opaque PRF Input TLS extension as specified in
8359 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8360 official specification yet and no extension type assignment by
8361 IANA exists, this extension (for now) will have to be explicitly
8362 enabled when building OpenSSL by providing the extension number
8363 to use. For example, specify an option
8365 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8367 to the "config" or "Configure" script to enable the extension,
8368 assuming extension number 0x9527 (which is a completely arbitrary
8369 and unofficial assignment based on the MD5 hash of the Internet
8370 Draft). Note that by doing so, you potentially lose
8371 interoperability with other TLS implementations since these might
8372 be using the same extension number for other purposes.
8374 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8375 opaque PRF input value to use in the handshake. This will create
8376 an internal copy of the length-'len' string at 'src', and will
8377 return non-zero for success.
8379 To get more control and flexibility, provide a callback function
8382 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8383 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8387 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8390 Callback function 'cb' will be called in handshakes, and is
8391 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8392 Argument 'arg' is for application purposes (the value as given to
8393 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8394 be provided to the callback function). The callback function
8395 has to return non-zero to report success: usually 1 to use opaque
8396 PRF input just if possible, or 2 to enforce use of the opaque PRF
8397 input. In the latter case, the library will abort the handshake
8398 if opaque PRF input is not successfully negotiated.
8400 Arguments 'peerinput' and 'len' given to the callback function
8401 will always be NULL and 0 in the case of a client. A server will
8402 see the client's opaque PRF input through these variables if
8403 available (NULL and 0 otherwise). Note that if the server
8404 provides an opaque PRF input, the length must be the same as the
8405 length of the client's opaque PRF input.
8407 Note that the callback function will only be called when creating
8408 a new session (session resumption can resume whatever was
8409 previously negotiated), and will not be called in SSL 2.0
8410 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8411 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8412 for applications that need to enforce opaque PRF input.
8416 * Update ssl code to support digests other than SHA1+MD5 for handshake
8419 *Victor B. Wagner <vitus@cryptocom.ru>*
8421 * Add RFC4507 support to OpenSSL. This includes the corrections in
8422 RFC4507bis. The encrypted ticket format is an encrypted encoded
8423 SSL_SESSION structure, that way new session features are automatically
8426 If a client application caches session in an SSL_SESSION structure
8427 support is transparent because tickets are now stored in the encoded
8430 The SSL_CTX structure automatically generates keys for ticket
8431 protection in servers so again support should be possible
8432 with no application modification.
8434 If a client or server wishes to disable RFC4507 support then the option
8435 SSL_OP_NO_TICKET can be set.
8437 Add a TLS extension debugging callback to allow the contents of any client
8438 or server extensions to be examined.
8440 This work was sponsored by Google.
8444 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8445 OpenSSL should now compile cleanly on gcc 4.2
8447 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8449 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8450 support including streaming MAC support: this is required for GOST
8451 ciphersuite support.
8453 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8455 * Add option -stream to use PKCS#7 streaming in smime utility. New
8456 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8457 to output in BER and PEM format.
8461 * Experimental support for use of HMAC via EVP_PKEY interface. This
8462 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8463 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8464 ENGINE support for HMAC keys which are unextractable. New -mac and
8465 -macopt options to dgst utility.
8469 * New option -sigopt to dgst utility. Update dgst to use
8470 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8471 alternative signing parameters such as X9.31 or PSS in the dgst
8476 * Change ssl_cipher_apply_rule(), the internal function that does
8477 the work each time a ciphersuite string requests enabling
8478 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8479 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8480 the order of disabled ciphersuites such that those ciphersuites
8481 that most recently went from enabled to disabled not only stay
8482 in order with respect to each other, but also have higher priority
8483 than other disabled ciphersuites the next time ciphersuites are
8486 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8487 the same ciphersuites as with "HIGH" alone, but in a specific
8488 order where the PSK ciphersuites come first (since they are the
8489 most recently disabled ciphersuites when "HIGH" is parsed).
8491 Also, change ssl_create_cipher_list() (using this new
8492 functionality) such that between otherwise identical
8493 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8498 * Change ssl_create_cipher_list() so that it automatically
8499 arranges the ciphersuites in reasonable order before starting
8500 to process the rule string. Thus, the definition for "DEFAULT"
8501 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8502 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8503 This makes it much easier to arrive at a reasonable default order
8504 in applications for which anonymous ciphers are OK (meaning
8505 that you can't actually use DEFAULT).
8507 *Bodo Moeller; suggested by Victor Duchovni*
8509 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8510 processing) into multiple integers instead of setting
8511 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8512 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8513 (These masks as well as the individual bit definitions are hidden
8514 away into the non-exported interface ssl/ssl_locl.h, so this
8515 change to the definition of the SSL_CIPHER structure shouldn't
8516 affect applications.) This give us more bits for each of these
8517 categories, so there is no longer a need to coagulate AES128 and
8518 AES256 into a single algorithm bit, and to coagulate Camellia128
8519 and Camellia256 into a single algorithm bit, which has led to all
8522 Thus, among other things, the kludge introduced in 0.9.7m and
8523 0.9.8e for masking out AES256 independently of AES128 or masking
8524 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8526 With the change, we also introduce new ciphersuite aliases that
8527 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8532 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8533 Use the leftmost N bytes of the signature input if the input is
8534 larger than the prime q (with N being the size in bytes of q).
8538 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8539 it yet and it is largely untested.
8543 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8547 * Initial incomplete changes to avoid need for function casts in OpenSSL
8548 some compilers (gcc 4.2 and later) reject their use. Safestack is
8549 reimplemented. Update ASN1 to avoid use of legacy functions.
8553 * Win32/64 targets are linked with Winsock2.
8557 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8558 to external functions. This can be used to increase CRL handling
8559 efficiency especially when CRLs are very large by (for example) storing
8560 the CRL revoked certificates in a database.
8564 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8565 new CRLs added to a directory can be used. New command line option
8566 -verify_return_error to s_client and s_server. This causes real errors
8567 to be returned by the verify callback instead of carrying on no matter
8568 what. This reflects the way a "real world" verify callback would behave.
8572 * GOST engine, supporting several GOST algorithms and public key formats.
8573 Kindly donated by Cryptocom.
8577 * Partial support for Issuing Distribution Point CRL extension. CRLs
8578 partitioned by DP are handled but no indirect CRL or reason partitioning
8579 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8580 selected via a scoring technique which handles IDP and AKID in CRLs.
8584 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8585 will ultimately be used for all verify operations: this will remove the
8586 X509_STORE dependency on certificate verification and allow alternative
8587 lookup methods. X509_STORE based implementations of these two callbacks.
8591 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8592 Modify get_crl() to find a valid (unexpired) CRL if possible.
8596 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8597 this would be called X509_CRL_cmp() but that name is already used by
8598 a function that just compares CRL issuer names. Cache several CRL
8599 extensions in X509_CRL structure and cache CRLDP in X509.
8603 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8604 this maps equivalent X509_NAME structures into a consistent structure.
8605 Name comparison can then be performed rapidly using memcmp().
8609 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8614 * Allow digests to supply their own micalg string for S/MIME type using
8615 the ctrl EVP_MD_CTRL_MICALG.
8619 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8620 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8621 ctrl. It can then customise the structure before and/or after signing
8626 * New function OBJ_add_sigid() to allow application defined signature OIDs
8627 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8628 to free up any added signature OIDs.
8632 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8633 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8634 digest and cipher tables. New options added to openssl utility:
8635 list-message-digest-algorithms and list-cipher-algorithms.
8639 * Change the array representation of binary polynomials: the list
8640 of degrees of non-zero coefficients is now terminated with -1.
8641 Previously it was terminated with 0, which was also part of the
8642 value; thus, the array representation was not applicable to
8643 polynomials where t^0 has coefficient zero. This change makes
8644 the array representation useful in a more general context.
8648 * Various modifications and fixes to SSL/TLS cipher string
8649 handling. For ECC, the code now distinguishes between fixed ECDH
8650 with RSA certificates on the one hand and with ECDSA certificates
8651 on the other hand, since these are separate ciphersuites. The
8652 unused code for Fortezza ciphersuites has been removed.
8654 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8655 (not "ECDHE"). For consistency with the code for DH
8656 certificates, use of ECDH certificates is now considered ECDH
8657 authentication, not RSA or ECDSA authentication (the latter is
8658 merely the CA's signing algorithm and not actively used in the
8661 The temporary ciphersuite alias "ECCdraft" is no longer
8662 available, and ECC ciphersuites are no longer excluded from "ALL"
8663 and "DEFAULT". The following aliases now exist for RFC 4492
8664 ciphersuites, most of these by analogy with the DH case:
8666 kECDHr - ECDH cert, signed with RSA
8667 kECDHe - ECDH cert, signed with ECDSA
8668 kECDH - ECDH cert (signed with either RSA or ECDSA)
8669 kEECDH - ephemeral ECDH
8670 ECDH - ECDH cert or ephemeral ECDH
8676 AECDH - anonymous ECDH
8677 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8681 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8682 Use correct micalg parameters depending on digest(s) in signed message.
8686 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8687 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8691 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8692 an engine to register a method. Add ENGINE lookups for methods and
8693 functional reference processing.
8697 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8698 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8703 * New -resign option to smime utility. This adds one or more signers
8704 to an existing PKCS#7 signedData structure. Also -md option to use an
8705 alternative message digest algorithm for signing.
8709 * Tidy up PKCS#7 routines and add new functions to make it easier to
8710 create PKCS7 structures containing multiple signers. Update smime
8711 application to support multiple signers.
8715 * New -macalg option to pkcs12 utility to allow setting of an alternative
8720 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8721 Reorganize PBE internals to lookup from a static table using NIDs,
8722 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8723 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8724 PRF which will be automatically used with PBES2.
8728 * Replace the algorithm specific calls to generate keys in "req" with the
8733 * Update PKCS#7 enveloped data routines to use new API. This is now
8734 supported by any public key method supporting the encrypt operation. A
8735 ctrl is added to allow the public key algorithm to examine or modify
8736 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8741 * Add a ctrl to asn1 method to allow a public key algorithm to express
8742 a default digest type to use. In most cases this will be SHA1 but some
8743 algorithms (such as GOST) need to specify an alternative digest. The
8744 return value indicates how strong the preference is 1 means optional and
8745 2 is mandatory (that is it is the only supported type). Modify
8746 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8747 use the default md. Update openssl utilities to use the default digest
8748 type for signing if it is not explicitly indicated.
8752 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8753 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8754 signing method from the key type. This effectively removes the link
8755 between digests and public key types.
8759 * Add an OID cross reference table and utility functions. Its purpose is to
8760 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8761 rsaEncryption. This will allow some of the algorithm specific hackery
8762 needed to use the correct OID to be removed.
8766 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8767 structures for PKCS7_sign(). They are now set up by the relevant public
8772 * Add provisional EC pkey method with support for ECDSA and ECDH.
8776 * Add support for key derivation (agreement) in the API, DH method and
8781 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8782 public and private key formats. As a side effect these add additional
8783 command line functionality not previously available: DSA signatures can be
8784 generated and verified using pkeyutl and DH key support and generation in
8791 *Oliver Tappe <zooey@hirschkaefer.de>*
8793 * New make target "install_html_docs" installs HTML renditions of the
8796 *Oliver Tappe <zooey@hirschkaefer.de>*
8798 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8799 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8800 support key and parameter generation and add initial key generation
8801 functionality for RSA.
8805 * Add functions for main EVP_PKEY_method operations. The undocumented
8806 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8807 `EVP_PKEY_{encrypt,decrypt}_old`.
8811 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8812 key API, doesn't do much yet.
8816 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8817 public key algorithms. New option to openssl utility:
8818 "list-public-key-algorithms" to print out info.
8822 * Implement the Supported Elliptic Curves Extension for
8823 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8827 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8828 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8832 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8833 utilities such as rsa, dsa, dsaparam etc except they process any key
8838 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8839 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8840 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8845 * Initial support for pluggable public key ASN1.
8846 De-spaghettify the public key ASN1 handling. Move public and private
8847 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8848 algorithm specific handling to a single module within the relevant
8849 algorithm directory. Add functions to allow (near) opaque processing
8850 of public and private key structures.
8854 * Implement the Supported Point Formats Extension for
8855 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8859 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8860 for the psk identity [hint] and the psk callback functions to the
8861 SSL_SESSION, SSL and SSL_CTX structure.
8864 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8868 SSL_CTX_use_psk_identity_hint
8869 SSL_get_psk_identity_hint
8870 SSL_get_psk_identity
8871 SSL_use_psk_identity_hint
8873 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8875 * Add RFC 3161 compliant time stamp request creation, response generation
8876 and response verification functionality.
8878 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8880 * Add initial support for TLS extensions, specifically for the server_name
8881 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8882 have new members for a host name. The SSL data structure has an
8883 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8884 stored in that context to allow for session resumption, even after the
8885 SSL has been switched to a new SSL_CTX in reaction to a client's
8886 server_name extension.
8888 New functions (subject to change):
8890 SSL_get_servername()
8891 SSL_get_servername_type()
8894 New CTRL codes and macros (subject to change):
8896 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8897 - SSL_CTX_set_tlsext_servername_callback()
8898 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8899 - SSL_CTX_set_tlsext_servername_arg()
8900 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8902 openssl s_client has a new '-servername ...' option.
8904 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8905 '-key2 ...', '-servername_fatal' (subject to change). This allows
8906 testing the HostName extension for a specific single host name ('-cert'
8907 and '-key' remain fallbacks for handshakes without HostName
8908 negotiation). If the unrecognized_name alert has to be sent, this by
8909 default is a warning; it becomes fatal with the '-servername_fatal'
8912 *Peter Sylvester, Remy Allais, Christophe Renou*
8914 * Whirlpool hash implementation is added.
8918 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8919 bn(64,32). Because of instruction set limitations it doesn't have
8920 any negative impact on performance. This was done mostly in order
8921 to make it possible to share assembler modules, such as bn_mul_mont
8922 implementations, between 32- and 64-bit builds without hassle.
8926 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8927 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8932 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8933 dedicated Montgomery multiplication procedure, is introduced.
8934 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8935 "64-bit" performance on certain 32-bit targets.
8939 * New option SSL_OP_NO_COMP to disable use of compression selectively
8940 in SSL structures. New SSL ctrl to set maximum send fragment size.
8941 Save memory by setting the I/O buffer sizes dynamically instead of
8942 using the maximum available value.
8946 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8947 in addition to the text details.
8951 * Very, very preliminary EXPERIMENTAL support for printing of general
8952 ASN1 structures. This currently produces rather ugly output and doesn't
8953 handle several customised structures at all.
8957 * Integrated support for PVK file format and some related formats such
8958 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8959 these in the 'rsa' and 'dsa' utilities.
8963 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8967 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8968 place for the (very old) "NETSCAPE" format certificates which are now
8969 handled using new ASN1 code equivalents.
8973 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8974 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8975 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8979 * Modify CRL distribution points extension code to print out previously
8980 unsupported fields. Enhance extension setting code to allow setting of
8985 * Add print and set support for Issuing Distribution Point CRL extension.
8989 * Change 'Configure' script to enable Camellia by default.
8996 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8998 * When rejecting SSL/TLS records due to an incorrect version number, never
8999 update s->server with a new major version number. As of
9000 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
9001 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
9002 the previous behavior could result in a read attempt at NULL when
9003 receiving specific incorrect SSL/TLS records once record payload
9004 protection is active. ([CVE-2010-0740])
9006 *Bodo Moeller, Adam Langley <agl@chromium.org>*
9008 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
9009 could be crashed if the relevant tables were not present (e.g. chrooted).
9011 *Tomas Hoger <thoger@redhat.com>*
9013 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
9015 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
9017 *Martin Olsson, Neel Mehta*
9019 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
9020 accommodate for stack sorting, always a write lock!).
9024 * On some versions of WIN32 Heap32Next is very slow. This can cause
9025 excessive delays in the RAND_poll(): over a minute. As a workaround
9026 include a time check in the inner Heap32Next loop too.
9030 * The code that handled flushing of data in SSL/TLS originally used the
9031 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9032 the problem outlined in PR#1949. The fix suggested there however can
9033 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9034 of Apache). So instead simplify the code to flush unconditionally.
9035 This should be fine since flushing with no data to flush is a no op.
9039 * Handle TLS versions 2.0 and later properly and correctly use the
9040 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9041 off ancient servers have a habit of sticking around for a while...
9045 * Modify compression code so it frees up structures without using the
9046 ex_data callbacks. This works around a problem where some applications
9047 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9048 restarting) then use compression (e.g. SSL with compression) later.
9049 This results in significant per-connection memory leaks and
9050 has caused some security issues including CVE-2008-1678 and
9055 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9056 change when encrypting or decrypting.
9060 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9061 connect and renegotiate with servers which do not support RI.
9062 Until RI is more widely deployed this option is enabled by default.
9066 * Add "missing" ssl ctrls to clear options and mode.
9070 * If client attempts to renegotiate and doesn't support RI respond with
9071 a no_renegotiation alert as required by RFC5746. Some renegotiating
9072 TLS clients will continue a connection gracefully when they receive
9073 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9074 waiting for a server hello which it will never receive. Now we treat a
9075 received no_renegotiation alert as a fatal error. This is because
9076 applications requesting a renegotiation might well expect it to succeed
9077 and would have no code in place to handle the server denying it so the
9078 only safe thing to do is to terminate the connection.
9082 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9083 peer supports secure renegotiation and 0 otherwise. Print out peer
9084 renegotiation support in s_client/s_server.
9088 * Replace the highly broken and deprecated SPKAC certification method with
9089 the updated NID creation version. This should correctly handle UTF8.
9093 * Implement RFC5746. Re-enable renegotiation but require the extension
9094 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9095 turns out to be a bad idea. It has been replaced by
9096 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9097 SSL_CTX_set_options(). This is really not recommended unless you
9098 know what you are doing.
9100 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9102 * Fixes to stateless session resumption handling. Use initial_ctx when
9103 issuing and attempting to decrypt tickets in case it has changed during
9104 servername handling. Use a non-zero length session ID when attempting
9105 stateless session resumption: this makes it possible to determine if
9106 a resumption has occurred immediately after receiving server hello
9107 (several places in OpenSSL subtly assume this) instead of later in
9112 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9113 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9114 fixes for a few places where the return code is not checked
9117 *Julia Lawall <julia@diku.dk>*
9119 * Add --strict-warnings option to Configure script to include devteam
9120 warnings in other configurations.
9124 * Add support for --libdir option and LIBDIR variable in makefiles. This
9125 makes it possible to install openssl libraries in locations which
9126 have names other than "lib", for example "/usr/lib64" which some
9129 *Steve Henson, based on patch from Jeremy Utley*
9131 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9132 X690 8.9.12 and can produce some misleading textual output of OIDs.
9134 *Steve Henson, reported by Dan Kaminsky*
9136 * Delete MD2 from algorithm tables. This follows the recommendation in
9137 several standards that it is not used in new applications due to
9138 several cryptographic weaknesses. For binary compatibility reasons
9139 the MD2 API is still compiled in by default.
9143 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9148 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9149 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9152 *Guenter <lists@gknw.net>*
9154 * Fix the server certificate chain building code to use X509_verify_cert(),
9155 it used to have an ad-hoc builder which was unable to cope with anything
9156 other than a simple chain.
9158 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9160 * Don't check self signed certificate signatures in X509_verify_cert()
9161 by default (a flag can override this): it just wastes time without
9162 adding any security. As a useful side effect self signed root CAs
9163 with non-FIPS digests are now usable in FIPS mode.
9167 * In dtls1_process_out_of_seq_message() the check if the current message
9168 is already buffered was missing. For every new message was memory
9169 allocated, allowing an attacker to perform an denial of service attack
9170 with sending out of seq handshake messages until there is no memory
9171 left. Additionally every future message was buffered, even if the
9172 sequence number made no sense and would be part of another handshake.
9173 So only messages with sequence numbers less than 10 in advance will be
9174 buffered. ([CVE-2009-1378])
9176 *Robin Seggelmann, discovered by Daniel Mentz*
9178 * Records are buffered if they arrive with a future epoch to be
9179 processed after finishing the corresponding handshake. There is
9180 currently no limitation to this buffer allowing an attacker to perform
9181 a DOS attack with sending records with future epochs until there is no
9182 memory left. This patch adds the pqueue_size() function to determine
9183 the size of a buffer and limits the record buffer to 100 entries.
9186 *Robin Seggelmann, discovered by Daniel Mentz*
9188 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9189 parent structure is freed. ([CVE-2009-1379])
9193 * Handle non-blocking I/O properly in SSL_shutdown() call.
9195 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9197 * Add `2.5.4.*` OIDs
9199 *Ilya O. <vrghost@gmail.com>*
9201 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9203 * Disable renegotiation completely - this fixes a severe security
9204 problem ([CVE-2009-3555]) at the cost of breaking all
9205 renegotiation. Renegotiation can be re-enabled by setting
9206 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9207 run-time. This is really not recommended unless you know what
9212 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9214 * Don't set val to NULL when freeing up structures, it is freed up by
9215 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9216 zeroing past the valid field. ([CVE-2009-0789])
9218 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9220 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9221 checked correctly. This would allow some invalid signed attributes to
9222 appear to verify correctly. ([CVE-2009-0591])
9224 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9226 * Reject UniversalString and BMPString types with invalid lengths. This
9227 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9228 a legal length. ([CVE-2009-0590])
9232 * Set S/MIME signing as the default purpose rather than setting it
9233 unconditionally. This allows applications to override it at the store
9238 * Permit restricted recursion of ASN1 strings. This is needed in practice
9239 to handle some structures.
9243 * Improve efficiency of mem_gets: don't search whole buffer each time
9246 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9248 * New -hex option for openssl rand.
9252 * Print out UTF8String and NumericString when parsing ASN1.
9256 * Support NumericString type for name components.
9260 * Allow CC in the environment to override the automatically chosen
9261 compiler. Note that nothing is done to ensure flags work with the
9266 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9268 * Properly check EVP_VerifyFinal() and similar return values
9271 *Ben Laurie, Bodo Moeller, Google Security Team*
9273 * Enable TLS extensions by default.
9277 * Allow the CHIL engine to be loaded, whether the application is
9278 multithreaded or not. (This does not release the developer from the
9279 obligation to set up the dynamic locking callbacks.)
9281 *Sander Temme <sander@temme.net>*
9283 * Use correct exit code if there is an error in dgst command.
9285 *Steve Henson; problem pointed out by Roland Dirlewanger*
9287 * Tweak Configure so that you need to say "experimental-jpake" to enable
9288 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9292 * Add experimental JPAKE support, including demo authentication in
9293 s_client and s_server.
9297 * Set the comparison function in v3_addr_canonize().
9299 *Rob Austein <sra@hactrn.net>*
9301 * Add support for XMPP STARTTLS in s_client.
9303 *Philip Paeps <philip@freebsd.org>*
9305 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9306 to ensure that even with this option, only ciphersuites in the
9307 server's preference list will be accepted. (Note that the option
9308 applies only when resuming a session, so the earlier behavior was
9309 just about the algorithm choice for symmetric cryptography.)
9313 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9315 * Fix NULL pointer dereference if a DTLS server received
9316 ChangeCipherSpec as first record ([CVE-2009-1386]).
9320 * Fix a state transition in s3_srvr.c and d1_srvr.c
9321 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9325 * The fix in 0.9.8c that supposedly got rid of unsafe
9326 double-checked locking was incomplete for RSA blinding,
9327 addressing just one layer of what turns out to have been
9328 doubly unsafe triple-checked locking.
9330 So now fix this for real by retiring the MONT_HELPER macro
9331 in crypto/rsa/rsa_eay.c.
9333 *Bodo Moeller; problem pointed out by Marius Schilder*
9335 * Various precautionary measures:
9337 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9339 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9340 (NB: This would require knowledge of the secret session ticket key
9341 to exploit, in which case you'd be SOL either way.)
9343 - Change bn_nist.c so that it will properly handle input BIGNUMs
9344 outside the expected range.
9346 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9349 *Neel Mehta, Bodo Moeller*
9351 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9352 the load fails. Useful for distros.
9354 *Ben Laurie and the FreeBSD team*
9356 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9360 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9364 * Expand ENGINE to support engine supplied SSL client certificate functions.
9366 This work was sponsored by Logica.
9370 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9371 keystores. Support for SSL/TLS client authentication too.
9372 Not compiled unless enable-capieng specified to Configure.
9374 This work was sponsored by Logica.
9378 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9379 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9380 attribute creation routines such as certificate requests and PKCS#12
9385 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9387 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9388 handshake which could lead to a client crash as found using the
9389 Codenomicon TLS test suite ([CVE-2008-1672])
9391 *Steve Henson, Mark Cox*
9393 * Fix double free in TLS server name extensions which could lead to
9394 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9398 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9400 Clear the error queue to ensure that error entries left from
9401 older function calls do not interfere with the correct operation.
9403 *Lutz Jaenicke, Erik de Castro Lopo*
9405 * Remove root CA certificates of commercial CAs:
9407 The OpenSSL project does not recommend any specific CA and does not
9408 have any policy with respect to including or excluding any CA.
9409 Therefore it does not make any sense to ship an arbitrary selection
9410 of root CA certificates with the OpenSSL software.
9414 * RSA OAEP patches to fix two separate invalid memory reads.
9415 The first one involves inputs when 'lzero' is greater than
9416 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9417 before the beginning of from). The second one involves inputs where
9418 the 'db' section contains nothing but zeroes (there is a one-byte
9419 invalid read after the end of 'db').
9421 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9423 * Partial backport from 0.9.9-dev:
9425 Introduce bn_mul_mont (dedicated Montgomery multiplication
9426 procedure) as a candidate for BIGNUM assembler implementation.
9427 While 0.9.9-dev uses assembler for various architectures, only
9428 x86_64 is available by default here in the 0.9.8 branch, and
9429 32-bit x86 is available through a compile-time setting.
9431 To try the 32-bit x86 assembler implementation, use Configure
9432 option "enable-montasm" (which exists only for this backport).
9434 As "enable-montasm" for 32-bit x86 disclaims code stability
9435 anyway, in this constellation we activate additional code
9436 backported from 0.9.9-dev for further performance improvements,
9437 namely BN_from_montgomery_word. (To enable this otherwise,
9438 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9440 *Andy Polyakov (backport partially by Bodo Moeller)*
9442 * Add TLS session ticket callback. This allows an application to set
9443 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9444 values. This is useful for key rollover for example where several key
9445 sets may exist with different names.
9449 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9450 This was broken until now in 0.9.8 releases, such that the only way
9451 a registered ENGINE could be used (assuming it initialises
9452 successfully on the host) was to explicitly set it as the default
9453 for the relevant algorithms. This is in contradiction with 0.9.7
9454 behaviour and the documentation. With this fix, when an ENGINE is
9455 registered into a given algorithm's table of implementations, the
9456 'uptodate' flag is reset so that auto-discovery will be used next
9457 time a new context for that algorithm attempts to select an
9460 *Ian Lister (tweaked by Geoff Thorpe)*
9462 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9463 implementation in the following ways:
9465 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9468 Lack of BER streaming support means one pass streaming processing is
9469 only supported if data is detached: setting the streaming flag is
9470 ignored for embedded content.
9472 CMS support is disabled by default and must be explicitly enabled
9473 with the enable-cms configuration option.
9477 * Update the GMP engine glue to do direct copies between BIGNUM and
9478 mpz_t when openssl and GMP use the same limb size. Otherwise the
9479 existing "conversion via a text string export" trick is still used.
9481 *Paul Sheer <paulsheer@gmail.com>*
9483 * Zlib compression BIO. This is a filter BIO which compressed and
9484 uncompresses any data passed through it.
9488 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9489 RFC3394 compatible AES key wrapping.
9493 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9494 sets string data without copying. X509_ALGOR_set0() and
9495 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9496 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9497 from an X509_ATTRIBUTE structure optionally checking it occurs only
9498 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9503 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9504 to get the expected BN_FLG_CONSTTIME behavior.
9506 *Bodo Moeller (Google)*
9510 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9511 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9512 - added some more tests to do_tests.pl
9513 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9514 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9515 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9516 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9517 - various changes to netware.pl to enable gcc-cross builds on Win32
9519 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9520 - various changes to fix missing prototype warnings
9521 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9522 - added AES, WHIRLPOOL and CPUID assembler code to build files
9523 - added missing AES assembler make rules to mk1mf.pl
9524 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9526 *Guenter Knauf <eflash@gmx.net>*
9528 * Implement certificate status request TLS extension defined in RFC3546.
9529 A client can set the appropriate parameters and receive the encoded
9530 OCSP response via a callback. A server can query the supplied parameters
9531 and set the encoded OCSP response in the callback. Add simplified examples
9532 to s_client and s_server.
9536 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9539 + Binary incompatibility of ssl_ctx_st structure
9540 + DTLS interoperation with non-compliant servers
9541 + Don't call get_session_cb() without proposed session
9542 + Fix ia64 assembler code
9544 *Andy Polyakov, Steve Henson*
9546 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9548 * DTLS Handshake overhaul. There were longstanding issues with
9549 OpenSSL DTLS implementation, which were making it impossible for
9550 RFC 4347 compliant client to communicate with OpenSSL server.
9551 Unfortunately just fixing these incompatibilities would "cut off"
9552 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9553 server keeps tolerating non RFC compliant syntax. The opposite is
9554 not true, 0.9.8f client can not communicate with earlier server.
9555 This update even addresses CVE-2007-4995.
9559 * Changes to avoid need for function casts in OpenSSL: some compilers
9560 (gcc 4.2 and later) reject their use.
9561 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9564 * Add RFC4507 support to OpenSSL. This includes the corrections in
9565 RFC4507bis. The encrypted ticket format is an encrypted encoded
9566 SSL_SESSION structure, that way new session features are automatically
9569 If a client application caches session in an SSL_SESSION structure
9570 support is transparent because tickets are now stored in the encoded
9573 The SSL_CTX structure automatically generates keys for ticket
9574 protection in servers so again support should be possible
9575 with no application modification.
9577 If a client or server wishes to disable RFC4507 support then the option
9578 SSL_OP_NO_TICKET can be set.
9580 Add a TLS extension debugging callback to allow the contents of any client
9581 or server extensions to be examined.
9583 This work was sponsored by Google.
9587 * Add initial support for TLS extensions, specifically for the server_name
9588 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9589 have new members for a host name. The SSL data structure has an
9590 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9591 stored in that context to allow for session resumption, even after the
9592 SSL has been switched to a new SSL_CTX in reaction to a client's
9593 server_name extension.
9595 New functions (subject to change):
9597 SSL_get_servername()
9598 SSL_get_servername_type()
9601 New CTRL codes and macros (subject to change):
9603 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9604 - SSL_CTX_set_tlsext_servername_callback()
9605 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9606 - SSL_CTX_set_tlsext_servername_arg()
9607 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9609 openssl s_client has a new '-servername ...' option.
9611 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9612 '-key2 ...', '-servername_fatal' (subject to change). This allows
9613 testing the HostName extension for a specific single host name ('-cert'
9614 and '-key' remain fallbacks for handshakes without HostName
9615 negotiation). If the unrecognized_name alert has to be sent, this by
9616 default is a warning; it becomes fatal with the '-servername_fatal'
9619 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9621 * Add AES and SSE2 assembly language support to VC++ build.
9625 * Mitigate attack on final subtraction in Montgomery reduction.
9629 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9630 (which previously caused an internal error).
9634 * Squeeze another 10% out of IGE mode when in != out.
9638 * AES IGE mode speedup.
9640 *Dean Gaudet (Google)*
9642 * Add the Korean symmetric 128-bit cipher SEED (see
9643 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9644 add SEED ciphersuites from RFC 4162:
9646 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9647 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9648 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9649 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9651 To minimize changes between patchlevels in the OpenSSL 0.9.8
9652 series, SEED remains excluded from compilation unless OpenSSL
9653 is configured with 'enable-seed'.
9655 *KISA, Bodo Moeller*
9657 * Mitigate branch prediction attacks, which can be practical if a
9658 single processor is shared, allowing a spy process to extract
9659 information. For detailed background information, see
9660 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9661 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9662 and Necessary Software Countermeasures"). The core of the change
9663 are new versions BN_div_no_branch() and
9664 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9665 respectively, which are slower, but avoid the security-relevant
9666 conditional branches. These are automatically called by BN_div()
9667 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9668 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9669 remove a conditional branch.
9671 BN_FLG_CONSTTIME is the new name for the previous
9672 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9673 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9674 in the exponent causes BN_mod_exp_mont() to use the alternative
9675 implementation in BN_mod_exp_mont_consttime().) The old name
9676 remains as a deprecated alias.
9678 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9679 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9680 constant-time implementations for more than just exponentiation.
9681 Here too the old name is kept as a deprecated alias.
9683 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9684 the BN_BLINDING structure gets an independent copy of the
9685 modulus. This means that the previous `BIGNUM *m` argument to
9686 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9687 essentially becomes `const BIGNUM *m`, although we can't actually
9688 change this in the header file before 0.9.9. It allows
9689 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9690 enable BN_FLG_CONSTTIME.
9692 *Matthew D Wood (Intel Corp)*
9694 * In the SSL/TLS server implementation, be strict about session ID
9695 context matching (which matters if an application uses a single
9696 external cache for different purposes). Previously,
9697 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9698 set. This did ensure strict client verification, but meant that,
9699 with applications using a single external cache for quite
9700 different requirements, clients could circumvent ciphersuite
9701 restrictions for a given session ID context by starting a session
9702 in a different context.
9706 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9707 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9708 authentication-only ciphersuites.
9712 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9713 not complete and could lead to a possible single byte overflow
9714 ([CVE-2007-5135]) [Ben Laurie]
9716 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9718 * Since AES128 and AES256 (and similarly Camellia128 and
9719 Camellia256) share a single mask bit in the logic of
9720 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9721 kludge to work properly if AES128 is available and AES256 isn't
9722 (or if Camellia128 is available and Camellia256 isn't).
9726 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9727 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9728 When a point or a seed is encoded in a BIT STRING, we need to
9729 prevent the removal of trailing zero bits to get the proper DER
9730 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9731 of a NamedBitList, for which trailing 0 bits need to be removed.)
9735 * Have SSL/TLS server implementation tolerate "mismatched" record
9736 protocol version while receiving ClientHello even if the
9737 ClientHello is fragmented. (The server can't insist on the
9738 particular protocol version it has chosen before the ServerHello
9739 message has informed the client about his choice.)
9743 * Add RFC 3779 support.
9745 *Rob Austein for ARIN, Ben Laurie*
9747 * Load error codes if they are not already present instead of using a
9748 static variable. This allows them to be cleanly unloaded and reloaded.
9749 Improve header file function name parsing.
9753 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9754 or CAPABILITY handshake as required by RFCs.
9758 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9760 * Introduce limits to prevent malicious keys being able to
9761 cause a denial of service. ([CVE-2006-2940])
9763 *Steve Henson, Bodo Moeller*
9765 * Fix ASN.1 parsing of certain invalid structures that can result
9766 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9768 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9769 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9771 * Fix SSL client code which could crash if connecting to a
9772 malicious SSLv2 server. ([CVE-2006-4343])
9774 *Tavis Ormandy and Will Drewry, Google Security Team*
9776 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9777 match only those. Before that, "AES256-SHA" would be interpreted
9778 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9779 the same strength classification in 0.9.7h) as we currently only
9780 have a single AES bit in the ciphersuite description bitmap.
9781 That change, however, also applied to ciphersuite strings such as
9782 "RC4-MD5" that intentionally matched multiple ciphersuites --
9783 namely, SSL 2.0 ciphersuites in addition to the more common ones
9784 from SSL 3.0/TLS 1.0.
9786 So we change the selection algorithm again: Naming an explicit
9787 ciphersuite selects this one ciphersuite, and any other similar
9788 ciphersuite (same bitmap) from *other* protocol versions.
9789 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9790 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9792 Since SSL 2.0 does not have any ciphersuites for which the
9793 128/256 bit distinction would be relevant, this works for now.
9794 The proper fix will be to use different bits for AES128 and
9795 AES256, which would have avoided the problems from the beginning;
9796 however, bits are scarce, so we can only do this in a new release
9797 (not just a patchlevel) when we can change the SSL_CIPHER
9798 definition to split the single 'unsigned long mask' bitmap into
9799 multiple values to extend the available space.
9803 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9805 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9806 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9808 * Add AES IGE and biIGE modes.
9812 * Change the Unix randomness entropy gathering to use poll() when
9813 possible instead of select(), since the latter has some
9814 undesirable limitations.
9816 *Darryl Miles via Richard Levitte and Bodo Moeller*
9818 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9819 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9820 cannot be implicitly activated as part of, e.g., the "AES" alias.
9821 However, please upgrade to OpenSSL 0.9.9[-dev] for
9822 non-experimental use of the ECC ciphersuites to get TLS extension
9823 support, which is required for curve and point format negotiation
9824 to avoid potential handshake problems.
9828 * Disable rogue ciphersuites:
9830 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9831 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9832 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9834 The latter two were purportedly from
9835 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9838 Also deactivate the remaining ciphersuites from
9839 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9840 unofficial, and the ID has long expired.
9844 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9845 dual-core machines) and other potential thread-safety issues.
9849 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9850 versions), which is now available for royalty-free use
9851 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9852 Also, add Camellia TLS ciphersuites from RFC 4132.
9854 To minimize changes between patchlevels in the OpenSSL 0.9.8
9855 series, Camellia remains excluded from compilation unless OpenSSL
9856 is configured with 'enable-camellia'.
9860 * Disable the padding bug check when compression is in use. The padding
9861 bug check assumes the first packet is of even length, this is not
9862 necessarily true if compression is enabled and can result in false
9863 positives causing handshake failure. The actual bug test is ancient
9864 code so it is hoped that implementations will either have fixed it by
9865 now or any which still have the bug do not support compression.
9869 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9871 * When applying a cipher rule check to see if string match is an explicit
9872 cipher suite and only match that one cipher suite if it is.
9876 * Link in manifests for VC++ if needed.
9878 *Austin Ziegler <halostatue@gmail.com>*
9880 * Update support for ECC-based TLS ciphersuites according to
9881 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9882 TLS extensions, which are supported starting with the 0.9.9
9883 branch, not in the OpenSSL 0.9.8 branch).
9887 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9888 opaque EVP_CIPHER_CTX handling.
9892 * Fixes and enhancements to zlib compression code. We now only use
9893 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9894 to conform with the standards mentioned here:
9895 <http://www.zlib.net/DLL_FAQ.txt>
9896 Static zlib linking now works on Windows and the new --with-zlib-include
9897 --with-zlib-lib options to Configure can be used to supply the location
9898 of the headers and library. Gracefully handle case where zlib library
9903 * Several fixes and enhancements to the OID generation code. The old code
9904 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9905 handle numbers larger than ULONG_MAX, truncated printing and had a
9906 non standard OBJ_obj2txt() behaviour.
9910 * Add support for building of engines under engine/ as shared libraries
9911 under VC++ build system.
9915 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9916 Hopefully, we will not see any false combination of paths any more.
9920 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9922 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9923 (part of SSL_OP_ALL). This option used to disable the
9924 countermeasure against man-in-the-middle protocol-version
9925 rollback in the SSL 2.0 server implementation, which is a bad
9926 idea. ([CVE-2005-2969])
9928 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9929 for Information Security, National Institute of Advanced Industrial
9930 Science and Technology [AIST], Japan)*
9932 * Add two function to clear and return the verify parameter flags.
9936 * Keep cipherlists sorted in the source instead of sorting them at
9937 runtime, thus removing the need for a lock.
9941 * Avoid some small subgroup attacks in Diffie-Hellman.
9943 *Nick Mathewson and Ben Laurie*
9945 * Add functions for well-known primes.
9949 * Extended Windows CE support.
9951 *Satoshi Nakamura and Andy Polyakov*
9953 * Initialize SSL_METHOD structures at compile time instead of during
9954 runtime, thus removing the need for a lock.
9958 * Make PKCS7_decrypt() work even if no certificate is supplied by
9959 attempting to decrypt each encrypted key in turn. Add support to
9964 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9966 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9969 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9973 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9974 key into the same file any more.
9978 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9982 * Add -utf8 command line and config file option to 'ca'.
9984 *Stefan <stf@udoma.org*
9986 * Removed the macro des_crypt(), as it seems to conflict with some
9987 libraries. Use DES_crypt().
9991 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9992 involves renaming the source and generated shared-libs for
9993 both. The engines will accept the corrected or legacy ids
9994 ('ncipher' and '4758_cca' respectively) when binding. NB,
9995 this only applies when building 'shared'.
9997 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9999 * Add attribute functions to EVP_PKEY structure. Modify
10000 PKCS12_create() to recognize a CSP name attribute and
10001 use it. Make -CSP option work again in pkcs12 utility.
10005 * Add new functionality to the bn blinding code:
10006 - automatic re-creation of the BN_BLINDING parameters after
10007 a fixed number of uses (currently 32)
10008 - add new function for parameter creation
10009 - introduce flags to control the update behaviour of the
10010 BN_BLINDING parameters
10011 - hide BN_BLINDING structure
10012 Add a second BN_BLINDING slot to the RSA structure to improve
10013 performance when a single RSA object is shared among several
10018 * Add support for DTLS.
10020 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
10022 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
10023 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
10027 * Remove buggy and incomplete DH cert support from
10028 ssl/ssl_rsa.c and ssl/s3_both.c
10032 * Use SHA-1 instead of MD5 as the default digest algorithm for
10033 the `apps/openssl` commands.
10037 * Compile clean with "-Wall -Wmissing-prototypes
10038 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10039 DEBUG_SAFESTACK must also be set.
10043 * Change ./Configure so that certain algorithms can be disabled by default.
10044 The new counterpiece to "no-xxx" is "enable-xxx".
10046 The patented RC5 and MDC2 algorithms will now be disabled unless
10047 "enable-rc5" and "enable-mdc2", respectively, are specified.
10049 (IDEA remains enabled despite being patented. This is because IDEA
10050 is frequently required for interoperability, and there is no license
10051 fee for non-commercial use. As before, "no-idea" can be used to
10052 avoid this algorithm.)
10056 * Add processing of proxy certificates (see RFC 3820). This work was
10057 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10058 EGEE (Enabling Grids for E-science in Europe).
10062 * RC4 performance overhaul on modern architectures/implementations, such
10063 as Intel P4, IA-64 and AMD64.
10067 * New utility extract-section.pl. This can be used specify an alternative
10068 section number in a pod file instead of having to treat each file as
10069 a separate case in Makefile. This can be done by adding two lines to the
10072 =for comment openssl_section:XXX
10074 The blank line is mandatory.
10078 * New arguments -certform, -keyform and -pass for s_client and s_server
10079 to allow alternative format key and certificate files and passphrase
10084 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10085 update associated structures and add various utility functions.
10087 Add new policy related verify parameters, include policy checking in
10088 standard verify code. Enhance 'smime' application with extra parameters
10089 to support policy checking and print out.
10093 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10094 Nehemiah processors. These extensions support AES encryption in hardware
10095 as well as RNG (though RNG support is currently disabled).
10097 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10099 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10103 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10105 *Andy Polyakov and a number of other people*
10107 * Improved PowerPC platform support. Most notably BIGNUM assembler
10108 implementation contributed by IBM.
10110 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10112 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10113 exponent rather than 'unsigned long'. There is a corresponding change to
10114 the new 'rsa_keygen' element of the RSA_METHOD structure.
10116 *Jelte Jansen, Geoff Thorpe*
10118 * Functionality for creating the initial serial number file is now
10119 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10121 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10122 number file to 1, which is bound to cause problems. To avoid
10123 the problems while respecting compatibility between different 0.9.7
10124 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10125 CA.pl for serial number initialization. With the new release 0.9.8,
10126 we can fix the problem directly in the 'ca' utility.)
10130 * Reduced header interdependencies by declaring more opaque objects in
10131 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10132 give fewer recursive includes, which could break lazy source code - so
10133 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10134 developers should define this symbol when building and using openssl to
10135 ensure they track the recommended behaviour, interfaces, [etc], but
10136 backwards-compatible behaviour prevails when this isn't defined.
10140 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10144 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10145 This will generate a random key of the appropriate length based on the
10146 cipher context. The EVP_CIPHER can provide its own random key generation
10147 routine to support keys of a specific form. This is used in the des and
10148 3des routines to generate a key of the correct parity. Update S/MIME
10149 code to use new functions and hence generate correct parity DES keys.
10150 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10151 valid (weak or incorrect parity).
10155 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10156 as looking them up. This is useful when the verified structure may contain
10157 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10158 present unless the new PKCS7_NO_CRL flag is asserted.
10162 * Extend ASN1 oid configuration module. It now additionally accepts the
10165 shortName = some long name, 1.2.3.4
10169 * Reimplemented the BN_CTX implementation. There is now no more static
10170 limitation on the number of variables it can handle nor the depth of the
10171 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10172 information can now expand as required, and rather than having a single
10173 static array of bignums, BN_CTX now uses a linked-list of such arrays
10174 allowing it to expand on demand whilst maintaining the usefulness of
10175 BN_CTX's "bundling".
10179 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10180 to allow all RSA operations to function using a single BN_CTX.
10184 * Preliminary support for certificate policy evaluation and checking. This
10185 is initially intended to pass the tests outlined in "Conformance Testing
10186 of Relying Party Client Certificate Path Processing Logic" v1.07.
10190 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10191 remained unused and not that useful. A variety of other little bignum
10192 tweaks and fixes have also been made continuing on from the audit (see
10197 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10198 associated ASN1, EVP and SSL functions and old ASN1 macros.
10202 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10203 and this should never fail. So the return value from the use of
10204 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10205 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10209 * BN_CTX_get() should return zero-valued bignums, providing the same
10210 initialised value as BN_new().
10212 *Geoff Thorpe, suggested by Ulf Möller*
10214 * Support for inhibitAnyPolicy certificate extension.
10218 * An audit of the BIGNUM code is underway, for which debugging code is
10219 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10220 is considered valid when processing BIGNUMs, and causes execution to
10221 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10222 further steps are taken to deliberately pollute unused data in BIGNUM
10223 structures to try and expose faulty code further on. For now, openssl will
10224 (in its default mode of operation) continue to tolerate the inconsistent
10225 forms that it has tolerated in the past, but authors and packagers should
10226 consider trying openssl and their own applications when compiled with
10227 these debugging symbols defined. It will help highlight potential bugs in
10228 their own code, and will improve the test coverage for OpenSSL itself. At
10229 some point, these tighter rules will become openssl's default to improve
10230 maintainability, though the assert()s and other overheads will remain only
10231 in debugging configurations. See bn.h for more details.
10233 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10235 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10236 that can only be obtained through BN_CTX_new() (which implicitly
10237 initialises it). The presence of this function only made it possible
10238 to overwrite an existing structure (and cause memory leaks).
10242 * Because of the callback-based approach for implementing LHASH as a
10243 template type, lh_insert() adds opaque objects to hash-tables and
10244 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10245 to clean up those corresponding objects before destroying the hash table
10246 (and losing the object pointers). So some over-zealous constifications in
10247 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10248 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10249 prototyped to have "const" restrictions on the object pointers they are
10250 given (and so aren't required to cast them away any more).
10254 * The tmdiff.h API was so ugly and minimal that our own timing utility
10255 (speed) prefers to use its own implementation. The two implementations
10256 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10257 its object type properly exposed (MS_TM) instead of casting to/from
10258 `char *`. This may still change yet if someone realises MS_TM and
10260 aren't necessarily the greatest nomenclatures - but this is what was used
10261 internally to the implementation so I've used that for now.
10265 * Ensure that deprecated functions do not get compiled when
10266 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10267 the self-tests were still using deprecated key-generation functions so
10268 these have been updated also.
10272 * Reorganise PKCS#7 code to separate the digest location functionality
10273 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10274 New function PKCS7_set_digest() to set the digest type for PKCS#7
10275 digestedData type. Add additional code to correctly generate the
10276 digestedData type and add support for this type in PKCS7 initialization
10281 * New function PKCS7_set0_type_other() this initializes a PKCS7
10282 structure of type "other".
10286 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10287 sure the loop does correctly stop and breaking ("division by zero")
10288 modulus operations are not performed. The (pre-generated) prime
10289 table crypto/bn/bn_prime.h was already correct, but it could not be
10290 re-generated on some platforms because of the "division by zero"
10291 situation in the script.
10293 *Ralf S. Engelschall*
10295 * Update support for ECC-based TLS ciphersuites according to
10296 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10297 SHA-1 now is only used for "small" curves (where the
10298 representation of a field element takes up to 24 bytes); for
10299 larger curves, the field element resulting from ECDH is directly
10300 used as premaster secret.
10302 *Douglas Stebila (Sun Microsystems Laboratories)*
10304 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10305 curve secp160r1 to the tests.
10307 *Douglas Stebila (Sun Microsystems Laboratories)*
10309 * Add the possibility to load symbols globally with DSO.
10311 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10313 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10314 control of the error stack.
10318 * Add support for STORE in ENGINE.
10322 * Add the STORE type. The intention is to provide a common interface
10323 to certificate and key stores, be they simple file-based stores, or
10324 HSM-type store, or LDAP stores, or...
10325 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10329 * Add a generic structure called OPENSSL_ITEM. This can be used to
10330 pass a list of arguments to any function as well as provide a way
10331 for a function to pass data back to the caller.
10335 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10336 works like BUF_strdup() but can be used to duplicate a portion of
10337 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10342 * Add the function sk_find_ex() which works like sk_find(), but will
10343 return an index to an element even if an exact match couldn't be
10344 found. The index is guaranteed to point at the element where the
10345 searched-for key would be inserted to preserve sorting order.
10349 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10350 takes an extra flags argument for optional functionality. Currently,
10351 the following flags are defined:
10353 OBJ_BSEARCH_VALUE_ON_NOMATCH
10354 This one gets OBJ_bsearch_ex() to return a pointer to the first
10355 element where the comparing function returns a negative or zero
10358 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10359 This one gets OBJ_bsearch_ex() to return a pointer to the first
10360 element where the comparing function returns zero. This is useful
10361 if there are more than one element where the comparing function
10366 * Make it possible to create self-signed certificates with 'openssl ca'
10367 in such a way that the self-signed certificate becomes part of the
10368 CA database and uses the same mechanisms for serial number generation
10369 as all other certificate signing. The new flag '-selfsign' enables
10370 this functionality. Adapt CA.sh and CA.pl.in.
10374 * Add functionality to check the public key of a certificate request
10375 against a given private. This is useful to check that a certificate
10376 request can be signed by that key (self-signing).
10380 * Make it possible to have multiple active certificates with the same
10381 subject in the CA index file. This is done only if the keyword
10382 'unique_subject' is set to 'no' in the main CA section (default
10383 if 'CA_default') of the configuration file. The value is saved
10384 with the database itself in a separate index attribute file,
10385 named like the index file with '.attr' appended to the name.
10389 * Generate multi-valued AVAs using '+' notation in config files for
10394 * Support for nameConstraints certificate extension.
10398 * Support for policyConstraints certificate extension.
10402 * Support for policyMappings certificate extension.
10406 * Make sure the default DSA_METHOD implementation only uses its
10407 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10408 and change its own handlers to be NULL so as to remove unnecessary
10409 indirection. This lets alternative implementations fallback to the
10410 default implementation more easily.
10414 * Support for directoryName in GeneralName related extensions
10419 * Make it possible to link applications using Makefile.shared.
10420 Make that possible even when linking against static libraries!
10424 * Support for single pass processing for S/MIME signing. This now
10425 means that S/MIME signing can be done from a pipe, in addition
10426 cleartext signing (multipart/signed type) is effectively streaming
10427 and the signed data does not need to be all held in memory.
10429 This is done with a new flag PKCS7_STREAM. When this flag is set
10430 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10431 is done after the data is output (and digests calculated) in
10432 SMIME_write_PKCS7().
10436 * Add full support for -rpath/-R, both in shared libraries and
10437 applications, at least on the platforms where it's known how
10442 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10443 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10444 will now compute a table of multiples of the generator that
10445 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10446 faster (notably in the case of a single point multiplication,
10447 scalar * generator).
10449 *Nils Larsch, Bodo Moeller*
10451 * IPv6 support for certificate extensions. The various extensions
10452 which use the IP:a.b.c.d can now take IPv6 addresses using the
10453 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10458 * Added an ENGINE that implements RSA by performing private key
10459 exponentiations with the GMP library. The conversions to and from
10460 GMP's mpz_t format aren't optimised nor are any montgomery forms
10461 cached, and on x86 it appears OpenSSL's own performance has caught up.
10462 However there are likely to be other architectures where GMP could
10463 provide a boost. This ENGINE is not built in by default, but it can be
10464 specified at Configure time and should be accompanied by the necessary
10465 linker additions, eg;
10466 ./config -DOPENSSL_USE_GMP -lgmp
10470 * "openssl engine" will not display ENGINE/DSO load failure errors when
10471 testing availability of engines with "-t" - the old behaviour is
10472 produced by increasing the feature's verbosity with "-tt".
10476 * ECDSA routines: under certain error conditions uninitialized BN objects
10477 could be freed. Solution: make sure initialization is performed early
10478 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10483 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10484 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10485 software implementations. For DSA and DH, parameter generation can
10486 also be overridden by providing the appropriate method callbacks.
10490 * Change the "progress" mechanism used in key-generation and
10491 primality testing to functions that take a new BN_GENCB pointer in
10492 place of callback/argument pairs. The new API functions have `_ex`
10493 postfixes and the older functions are reimplemented as wrappers for
10494 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10495 declarations of the old functions to help (graceful) attempts to
10496 migrate to the new functions. Also, the new key-generation API
10497 functions operate on a caller-supplied key-structure and return
10498 success/failure rather than returning a key or NULL - this is to
10499 help make "keygen" another member function of RSA_METHOD etc.
10501 Example for using the new callback interface:
10503 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10504 void *my_arg = ...;
10507 BN_GENCB_set(&my_cb, my_callback, my_arg);
10509 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10510 /* For the meaning of a, b in calls to my_callback(), see the
10511 * documentation of the function that calls the callback.
10512 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10513 * my_callback should return 1 if it wants BN_is_prime_ex()
10514 * to continue, or 0 to stop.
10519 * Change the ZLIB compression method to be stateful, and make it
10520 available to TLS with the number defined in
10521 draft-ietf-tls-compression-04.txt.
10525 * Add the ASN.1 structures and functions for CertificatePair, which
10526 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10528 CertificatePair ::= SEQUENCE {
10529 forward [0] Certificate OPTIONAL,
10530 reverse [1] Certificate OPTIONAL,
10531 -- at least one of the pair shall be present -- }
10533 Also implement the PEM functions to read and write certificate
10534 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10536 This needed to be defined, mostly for the sake of the LDAP
10537 attribute crossCertificatePair, but may prove useful elsewhere as
10542 * Make it possible to inhibit symlinking of shared libraries in
10543 Makefile.shared, for Cygwin's sake.
10547 * Extend the BIGNUM API by creating a function
10548 void BN_set_negative(BIGNUM *a, int neg);
10549 and a macro that behave like
10550 int BN_is_negative(const BIGNUM *a);
10552 to avoid the need to access 'a->neg' directly in applications.
10556 * Implement fast modular reduction for pseudo-Mersenne primes
10557 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10558 EC_GROUP_new_curve_GFp() will now automatically use this
10561 *Nils Larsch <nla@trustcenter.de>*
10563 * Add new lock type (CRYPTO_LOCK_BN).
10567 * Change the ENGINE framework to automatically load engines
10568 dynamically from specific directories unless they could be
10569 found to already be built in or loaded. Move all the
10570 current engines except for the cryptodev one to a new
10571 directory engines/.
10572 The engines in engines/ are built as shared libraries if
10573 the "shared" options was given to ./Configure or ./config.
10574 Otherwise, they are inserted in libcrypto.a.
10575 /usr/local/ssl/engines is the default directory for dynamic
10576 engines, but that can be overridden at configure time through
10577 the usual use of --prefix and/or --openssldir, and at run
10578 time with the environment variable OPENSSL_ENGINES.
10580 *Geoff Thorpe and Richard Levitte*
10582 * Add Makefile.shared, a helper makefile to build shared
10583 libraries. Adapt Makefile.org.
10587 * Add version info to Win32 DLLs.
10589 *Peter 'Luna' Runestig" <peter@runestig.com>*
10591 * Add new 'medium level' PKCS#12 API. Certificates and keys
10592 can be added using this API to created arbitrary PKCS#12
10593 files while avoiding the low level API.
10595 New options to PKCS12_create(), key or cert can be NULL and
10596 will then be omitted from the output file. The encryption
10597 algorithm NIDs can be set to -1 for no encryption, the mac
10598 iteration count can be set to 0 to omit the mac.
10600 Enhance pkcs12 utility by making the -nokeys and -nocerts
10601 options work when creating a PKCS#12 file. New option -nomac
10602 to omit the mac, NONE can be set for an encryption algorithm.
10603 New code is modified to use the enhanced PKCS12_create()
10604 instead of the low level API.
10608 * Extend ASN1 encoder to support indefinite length constructed
10609 encoding. This can output sequences tags and octet strings in
10610 this form. Modify pk7_asn1.c to support indefinite length
10611 encoding. This is experimental and needs additional code to
10612 be useful, such as an ASN1 bio and some enhanced streaming
10615 Extend template encode functionality so that tagging is passed
10616 down to the template encoder.
10620 * Let 'openssl req' fail if an argument to '-newkey' is not
10621 recognized instead of using RSA as a default.
10625 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10626 As these are not official, they are not included in "ALL";
10627 the "ECCdraft" ciphersuite group alias can be used to select them.
10629 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10631 * Add ECDH engine support.
10633 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10635 * Add ECDH in new directory crypto/ecdh/.
10637 *Douglas Stebila (Sun Microsystems Laboratories)*
10639 * Let BN_rand_range() abort with an error after 100 iterations
10640 without success (which indicates a broken PRNG).
10644 * Change BN_mod_sqrt() so that it verifies that the input value
10645 is really the square of the return value. (Previously,
10646 BN_mod_sqrt would show GIGO behaviour.)
10650 * Add named elliptic curves over binary fields from X9.62, SECG,
10651 and WAP/WTLS; add OIDs that were still missing.
10653 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10655 * Extend the EC library for elliptic curves over binary fields
10656 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10659 EC_GF2m_simple_method
10663 EC_GROUP_new_curve_GF2m
10664 EC_GROUP_set_curve_GF2m
10665 EC_GROUP_get_curve_GF2m
10666 EC_POINT_set_affine_coordinates_GF2m
10667 EC_POINT_get_affine_coordinates_GF2m
10668 EC_POINT_set_compressed_coordinates_GF2m
10670 Point compression for binary fields is disabled by default for
10671 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10674 As binary polynomials are represented as BIGNUMs, various members
10675 of the EC_GROUP and EC_POINT data structures can be shared
10676 between the implementations for prime fields and binary fields;
10677 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10678 are essentially identical to their `..._GFp` counterparts.
10679 (For simplicity, the `..._GFp` prefix has been dropped from
10680 various internal method names.)
10682 An internal 'field_div' method (similar to 'field_mul' and
10683 'field_sqr') has been added; this is used only for binary fields.
10685 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10687 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10688 through methods ('mul', 'precompute_mult').
10690 The generic implementations (now internally called 'ec_wNAF_mul'
10691 and 'ec_wNAF_precomputed_mult') remain the default if these
10692 methods are undefined.
10694 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10696 * New function EC_GROUP_get_degree, which is defined through
10697 EC_METHOD. For curves over prime fields, this returns the bit
10698 length of the modulus.
10700 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10702 * New functions EC_GROUP_dup, EC_POINT_dup.
10703 (These simply call ..._new and ..._copy).
10705 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10707 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10708 Polynomials are represented as BIGNUMs (where the sign bit is not
10709 used) in the following functions [macros]:
10712 BN_GF2m_sub [= BN_GF2m_add]
10713 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10714 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10715 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10717 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10718 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10719 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10720 BN_GF2m_cmp [= BN_ucmp]
10722 (Note that only the 'mod' functions are actually for fields GF(2^m).
10723 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10725 For some functions, an the irreducible polynomial defining a
10726 field can be given as an 'unsigned int[]' with strictly
10727 decreasing elements giving the indices of those bits that are set;
10728 i.e., p[] represents the polynomial
10729 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10731 p[0] > p[1] > ... > p[k] = 0.
10732 This applies to the following functions:
10735 BN_GF2m_mod_mul_arr
10736 BN_GF2m_mod_sqr_arr
10737 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10738 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10739 BN_GF2m_mod_exp_arr
10740 BN_GF2m_mod_sqrt_arr
10741 BN_GF2m_mod_solve_quad_arr
10745 Conversion can be performed by the following functions:
10750 bntest.c has additional tests for binary polynomial arithmetic.
10752 Two implementations for BN_GF2m_mod_div() are available.
10753 The default algorithm simply uses BN_GF2m_mod_inv() and
10754 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10755 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10756 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10758 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10760 * Add new error code 'ERR_R_DISABLED' that can be used when some
10761 functionality is disabled at compile-time.
10763 *Douglas Stebila <douglas.stebila@sun.com>*
10765 * Change default behaviour of 'openssl asn1parse' so that more
10766 information is visible when viewing, e.g., a certificate:
10768 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10769 mode the content of non-printable OCTET STRINGs is output in a
10770 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10771 avoid the appearance of a printable string.
10773 *Nils Larsch <nla@trustcenter.de>*
10775 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10777 EC_GROUP_set_asn1_flag()
10778 EC_GROUP_get_asn1_flag()
10779 EC_GROUP_set_point_conversion_form()
10780 EC_GROUP_get_point_conversion_form()
10781 These control ASN1 encoding details:
10782 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10783 has been set to OPENSSL_EC_NAMED_CURVE.
10784 - Points are encoded in uncompressed form by default; options for
10785 asn1_for are as for point2oct, namely
10786 POINT_CONVERSION_COMPRESSED
10787 POINT_CONVERSION_UNCOMPRESSED
10788 POINT_CONVERSION_HYBRID
10790 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10792 EC_GROUP_set_seed()
10793 EC_GROUP_get0_seed()
10794 EC_GROUP_get_seed_len()
10795 This is used only for ASN1 purposes (so far).
10797 *Nils Larsch <nla@trustcenter.de>*
10799 * Add 'field_type' member to EC_METHOD, which holds the NID
10800 of the appropriate field type OID. The new function
10801 EC_METHOD_get_field_type() returns this value.
10803 *Nils Larsch <nla@trustcenter.de>*
10806 EC_POINT_point2bn()
10807 EC_POINT_bn2point()
10808 EC_POINT_point2hex()
10809 EC_POINT_hex2point()
10810 providing useful interfaces to EC_POINT_point2oct() and
10811 EC_POINT_oct2point().
10813 *Nils Larsch <nla@trustcenter.de>*
10815 * Change internals of the EC library so that the functions
10816 EC_GROUP_set_generator()
10817 EC_GROUP_get_generator()
10818 EC_GROUP_get_order()
10819 EC_GROUP_get_cofactor()
10820 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10821 to methods, which would lead to unnecessary code duplication when
10822 adding different types of curves.
10824 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10826 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10827 arithmetic, and such that modified wNAFs are generated
10828 (which avoid length expansion in many cases).
10832 * Add a function EC_GROUP_check_discriminant() (defined via
10833 EC_METHOD) that verifies that the curve discriminant is non-zero.
10835 Add a function EC_GROUP_check() that makes some sanity tests
10836 on a EC_GROUP, its generator and order. This includes
10837 EC_GROUP_check_discriminant().
10839 *Nils Larsch <nla@trustcenter.de>*
10841 * Add ECDSA in new directory crypto/ecdsa/.
10843 Add applications 'openssl ecparam' and 'openssl ecdsa'
10844 (these are based on 'openssl dsaparam' and 'openssl dsa').
10846 ECDSA support is also included in various other files across the
10847 library. Most notably,
10848 - 'openssl req' now has a '-newkey ecdsa:file' option;
10849 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10850 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10851 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10852 them suitable for ECDSA where domain parameters must be
10853 extracted before the specific public key;
10854 - ECDSA engine support has been added.
10856 *Nils Larsch <nla@trustcenter.de>*
10858 * Include some named elliptic curves, and add OIDs from X9.62,
10859 SECG, and WAP/WTLS. Each curve can be obtained from the new
10861 EC_GROUP_new_by_curve_name(),
10862 and the list of available named curves can be obtained with
10863 EC_get_builtin_curves().
10864 Also add a 'curve_name' member to EC_GROUP objects, which can be
10866 EC_GROUP_set_curve_name()
10867 EC_GROUP_get_curve_name()
10869 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10871 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10872 was actually never needed) and in BN_mul(). The removal in BN_mul()
10873 required a small change in bn_mul_part_recursive() and the addition
10874 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10875 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10876 bn_sub_words() and bn_add_words() except they take arrays with
10881 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10883 * Cleanse PEM buffers before freeing them since they may contain
10886 *Benjamin Bennett <ben@psc.edu>*
10888 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10889 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10890 authentication-only ciphersuites.
10894 * Since AES128 and AES256 share a single mask bit in the logic of
10895 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10896 kludge to work properly if AES128 is available and AES256 isn't.
10900 * Expand security boundary to match 1.1.1 module.
10904 * Remove redundant features: hash file source, editing of test vectors
10905 modify fipsld to use external fips_premain.c signature.
10909 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10910 run algorithm test programs.
10914 * Make algorithm test programs more tolerant of whitespace.
10918 * Have SSL/TLS server implementation tolerate "mismatched" record
10919 protocol version while receiving ClientHello even if the
10920 ClientHello is fragmented. (The server can't insist on the
10921 particular protocol version it has chosen before the ServerHello
10922 message has informed the client about his choice.)
10926 * Load error codes if they are not already present instead of using a
10927 static variable. This allows them to be cleanly unloaded and reloaded.
10931 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10933 * Introduce limits to prevent malicious keys being able to
10934 cause a denial of service. ([CVE-2006-2940])
10936 *Steve Henson, Bodo Moeller*
10938 * Fix ASN.1 parsing of certain invalid structures that can result
10939 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10941 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10942 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10944 * Fix SSL client code which could crash if connecting to a
10945 malicious SSLv2 server. ([CVE-2006-4343])
10947 *Tavis Ormandy and Will Drewry, Google Security Team*
10949 * Change ciphersuite string processing so that an explicit
10950 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10951 will no longer include "AES128-SHA"), and any other similar
10952 ciphersuite (same bitmap) from *other* protocol versions (so that
10953 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10954 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10955 changes from 0.9.8b and 0.9.8d.
10959 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10961 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10962 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10964 * Change the Unix randomness entropy gathering to use poll() when
10965 possible instead of select(), since the latter has some
10966 undesirable limitations.
10968 *Darryl Miles via Richard Levitte and Bodo Moeller*
10970 * Disable rogue ciphersuites:
10972 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10973 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10974 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10976 The latter two were purportedly from
10977 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10980 Also deactivate the remaining ciphersuites from
10981 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10982 unofficial, and the ID has long expired.
10986 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10987 dual-core machines) and other potential thread-safety issues.
10991 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10993 * Adapt fipsld and the build system to link against the validated FIPS
10994 module in FIPS mode.
10998 * Fixes for VC++ 2005 build under Windows.
11002 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
11003 from a Windows bash shell such as MSYS. It is autodetected from the
11004 "config" script when run from a VC++ environment. Modify standard VC++
11005 build to use fipscanister.o from the GNU make build.
11009 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
11011 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
11012 The value now differs depending on if you build for FIPS or not.
11013 BEWARE! A program linked with a shared FIPSed libcrypto can't be
11014 safely run with a non-FIPSed libcrypto, as it may crash because of
11015 the difference induced by this change.
11019 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
11021 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
11022 (part of SSL_OP_ALL). This option used to disable the
11023 countermeasure against man-in-the-middle protocol-version
11024 rollback in the SSL 2.0 server implementation, which is a bad
11025 idea. ([CVE-2005-2969])
11027 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
11028 for Information Security, National Institute of Advanced Industrial
11029 Science and Technology [AIST, Japan)]*
11031 * Minimal support for X9.31 signatures and PSS padding modes. This is
11032 mainly for FIPS compliance and not fully integrated at this stage.
11036 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11037 the exponentiation using a fixed-length exponent. (Otherwise,
11038 the information leaked through timing could expose the secret key
11039 after many signatures; cf. Bleichenbacher's attack on DSA with
11044 * Make a new fixed-window mod_exp implementation the default for
11045 RSA, DSA, and DH private-key operations so that the sequence of
11046 squares and multiplies and the memory access pattern are
11047 independent of the particular secret key. This will mitigate
11048 cache-timing and potential related attacks.
11050 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11051 and this is automatically used by BN_mod_exp_mont() if the new flag
11052 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11053 will use this BN flag for private exponents unless the flag
11054 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11055 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11057 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11059 * Change the client implementation for SSLv23_method() and
11060 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11061 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11062 (Previously, the SSL 2.0 backwards compatible Client Hello
11063 message format would be used even with SSL_OP_NO_SSLv2.)
11067 * Add support for smime-type MIME parameter in S/MIME messages which some
11072 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11073 a threadsafe manner. Modify rsa code to use new function and add calls
11074 to dsa and dh code (which had race conditions before).
11078 * Include the fixed error library code in the C error file definitions
11079 instead of fixing them up at runtime. This keeps the error code
11080 structures constant.
11084 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11086 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11089 * Fixes for newer kerberos headers. NB: the casts are needed because
11090 the 'length' field is signed on one version and unsigned on another
11091 with no (?) obvious way to tell the difference, without these VC++
11092 complains. Also the "definition" of FAR (blank) is no longer included
11093 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11094 some needed definitions.
11098 * Undo Cygwin change.
11102 * Added support for proxy certificates according to RFC 3820.
11103 Because they may be a security thread to unaware applications,
11104 they must be explicitly allowed in run-time. See
11105 docs/HOWTO/proxy_certificates.txt for further information.
11109 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11111 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11112 server and client random values. Previously
11113 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11114 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11116 This change has negligible security impact because:
11118 1. Server and client random values still have 24 bytes of pseudo random
11121 2. Server and client random values are sent in the clear in the initial
11124 3. The master secret is derived using the premaster secret (48 bytes in
11125 size for static RSA ciphersuites) as well as client server and random
11128 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11131 *Stephen Henson, reported by UK NISCC*
11133 * Use Windows randomness collection on Cygwin.
11137 * Fix hang in EGD/PRNGD query when communication socket is closed
11138 prematurely by EGD/PRNGD.
11140 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11142 * Prompt for pass phrases when appropriate for PKCS12 input format.
11146 * Back-port of selected performance improvements from development
11147 branch, as well as improved support for PowerPC platforms.
11151 * Add lots of checks for memory allocation failure, error codes to indicate
11152 failure and freeing up memory if a failure occurs.
11154 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11156 * Add new -passin argument to dgst.
11160 * Perform some character comparisons of different types in X509_NAME_cmp:
11161 this is needed for some certificates that re-encode DNs into UTF8Strings
11162 (in violation of RFC3280) and can't or won't issue name rollover
11167 * Make an explicit check during certificate validation to see that
11168 the CA setting in each certificate on the chain is correct. As a
11169 side effect always do the following basic checks on extensions,
11170 not just when there's an associated purpose to the check:
11172 - if there is an unhandled critical extension (unless the user
11173 has chosen to ignore this fault)
11174 - if the path length has been exceeded (if one is set at all)
11175 - that certain extensions fit the associated purpose (if one has
11180 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11182 * Avoid a race condition when CRLs are checked in a multi threaded
11183 environment. This would happen due to the reordering of the revoked
11184 entries during signature checking and serial number lookup. Now the
11185 encoding is cached and the serial number sort performed under a lock.
11186 Add new STACK function sk_is_sorted().
11190 * Add Delta CRL to the extension code.
11194 * Various fixes to s3_pkt.c so alerts are sent properly.
11196 *David Holmes <d.holmes@f5.com>*
11198 * Reduce the chances of duplicate issuer name and serial numbers (in
11199 violation of RFC3280) using the OpenSSL certificate creation utilities.
11200 This is done by creating a random 64 bit value for the initial serial
11201 number when a serial number file is created or when a self signed
11202 certificate is created using 'openssl req -x509'. The initial serial
11203 number file is created using 'openssl x509 -next_serial' in CA.pl
11204 rather than being initialized to 1.
11208 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11210 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11211 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11213 *Joe Orton, Steve Henson*
11215 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11218 *Joe Orton, Steve Henson*
11220 * Make it possible to have multiple active certificates with the same
11221 subject in the CA index file. This is done only if the keyword
11222 'unique_subject' is set to 'no' in the main CA section (default
11223 if 'CA_default') of the configuration file. The value is saved
11224 with the database itself in a separate index attribute file,
11225 named like the index file with '.attr' appended to the name.
11229 * X509 verify fixes. Disable broken certificate workarounds when
11230 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11231 keyUsage extension present. Don't accept CRLs with unhandled critical
11232 extensions: since verify currently doesn't process CRL extensions this
11233 rejects a CRL with *any* critical extensions. Add new verify error codes
11238 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11239 A clarification of RFC2560 will require the use of OCTET STRINGs and
11240 some implementations cannot handle the current raw format. Since OpenSSL
11241 copies and compares OCSP nonces as opaque blobs without any attempt at
11242 parsing them this should not create any compatibility issues.
11246 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11247 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11248 this HMAC (and other) operations are several times slower than OpenSSL
11253 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11255 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11257 * Use the correct content when signing type "other".
11261 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11263 * Fix various bugs revealed by running the NISCC test suite:
11265 Stop out of bounds reads in the ASN1 code when presented with
11266 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11268 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11270 If verify callback ignores invalid public key errors don't try to check
11271 certificate signature with the NULL public key.
11275 * New -ignore_err option in ocsp application to stop the server
11276 exiting on the first error in a request.
11280 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11281 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11286 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11287 extra data after the compression methods not only for TLS 1.0
11288 but also for SSL 3.0 (as required by the specification).
11290 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11292 * Change X509_certificate_type() to mark the key as exported/exportable
11293 when it's 512 *bits* long, not 512 bytes.
11297 * Change AES_cbc_encrypt() so it outputs exact multiple of
11298 blocks during encryption.
11302 * Various fixes to base64 BIO and non blocking I/O. On write
11303 flushes were not handled properly if the BIO retried. On read
11304 data was not being buffered properly and had various logic bugs.
11305 This also affects blocking I/O when the data being decoded is a
11310 * Various S/MIME bugfixes and compatibility changes:
11311 output correct application/pkcs7 MIME type if
11312 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11313 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11314 of files as .eml work). Correctly handle very long lines in MIME
11319 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11321 * Countermeasure against the Klima-Pokorny-Rosa extension of
11322 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11323 a protocol version number mismatch like a decryption error
11324 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11328 * Turn on RSA blinding by default in the default implementation
11329 to avoid a timing attack. Applications that don't want it can call
11330 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11331 They would be ill-advised to do so in most cases.
11333 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11335 * Change RSA blinding code so that it works when the PRNG is not
11336 seeded (in this case, the secret RSA exponent is abused as
11337 an unpredictable seed -- if it is not unpredictable, there
11338 is no point in blinding anyway). Make RSA blinding thread-safe
11339 by remembering the creator's thread ID in rsa->blinding and
11340 having all other threads use local one-time blinding factors
11341 (this requires more computation than sharing rsa->blinding, but
11342 avoids excessive locking; and if an RSA object is not shared
11343 between threads, blinding will still be very fast).
11347 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11348 ENGINE as defaults for all supported algorithms irrespective of
11349 the 'flags' parameter. 'flags' is now honoured, so applications
11350 should make sure they are passing it correctly.
11354 * Target "mingw" now allows native Windows code to be generated in
11355 the Cygwin environment as well as with the MinGW compiler.
11359 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11361 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11362 via timing by performing a MAC computation even if incorrect
11363 block cipher padding has been found. This is a countermeasure
11364 against active attacks where the attacker has to distinguish
11365 between bad padding and a MAC verification error. ([CVE-2003-0078])
11367 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11368 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11369 Martin Vuagnoux (EPFL, Ilion)*
11371 * Make the no-err option work as intended. The intention with no-err
11372 is not to have the whole error stack handling routines removed from
11373 libcrypto, it's only intended to remove all the function name and
11374 reason texts, thereby removing some of the footprint that may not
11375 be interesting if those errors aren't displayed anyway.
11377 NOTE: it's still possible for any application or module to have its
11378 own set of error texts inserted. The routines are there, just not
11379 used by default when no-err is given.
11383 * Add support for FreeBSD on IA64.
11385 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11387 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11388 Kerberos function mit_des_cbc_cksum(). Before this change,
11389 the value returned by DES_cbc_cksum() was like the one from
11390 mit_des_cbc_cksum(), except the bytes were swapped.
11392 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11394 * Allow an application to disable the automatic SSL chain building.
11395 Before this a rather primitive chain build was always performed in
11396 ssl3_output_cert_chain(): an application had no way to send the
11397 correct chain if the automatic operation produced an incorrect result.
11399 Now the chain builder is disabled if either:
11401 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11403 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11405 The reasoning behind this is that an application would not want the
11406 auto chain building to take place if extra chain certificates are
11407 present and it might also want a means of sending no additional
11408 certificates (for example the chain has two certificates and the
11413 * Add the possibility to build without the ENGINE framework.
11415 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11417 * Under Win32 gmtime() can return NULL: check return value in
11418 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11422 * DSA routines: under certain error conditions uninitialized BN objects
11423 could be freed. Solution: make sure initialization is performed early
11424 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11425 Nils Larsch <nla@trustcenter.de> via PR#459)
11429 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11430 checked on reconnect on the client side, therefore session resumption
11431 could still fail with a "ssl session id is different" error. This
11432 behaviour is masked when SSL_OP_ALL is used due to
11433 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11434 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11435 followup to PR #377.
11439 * IA-32 assembler support enhancements: unified ELF targets, support
11440 for SCO/Caldera platforms, fix for Cygwin shared build.
11444 * Add support for FreeBSD on sparc64. As a consequence, support for
11445 FreeBSD on non-x86 processors is separate from x86 processors on
11446 the config script, much like the NetBSD support.
11448 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11450 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11452 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11455 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11456 code (06) was taken as the first octet of the session ID and the last
11457 octet was ignored consequently. As a result SSLv2 client side session
11458 caching could not have worked due to the session ID mismatch between
11460 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11465 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11466 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11471 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11472 seems that in spite of existing for more than a year, many application
11473 author have done nothing to provide the necessary callbacks, which
11474 means that this particular engine will not work properly anywhere.
11475 This is a very unfortunate situation which forces us, in the name
11476 of usability, to give the hw_ncipher.c a static lock, which is part
11478 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11479 appear in 0.9.8 or later. We EXPECT application authors to have
11480 dealt properly with this when 0.9.8 is released (unless we actually
11481 make such changes in the libcrypto locking code that changes will
11482 have to be made anyway).
11486 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11487 octets have been read, EOF or an error occurs. Without this change
11488 some truncated ASN1 structures will not produce an error.
11492 * Disable Heimdal support, since it hasn't been fully implemented.
11493 Still give the possibility to force the use of Heimdal, but with
11494 warnings and a request that patches get sent to openssl-dev.
11498 * Add the VC-CE target, introduce the WINCE sysname, and add
11499 INSTALL.WCE and appropriate conditionals to make it build.
11501 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11503 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11504 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11505 edit numbers of the version.
11507 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11509 * Introduce safe string copy and catenation functions
11510 (BUF_strlcpy() and BUF_strlcat()).
11512 *Ben Laurie (CHATS) and Richard Levitte*
11514 * Avoid using fixed-size buffers for one-line DNs.
11516 *Ben Laurie (CHATS)*
11518 * Add BUF_MEM_grow_clean() to avoid information leakage when
11519 resizing buffers containing secrets, and use where appropriate.
11521 *Ben Laurie (CHATS)*
11523 * Avoid using fixed size buffers for configuration file location.
11525 *Ben Laurie (CHATS)*
11527 * Avoid filename truncation for various CA files.
11529 *Ben Laurie (CHATS)*
11531 * Use sizeof in preference to magic numbers.
11533 *Ben Laurie (CHATS)*
11535 * Avoid filename truncation in cert requests.
11537 *Ben Laurie (CHATS)*
11539 * Add assertions to check for (supposedly impossible) buffer
11542 *Ben Laurie (CHATS)*
11544 * Don't cache truncated DNS entries in the local cache (this could
11545 potentially lead to a spoofing attack).
11547 *Ben Laurie (CHATS)*
11549 * Fix various buffers to be large enough for hex/decimal
11550 representations in a platform independent manner.
11552 *Ben Laurie (CHATS)*
11554 * Add CRYPTO_realloc_clean() to avoid information leakage when
11555 resizing buffers containing secrets, and use where appropriate.
11557 *Ben Laurie (CHATS)*
11559 * Add BIO_indent() to avoid much slightly worrying code to do
11562 *Ben Laurie (CHATS)*
11564 * Convert sprintf()/BIO_puts() to BIO_printf().
11566 *Ben Laurie (CHATS)*
11568 * buffer_gets() could terminate with the buffer only half
11571 *Ben Laurie (CHATS)*
11573 * Add assertions to prevent user-supplied crypto functions from
11574 overflowing internal buffers by having large block sizes, etc.
11576 *Ben Laurie (CHATS)*
11578 * New OPENSSL_assert() macro (similar to assert(), but enabled
11581 *Ben Laurie (CHATS)*
11583 * Eliminate unused copy of key in RC4.
11585 *Ben Laurie (CHATS)*
11587 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11589 *Ben Laurie (CHATS)*
11591 * Fix off-by-one error in EGD path.
11593 *Ben Laurie (CHATS)*
11595 * If RANDFILE path is too long, ignore instead of truncating.
11597 *Ben Laurie (CHATS)*
11599 * Eliminate unused and incorrectly sized X.509 structure
11602 *Ben Laurie (CHATS)*
11604 * Eliminate unused and dangerous function knumber().
11606 *Ben Laurie (CHATS)*
11608 * Eliminate unused and dangerous structure, KSSL_ERR.
11610 *Ben Laurie (CHATS)*
11612 * Protect against overlong session ID context length in an encoded
11613 session object. Since these are local, this does not appear to be
11616 *Ben Laurie (CHATS)*
11618 * Change from security patch (see 0.9.6e below) that did not affect
11619 the 0.9.6 release series:
11621 Remote buffer overflow in SSL3 protocol - an attacker could
11622 supply an oversized master key in Kerberos-enabled versions.
11625 *Ben Laurie (CHATS)*
11627 * Change the SSL kerb5 codes to match RFC 2712.
11631 * Make -nameopt work fully for req and add -reqopt switch.
11633 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11635 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11637 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11639 * Make sure tests can be performed even if the corresponding algorithms
11640 have been removed entirely. This was also the last step to make
11641 OpenSSL compilable with DJGPP under all reasonable conditions.
11643 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11645 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11646 to allow version independent disabling of normally unselected ciphers,
11647 which may be activated as a side-effect of selecting a single cipher.
11649 (E.g., cipher list string "RSA" enables ciphersuites that are left
11650 out of "ALL" because they do not provide symmetric encryption.
11651 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11653 *Lutz Jaenicke, Bodo Moeller*
11655 * Add appropriate support for separate platform-dependent build
11656 directories. The recommended way to make a platform-dependent
11657 build directory is the following (tested on Linux), maybe with
11660 # Place yourself outside of the OpenSSL source tree. In
11661 # this example, the environment variable OPENSSL_SOURCE
11662 # is assumed to contain the absolute OpenSSL source directory.
11663 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11664 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11665 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11666 mkdir -p `dirname $F`
11667 ln -s $OPENSSL_SOURCE/$F $F
11670 To be absolutely sure not to disturb the source tree, a "make clean"
11671 is a good thing. If it isn't successful, don't worry about it,
11672 it probably means the source directory is very clean.
11676 * Make sure any ENGINE control commands make local copies of string
11677 pointers passed to them whenever necessary. Otherwise it is possible
11678 the caller may have overwritten (or deallocated) the original string
11679 data when a later ENGINE operation tries to use the stored values.
11681 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11683 * Improve diagnostics in file reading and command-line digests.
11685 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11687 * Add AES modes CFB and OFB to the object database. Correct an
11688 error in AES-CFB decryption.
11692 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11693 allows existing EVP_CIPHER_CTX structures to be reused after
11694 calling `EVP_*Final()`. This behaviour is used by encryption
11695 BIOs and some applications. This has the side effect that
11696 applications must explicitly clean up cipher contexts with
11697 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11701 * Check the values of dna and dnb in bn_mul_recursive before calling
11702 bn_mul_comba (a non zero value means the a or b arrays do not contain
11703 n2 elements) and fallback to bn_mul_normal if either is not zero.
11707 * Fix escaping of non-ASCII characters when using the -subj option
11708 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11712 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11713 form for "surname", serialNumber has no short form.
11714 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11715 therefore remove "mail" short name for "internet 7".
11716 The OID for unique identifiers in X509 certificates is
11717 x500UniqueIdentifier, not uniqueIdentifier.
11718 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11722 * Add an "init" command to the ENGINE config module and auto initialize
11723 ENGINEs. Without any "init" command the ENGINE will be initialized
11724 after all ctrl commands have been executed on it. If init=1 the
11725 ENGINE is initialized at that point (ctrls before that point are run
11726 on the uninitialized ENGINE and after on the initialized one). If
11727 init=0 then the ENGINE will not be initialized at all.
11731 * Fix the 'app_verify_callback' interface so that the user-defined
11732 argument is actually passed to the callback: In the
11733 SSL_CTX_set_cert_verify_callback() prototype, the callback
11734 declaration has been changed from
11737 int (*cb)(X509_STORE_CTX *,void *);
11738 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11739 i=s->ctx->app_verify_callback(&ctx)
11740 has been changed into
11741 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11743 To update applications using SSL_CTX_set_cert_verify_callback(),
11744 a dummy argument can be added to their callback functions.
11746 *D. K. Smetters <smetters@parc.xerox.com>*
11748 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11750 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11752 * Add and OPENSSL_LOAD_CONF define which will cause
11753 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11754 This allows older applications to transparently support certain
11755 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11756 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11757 load the config file and OPENSSL_add_all_algorithms_conf() which will
11758 always load it have also been added.
11762 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11763 Adjust NIDs and EVP layer.
11765 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11767 * Config modules support in openssl utility.
11769 Most commands now load modules from the config file,
11770 though in a few (such as version) this isn't done
11771 because it couldn't be used for anything.
11773 In the case of ca and req the config file used is
11774 the same as the utility itself: that is the -config
11775 command line option can be used to specify an
11780 * Move default behaviour from OPENSSL_config(). If appname is NULL
11781 use "openssl_conf" if filename is NULL use default openssl config file.
11785 * Add an argument to OPENSSL_config() to allow the use of an alternative
11786 config section name. Add a new flag to tolerate a missing config file
11787 and move code to CONF_modules_load_file().
11791 * Support for crypto accelerator cards from Accelerated Encryption
11792 Processing, www.aep.ie. (Use engine 'aep')
11793 The support was copied from 0.9.6c [engine] and adapted/corrected
11794 to work with the new engine framework.
11796 *AEP Inc. and Richard Levitte*
11798 * Support for SureWare crypto accelerator cards from Baltimore
11799 Technologies. (Use engine 'sureware')
11800 The support was copied from 0.9.6c [engine] and adapted
11801 to work with the new engine framework.
11805 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11806 make the newer ENGINE framework commands for the CHIL engine work.
11808 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11810 * Make it possible to produce shared libraries on ReliantUNIX.
11812 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11814 * Add the configuration target debug-linux-ppro.
11815 Make 'openssl rsa' use the general key loading routines
11816 implemented in `apps.c`, and make those routines able to
11817 handle the key format FORMAT_NETSCAPE and the variant
11820 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11822 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11824 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11826 * Add -keyform to rsautl, and document -engine.
11828 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11830 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11831 BIO_R_NO_SUCH_FILE error code rather than the generic
11832 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11836 * Add new functions
11837 ERR_peek_last_error
11838 ERR_peek_last_error_line
11839 ERR_peek_last_error_line_data.
11840 These are similar to
11842 ERR_peek_error_line
11843 ERR_peek_error_line_data,
11844 but report on the latest error recorded rather than the first one
11845 still in the error queue.
11847 *Ben Laurie, Bodo Moeller*
11849 * default_algorithms option in ENGINE config module. This allows things
11851 default_algorithms = ALL
11852 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11856 * Preliminary ENGINE config module.
11860 * New experimental application configuration code.
11864 * Change the AES code to follow the same name structure as all other
11865 symmetric ciphers, and behave the same way. Move everything to
11866 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11868 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11870 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11872 *Ben Laurie and Theo de Raadt*
11874 * Add option to output public keys in req command.
11876 *Massimiliano Pala madwolf@openca.org*
11878 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11879 (up to about 10% better than before for P-192 and P-224).
11883 * New functions/macros
11885 SSL_CTX_set_msg_callback(ctx, cb)
11886 SSL_CTX_set_msg_callback_arg(ctx, arg)
11887 SSL_set_msg_callback(ssl, cb)
11888 SSL_set_msg_callback_arg(ssl, arg)
11890 to request calling a callback function
11892 void cb(int write_p, int version, int content_type,
11893 const void *buf, size_t len, SSL *ssl, void *arg)
11895 whenever a protocol message has been completely received
11896 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11897 protocol version according to which the SSL library interprets
11898 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11899 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11900 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11901 specification (change_cipher_spec(20), alert(21), handshake(22)).
11902 'buf' and 'len' point to the actual message, 'ssl' to the
11903 SSL object, and 'arg' is the application-defined value set by
11904 SSL[_CTX]_set_msg_callback_arg().
11906 'openssl s_client' and 'openssl s_server' have new '-msg' options
11907 to enable a callback that displays all protocol messages.
11911 * Change the shared library support so shared libraries are built as
11912 soon as the corresponding static library is finished, and thereby get
11913 openssl and the test programs linked against the shared library.
11914 This still only happens when the keyword "shard" has been given to
11915 the configuration scripts.
11917 NOTE: shared library support is still an experimental thing, and
11918 backward binary compatibility is still not guaranteed.
11920 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11922 * Add support for Subject Information Access extension.
11924 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11926 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11927 additional bytes when new memory had to be allocated, not just
11928 when reusing an existing buffer.
11932 * New command line and configuration option 'utf8' for the req command.
11933 This allows field values to be specified as UTF8 strings.
11937 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11938 runs for the former and machine-readable output for the latter.
11942 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11943 of the e-mail address in the DN (i.e., it will go into a certificate
11944 extension only). The new configuration file option 'email_in_dn = no'
11945 has the same effect.
11947 *Massimiliano Pala madwolf@openca.org*
11949 * Change all functions with names starting with `des_` to be starting
11950 with `DES_` instead. Add wrappers that are compatible with libdes,
11951 but are named `_ossl_old_des_*`. Finally, add macros that map the
11952 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11953 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11954 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11957 Since we provide two compatibility mappings, the user needs to
11958 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11959 compatibility is desired. The default (i.e., when that macro
11960 isn't defined) is OpenSSL 0.9.6c compatibility.
11962 There are also macros that enable and disable the support of old
11963 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11964 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11965 are defined, the default will apply: to support the old des routines.
11967 In either case, one must include openssl/des.h to get the correct
11968 definitions. Do not try to just include openssl/des_old.h, that
11971 NOTE: This is a major break of an old API into a new one. Software
11972 authors are encouraged to switch to the `DES_` style functions. Some
11973 time in the future, des_old.h and the libdes compatibility functions
11974 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11975 default), and then completely removed.
11979 * Test for certificates which contain unsupported critical extensions.
11980 If such a certificate is found during a verify operation it is
11981 rejected by default: this behaviour can be overridden by either
11982 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11983 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11984 X509_supported_extension() has also been added which returns 1 if a
11985 particular extension is supported.
11989 * Modify the behaviour of EVP cipher functions in similar way to digests
11990 to retain compatibility with existing code.
11994 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11995 compatibility with existing code. In particular the 'ctx' parameter does
11996 not have to be to be initialized before the call to EVP_DigestInit() and
11997 it is tidied up after a call to EVP_DigestFinal(). New function
11998 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11999 EVP_MD_CTX_copy() changed to not require the destination to be
12000 initialized valid and new function EVP_MD_CTX_copy_ex() added which
12001 requires the destination to be valid.
12003 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
12004 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
12008 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
12009 so that complete 'Handshake' protocol structures are kept in memory
12010 instead of overwriting 'msg_type' and 'length' with 'body' data.
12014 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
12016 *Massimo Santin via Richard Levitte*
12018 * Major restructuring to the underlying ENGINE code. This includes
12019 reduction of linker bloat, separation of pure "ENGINE" manipulation
12020 (initialisation, etc) from functionality dealing with implementations
12021 of specific crypto interfaces. This change also introduces integrated
12022 support for symmetric ciphers and digest implementations - so ENGINEs
12023 can now accelerate these by providing EVP_CIPHER and EVP_MD
12024 implementations of their own. This is detailed in
12025 [crypto/engine/README.md](crypto/engine/README.md)
12026 as it couldn't be adequately described here. However, there are a few
12027 API changes worth noting - some RSA, DSA, DH, and RAND functions that
12028 were changed in the original introduction of ENGINE code have now
12029 reverted back - the hooking from this code to ENGINE is now a good
12030 deal more passive and at run-time, operations deal directly with
12031 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12032 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12033 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12034 they were not being used by the framework as there is no concept of a
12035 BIGNUM_METHOD and they could not be generalised to the new
12036 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12037 ENGINE_cpy() has been removed as it cannot be consistently defined in
12042 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12046 * Change mkdef.pl to sort symbols that get the same entry number,
12047 and make sure the automatically generated functions `ERR_load_*`
12048 become part of libeay.num as well.
12052 * New function SSL_renegotiate_pending(). This returns true once
12053 renegotiation has been requested (either SSL_renegotiate() call
12054 or HelloRequest/ClientHello received from the peer) and becomes
12055 false once a handshake has been completed.
12056 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12057 sends a HelloRequest, but does not ensure that a handshake takes
12058 place. SSL_renegotiate_pending() is useful for checking if the
12059 client has followed the request.)
12063 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12064 By default, clients may request session resumption even during
12065 renegotiation (if session ID contexts permit); with this option,
12066 session resumption is possible only in the first handshake.
12068 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12069 more bits available for options that should not be part of
12070 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12074 * Add some demos for certificate and certificate request creation.
12078 * Make maximum certificate chain size accepted from the peer application
12079 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12080 "Douglas E. Engert" <deengert@anl.gov>.
12084 * Add support for shared libraries for Unixware-7
12085 (Boyd Lynn Gerber <gerberb@zenez.com>).
12089 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12090 be done prior to destruction. Use this to unload error strings from
12091 ENGINEs that load their own error strings. NB: This adds two new API
12092 functions to "get" and "set" this destroy handler in an ENGINE.
12096 * Alter all existing ENGINE implementations (except "openssl" and
12097 "openbsd") to dynamically instantiate their own error strings. This
12098 makes them more flexible to be built both as statically-linked ENGINEs
12099 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12100 Also, add stub code to each that makes building them as self-contained
12101 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12105 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12106 implementations into applications that are completely implemented in
12107 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12108 commands that can be used to configure what shared-library to load and
12109 to control aspects of the way it is handled. Also, made an update to
12110 the [README-Engine.md](README-Engine.md) file
12111 that brings its information up-to-date and
12112 provides some information and instructions on the "dynamic" ENGINE
12113 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12117 * Make it possible to unload ranges of ERR strings with a new
12118 "ERR_unload_strings" function.
12122 * Add a copy() function to EVP_MD.
12126 * Make EVP_MD routines take a context pointer instead of just the
12127 md_data void pointer.
12131 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12132 that the digest can only process a single chunk of data
12133 (typically because it is provided by a piece of
12134 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12135 is only going to provide a single chunk of data, and hence the
12136 framework needn't accumulate the data for oneshot drivers.
12140 * As with "ERR", make it possible to replace the underlying "ex_data"
12141 functions. This change also alters the storage and management of global
12142 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12143 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12144 index counters. The API functions that use this state have been changed
12145 to take a "class_index" rather than pointers to the class's local STACK
12146 and counter, and there is now an API function to dynamically create new
12147 classes. This centralisation allows us to (a) plug a lot of the
12148 thread-safety problems that existed, and (b) makes it possible to clean
12149 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12150 such data would previously have always leaked in application code and
12151 workarounds were in place to make the memory debugging turn a blind eye
12152 to it. Application code that doesn't use this new function will still
12153 leak as before, but their memory debugging output will announce it now
12154 rather than letting it slide.
12156 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12157 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12158 has a return value to indicate success or failure.
12162 * Make it possible to replace the underlying "ERR" functions such that the
12163 global state (2 LHASH tables and 2 locks) is only used by the "default"
12164 implementation. This change also adds two functions to "get" and "set"
12165 the implementation prior to it being automatically set the first time
12166 any other ERR function takes place. Ie. an application can call "get",
12167 pass the return value to a module it has just loaded, and that module
12168 can call its own "set" function using that value. This means the
12169 module's "ERR" operations will use (and modify) the error state in the
12170 application and not in its own statically linked copy of OpenSSL code.
12174 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12175 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12176 the operation, and provides a more encapsulated way for external code
12177 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12178 to use these functions rather than manually incrementing the counts.
12180 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12184 * Add EVP test program.
12188 * Add symmetric cipher support to ENGINE. Expect the API to change!
12192 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12193 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12194 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12195 These allow a CRL to be built without having to access X509_CRL fields
12196 directly. Modify 'ca' application to use new functions.
12200 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12201 bug workarounds. Rollback attack detection is a security feature.
12202 The problem will only arise on OpenSSL servers when TLSv1 is not
12203 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12204 Software authors not wanting to support TLSv1 will have special reasons
12205 for their choice and can explicitly enable this option.
12207 *Bodo Moeller, Lutz Jaenicke*
12209 * Rationalise EVP so it can be extended: don't include a union of
12210 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12211 (similar to those existing for EVP_CIPHER_CTX).
12216 EVP_MD_CTX_init(&md); /* new function call */
12217 EVP_DigestInit(&md, EVP_sha1());
12218 EVP_DigestUpdate(&md, in, len);
12219 EVP_DigestFinal(&md, out, NULL);
12220 EVP_MD_CTX_cleanup(&md); /* new function call */
12224 * Make DES key schedule conform to the usual scheme, as well as
12225 correcting its structure. This means that calls to DES functions
12226 now have to pass a pointer to a des_key_schedule instead of a
12227 plain des_key_schedule (which was actually always a pointer
12230 des_key_schedule ks;
12232 des_set_key_checked(..., &ks);
12233 des_ncbc_encrypt(..., &ks, ...);
12235 (Note that a later change renames 'des_...' into 'DES_...'.)
12239 * Initial reduction of linker bloat: the use of some functions, such as
12240 PEM causes large amounts of unused functions to be linked in due to
12241 poor organisation. For example pem_all.c contains every PEM function
12242 which has a knock on effect of linking in large amounts of (unused)
12243 ASN1 code. Grouping together similar functions and splitting unrelated
12244 functions prevents this.
12248 * Cleanup of EVP macros.
12252 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12253 correct `_ecb suffix`.
12257 * Add initial OCSP responder support to ocsp application. The
12258 revocation information is handled using the text based index
12259 use by the ca application. The responder can either handle
12260 requests generated internally, supplied in files (for example
12261 via a CGI script) or using an internal minimal server.
12265 * Add configuration choices to get zlib compression for TLS.
12269 * Changes to Kerberos SSL for RFC 2712 compliance:
12270 1. Implemented real KerberosWrapper, instead of just using
12271 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12272 2. Implemented optional authenticator field of KerberosWrapper.
12274 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12275 and authenticator structs; see crypto/krb5/.
12277 Generalized Kerberos calls to support multiple Kerberos libraries.
12278 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12279 via Richard Levitte*
12281 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12282 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12283 values for each of the key sizes rather than having just
12284 parameters (and 'speed' generating keys each time).
12288 * Speed up EVP routines.
12291 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12292 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12293 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12294 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12296 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12297 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12298 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12301 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12303 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12307 * Added the OS2-EMX target.
12309 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12311 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12312 New functions to support `NCONF `routines in extension code.
12313 New function `CONF_set_nconf()`
12314 to allow functions which take an `NCONF` to also handle the old `LHASH`
12315 structure: this means that the old `CONF` compatible routines can be
12316 retained (in particular w.rt. extensions) without having to duplicate the
12317 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12321 * Enhance the general user interface with mechanisms for inner control
12322 and with possibilities to have yes/no kind of prompts.
12326 * Change all calls to low level digest routines in the library and
12327 applications to use EVP. Add missing calls to HMAC_cleanup() and
12328 don't assume HMAC_CTX can be copied using memcpy().
12330 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12332 * Add the possibility to control engines through control names but with
12333 arbitrary arguments instead of just a string.
12334 Change the key loaders to take a UI_METHOD instead of a callback
12335 function pointer. NOTE: this breaks binary compatibility with earlier
12336 versions of OpenSSL [engine].
12337 Adapt the nCipher code for these new conditions and add a card insertion
12342 * Enhance the general user interface with mechanisms to better support
12343 dialog box interfaces, application-defined prompts, the possibility
12344 to use defaults (for example default passwords from somewhere else)
12345 and interrupts/cancellations.
12349 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12350 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12354 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12355 tidy up some unnecessarily weird code in 'sk_new()').
12357 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12359 * Change the key loading routines for ENGINEs to use the same kind
12360 callback (pem_password_cb) as all other routines that need this
12365 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12366 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12367 than this minimum value is recommended.
12371 * New random seeder for OpenVMS, using the system process statistics
12372 that are easily reachable.
12376 * Windows apparently can't transparently handle global
12377 variables defined in DLLs. Initialisations such as:
12379 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12381 won't compile. This is used by the any applications that need to
12382 declare their own ASN1 modules. This was fixed by adding the option
12383 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12384 needed for static libraries under Win32.
12388 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12389 setting of purpose and trust fields. New X509_STORE trust and
12390 purpose functions and tidy up setting in other SSL functions.
12394 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12395 structure. These are inherited by X509_STORE_CTX when it is
12396 initialised. This allows various defaults to be set in the
12397 X509_STORE structure (such as flags for CRL checking and custom
12398 purpose or trust settings) for functions which only use X509_STORE_CTX
12399 internally such as S/MIME.
12401 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12402 trust settings if they are not set in X509_STORE. This allows X509_STORE
12403 purposes and trust (in S/MIME for example) to override any set by default.
12405 Add command line options for CRL checking to smime, s_client and s_server
12410 * Initial CRL based revocation checking. If the CRL checking flag(s)
12411 are set then the CRL is looked up in the X509_STORE structure and
12412 its validity and signature checked, then if the certificate is found
12413 in the CRL the verify fails with a revoked error.
12415 Various new CRL related callbacks added to X509_STORE_CTX structure.
12417 Command line options added to 'verify' application to support this.
12419 This needs some additional work, such as being able to handle multiple
12420 CRLs with different times, extension based lookup (rather than just
12421 by subject name) and ultimately more complete V2 CRL extension
12426 * Add a general user interface API (crypto/ui/). This is designed
12427 to replace things like des_read_password and friends (backward
12428 compatibility functions using this new API are provided).
12429 The purpose is to remove prompting functions from the DES code
12430 section as well as provide for prompting through dialog boxes in
12431 a window system and the like.
12435 * Add "ex_data" support to ENGINE so implementations can add state at a
12436 per-structure level rather than having to store it globally.
12440 * Make it possible for ENGINE structures to be copied when retrieved by
12441 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12442 This causes the "original" ENGINE structure to act like a template,
12443 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12444 operational state can be localised to each ENGINE structure, despite the
12445 fact they all share the same "methods". New ENGINE structures returned in
12446 this case have no functional references and the return value is the single
12447 structural reference. This matches the single structural reference returned
12448 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12453 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12454 needs to match any other type at all we need to manually clear the
12459 * Changes to the "openssl engine" utility to include;
12460 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12461 about an ENGINE's available control commands.
12462 - executing control commands from command line arguments using the
12463 '-pre' and '-post' switches. '-post' is only used if '-t' is
12464 specified and the ENGINE is successfully initialised. The syntax for
12465 the individual commands are colon-separated, for example;
12466 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12470 * New dynamic control command support for ENGINEs. ENGINEs can now
12471 declare their own commands (numbers), names (strings), descriptions,
12472 and input types for run-time discovery by calling applications. A
12473 subset of these commands are implicitly classed as "executable"
12474 depending on their input type, and only these can be invoked through
12475 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12476 can be based on user input, config files, etc). The distinction is
12477 that "executable" commands cannot return anything other than a boolean
12478 result and can only support numeric or string input, whereas some
12479 discoverable commands may only be for direct use through
12480 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12481 pointers, or other custom uses. The "executable" commands are to
12482 support parameterisations of ENGINE behaviour that can be
12483 unambiguously defined by ENGINEs and used consistently across any
12484 OpenSSL-based application. Commands have been added to all the
12485 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12486 control over shared-library paths without source code alterations.
12490 * Changed all ENGINE implementations to dynamically allocate their
12491 ENGINEs rather than declaring them statically. Apart from this being
12492 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12493 this also allows the implementations to compile without using the
12494 internal engine_int.h header.
12498 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12499 'const' value. Any code that should be able to modify a RAND_METHOD
12500 should already have non-const pointers to it (ie. they should only
12501 modify their own ones).
12505 * Made a variety of little tweaks to the ENGINE code.
12506 - "atalla" and "ubsec" string definitions were moved from header files
12507 to C code. "nuron" string definitions were placed in variables
12508 rather than hard-coded - allowing parameterisation of these values
12509 later on via ctrl() commands.
12510 - Removed unused "#if 0"'d code.
12511 - Fixed engine list iteration code so it uses ENGINE_free() to release
12512 structural references.
12513 - Constified the RAND_METHOD element of ENGINE structures.
12514 - Constified various get/set functions as appropriate and added
12515 missing functions (including a catch-all ENGINE_cpy that duplicates
12516 all ENGINE values onto a new ENGINE except reference counts/state).
12517 - Removed NULL parameter checks in get/set functions. Setting a method
12518 or function to NULL is a way of cancelling out a previously set
12519 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12520 and doesn't justify the extra error symbols and code.
12521 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12522 flags from engine_int.h to engine.h.
12523 - Changed prototypes for ENGINE handler functions (init(), finish(),
12524 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12528 * Implement binary inversion algorithm for BN_mod_inverse in addition
12529 to the algorithm using long division. The binary algorithm can be
12530 used only if the modulus is odd. On 32-bit systems, it is faster
12531 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12532 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12533 up to 450 bits. In 64-bit environments, the binary algorithm
12534 appears to be advantageous for much longer moduli; here we use it
12535 for moduli up to 2048 bits.
12539 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12540 could not support the combine flag in choice fields.
12544 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12545 extensions from a certificate request to the certificate.
12549 * Allow multiple 'certopt' and 'nameopt' options to be separated
12550 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12551 file: this allows the display of the certificate about to be
12552 signed to be customised, to allow certain fields to be included
12553 or excluded and extension details. The old system didn't display
12554 multicharacter strings properly, omitted fields not in the policy
12555 and couldn't display additional details such as extensions.
12559 * Function EC_POINTs_mul for multiple scalar multiplication
12560 of an arbitrary number of elliptic curve points
12561 \sum scalars[i]*points[i],
12562 optionally including the generator defined for the EC_GROUP:
12563 scalar*generator + \sum scalars[i]*points[i].
12565 EC_POINT_mul is a simple wrapper function for the typical case
12566 that the point list has just one item (besides the optional
12571 * First EC_METHODs for curves over GF(p):
12573 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12574 operations and provides various method functions that can also
12575 operate with faster implementations of modular arithmetic.
12577 EC_GFp_mont_method() reuses most functions that are part of
12578 EC_GFp_simple_method, but uses Montgomery arithmetic.
12580 *Bodo Moeller; point addition and point doubling
12581 implementation directly derived from source code provided by
12582 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12584 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12585 crypto/ec/ec_lib.c):
12587 Curves are EC_GROUP objects (with an optional group generator)
12588 based on EC_METHODs that are built into the library.
12590 Points are EC_POINT objects based on EC_GROUP objects.
12592 Most of the framework would be able to handle curves over arbitrary
12593 finite fields, but as there are no obvious types for fields other
12594 than GF(p), some functions are limited to that for now.
12598 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12599 that the file contains a complete HTTP response.
12603 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12604 change the def and num file printf format specifier from "%-40sXXX"
12605 to "%-39s XXX". The latter will always guarantee a space after the
12606 field while the former will cause them to run together if the field
12607 is 40 of more characters long.
12611 * Constify the cipher and digest 'method' functions and structures
12612 and modify related functions to take constant EVP_MD and EVP_CIPHER
12617 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12618 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12622 * Modify `EVP_Digest*()` routines so they now return values. Although the
12623 internal software routines can never fail additional hardware versions
12628 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12630 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12631 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12634 ERR_R_NESTED_ASN1_ERROR
12636 ERR_R_MISSING_ASN1_EOS
12637 were 4 .. 9, conflicting with
12638 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12640 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12641 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12643 Add new error code 'ERR_R_INTERNAL_ERROR'.
12647 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12652 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12653 sets the subject name for a new request or supersedes the
12654 subject name in a given request. Formats that can be parsed are
12655 'CN=Some Name, OU=myOU, C=IT'
12657 'CN=Some Name/OU=myOU/C=IT'.
12659 Add options '-batch' and '-verbose' to 'openssl req'.
12661 *Massimiliano Pala <madwolf@hackmasters.net>*
12663 * Introduce the possibility to access global variables through
12664 functions on platform were that's the best way to handle exporting
12665 global variables in shared libraries. To enable this functionality,
12666 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12667 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12668 is normally done by Configure or something similar).
12670 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12671 in the source file (foo.c) like this:
12673 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12674 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12676 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12677 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12679 OPENSSL_DECLARE_GLOBAL(int,foo);
12680 #define foo OPENSSL_GLOBAL_REF(foo)
12681 OPENSSL_DECLARE_GLOBAL(double,bar);
12682 #define bar OPENSSL_GLOBAL_REF(bar)
12684 The #defines are very important, and therefore so is including the
12685 header file everywhere where the defined globals are used.
12687 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12688 of ASN.1 items, but that structure is a bit different.
12690 The largest change is in util/mkdef.pl which has been enhanced with
12691 better and easier to understand logic to choose which symbols should
12692 go into the Windows .def files as well as a number of fixes and code
12693 cleanup (among others, algorithm keywords are now sorted
12694 lexicographically to avoid constant rewrites).
12698 * In BN_div() keep a copy of the sign of 'num' before writing the
12699 result to 'rm' because if rm==num the value will be overwritten
12700 and produce the wrong result if 'num' is negative: this caused
12701 problems with BN_mod() and BN_nnmod().
12705 * Function OCSP_request_verify(). This checks the signature on an
12706 OCSP request and verifies the signer certificate. The signer
12707 certificate is just checked for a generic purpose and OCSP request
12712 * Add OCSP_check_validity() function to check the validity of OCSP
12713 responses. OCSP responses are prepared in real time and may only
12714 be a few seconds old. Simply checking that the current time lies
12715 between thisUpdate and nextUpdate max reject otherwise valid responses
12716 caused by either OCSP responder or client clock inaccuracy. Instead
12717 we allow thisUpdate and nextUpdate to fall within a certain period of
12718 the current time. The age of the response can also optionally be
12719 checked. Two new options -validity_period and -status_age added to
12724 * If signature or public key algorithm is unrecognized print out its
12725 OID rather that just UNKNOWN.
12729 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12730 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12731 ID to be generated from the issuer certificate alone which can then be
12732 passed to OCSP_id_issuer_cmp().
12736 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12737 ASN1 modules to export functions returning ASN1_ITEM pointers
12738 instead of the ASN1_ITEM structures themselves. This adds several
12739 new macros which allow the underlying ASN1 function/structure to
12740 be accessed transparently. As a result code should not use ASN1_ITEM
12741 references directly (such as &X509_it) but instead use the relevant
12742 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12743 use of the new ASN1 code on platforms where exporting structures
12744 is problematical (for example in shared libraries) but exporting
12745 functions returning pointers to structures is not.
12749 * Add support for overriding the generation of SSL/TLS session IDs.
12750 These callbacks can be registered either in an SSL_CTX or per SSL.
12751 The purpose of this is to allow applications to control, if they wish,
12752 the arbitrary values chosen for use as session IDs, particularly as it
12753 can be useful for session caching in multiple-server environments. A
12754 command-line switch for testing this (and any client code that wishes
12755 to use such a feature) has been added to "s_server".
12757 *Geoff Thorpe, Lutz Jaenicke*
12759 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12760 of the form `#if defined(...) || defined(...) || ...` and
12761 `#if !defined(...) && !defined(...) && ...`. This also avoids
12762 the growing number of special cases it was previously handling.
12766 * Make all configuration macros available for application by making
12767 sure they are available in opensslconf.h, by giving them names starting
12768 with `OPENSSL_` to avoid conflicts with other packages and by making
12769 sure e_os2.h will cover all platform-specific cases together with
12771 Additionally, it is now possible to define configuration/platform-
12772 specific names (called "system identities"). In the C code, these
12773 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12774 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12775 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12780 * New option -set_serial to 'req' and 'x509' this allows the serial
12781 number to use to be specified on the command line. Previously self
12782 signed certificates were hard coded with serial number 0 and the
12783 CA options of 'x509' had to use a serial number in a file which was
12788 * New options to 'ca' utility to support V2 CRL entry extensions.
12789 Currently CRL reason, invalidity date and hold instruction are
12790 supported. Add new CRL extensions to V3 code and some new objects.
12794 * New function EVP_CIPHER_CTX_set_padding() this is used to
12795 disable standard block padding (aka PKCS#5 padding) in the EVP
12796 API, which was previously mandatory. This means that the data is
12797 not padded in any way and so the total length much be a multiple
12798 of the block size, otherwise an error occurs.
12802 * Initial (incomplete) OCSP SSL support.
12806 * New function OCSP_parse_url(). This splits up a URL into its host,
12807 port and path components: primarily to parse OCSP URLs. New -url
12808 option to ocsp utility.
12812 * New nonce behavior. The return value of OCSP_check_nonce() now
12813 reflects the various checks performed. Applications can decide
12814 whether to tolerate certain situations such as an absent nonce
12815 in a response when one was present in a request: the ocsp application
12816 just prints out a warning. New function OCSP_add1_basic_nonce()
12817 this is to allow responders to include a nonce in a response even if
12818 the request is nonce-less.
12822 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12823 skipped when using openssl x509 multiple times on a single input file,
12824 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12828 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12829 set string type: to handle setting ASN1_TIME structures. Fix ca
12830 utility to correctly initialize revocation date of CRLs.
12834 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12835 the clients preferred ciphersuites and rather use its own preferences.
12836 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12837 Internet Explorer by ensuring unchanged hash method during stepup.
12838 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12842 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12843 to aes and add a new 'exist' option to print out symbols that don't
12848 * Additional options to ocsp utility to allow flags to be set and
12849 additional certificates supplied.
12853 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12854 OCSP client a number of certificate to only verify the response
12859 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12860 handle the new API. Currently only ECB, CBC modes supported. Add new
12863 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12864 Encryption Standard (AES) Ciphersuites for Transport Layer
12865 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12866 not enabled by default and were not part of the "ALL" ciphersuite
12867 alias because they were not yet official; they could be
12868 explicitly requested by specifying the "AESdraft" ciphersuite
12869 group alias. In the final release of OpenSSL 0.9.7, the group
12870 alias is called "AES" and is part of "ALL".)
12872 *Ben Laurie, Steve Henson, Bodo Moeller*
12874 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12875 request to response.
12879 * Functions for OCSP responders. OCSP_request_onereq_count(),
12880 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12881 extract information from a certificate request. OCSP_response_create()
12882 creates a response and optionally adds a basic response structure.
12883 OCSP_basic_add1_status() adds a complete single response to a basic
12884 response and returns the OCSP_SINGLERESP structure just added (to allow
12885 extensions to be included for example). OCSP_basic_add1_cert() adds a
12886 certificate to a basic response and OCSP_basic_sign() signs a basic
12887 response with various flags. New helper functions ASN1_TIME_check()
12888 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12889 (converts ASN1_TIME to GeneralizedTime).
12893 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12894 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12895 structure from a certificate. X509_pubkey_digest() digests the public_key
12896 contents: this is used in various key identifiers.
12900 * Make sk_sort() tolerate a NULL argument.
12902 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12904 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12905 passed by the function are trusted implicitly. If any of them signed the
12906 response then it is assumed to be valid and is not verified.
12910 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12911 to data. This was previously part of the PKCS7 ASN1 code. This
12912 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12913 *Steve Henson, reported by Kenneth R. Robinette
12914 <support@securenetterm.com>*
12916 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12917 routines: without these tracing memory leaks is very painful.
12918 Fix leaks in PKCS12 and PKCS7 routines.
12922 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12923 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12924 effectively meant GeneralizedTime would never be used. Now it
12925 is initialised to -1 but X509_time_adj() now has to check the value
12926 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12927 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12928 *Steve Henson, reported by Kenneth R. Robinette
12929 <support@securenetterm.com>*
12931 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12932 result in a zero length in the ASN1_INTEGER structure which was
12933 not consistent with the structure when d2i_ASN1_INTEGER() was used
12934 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12935 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12936 where it did not print out a minus for negative ASN1_INTEGER.
12940 * Add summary printout to ocsp utility. The various functions which
12941 convert status values to strings have been renamed to:
12942 OCSP_response_status_str(), OCSP_cert_status_str() and
12943 OCSP_crl_reason_str() and are no longer static. New options
12944 to verify nonce values and to disable verification. OCSP response
12945 printout format cleaned up.
12949 * Add additional OCSP certificate checks. These are those specified
12950 in RFC2560. This consists of two separate checks: the CA of the
12951 certificate being checked must either be the OCSP signer certificate
12952 or the issuer of the OCSP signer certificate. In the latter case the
12953 OCSP signer certificate must contain the OCSP signing extended key
12954 usage. This check is performed by attempting to match the OCSP
12955 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12956 in the OCSP_CERTID structures of the response.
12960 * Initial OCSP certificate verification added to OCSP_basic_verify()
12961 and related routines. This uses the standard OpenSSL certificate
12962 verify routines to perform initial checks (just CA validity) and
12963 to obtain the certificate chain. Then additional checks will be
12964 performed on the chain. Currently the root CA is checked to see
12965 if it is explicitly trusted for OCSP signing. This is used to set
12966 a root CA as a global signing root: that is any certificate that
12967 chains to that CA is an acceptable OCSP signing certificate.
12971 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12972 extensions from a separate configuration file.
12973 As when reading extensions from the main configuration file,
12974 the '-extensions ...' option may be used for specifying the
12977 *Massimiliano Pala <madwolf@comune.modena.it>*
12979 * New OCSP utility. Allows OCSP requests to be generated or
12980 read. The request can be sent to a responder and the output
12981 parsed, outputted or printed in text form. Not complete yet:
12982 still needs to check the OCSP response validity.
12986 * New subcommands for 'openssl ca':
12987 `openssl ca -status <serial>` prints the status of the cert with
12988 the given serial number (according to the index file).
12989 `openssl ca -updatedb` updates the expiry status of certificates
12992 *Massimiliano Pala <madwolf@comune.modena.it>*
12994 * New '-newreq-nodes' command option to CA.pl. This is like
12995 '-newreq', but calls 'openssl req' with the '-nodes' option
12996 so that the resulting key is not encrypted.
12998 *Damien Miller <djm@mindrot.org>*
13000 * New configuration for the GNU Hurd.
13002 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
13004 * Initial code to implement OCSP basic response verify. This
13005 is currently incomplete. Currently just finds the signer's
13006 certificate and verifies the signature on the response.
13010 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
13011 value of OPENSSLDIR. This is available via the new '-d' option
13012 to 'openssl version', and is also included in 'openssl version -a'.
13016 * Allowing defining memory allocation callbacks that will be given
13017 file name and line number information in additional arguments
13018 (a `const char*` and an int). The basic functionality remains, as
13019 well as the original possibility to just replace malloc(),
13020 realloc() and free() by functions that do not know about these
13021 additional arguments. To register and find out the current
13022 settings for extended allocation functions, the following
13023 functions are provided:
13025 CRYPTO_set_mem_ex_functions
13026 CRYPTO_set_locked_mem_ex_functions
13027 CRYPTO_get_mem_ex_functions
13028 CRYPTO_get_locked_mem_ex_functions
13030 These work the same way as CRYPTO_set_mem_functions and friends.
13031 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13032 extended allocation function is enabled.
13033 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13034 a conventional allocation function is enabled.
13036 *Richard Levitte, Bodo Moeller*
13038 * Finish off removing the remaining LHASH function pointer casts.
13039 There should no longer be any prototype-casting required when using
13040 the LHASH abstraction, and any casts that remain are "bugs". See
13041 the callback types and macros at the head of lhash.h for details
13042 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13046 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13047 If /dev/[u]random devices are not available or do not return enough
13048 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13050 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13051 /etc/entropy will be queried once each in this sequence, querying stops
13052 when enough entropy was collected without querying more sockets.
13056 * Change the Unix RAND_poll() variant to be able to poll several
13057 random devices, as specified by DEVRANDOM, until a sufficient amount
13058 of data has been collected. We spend at most 10 ms on each file
13059 (select timeout) and read in non-blocking mode. DEVRANDOM now
13060 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13061 (previously it was just the string "/dev/urandom"), so on typical
13062 platforms the 10 ms delay will never occur.
13063 Also separate out the Unix variant to its own file, rand_unix.c.
13064 For VMS, there's a currently-empty rand_vms.c.
13068 * Move OCSP client related routines to ocsp_cl.c. These
13069 provide utility functions which an application needing
13070 to issue a request to an OCSP responder and analyse the
13071 response will typically need: as opposed to those which an
13072 OCSP responder itself would need which will be added later.
13074 OCSP_request_sign() signs an OCSP request with an API similar
13075 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13076 response. OCSP_response_get1_basic() extracts basic response
13077 from response. OCSP_resp_find_status(): finds and extracts status
13078 information from an OCSP_CERTID structure (which will be created
13079 when the request structure is built). These are built from lower
13080 level functions which work on OCSP_SINGLERESP structures but
13081 won't normally be used unless the application wishes to examine
13082 extensions in the OCSP response for example.
13084 Replace nonce routines with a pair of functions.
13085 OCSP_request_add1_nonce() adds a nonce value and optionally
13086 generates a random value. OCSP_check_nonce() checks the
13087 validity of the nonce in an OCSP response.
13091 * Change function OCSP_request_add() to OCSP_request_add0_id().
13092 This doesn't copy the supplied OCSP_CERTID and avoids the
13093 need to free up the newly created id. Change return type
13094 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13095 This can then be used to add extensions to the request.
13096 Deleted OCSP_request_new(), since most of its functionality
13097 is now in OCSP_REQUEST_new() (and the case insensitive name
13098 clash) apart from the ability to set the request name which
13099 will be added elsewhere.
13103 * Update OCSP API. Remove obsolete extensions argument from
13104 various functions. Extensions are now handled using the new
13105 OCSP extension code. New simple OCSP HTTP function which
13106 can be used to send requests and parse the response.
13110 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13111 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13112 uses the special reorder version of SET OF to sort the attributes
13113 and reorder them to match the encoded order. This resolves a long
13114 standing problem: a verify on a PKCS7 structure just after signing
13115 it used to fail because the attribute order did not match the
13116 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13117 it uses the received order. This is necessary to tolerate some broken
13118 software that does not order SET OF. This is handled by encoding
13119 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13120 to produce the required SET OF.
13124 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13125 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13126 files to get correct declarations of the ASN.1 item variables.
13130 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13131 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13132 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13133 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13134 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13135 ASN1_ITEM and no wrapper functions.
13139 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13140 replace the old function pointer based I/O routines. Change most of
13141 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13145 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13146 lines, recognize more "algorithms" that can be deselected, and make
13147 it complain about algorithm deselection that isn't recognised.
13151 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13152 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13153 to use new functions. Add NO_ASN1_OLD which can be set to remove
13154 some old style ASN1 functions: this can be used to determine if old
13155 code will still work when these eventually go away.
13159 * New extension functions for OCSP structures, these follow the
13160 same conventions as certificates and CRLs.
13164 * New function X509V3_add1_i2d(). This automatically encodes and
13165 adds an extension. Its behaviour can be customised with various
13166 flags to append, replace or delete. Various wrappers added for
13167 certificates and CRLs.
13171 * Fix to avoid calling the underlying ASN1 print routine when
13172 an extension cannot be parsed. Correct a typo in the
13173 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13177 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13178 entries for variables.
13182 * Add functionality to `apps/openssl.c` for detecting locking
13183 problems: As the program is single-threaded, all we have
13184 to do is register a locking callback using an array for
13185 storing which locks are currently held by the program.
13189 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13190 SSL_get_ex_data_X509_STORE_idx(), which is used in
13191 ssl_verify_cert_chain() and thus can be called at any time
13192 during TLS/SSL handshakes so that thread-safety is essential.
13193 Unfortunately, the ex_data design is not at all suited
13194 for multi-threaded use, so it probably should be abolished.
13198 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13200 *Broadcom, tweaked and integrated by Geoff Thorpe*
13202 * Move common extension printing code to new function
13203 X509V3_print_extensions(). Reorganise OCSP print routines and
13204 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13208 * New function X509_signature_print() to remove duplication in some
13213 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13214 set (this was treated exactly the same as SET OF previously). This
13215 is used to reorder the STACK representing the structure to match the
13216 encoding. This will be used to get round a problem where a PKCS7
13217 structure which was signed could not be verified because the STACK
13218 order did not reflect the encoded order.
13222 * Reimplement the OCSP ASN1 module using the new code.
13226 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13227 for its ASN1 operations. The old style function pointers still exist
13228 for now but they will eventually go away.
13232 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13233 completely replaces the old ASN1 functionality with a table driven
13234 encoder and decoder which interprets an ASN1_ITEM structure describing
13235 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13236 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13237 has also been converted to the new form.
13241 * Change BN_mod_exp_recp so that negative moduli are tolerated
13242 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13243 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13244 for negative moduli.
13248 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13249 of not touching the result's sign bit.
13253 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13258 * Changed the LHASH code to use prototypes for callbacks, and created
13259 macros to declare and implement thin (optionally static) functions
13260 that provide type-safety and avoid function pointer casting for the
13261 type-specific callbacks.
13265 * Added Kerberos Cipher Suites to be used with TLS, as written in
13267 *Veers Staats <staatsvr@asc.hpc.mil>,
13268 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13270 * Reformat the FAQ so the different questions and answers can be divided
13271 in sections depending on the subject.
13275 * Have the zlib compression code load ZLIB.DLL dynamically under
13280 * New function BN_mod_sqrt for computing square roots modulo a prime
13281 (using the probabilistic Tonelli-Shanks algorithm unless
13282 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13283 be handled deterministically).
13285 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13287 * Make BN_mod_inverse faster by explicitly handling small quotients
13288 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13289 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13293 * New function BN_kronecker.
13297 * Fix BN_gcd so that it works on negative inputs; the result is
13298 positive unless both parameters are zero.
13299 Previously something reasonably close to an infinite loop was
13300 possible because numbers could be growing instead of shrinking
13301 in the implementation of Euclid's algorithm.
13305 * Fix BN_is_word() and BN_is_one() macros to take into account the
13306 sign of the number in question.
13308 Fix BN_is_word(a,w) to work correctly for w == 0.
13310 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13311 because its test if the absolute value of 'a' equals 'w'.
13312 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13313 it exists mostly for use in the implementations of BN_is_zero(),
13314 BN_is_one(), and BN_is_word().
13318 * New function BN_swap.
13322 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13323 the exponentiation functions are more likely to produce reasonable
13324 results on negative inputs.
13328 * Change BN_mod_mul so that the result is always non-negative.
13329 Previously, it could be negative if one of the factors was negative;
13330 I don't think anyone really wanted that behaviour.
13334 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13335 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13336 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13337 and add new functions:
13346 BN_mod_lshift1_quick
13348 BN_mod_lshift_quick
13350 These functions always generate non-negative results.
13352 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13353 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13355 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13356 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13357 be reduced modulo `m`.
13359 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13362 The following entry accidentally appeared in the CHANGES file
13363 distributed with OpenSSL 0.9.7. The modifications described in
13364 it do *not* apply to OpenSSL 0.9.7.
13366 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13367 was actually never needed) and in BN_mul(). The removal in BN_mul()
13368 required a small change in bn_mul_part_recursive() and the addition
13369 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13370 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13371 bn_sub_words() and bn_add_words() except they take arrays with
13377 * In 'openssl passwd', verify passwords read from the terminal
13378 unless the '-salt' option is used (which usually means that
13379 verification would just waste user's time since the resulting
13380 hash is going to be compared with some given password hash)
13381 or the new '-noverify' option is used.
13383 This is an incompatible change, but it does not affect
13384 non-interactive use of 'openssl passwd' (passwords on the command
13385 line, '-stdin' option, '-in ...' option) and thus should not
13386 cause any problems.
13390 * Remove all references to RSAref, since there's no more need for it.
13394 * Make DSO load along a path given through an environment variable
13395 (SHLIB_PATH) with shl_load().
13399 * Constify the ENGINE code as a result of BIGNUM constification.
13400 Also constify the RSA code and most things related to it. In a
13401 few places, most notable in the depth of the ASN.1 code, ugly
13402 casts back to non-const were required (to be solved at a later
13407 * Make it so the openssl application has all engines loaded by default.
13411 * Constify the BIGNUM routines a little more.
13415 * Add the following functions:
13417 ENGINE_load_cswift()
13419 ENGINE_load_atalla()
13420 ENGINE_load_nuron()
13421 ENGINE_load_builtin_engines()
13423 That way, an application can itself choose if external engines that
13424 are built-in in OpenSSL shall ever be used or not. The benefit is
13425 that applications won't have to be linked with libdl or other dso
13426 libraries unless it's really needed.
13428 Changed 'openssl engine' to load all engines on demand.
13429 Changed the engine header files to avoid the duplication of some
13430 declarations (they differed!).
13434 * 'openssl engine' can now list capabilities.
13438 * Better error reporting in 'openssl engine'.
13442 * Never call load_dh_param(NULL) in s_server.
13446 * Add engine application. It can currently list engines by name and
13447 identity, and test if they are actually available.
13451 * Improve RPM specification file by forcing symbolic linking and making
13452 sure the installed documentation is also owned by root.root.
13454 *Damien Miller <djm@mindrot.org>*
13456 * Give the OpenSSL applications more possibilities to make use of
13457 keys (public as well as private) handled by engines.
13461 * Add OCSP code that comes from CertCo.
13465 * Add VMS support for the Rijndael code.
13469 * Added untested support for Nuron crypto accelerator.
13473 * Add support for external cryptographic devices. This code was
13474 previously distributed separately as the "engine" branch.
13476 *Geoff Thorpe, Richard Levitte*
13478 * Rework the filename-translation in the DSO code. It is now possible to
13479 have far greater control over how a "name" is turned into a filename
13480 depending on the operating environment and any oddities about the
13481 different shared library filenames on each system.
13485 * Support threads on FreeBSD-elf in Configure.
13489 * Fix for SHA1 assembly problem with MASM: it produces
13490 warnings about corrupt line number information when assembling
13491 with debugging information. This is caused by the overlapping
13494 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13497 NCONF_get_number() has no error checking at all. As a replacement,
13498 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13499 promoted strongly. The old NCONF_get_number is kept around for
13500 binary backward compatibility.
13501 Make it possible for methods to load from something other than a BIO,
13502 by providing a function pointer that is given a name instead of a BIO.
13503 For example, this could be used to load configuration data from an
13508 * Fix for non blocking accept BIOs. Added new I/O special reason
13509 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13510 with non blocking I/O was not possible because no retry code was
13511 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13516 * Added the beginnings of Rijndael support.
13520 * Fix for bug in DirectoryString mask setting. Add support for
13521 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13522 to allow certificate printing to more controllable, additional
13523 'certopt' option to 'x509' to allow new printing options to be
13528 * Clean old EAY MD5 hack from e_os.h.
13532 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13534 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13535 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13537 *Joe Orton, Steve Henson*
13539 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13541 * Fix additional bug revealed by the NISCC test suite:
13543 Stop bug triggering large recursion when presented with
13544 certain ASN.1 tags ([CVE-2003-0851])
13548 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13550 * Fix various bugs revealed by running the NISCC test suite:
13552 Stop out of bounds reads in the ASN1 code when presented with
13553 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13555 If verify callback ignores invalid public key errors don't try to check
13556 certificate signature with the NULL public key.
13560 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13561 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13566 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13567 extra data after the compression methods not only for TLS 1.0
13568 but also for SSL 3.0 (as required by the specification).
13570 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13572 * Change X509_certificate_type() to mark the key as exported/exportable
13573 when it's 512 *bits* long, not 512 bytes.
13577 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13579 * Countermeasure against the Klima-Pokorny-Rosa extension of
13580 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13581 a protocol version number mismatch like a decryption error
13582 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13586 * Turn on RSA blinding by default in the default implementation
13587 to avoid a timing attack. Applications that don't want it can call
13588 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13589 They would be ill-advised to do so in most cases.
13591 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13593 * Change RSA blinding code so that it works when the PRNG is not
13594 seeded (in this case, the secret RSA exponent is abused as
13595 an unpredictable seed -- if it is not unpredictable, there
13596 is no point in blinding anyway). Make RSA blinding thread-safe
13597 by remembering the creator's thread ID in rsa->blinding and
13598 having all other threads use local one-time blinding factors
13599 (this requires more computation than sharing rsa->blinding, but
13600 avoids excessive locking; and if an RSA object is not shared
13601 between threads, blinding will still be very fast).
13605 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13607 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13608 via timing by performing a MAC computation even if incorrect
13609 block cipher padding has been found. This is a countermeasure
13610 against active attacks where the attacker has to distinguish
13611 between bad padding and a MAC verification error. ([CVE-2003-0078])
13613 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13614 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13615 Martin Vuagnoux (EPFL, Ilion)*
13617 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13619 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13620 memory from its contents. This is done with a counter that will
13621 place alternating values in each byte. This can be used to solve
13622 two issues: 1) the removal of calls to memset() by highly optimizing
13623 compilers, and 2) cleansing with other values than 0, since those can
13624 be read through on certain media, for example a swap space on disk.
13628 * Bugfix: client side session caching did not work with external caching,
13629 because the session->cipher setting was not restored when reloading
13630 from the external cache. This problem was masked, when
13631 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13632 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13636 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13637 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13639 *Zeev Lieber <zeev-l@yahoo.com>*
13641 * Undo an undocumented change introduced in 0.9.6e which caused
13642 repeated calls to OpenSSL_add_all_ciphers() and
13643 OpenSSL_add_all_digests() to be ignored, even after calling
13648 * Change the default configuration reader to deal with last line not
13649 being properly terminated.
13653 * Change X509_NAME_cmp() so it applies the special rules on handling
13654 DN values that are of type PrintableString, as well as RDNs of type
13655 emailAddress where the value has the type ia5String.
13657 *stefank@valicert.com via Richard Levitte*
13659 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13660 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13661 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13662 the bitwise-OR of the two for use by the majority of applications
13663 wanting this behaviour, and update the docs. The documented
13664 behaviour and actual behaviour were inconsistent and had been
13665 changing anyway, so this is more a bug-fix than a behavioural
13668 *Geoff Thorpe, diagnosed by Nadav Har'El*
13670 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13671 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13675 * Fix initialization code race conditions in
13676 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13677 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13678 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13679 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13680 ssl2_get_cipher_by_char(),
13681 ssl3_get_cipher_by_char().
13683 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13685 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13686 the cached sessions are flushed, as the remove_cb() might use ex_data
13687 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13688 (see [openssl.org #212]).
13690 *Geoff Thorpe, Lutz Jaenicke*
13692 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13693 length, instead of the encoding length to d2i_ASN1_OBJECT.
13697 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13699 * [In 0.9.6g-engine release:]
13700 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13702 *Lynn Gazis <lgazis@rainbow.com>*
13704 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13706 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13707 and get fix the header length calculation.
13708 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13709 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13711 * Use proper error handling instead of 'assertions' in buffer
13712 overflow checks added in 0.9.6e. This prevents DoS (the
13713 assertions could call abort()).
13715 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13717 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13719 * Add various sanity checks to asn1_get_length() to reject
13720 the ASN1 length bytes if they exceed sizeof(long), will appear
13721 negative or the content length exceeds the length of the
13724 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13726 * Fix cipher selection routines: ciphers without encryption had no flags
13727 for the cipher strength set and where therefore not handled correctly
13728 by the selection routines (PR #130).
13732 * Fix EVP_dsa_sha macro.
13737 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13738 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13739 that was added in OpenSSL 0.9.6d.
13741 As the countermeasure turned out to be incompatible with some
13742 broken SSL implementations, the new option is part of SSL_OP_ALL.
13743 SSL_OP_ALL is usually employed when compatibility with weird SSL
13744 implementations is desired (e.g. '-bugs' option to 's_client' and
13745 's_server'), so the new option is automatically set in many
13750 * Changes in security patch:
13752 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13753 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13754 Air Force Materiel Command, USAF, under agreement number
13757 * Add various sanity checks to asn1_get_length() to reject
13758 the ASN1 length bytes if they exceed sizeof(long), will appear
13759 negative or the content length exceeds the length of the
13760 supplied buffer. ([CVE-2002-0659])
13762 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13764 * Assertions for various potential buffer overflows, not known to
13765 happen in practice.
13767 *Ben Laurie (CHATS)*
13769 * Various temporary buffers to hold ASCII versions of integers were
13770 too small for 64 bit platforms. ([CVE-2002-0655])
13771 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13773 * Remote buffer overflow in SSL3 protocol - an attacker could
13774 supply an oversized session ID to a client. ([CVE-2002-0656])
13776 *Ben Laurie (CHATS)*
13778 * Remote buffer overflow in SSL2 protocol - an attacker could
13779 supply an oversized client master key. ([CVE-2002-0656])
13781 *Ben Laurie (CHATS)*
13783 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13785 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13786 encoded as NULL) with id-dsa-with-sha1.
13788 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13790 * Check various `X509_...()` return values in `apps/req.c`.
13792 *Nils Larsch <nla@trustcenter.de>*
13794 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13795 an end-of-file condition would erroneously be flagged, when the CRLF
13796 was just at the end of a processed block. The bug was discovered when
13797 processing data through a buffering memory BIO handing the data to a
13798 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13799 <ptsekov@syntrex.com> and Nedelcho Stanev.
13803 * Implement a countermeasure against a vulnerability recently found
13804 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13805 before application data chunks to avoid the use of known IVs
13806 with data potentially chosen by the attacker.
13810 * Fix length checks in ssl3_get_client_hello().
13814 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13815 to prevent ssl3_read_internal() from incorrectly assuming that
13816 ssl3_read_bytes() found application data while handshake
13817 processing was enabled when in fact s->s3->in_read_app_data was
13818 merely automatically cleared during the initial handshake.
13820 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13822 * Fix object definitions for Private and Enterprise: they were not
13823 recognized in their shortname (=lowercase) representation. Extend
13824 obj_dat.pl to issue an error when using undefined keywords instead
13825 of silently ignoring the problem (Svenning Sorensen
13826 <sss@sss.dnsalias.net>).
13830 * Fix DH_generate_parameters() so that it works for 'non-standard'
13831 generators, i.e. generators other than 2 and 5. (Previously, the
13832 code did not properly initialise the 'add' and 'rem' values to
13833 BN_generate_prime().)
13835 In the new general case, we do not insist that 'generator' is
13836 actually a primitive root: This requirement is rather pointless;
13837 a generator of the order-q subgroup is just as good, if not
13842 * Map new X509 verification errors to alerts. Discovered and submitted by
13843 Tom Wu <tom@arcot.com>.
13847 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13848 returning non-zero before the data has been completely received
13849 when using non-blocking I/O.
13851 *Bodo Moeller; problem pointed out by John Hughes*
13853 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13855 *Ben Laurie, Lutz Jaenicke*
13857 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13858 Yoram Zahavi <YoramZ@gilian.com>).
13862 * Add information about CygWin 1.3 and on, and preserve proper
13863 configuration for the versions before that.
13865 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13867 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13868 check whether we deal with a copy of a session and do not delete from
13869 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13870 <izhar@checkpoint.com>.
13874 * Do not store session data into the internal session cache, if it
13875 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13876 flag is set). Proposed by Aslam <aslam@funk.com>.
13880 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13885 * [In 0.9.6d-engine release:]
13886 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13888 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13890 * Add the configuration target linux-s390x.
13892 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13894 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13895 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13896 variable as an indication that a ClientHello message has been
13897 received. As the flag value will be lost between multiple
13898 invocations of ssl3_accept when using non-blocking I/O, the
13899 function may not be aware that a handshake has actually taken
13900 place, thus preventing a new session from being added to the
13903 To avoid this problem, we now set s->new_session to 2 instead of
13904 using a local variable.
13906 *Lutz Jaenicke, Bodo Moeller*
13908 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13909 if the SSL_R_LENGTH_MISMATCH error is detected.
13911 *Geoff Thorpe, Bodo Moeller*
13913 * New 'shared_ldflag' column in Configure platform table.
13917 * Fix EVP_CIPHER_mode macro.
13919 *"Dan S. Camper" <dan@bti.net>*
13921 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13922 type, we must throw them away by setting rr->length to 0.
13924 *D P Chang <dpc@qualys.com>*
13926 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13928 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13929 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13930 worked incorrectly for those cases where range = `10..._2` and
13931 `3*range` is two bits longer than range.)
13935 * Only add signing time to PKCS7 structures if it is not already
13940 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13941 OBJ_ld_ce should be OBJ_id_ce.
13942 Also some ip-pda OIDs in crypto/objects/objects.txt were
13943 incorrect (cf. RFC 3039).
13945 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13947 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13948 returns early because it has nothing to do.
13950 *Andy Schneider <andy.schneider@bjss.co.uk>*
13952 * [In 0.9.6c-engine release:]
13953 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13955 *Andy Schneider <andy.schneider@bjss.co.uk>*
13957 * [In 0.9.6c-engine release:]
13958 Add support for Cryptographic Appliance's keyserver technology.
13959 (Use engine 'keyclient')
13961 *Cryptographic Appliances and Geoff Thorpe*
13963 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13964 is called via tools/c89.sh because arguments have to be
13965 rearranged (all '-L' options must appear before the first object
13968 *Richard Shapiro <rshapiro@abinitio.com>*
13970 * [In 0.9.6c-engine release:]
13971 Add support for Broadcom crypto accelerator cards, backported
13974 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13976 * [In 0.9.6c-engine release:]
13977 Add support for SureWare crypto accelerator cards from
13978 Baltimore Technologies. (Use engine 'sureware')
13980 *Baltimore Technologies and Mark Cox*
13982 * [In 0.9.6c-engine release:]
13983 Add support for crypto accelerator cards from Accelerated
13984 Encryption Processing, www.aep.ie. (Use engine 'aep')
13986 *AEP Inc. and Mark Cox*
13988 * Add a configuration entry for gcc on UnixWare.
13990 *Gary Benson <gbenson@redhat.com>*
13992 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13993 messages are stored in a single piece (fixed-length part and
13994 variable-length part combined) and fix various bugs found on the way.
13998 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13999 instead. BIO_gethostbyname() does not know what timeouts are
14000 appropriate, so entries would stay in cache even when they have
14002 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
14004 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
14005 faced with a pathologically small ClientHello fragment that does
14006 not contain client_version: Instead of aborting with an error,
14007 simply choose the highest available protocol version (i.e.,
14008 TLS 1.0 unless it is disabled). In practice, ClientHello
14009 messages are never sent like this, but this change gives us
14010 strictly correct behaviour at least for TLS.
14014 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
14015 never resets s->method to s->ctx->method when called from within
14016 one of the SSL handshake functions.
14018 *Bodo Moeller; problem pointed out by Niko Baric*
14020 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
14021 (sent using the client's version number) if client_version is
14022 smaller than the protocol version in use. Also change
14023 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
14024 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
14025 the client will at least see that alert.
14029 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14034 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14035 client receives HelloRequest while in a handshake.
14037 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14039 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14040 should end in 'break', not 'goto end' which circumvents various
14041 cleanups done in state SSL_ST_OK. But session related stuff
14042 must be disabled for SSL_ST_OK in the case that we just sent a
14045 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14046 before just sending a HelloRequest.
14048 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14050 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14051 reveal whether illegal block cipher padding was found or a MAC
14052 verification error occurred. (Neither SSLerr() codes nor alerts
14053 are directly visible to potential attackers, but the information
14054 may leak via logfiles.)
14056 Similar changes are not required for the SSL 2.0 implementation
14057 because the number of padding bytes is sent in clear for SSL 2.0,
14058 and the extra bytes are just ignored. However ssl/s2_pkt.c
14059 failed to verify that the purported number of padding bytes is in
14064 * Add OpenUNIX-8 support including shared libraries
14065 (Boyd Lynn Gerber <gerberb@zenez.com>).
14069 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14070 'wristwatch attack' using huge encoding parameters (cf.
14071 James H. Manger's CRYPTO 2001 paper). Note that the
14072 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14073 encoding parameters and hence was not vulnerable.
14077 * BN_sqr() bug fix.
14079 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14081 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14082 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14083 followed by modular reduction.
14085 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14087 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14088 equivalent based on BN_pseudo_rand() instead of BN_rand().
14092 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14093 This function was broken, as the check for a new client hello message
14094 to handle SGC did not allow these large messages.
14095 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14099 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14103 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14104 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14108 * Rework the configuration and shared library support for Tru64 Unix.
14109 The configuration part makes use of modern compiler features and
14110 still retains old compiler behavior for those that run older versions
14111 of the OS. The shared library support part includes a variant that
14112 uses the RPATH feature, and is available through the special
14113 configuration target "alpha-cc-rpath", which will never be selected
14116 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14118 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14119 with the same message size as in ssl3_get_certificate_request().
14120 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14121 messages might inadvertently be reject as too long.
14123 *Petr Lampa <lampa@fee.vutbr.cz>*
14125 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14129 * Modified SSL library such that the verify_callback that has been set
14130 specifically for an SSL object with SSL_set_verify() is actually being
14131 used. Before the change, a verify_callback set with this function was
14132 ignored and the verify_callback() set in the SSL_CTX at the time of
14133 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14134 to allow the necessary settings.
14138 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14139 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14140 done automatically (in contradiction to the requirements of the C
14141 standard). This made problems when used from OpenSSH.
14145 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14146 dh->length and always used
14148 BN_rand_range(priv_key, dh->p).
14150 BN_rand_range() is not necessary for Diffie-Hellman, and this
14151 specific range makes Diffie-Hellman unnecessarily inefficient if
14152 dh->length (recommended exponent length) is much smaller than the
14153 length of dh->p. We could use BN_rand_range() if the order of
14154 the subgroup was stored in the DH structure, but we only have
14159 BN_rand(priv_key, l, ...)
14161 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14168 RSA_eay_public_encrypt
14169 RSA_eay_private_decrypt
14170 RSA_eay_private_encrypt (signing)
14171 RSA_eay_public_decrypt (signature verification)
14173 (default implementations for RSA_public_encrypt,
14174 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14175 always reject numbers >= n.
14179 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14180 to synchronize access to 'locking_thread'. This is necessary on
14181 systems where access to 'locking_thread' (an 'unsigned long'
14182 variable) is not atomic.
14186 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14187 *before* setting the 'crypto_lock_rand' flag. The previous code had
14188 a race condition if 0 is a valid thread ID.
14190 *Travis Vitek <vitek@roguewave.com>*
14192 * Add support for shared libraries under Irix.
14194 *Albert Chin-A-Young <china@thewrittenword.com>*
14196 * Add configuration option to build on Linux on both big-endian and
14197 little-endian MIPS.
14199 *Ralf Baechle <ralf@uni-koblenz.de>*
14201 * Add the possibility to create shared libraries on HP-UX.
14205 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14207 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14208 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14209 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14210 PRNG state recovery was possible based on the output of
14211 one PRNG request appropriately sized to gain knowledge on
14212 'md' followed by enough consecutive 1-byte PRNG requests
14213 to traverse all of 'state'.
14215 1. When updating 'md_local' (the current thread's copy of 'md')
14216 during PRNG output generation, hash all of the previous
14217 'md_local' value, not just the half used for PRNG output.
14219 2. Make the number of bytes from 'state' included into the hash
14220 independent from the number of PRNG bytes requested.
14222 The first measure alone would be sufficient to avoid
14223 Markku-Juhani's attack. (Actually it had never occurred
14224 to me that the half of 'md_local' used for chaining was the
14225 half from which PRNG output bytes were taken -- I had always
14226 assumed that the secret half would be used.) The second
14227 measure makes sure that additional data from 'state' is never
14228 mixed into 'md_local' in small portions; this heuristically
14229 further strengthens the PRNG.
14233 * Fix crypto/bn/asm/mips3.s.
14237 * When only the key is given to "enc", the IV is undefined. Print out
14238 an error message in this case.
14242 * Handle special case when X509_NAME is empty in X509 printing routines.
14246 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14247 positive and less than q.
14251 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14252 used: it isn't thread safe and the add_lock_callback should handle
14255 *Paul Rose <Paul.Rose@bridge.com>*
14257 * Verify that incoming data obeys the block size in
14258 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14264 *Ulf Möller, Bodo Möller*
14266 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14267 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14268 when fixing the server behaviour for backwards-compatible 'client
14269 hello' messages. (Note that the attack is impractical against
14270 SSL 3.0 and TLS 1.0 anyway because length and version checking
14271 means that the probability of guessing a valid ciphertext is
14272 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14275 Before 0.9.5, the countermeasure (hide the error by generating a
14276 random 'decryption result') did not work properly because
14277 ERR_clear_error() was missing, meaning that SSL_get_error() would
14278 detect the supposedly ignored error.
14280 Both problems are now fixed.
14284 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14285 (previously it was 1024).
14289 * Fix for compatibility mode trust settings: ignore trust settings
14290 unless some valid trust or reject settings are present.
14294 * Fix for blowfish EVP: its a variable length cipher.
14298 * Fix various bugs related to DSA S/MIME verification. Handle missing
14299 parameters in DSA public key structures and return an error in the
14300 DSA routines if parameters are absent.
14304 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14305 in the current directory if neither $RANDFILE nor $HOME was set.
14306 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14307 caused some confusion to Windows users who haven't defined $HOME.
14308 Thus RAND_file_name() is changed again: e_os.h can define a
14309 DEFAULT_HOME, which will be used if $HOME is not set.
14310 For Windows, we use "C:"; on other platforms, we still require
14311 environment variables.
14313 * Move 'if (!initialized) RAND_poll()' into regions protected by
14314 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14315 having multiple threads call RAND_poll() concurrently.
14319 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14320 combination of a flag and a thread ID variable.
14321 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14322 flag), *other* threads can enter ssleay_add_bytes without obeying
14323 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14324 that they do not hold after the first thread unsets add_do_not_lock).
14328 * Change bctest again: '-x' expressions are not available in all
14329 versions of 'test'.
14333 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14335 * Fix a couple of memory leaks in PKCS7_dataDecode()
14337 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14339 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14340 the default extension for executables, if any. Also, make the perl
14341 scripts that use symlink() to test if it really exists and use "cp"
14342 if it doesn't. All this made OpenSSL compilable and installable in
14347 * Fix for asn1_GetSequence() for indefinite length constructed data.
14348 If SEQUENCE is length is indefinite just set c->slen to the total
14349 amount of data available.
14351 *Steve Henson, reported by shige@FreeBSD.org*
14353 *This change does not apply to 0.9.7.*
14355 * Change bctest to avoid here-documents inside command substitution
14356 (workaround for FreeBSD /bin/sh bug).
14357 For compatibility with Ultrix, avoid shell functions (introduced
14358 in the bctest version that searches along $PATH).
14362 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14363 with des_encrypt() defined on some operating systems, like Solaris
14368 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14369 On the Importance of Eliminating Errors in Cryptographic
14370 Computations, J. Cryptology 14 (2001) 2, 101-119,
14371 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14375 * MIPS assembler BIGNUM division bug fix.
14379 * Disabled incorrect Alpha assembler code.
14383 * Fix PKCS#7 decode routines so they correctly update the length
14384 after reading an EOC for the EXPLICIT tag.
14388 *This change does not apply to 0.9.7.*
14390 * Fix bug in PKCS#12 key generation routines. This was triggered
14391 if a 3DES key was generated with a 0 initial byte. Include
14392 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14393 (but broken) behaviour.
14397 * Enhance bctest to search for a working bc along $PATH and print
14400 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14402 * Fix memory leaks in err.c: free err_data string if necessary;
14403 don't write to the wrong index in ERR_set_error_data.
14407 * Implement ssl23_peek (analogous to ssl23_read), which previously
14412 * Replace rdtsc with `_emit` statements for VC++ version 5.
14414 *Jeremy Cooper <jeremy@baymoo.org>*
14416 * Make it possible to reuse SSLv2 sessions.
14420 * In copy_email() check for >= 0 as a return value for
14421 X509_NAME_get_index_by_NID() since 0 is a valid index.
14423 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14425 * Avoid coredump with unsupported or invalid public keys by checking if
14426 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14427 PKCS7_verify() fails with non detached data.
14431 * Don't use getenv in library functions when run as setuid/setgid.
14432 New function OPENSSL_issetugid().
14436 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14437 due to incorrect handling of multi-threading:
14439 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14441 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14443 3. Count how many times MemCheck_off() has been called so that
14444 nested use can be treated correctly. This also avoids
14445 inband-signalling in the previous code (which relied on the
14446 assumption that thread ID 0 is impossible).
14450 * Add "-rand" option also to s_client and s_server.
14454 * Fix CPU detection on Irix 6.x.
14455 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14456 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14458 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14463 *This change does not apply to 0.9.7.*
14465 * Use the cached encoding of an X509_NAME structure rather than
14466 copying it. This is apparently the reason for the libsafe "errors"
14467 but the code is actually correct.
14471 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14472 Bleichenbacher's DSA attack.
14473 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14474 to be set and top=0 forces the highest bit to be set; top=-1 is new
14475 and leaves the highest bit random.
14477 *Ulf Moeller, Bodo Moeller*
14479 * In the `NCONF_...`-based implementations for `CONF_...` queries
14480 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14481 a temporary CONF structure with the data component set to NULL
14482 (which gives segmentation faults in lh_retrieve).
14483 Instead, use NULL for the CONF pointer in CONF_get_string and
14484 CONF_get_number (which may use environment variables) and directly
14485 return NULL from CONF_get_section.
14489 * Fix potential buffer overrun for EBCDIC.
14493 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14494 keyUsage if basicConstraints absent for a CA.
14498 * Make SMIME_write_PKCS7() write mail header values with a format that
14499 is more generally accepted (no spaces before the semicolon), since
14500 some programs can't parse those values properly otherwise. Also make
14501 sure BIO's that break lines after each write do not create invalid
14506 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14507 macros previously used would not encode an empty SEQUENCE OF
14508 and break the signature.
14512 *This change does not apply to 0.9.7.*
14514 * Zero the premaster secret after deriving the master secret in
14519 * Add some EVP_add_digest_alias registrations (as found in
14520 OpenSSL_add_all_digests()) to SSL_library_init()
14521 aka OpenSSL_add_ssl_algorithms(). This provides improved
14522 compatibility with peers using X.509 certificates
14523 with unconventional AlgorithmIdentifier OIDs.
14527 * Fix for Irix with NO_ASM.
14529 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14531 * ./config script fixes.
14533 *Ulf Moeller, Richard Levitte*
14535 * Fix 'openssl passwd -1'.
14539 * Change PKCS12_key_gen_asc() so it can cope with non null
14540 terminated strings whose length is passed in the passlen
14541 parameter, for example from PEM callbacks. This was done
14542 by adding an extra length parameter to asc2uni().
14544 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14546 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14547 call failed, free the DSA structure.
14551 * Fix to uni2asc() to cope with zero length Unicode strings.
14552 These are present in some PKCS#12 files.
14556 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14557 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14558 when writing a 32767 byte record.
14560 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14562 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14563 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14565 (RSA objects have a reference count access to which is protected
14566 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14567 so they are meant to be shared between threads.)
14568 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14569 "Reddie, Steven" <Steven.Reddie@ca.com>*
14571 * Fix a deadlock in CRYPTO_mem_leaks().
14575 * Use better test patterns in bntest.
14579 * rand_win.c fix for Borland C.
14583 * BN_rshift bugfix for n == 0.
14587 * Add a 'bctest' script that checks for some known 'bc' bugs
14588 so that 'make test' does not abort just because 'bc' is broken.
14592 * Store verify_result within SSL_SESSION also for client side to
14593 avoid potential security hole. (Re-used sessions on the client side
14594 always resulted in verify_result==X509_V_OK, not using the original
14595 result of the server certificate verification.)
14599 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14600 SSL3_RT_APPLICATION_DATA, return 0.
14601 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14606 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14607 releases, have been re-implemented by renaming the previous
14608 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14609 and ssl3_read_internal, respectively, and adding 'peek' parameters
14610 to them. The new ssl[23]_{read,peek} functions are calls to
14611 ssl[23]_read_internal with the 'peek' flag set appropriately.
14612 A 'peek' parameter has also been added to ssl3_read_bytes, which
14613 does the actual work for ssl3_read_internal.
14617 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14618 the method-specific "init()" handler. Also clean up ex_data after
14619 calling the method-specific "finish()" handler. Previously, this was
14620 happening the other way round.
14624 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14625 The previous value, 12, was not always sufficient for BN_mod_exp().
14629 * Make sure that shared libraries get the internal name engine with
14630 the full version number and not just 0. This should mark the
14631 shared libraries as not backward compatible. Of course, this should
14632 be changed again when we can guarantee backward binary compatibility.
14636 * Fix typo in get_cert_by_subject() in by_dir.c
14638 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14640 * Rework the system to generate shared libraries:
14642 - Make note of the expected extension for the shared libraries and
14643 if there is a need for symbolic links from for example libcrypto.so.0
14644 to libcrypto.so.0.9.7. There is extended info in Configure for
14647 - Make as few rebuilds of the shared libraries as possible.
14649 - Still avoid linking the OpenSSL programs with the shared libraries.
14651 - When installing, install the shared libraries separately from the
14656 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14658 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14659 and not in SSL_clear because the latter is also used by the
14660 accept/connect functions; previously, the settings made by
14661 SSL_set_read_ahead would be lost during the handshake.
14663 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14665 * Correct util/mkdef.pl to be selective about disabled algorithms.
14666 Previously, it would create entries for disabled algorithms no
14671 * Added several new manual pages for SSL_* function.
14675 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14677 * In ssl23_get_client_hello, generate an error message when faced
14678 with an initial SSL 3.0/TLS record that is too small to contain the
14679 first two bytes of the ClientHello message, i.e. client_version.
14680 (Note that this is a pathologic case that probably has never happened
14681 in real life.) The previous approach was to use the version number
14682 from the record header as a substitute; but our protocol choice
14683 should not depend on that one because it is not authenticated
14684 by the Finished messages.
14688 * More robust randomness gathering functions for Windows.
14690 *Jeffrey Altman <jaltman@columbia.edu>*
14692 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14693 not set then we don't setup the error code for issuer check errors
14694 to avoid possibly overwriting other errors which the callback does
14695 handle. If an application does set the flag then we assume it knows
14696 what it is doing and can handle the new informational codes
14701 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14702 a general "ANY" type, as such it should be able to decode anything
14703 including tagged types. However it didn't check the class so it would
14704 wrongly interpret tagged types in the same way as their universal
14705 counterpart and unknown types were just rejected. Changed so that the
14706 tagged and unknown types are handled in the same way as a SEQUENCE:
14707 that is the encoding is stored intact. There is also a new type
14708 "V_ASN1_OTHER" which is used when the class is not universal, in this
14709 case we have no idea what the actual type is so we just lump them all
14714 * On VMS, stdout may very well lead to a file that is written to
14715 in a record-oriented fashion. That means that every write() will
14716 write a separate record, which will be read separately by the
14717 programs trying to read from it. This can be very confusing.
14719 The solution is to put a BIO filter in the way that will buffer
14720 text until a linefeed is reached, and then write everything a
14721 line at a time, so every record written will be an actual line,
14722 not chunks of lines and not (usually doesn't happen, but I've
14723 seen it once) several lines in one record. BIO_f_linebuffer() is
14726 Currently, it's a VMS-only method, because that's where it has
14727 been tested well enough.
14731 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14732 it can return incorrect results.
14733 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14734 but it was in 0.9.6-beta[12].)
14738 * Disable the check for content being present when verifying detached
14739 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14740 include zero length content when signing messages.
14744 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14745 BIO_ctrl (for BIO pairs).
14749 * Add DSO method for VMS.
14753 * Bug fix: Montgomery multiplication could produce results with the
14758 * Add RPM specification openssl.spec and modify it to build three
14759 packages. The default package contains applications, application
14760 documentation and run-time libraries. The devel package contains
14761 include files, static libraries and function documentation. The
14762 doc package contains the contents of the doc directory. The original
14763 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14767 * Add a large number of documentation files for many SSL routines.
14769 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14771 * Add a configuration entry for Sony News 4.
14773 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14775 * Don't set the two most significant bits to one when generating a
14776 random number < q in the DSA library.
14780 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14781 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14782 the underlying transport is blocking) if a handshake took place.
14783 (The default behaviour is needed by applications such as s_client
14784 and s_server that use select() to determine when to use SSL_read;
14785 but for applications that know in advance when to expect data, it
14786 just makes things more complicated.)
14790 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14795 * Add a few more EBCDIC conditionals that make `req` and `x509`
14796 work better on such systems.
14798 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14800 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14801 Update PKCS12_parse() so it copies the friendlyName and the
14802 keyid to the certificates aux info.
14806 * Fix bug in PKCS7_verify() which caused an infinite loop
14807 if there was more than one signature.
14809 *Sven Uszpelkat <su@celocom.de>*
14811 * Major change in util/mkdef.pl to include extra information
14812 about each symbol, as well as presenting variables as well
14813 as functions. This change means that there's n more need
14814 to rebuild the .num files when some algorithms are excluded.
14818 * Allow the verify time to be set by an application,
14819 rather than always using the current time.
14823 * Phase 2 verify code reorganisation. The certificate
14824 verify code now looks up an issuer certificate by a
14825 number of criteria: subject name, authority key id
14826 and key usage. It also verifies self signed certificates
14827 by the same criteria. The main comparison function is
14828 X509_check_issued() which performs these checks.
14830 Lot of changes were necessary in order to support this
14831 without completely rewriting the lookup code.
14833 Authority and subject key identifier are now cached.
14835 The LHASH 'certs' is X509_STORE has now been replaced
14836 by a STACK_OF(X509_OBJECT). This is mainly because an
14837 LHASH can't store or retrieve multiple objects with
14838 the same hash value.
14840 As a result various functions (which were all internal
14841 use only) have changed to handle the new X509_STORE
14842 structure. This will break anything that messed round
14843 with X509_STORE internally.
14845 The functions X509_STORE_add_cert() now checks for an
14846 exact match, rather than just subject name.
14848 The X509_STORE API doesn't directly support the retrieval
14849 of multiple certificates matching a given criteria, however
14850 this can be worked round by performing a lookup first
14851 (which will fill the cache with candidate certificates)
14852 and then examining the cache for matches. This is probably
14853 the best we can do without throwing out X509_LOOKUP
14854 entirely (maybe later...).
14856 The X509_VERIFY_CTX structure has been enhanced considerably.
14858 All certificate lookup operations now go via a get_issuer()
14859 callback. Although this currently uses an X509_STORE it
14860 can be replaced by custom lookups. This is a simple way
14861 to bypass the X509_STORE hackery necessary to make this
14862 work and makes it possible to use more efficient techniques
14863 in future. A very simple version which uses a simple
14864 STACK for its trusted certificate store is also provided
14865 using X509_STORE_CTX_trusted_stack().
14867 The verify_cb() and verify() callbacks now have equivalents
14868 in the X509_STORE_CTX structure.
14870 X509_STORE_CTX also has a 'flags' field which can be used
14871 to customise the verify behaviour.
14875 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14876 excludes S/MIME capabilities.
14880 * When a certificate request is read in keep a copy of the
14881 original encoding of the signed data and use it when outputting
14882 again. Signatures then use the original encoding rather than
14883 a decoded, encoded version which may cause problems if the
14884 request is improperly encoded.
14888 * For consistency with other BIO_puts implementations, call
14889 buffer_write(b, ...) directly in buffer_puts instead of calling
14892 In BIO_puts, increment b->num_write as in BIO_write.
14894 *Peter.Sylvester@EdelWeb.fr*
14896 * Fix BN_mul_word for the case where the word is 0. (We have to use
14897 BN_zero, we may not return a BIGNUM with an array consisting of
14898 words set to zero.)
14902 * Avoid calling abort() from within the library when problems are
14903 detected, except if preprocessor symbols have been defined
14904 (such as REF_CHECK, BN_DEBUG etc.).
14908 * New openssl application 'rsautl'. This utility can be
14909 used for low level RSA operations. DER public key
14910 BIO/fp routines also added.
14914 * New Configure entry and patches for compiling on QNX 4.
14916 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14918 * A demo state-machine implementation was sponsored by
14919 Nuron (<http://www.nuron.com/>) and is now available in
14920 demos/state_machine.
14924 * New options added to the 'dgst' utility for signature
14925 generation and verification.
14929 * Unrecognized PKCS#7 content types are now handled via a
14930 catch all ASN1_TYPE structure. This allows unsupported
14931 types to be stored as a "blob" and an application can
14932 encode and decode it manually.
14936 * Fix various signed/unsigned issues to make a_strex.c
14937 compile under VC++.
14939 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14941 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14942 length if passed a buffer. ASN1_INTEGER_to_BN failed
14943 if passed a NULL BN and its argument was negative.
14945 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14947 * Modification to PKCS#7 encoding routines to output definite
14948 length encoding. Since currently the whole structures are in
14949 memory there's not real point in using indefinite length
14950 constructed encoding. However if OpenSSL is compiled with
14951 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14955 * Added BIO_vprintf() and BIO_vsnprintf().
14959 * Added more prefixes to parse for in the strings written
14960 through a logging bio, to cover all the levels that are available
14961 through syslog. The prefixes are now:
14963 PANIC, EMERG, EMR => LOG_EMERG
14964 ALERT, ALR => LOG_ALERT
14965 CRIT, CRI => LOG_CRIT
14966 ERROR, ERR => LOG_ERR
14967 WARNING, WARN, WAR => LOG_WARNING
14968 NOTICE, NOTE, NOT => LOG_NOTICE
14969 INFO, INF => LOG_INFO
14970 DEBUG, DBG => LOG_DEBUG
14972 and as before, if none of those prefixes are present at the
14973 beginning of the string, LOG_ERR is chosen.
14975 On Win32, the `LOG_*` levels are mapped according to this:
14977 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14978 LOG_WARNING => EVENTLOG_WARNING_TYPE
14979 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14983 * Made it possible to reconfigure with just the configuration
14984 argument "reconf" or "reconfigure". The command line arguments
14985 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14986 and are retrieved from there when reconfiguring.
14992 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14994 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14998 * The obj_dat.pl script was messing up the sorting of object
14999 names. The reason was that it compared the quoted version
15000 of strings as a result "OCSP" > "OCSP Signing" because
15001 " > SPACE. Changed script to store unquoted versions of
15002 names and add quotes on output. It was also omitting some
15003 names from the lookup table if they were given a default
15004 value (that is if SN is missing it is given the same
15005 value as LN and vice versa), these are now added on the
15006 grounds that if an object has a name we should be able to
15007 look it up. Finally added warning output when duplicate
15008 short or long names are found.
15012 * Changes needed for Tandem NSK.
15014 *Scott Uroff <scott@xypro.com>*
15016 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
15017 RSA_padding_check_SSLv23(), special padding was never detected
15018 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
15019 version rollback attacks was not effective.
15021 In s23_clnt.c, don't use special rollback-attack detection padding
15022 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
15023 client; similarly, in s23_srvr.c, don't do the rollback check if
15024 SSL 2.0 is the only protocol enabled in the server.
15028 * Make it possible to get hexdumps of unprintable data with 'openssl
15029 asn1parse'. By implication, the functions ASN1_parse_dump() and
15030 BIO_dump_indent() are added.
15034 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15035 these print out strings and name structures based on various
15036 flags including RFC2253 support and proper handling of
15037 multibyte characters. Added options to the 'x509' utility
15038 to allow the various flags to be set.
15042 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15043 Also change the functions X509_cmp_current_time() and
15044 X509_gmtime_adj() work with an ASN1_TIME structure,
15045 this will enable certificates using GeneralizedTime in validity
15046 dates to be checked.
15050 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15051 negative public key encodings) on by default,
15052 NO_NEG_PUBKEY_BUG can be set to disable it.
15056 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15057 content octets. An i2c_ASN1_OBJECT is unnecessary because
15058 the encoding can be trivially obtained from the structure.
15062 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15063 not read locks (`CRYPTO_r_[un]lock`).
15067 * A first attempt at creating official support for shared
15068 libraries through configuration. I've kept it so the
15069 default is static libraries only, and the OpenSSL programs
15070 are always statically linked for now, but there are
15071 preparations for dynamic linking in place.
15072 This has been tested on Linux and Tru64.
15076 * Randomness polling function for Win9x, as described in:
15077 Peter Gutmann, Software Generation of Practically Strong
15082 * Fix so PRNG is seeded in req if using an already existing
15087 * New options to smime application. -inform and -outform
15088 allow alternative formats for the S/MIME message including
15089 PEM and DER. The -content option allows the content to be
15090 specified separately. This should allow things like Netscape
15091 form signing output easier to verify.
15095 * Fix the ASN1 encoding of tags using the 'long form'.
15099 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15100 STRING types. These convert content octets to and from the
15101 underlying type. The actual tag and length octets are
15102 already assumed to have been read in and checked. These
15103 are needed because all other string types have virtually
15104 identical handling apart from the tag. By having versions
15105 of the ASN1 functions that just operate on content octets
15106 IMPLICIT tagging can be handled properly. It also allows
15107 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15108 and ASN1_INTEGER are identical apart from the tag.
15112 * Change the handling of OID objects as follows:
15114 - New object identifiers are inserted in objects.txt, following
15115 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15116 - objects.pl is used to process obj_mac.num and create a new
15118 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15121 This is currently kind of a hack, and the perl code in objects.pl
15122 isn't very elegant, but it works as I intended. The simplest way
15123 to check that it worked correctly is to look in obj_dat.h and
15124 check the array nid_objs and make sure the objects haven't moved
15125 around (this is important!). Additions are OK, as well as
15126 consistent name changes.
15130 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15134 * Addition of the command line parameter '-rand file' to 'openssl req'.
15135 The given file adds to whatever has already been seeded into the
15136 random pool through the RANDFILE configuration file option or
15137 environment variable, or the default random state file.
15141 * mkstack.pl now sorts each macro group into lexical order.
15142 Previously the output order depended on the order the files
15143 appeared in the directory, resulting in needless rewriting
15148 * Patches to make OpenSSL compile under Win32 again. Mostly
15149 work arounds for the VC++ problem that it treats func() as
15150 func(void). Also stripped out the parts of mkdef.pl that
15151 added extra typesafe functions: these no longer exist.
15155 * Reorganisation of the stack code. The macros are now all
15156 collected in safestack.h . Each macro is defined in terms of
15157 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15158 DEBUG_SAFESTACK is now handled in terms of function casts,
15159 this has the advantage of retaining type safety without the
15160 use of additional functions. If DEBUG_SAFESTACK is not defined
15161 then the non typesafe macros are used instead. Also modified the
15162 mkstack.pl script to handle the new form. Needs testing to see
15163 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15164 the default if no major problems. Similar behaviour for ASN1_SET_OF
15165 and PKCS12_STACK_OF.
15169 * When some versions of IIS use the 'NET' form of private key the
15170 key derivation algorithm is different. Normally MD5(password) is
15171 used as a 128 bit RC4 key. In the modified case
15172 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15173 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15174 as the old Netscape_RSA functions except they have an additional
15175 'sgckey' parameter which uses the modified algorithm. Also added
15176 an -sgckey command line option to the rsa utility. Thanks to
15177 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15178 algorithm to openssl-dev.
15182 * The evp_local.h macros were using 'c.##kname' which resulted in
15183 invalid expansion on some systems (SCO 5.0.5 for example).
15184 Corrected to 'c.kname'.
15186 *Phillip Porch <root@theporch.com>*
15188 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15189 a STACK of email addresses from a certificate or request, these look
15190 in the subject name and the subject alternative name extensions and
15191 omit any duplicate addresses.
15195 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15196 This makes DSA verification about 2 % faster.
15200 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15201 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15202 plus overhead for 1024 bit moduli).
15203 This makes exponentiations about 0.5 % faster for 1024 bit
15204 exponents (as measured by "openssl speed rsa2048").
15208 * Rename memory handling macros to avoid conflicts with other
15210 Malloc => OPENSSL_malloc
15211 Malloc_locked => OPENSSL_malloc_locked
15212 Realloc => OPENSSL_realloc
15213 Free => OPENSSL_free
15217 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15218 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15222 * CygWin32 support.
15224 *John Jarvie <jjarvie@newsguy.com>*
15226 * The type-safe stack code has been rejigged. It is now only compiled
15227 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15228 by default all type-specific stack functions are "#define"d back to
15229 standard stack functions. This results in more streamlined output
15230 but retains the type-safety checking possibilities of the original
15235 * The STACK code has been cleaned up, and certain type declarations
15236 that didn't make a lot of sense have been brought in line. This has
15237 also involved a cleanup of sorts in safestack.h to more correctly
15238 map type-safe stack functions onto their plain stack counterparts.
15239 This work has also resulted in a variety of "const"ifications of
15240 lots of the code, especially `_cmp` operations which should normally
15241 be prototyped with "const" parameters anyway.
15245 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15246 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15247 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15248 where all of 'md' is used each time the PRNG is used, but 'state'
15249 is used only indexed by a cyclic counter. As entropy may not be
15250 well distributed from the beginning, 'md' is important as a
15251 chaining variable. However, the output function chains only half
15252 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15253 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15254 in all of 'state' being rewritten, with the new values depending
15255 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15259 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15260 the handshake is continued after ssl_verify_cert_chain();
15261 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15262 can lead to 'unexplainable' connection aborts later.
15264 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15266 * Major EVP API cipher revision.
15267 Add hooks for extra EVP features. This allows various cipher
15268 parameters to be set in the EVP interface. Support added for variable
15269 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15270 setting of RC2 and RC5 parameters.
15272 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15275 Remove lots of duplicated code from the EVP library. For example *every*
15276 cipher init() function handles the 'iv' in the same way according to the
15277 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15278 for CFB and OFB modes they zero ctx->num.
15280 New functionality allows removal of S/MIME code RC2 hack.
15282 Most of the routines have the same form and so can be declared in terms
15285 By shifting this to the top level EVP_CipherInit() it can be removed from
15286 all individual ciphers. If the cipher wants to handle IVs or keys
15287 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15290 Change lots of functions like EVP_EncryptUpdate() to now return a
15291 value: although software versions of the algorithms cannot fail
15292 any installed hardware versions can.
15296 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15297 this option is set, tolerate broken clients that send the negotiated
15298 protocol version number instead of the requested protocol version
15303 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15304 i.e. non-zero for export ciphersuites, zero otherwise.
15305 Previous versions had this flag inverted, inconsistent with
15306 rsa_tmp_cb (..._TMP_RSA_CB).
15308 *Bodo Moeller; problem reported by Amit Chopra*
15310 * Add missing DSA library text string. Work around for some IIS
15311 key files with invalid SEQUENCE encoding.
15315 * Add a document (doc/standards.txt) that list all kinds of standards
15316 and so on that are implemented in OpenSSL.
15320 * Enhance c_rehash script. Old version would mishandle certificates
15321 with the same subject name hash and wouldn't handle CRLs at all.
15322 Added -fingerprint option to crl utility, to support new c_rehash
15327 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15331 * Fix for SSL server purpose checking. Server checking was
15332 rejecting certificates which had extended key usage present
15333 but no ssl client purpose.
15335 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15337 * Make PKCS#12 code work with no password. The PKCS#12 spec
15338 is a little unclear about how a blank password is handled.
15339 Since the password in encoded as a BMPString with terminating
15340 double NULL a zero length password would end up as just the
15341 double NULL. However no password at all is different and is
15342 handled differently in the PKCS#12 key generation code. NS
15343 treats a blank password as zero length. MSIE treats it as no
15344 password on export: but it will try both on import. We now do
15345 the same: PKCS12_parse() tries zero length and no password if
15346 the password is set to "" or NULL (NULL is now a valid password:
15347 it wasn't before) as does the pkcs12 application.
15351 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15352 perror when PEM_read_bio_X509_REQ fails, the error message must
15353 be obtained from the error queue.
15357 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15358 it in ERR_remove_state if appropriate, and change ERR_get_state
15359 accordingly to avoid race conditions (this is necessary because
15360 thread_hash is no longer constant once set).
15364 * Bugfix for linux-elf makefile.one.
15368 * RSA_get_default_method() will now cause a default
15369 RSA_METHOD to be chosen if one doesn't exist already.
15370 Previously this was only set during a call to RSA_new()
15371 or RSA_new_method(NULL) meaning it was possible for
15372 RSA_get_default_method() to return NULL.
15376 * Added native name translation to the existing DSO code
15377 that will convert (if the flag to do so is set) filenames
15378 that are sufficiently small and have no path information
15379 into a canonical native form. Eg. "blah" converted to
15380 "libblah.so" or "blah.dll" etc.
15384 * New function ERR_error_string_n(e, buf, len) which is like
15385 ERR_error_string(e, buf), but writes at most 'len' bytes
15386 including the 0 terminator. For ERR_error_string_n, 'buf'
15389 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15391 * CONF library reworked to become more general. A new CONF
15392 configuration file reader "class" is implemented as well as a
15393 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15394 old `CONF_*` functions are still there, but are reimplemented to
15395 work in terms of the new functions. Also, a set of functions
15396 to handle the internal storage of the configuration data is
15397 provided to make it easier to write new configuration file
15398 reader "classes" (I can definitely see something reading a
15399 configuration file in XML format, for example), called `_CONF_*`,
15400 or "the configuration storage API"...
15402 The new configuration file reading functions are:
15404 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15405 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15407 NCONF_default, NCONF_WIN32
15409 NCONF_dump_fp, NCONF_dump_bio
15411 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15412 NCONF_new creates a new CONF object. This works in the same way
15413 as other interfaces in OpenSSL, like the BIO interface.
15414 `NCONF_dump_*` dump the internal storage of the configuration file,
15415 which is useful for debugging. All other functions take the same
15416 arguments as the old `CONF_*` functions with the exception of the
15417 first that must be a `CONF *` instead of a `LHASH *`.
15419 To make it easier to use the new classes with the old `CONF_*` functions,
15420 the function CONF_set_default_method is provided.
15424 * Add '-tls1' option to 'openssl ciphers', which was already
15425 mentioned in the documentation but had not been implemented.
15426 (This option is not yet really useful because even the additional
15427 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15431 * Initial DSO code added into libcrypto for letting OpenSSL (and
15432 OpenSSL-based applications) load shared libraries and bind to
15433 them in a portable way.
15435 *Geoff Thorpe, with contributions from Richard Levitte*
15437 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15439 * Make sure _lrotl and _lrotr are only used with MSVC.
15441 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15442 (the default implementation of RAND_status).
15444 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15445 to '-clrext' (= clear extensions), as intended and documented.
15446 *Bodo Moeller; inconsistency pointed out by Michael Attili
15447 <attili@amaxo.com>*
15449 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15450 was larger than the MD block size.
15452 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15454 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15455 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15456 using the passed key: if the passed key was a private key the result
15457 of X509_print(), for example, would be to print out all the private key
15462 * des_quad_cksum() byte order bug fix.
15463 *Ulf Möller, using the problem description in krb4-0.9.7, where
15464 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15466 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15469 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15471 * For easily testing in shell scripts whether some command
15472 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15473 returns with exit code 0 iff no command of the given name is available.
15474 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15475 the output goes to stdout and nothing is printed to stderr.
15476 Additional arguments are always ignored.
15478 Since for each cipher there is a command of the same name,
15479 the 'no-cipher' compilation switches can be tested this way.
15481 ('openssl no-XXX' is not able to detect pseudo-commands such
15482 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15486 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15490 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15491 is set; it will be thrown away anyway because each handshake creates
15493 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15494 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15495 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15496 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15500 * New s_client option -ign_eof: EOF at stdin is ignored, and
15501 'Q' and 'R' lose their special meanings (quit/renegotiate).
15502 This is part of what -quiet does; unlike -quiet, -ign_eof
15503 does not suppress any output.
15507 * Add compatibility options to the purpose and trust code. The
15508 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15509 accepts a certificate or CA, this was the previous behaviour,
15510 with all the associated security issues.
15512 X509_TRUST_COMPAT is the old trust behaviour: only and
15513 automatically trust self signed roots in certificate store. A
15514 new trust setting X509_TRUST_DEFAULT is used to specify that
15515 a purpose has no associated trust setting and it should instead
15516 use the value in the default purpose.
15520 * Fix the PKCS#8 DSA private key code so it decodes keys again
15521 and fix a memory leak.
15525 * In util/mkerr.pl (which implements 'make errors'), preserve
15526 reason strings from the previous version of the .c file, as
15527 the default to have only downcase letters (and digits) in
15528 automatically generated reasons codes is not always appropriate.
15532 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15533 using strerror. Previously, ERR_reason_error_string() returned
15534 library names as reason strings for SYSerr; but SYSerr is a special
15535 case where small numbers are errno values, not library numbers.
15539 * Add '-dsaparam' option to 'openssl dhparam' application. This
15540 converts DSA parameters into DH parameters. (When creating parameters,
15541 DSA_generate_parameters is used.)
15545 * Include 'length' (recommended exponent length) in C code generated
15546 by 'openssl dhparam -C'.
15550 * The second argument to set_label in perlasm was already being used
15551 so couldn't be used as a "file scope" flag. Moved to third argument
15556 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15557 instead of RAND_bytes for encryption IVs and salts.
15561 * Include RAND_status() into RAND_METHOD instead of implementing
15562 it only for md_rand.c Otherwise replacing the PRNG by calling
15563 RAND_set_rand_method would be impossible.
15567 * Don't let DSA_generate_key() enter an infinite loop if the random
15568 number generation fails.
15572 * New 'rand' application for creating pseudo-random output.
15576 * Added configuration support for Linux/IA64
15578 *Rolf Haberrecker <rolf@suse.de>*
15580 * Assembler module support for Mingw32.
15584 * Shared library support for HPUX (in shlib/).
15586 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15588 * Shared library support for Solaris gcc.
15590 *Lutz Behnke <behnke@trustcenter.de>*
15592 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15594 * PKCS7_encrypt() was adding text MIME headers twice because they
15595 were added manually and by SMIME_crlf_copy().
15599 * In bntest.c don't call BN_rand with zero bits argument.
15601 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15603 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15604 case was implemented. This caused BN_div_recp() to fail occasionally.
15608 * Add an optional second argument to the set_label() in the perl
15609 assembly language builder. If this argument exists and is set
15610 to 1 it signals that the assembler should use a symbol whose
15611 scope is the entire file, not just the current function. This
15612 is needed with MASM which uses the format label:: for this scope.
15614 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15616 * Change the ASN1 types so they are typedefs by default. Before
15617 almost all types were #define'd to ASN1_STRING which was causing
15618 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15623 * Change names of new functions to the new get1/get0 naming
15624 convention: After 'get1', the caller owns a reference count
15625 and has to call `..._free`; 'get0' returns a pointer to some
15626 data structure without incrementing reference counters.
15627 (Some of the existing 'get' functions increment a reference
15628 counter, some don't.)
15629 Similarly, 'set1' and 'add1' functions increase reference
15630 counters or duplicate objects.
15634 * Allow for the possibility of temp RSA key generation failure:
15635 the code used to assume it always worked and crashed on failure.
15639 * Fix potential buffer overrun problem in BIO_printf().
15640 *Ulf Möller, using public domain code by Patrick Powell; problem
15641 pointed out by David Sacerdote <das33@cornell.edu>*
15643 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15644 RAND_egd() and RAND_status(). In the command line application,
15645 the EGD socket can be specified like a seed file using RANDFILE
15650 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15651 Some CAs (e.g. Verisign) distribute certificates in this form.
15655 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15656 list to exclude them. This means that no special compilation option
15657 is needed to use anonymous DH: it just needs to be included in the
15662 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15663 EVP_MD_type. The old functionality is available in a new macro called
15664 EVP_MD_md(). Change code that uses it and update docs.
15668 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15669 where the `void *` argument is replaced by a function pointer argument.
15670 Previously `void *` was abused to point to functions, which works on
15671 many platforms, but is not correct. As these functions are usually
15672 called by macros defined in OpenSSL header files, most source code
15673 should work without changes.
15677 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15678 sections with information on -D... compiler switches used for
15679 compiling the library so that applications can see them. To enable
15680 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15681 must be defined. E.g.,
15682 #define OPENSSL_ALGORITHM_DEFINES
15683 #include <openssl/opensslconf.h>
15684 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15686 *Richard Levitte, Ulf and Bodo Möller*
15688 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15693 * Change the 'other' type in certificate aux info to a STACK_OF
15694 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15695 the required ASN1 format: arbitrary types determined by an OID.
15699 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15700 argument to 'req'. This is not because the function is newer or
15701 better than others it just uses the work 'NEW' in the certificate
15702 request header lines. Some software needs this.
15706 * Reorganise password command line arguments: now passwords can be
15707 obtained from various sources. Delete the PEM_cb function and make
15708 it the default behaviour: i.e. if the callback is NULL and the
15709 usrdata argument is not NULL interpret it as a null terminated pass
15710 phrase. If usrdata and the callback are NULL then the pass phrase
15711 is prompted for as usual.
15715 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15716 the support is automatically enabled. The resulting binaries will
15717 autodetect the card and use it if present.
15719 *Ben Laurie and Compaq Inc.*
15721 * Work around for Netscape hang bug. This sends certificate request
15722 and server done in one record. Since this is perfectly legal in the
15723 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15724 the bugs/SSLv3 entry for more info.
15728 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15732 * Add -rand argument to smime and pkcs12 applications and read/write
15737 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15741 * Add command line password options to the remaining applications.
15745 * Bug fix for BN_div_recp() for numerators with an even number of
15750 * More tests in bntest.c, and changed test_bn output.
15754 * ./config recognizes MacOS X now.
15758 * Bug fix for BN_div() when the first words of num and divisor are
15759 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15763 * Add support for various broken PKCS#8 formats, and command line
15764 options to produce them.
15768 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15769 get temporary BIGNUMs from a BN_CTX.
15773 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15778 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15779 include a #define from the old name to the new. The original intent
15780 was that statically linked binaries could for example just call
15781 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15782 link with digests. This never worked because SSLeay_add_all_digests()
15783 and SSLeay_add_all_ciphers() were in the same source file so calling
15784 one would link with the other. They are now in separate source files.
15788 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15792 * Use a less unusual form of the Miller-Rabin primality test (it used
15793 a binary algorithm for exponentiation integrated into the Miller-Rabin
15794 loop, our standard modexp algorithms are faster).
15798 * Support for the EBCDIC character set completed.
15800 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15802 * Source code cleanups: use const where appropriate, eliminate casts,
15803 use `void *` instead of `char *` in lhash.
15807 * Bugfix: ssl3_send_server_key_exchange was not restartable
15808 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15809 this the server could overwrite ephemeral keys that the client
15814 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15815 using 50 iterations of the Rabin-Miller test.
15817 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15818 iterations of the Rabin-Miller test as required by the appendix
15819 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15820 As BN_is_prime_fasttest includes trial division, DSA parameter
15821 generation becomes much faster.
15823 This implies a change for the callback functions in DSA_is_prime
15824 and DSA_generate_parameters: The callback function is called once
15825 for each positive witness in the Rabin-Miller test, not just
15826 occasionally in the inner loop; and the parameters to the
15827 callback function now provide an iteration count for the outer
15828 loop rather than for the current invocation of the inner loop.
15829 DSA_generate_parameters additionally can call the callback
15830 function with an 'iteration count' of -1, meaning that a
15831 candidate has passed the trial division test (when q is generated
15832 from an application-provided seed, trial division is skipped).
15836 * New function BN_is_prime_fasttest that optionally does trial
15837 division before starting the Rabin-Miller test and has
15838 an additional BN_CTX * argument (whereas BN_is_prime always
15839 has to allocate at least one BN_CTX).
15840 'callback(1, -1, cb_arg)' is called when a number has passed the
15841 trial division stage.
15845 * Fix for bug in CRL encoding. The validity dates weren't being handled
15850 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15854 * New function BN_pseudo_rand().
15858 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15859 bignum version of BN_from_montgomery() with the working code from
15860 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15865 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15866 made it impossible to use the same SSL_SESSION data structure in
15867 SSL2 clients in multiple threads.
15871 * The return value of RAND_load_file() no longer counts bytes obtained
15872 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15873 to seed the PRNG (previously an explicit byte count was required).
15875 *Ulf Möller, Bodo Möller*
15877 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15878 used `char *` instead of `void *` and had casts all over the place.
15882 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15886 * Retain source code compatibility for BN_prime_checks macro:
15887 BN_is_prime(..., BN_prime_checks, ...) now uses
15888 BN_prime_checks_for_size to determine the appropriate number of
15889 Rabin-Miller iterations.
15893 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15894 DH_CHECK_P_NOT_SAFE_PRIME.
15895 (Check if this is true? OpenPGP calls them "strong".)
15899 * Merge the functionality of "dh" and "gendh" programs into a new program
15900 "dhparam". The old programs are retained for now but will handle DH keys
15901 (instead of parameters) in future.
15905 * Make the ciphers, s_server and s_client programs check the return values
15906 when a new cipher list is set.
15910 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15911 ciphers. Before when the 56bit ciphers were enabled the sorting was
15914 The syntax for the cipher sorting has been extended to support sorting by
15915 cipher-strength (using the strength_bits hard coded in the tables).
15916 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15918 Fix a bug in the cipher-command parser: when supplying a cipher command
15919 string with an "undefined" symbol (neither command nor alphanumeric
15920 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15921 an error is flagged.
15923 Due to the strength-sorting extension, the code of the
15924 ssl_create_cipher_list() function was completely rearranged. I hope that
15925 the readability was also increased :-)
15927 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15929 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15930 for the first serial number and places 2 in the serial number file. This
15931 avoids problems when the root CA is created with serial number zero and
15932 the first user certificate has the same issuer name and serial number
15937 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15938 the new code. Add documentation for this stuff.
15942 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15943 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15944 structures and behave in an analogous way to the X509v3 functions:
15945 they shouldn't be called directly but wrapper functions should be used
15948 So we also now have some wrapper functions that call the X509at functions
15949 when passed certificate requests. (TO DO: similar things can be done with
15950 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15951 things. Some of these need some d2i or i2d and print functionality
15952 because they handle more complex structures.)
15956 * Add missing #ifndefs that caused missing symbols when building libssl
15957 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15958 NO_RSA in `ssl/s2*.c`.
15960 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15962 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15963 has a return value which indicates the quality of the random data
15964 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15965 error queue. New function RAND_pseudo_bytes() generates output that is
15966 guaranteed to be unique but not unpredictable. RAND_add is like
15967 RAND_seed, but takes an extra argument for an entropy estimate
15968 (RAND_seed always assumes full entropy).
15972 * Do more iterations of Rabin-Miller probable prime test (specifically,
15973 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15974 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15975 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15976 false-positive rate of at most 2^-80 for random input.
15980 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15984 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15985 in the 0.9.5 release), this returns the chain
15986 from an X509_CTX structure with a dup of the stack and all
15987 the X509 reference counts upped: so the stack will exist
15988 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15991 Also make SSL_SESSION_print() print out the verify return
15996 * Add manpage for the pkcs12 command. Also change the default
15997 behaviour so MAC iteration counts are used unless the new
15998 -nomaciter option is used. This improves file security and
15999 only older versions of MSIE (4.0 for example) need it.
16003 * Honor the no-xxx Configure options when creating .DEF files.
16007 * Add PKCS#10 attributes to field table: challengePassword,
16008 unstructuredName and unstructuredAddress. These are taken from
16009 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
16010 international characters are used.
16012 More changes to X509_ATTRIBUTE code: allow the setting of types
16013 based on strings. Remove the 'loc' parameter when adding
16014 attributes because these will be a SET OF encoding which is sorted
16019 * Initial changes to the 'req' utility to allow request generation
16020 automation. This will allow an application to just generate a template
16021 file containing all the field values and have req construct the
16024 Initial support for X509_ATTRIBUTE handling. Stacks of these are
16025 used all over the place including certificate requests and PKCS#7
16026 structures. They are currently handled manually where necessary with
16027 some primitive wrappers for PKCS#7. The new functions behave in a
16028 manner analogous to the X509 extension functions: they allow
16029 attributes to be looked up by NID and added.
16031 Later something similar to the X509V3 code would be desirable to
16032 automatically handle the encoding, decoding and printing of the
16033 more complex types. The string types like challengePassword can
16034 be handled by the string table functions.
16036 Also modified the multi byte string table handling. Now there is
16037 a 'global mask' which masks out certain types. The table itself
16038 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16039 is useful when for example there is only one permissible type
16040 (as in countryName) and using the mask might result in no valid
16045 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16046 SSL_get_peer_finished to allow applications to obtain the latest
16047 Finished messages sent to the peer or expected from the peer,
16048 respectively. (SSL_get_peer_finished is usually the Finished message
16049 actually received from the peer, otherwise the protocol will be aborted.)
16051 As the Finished message are message digests of the complete handshake
16052 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16053 be used for external authentication procedures when the authentication
16054 provided by SSL/TLS is not desired or is not enough.
16058 * Enhanced support for Alpha Linux is added. Now ./config checks if
16059 the host supports BWX extension and if Compaq C is present on the
16060 $PATH. Just exploiting of the BWX extension results in 20-30%
16061 performance kick for some algorithms, e.g. DES and RC4 to mention
16062 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16067 * Add support for MS "fast SGC". This is arguably a violation of the
16068 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16069 weak crypto and after checking the certificate is SGC a second one
16070 with strong crypto. MS SGC stops the first handshake after receiving
16071 the server certificate message and sends a second client hello. Since
16072 a server will typically do all the time consuming operations before
16073 expecting any further messages from the client (server key exchange
16074 is the most expensive) there is little difference between the two.
16076 To get OpenSSL to support MS SGC we have to permit a second client
16077 hello message after we have sent server done. In addition we have to
16078 reset the MAC if we do get this second client hello.
16082 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16083 if a DER encoded private key is RSA or DSA traditional format. Changed
16084 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16085 format DER encoded private key. Newer code should use PKCS#8 format which
16086 has the key type encoded in the ASN1 structure. Added DER private key
16087 support to pkcs8 application.
16091 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16092 ciphersuites has been selected (as required by the SSL 3/TLS 1
16093 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16094 is set, we interpret this as a request to violate the specification
16095 (the worst that can happen is a handshake failure, and 'correct'
16096 behaviour would result in a handshake failure anyway).
16100 * In SSL_CTX_add_session, take into account that there might be multiple
16101 SSL_SESSION structures with the same session ID (e.g. when two threads
16102 concurrently obtain them from an external cache).
16103 The internal cache can handle only one SSL_SESSION with a given ID,
16104 so if there's a conflict, we now throw out the old one to achieve
16109 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16110 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16111 some routines that use cipher OIDs: some ciphers do not have OIDs
16112 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16117 * Simplify the trust setting structure and code. Now we just have
16118 two sequences of OIDs for trusted and rejected settings. These will
16119 typically have values the same as the extended key usage extension
16120 and any application specific purposes.
16122 The trust checking code now has a default behaviour: it will just
16123 check for an object with the same NID as the passed id. Functions can
16124 be provided to override either the default behaviour or the behaviour
16125 for a given id. SSL client, server and email already have functions
16126 in place for compatibility: they check the NID and also return "trusted"
16127 if the certificate is self signed.
16131 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16132 traditional format into an EVP_PKEY structure.
16136 * Add a password callback function PEM_cb() which either prompts for
16137 a password if usr_data is NULL or otherwise assumes it is a null
16138 terminated password. Allow passwords to be passed on command line
16139 environment or config files in a few more utilities.
16143 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16144 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16145 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16146 Update documentation.
16150 * Support for ASN1 "NULL" type. This could be handled before by using
16151 ASN1_TYPE but there wasn't any function that would try to read a NULL
16152 and produce an error if it couldn't. For compatibility we also have
16153 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16154 don't allocate anything because they don't need to.
16158 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16161 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16163 * Rebuild of the memory allocation routines used by OpenSSL code and
16164 possibly others as well. The purpose is to make an interface that
16165 provide hooks so anyone can build a separate set of allocation and
16166 deallocation routines to be used by OpenSSL, for example memory
16167 pool implementations, or something else, which was previously hard
16168 since Malloc(), Realloc() and Free() were defined as macros having
16169 the values malloc, realloc and free, respectively (except for Win32
16170 compilations). The same is provided for memory debugging code.
16171 OpenSSL already comes with functionality to find memory leaks, but
16172 this gives people a chance to debug other memory problems.
16174 With these changes, a new set of functions and macros have appeared:
16176 CRYPTO_set_mem_debug_functions() [F]
16177 CRYPTO_get_mem_debug_functions() [F]
16178 CRYPTO_dbg_set_options() [F]
16179 CRYPTO_dbg_get_options() [F]
16180 CRYPTO_malloc_debug_init() [M]
16182 The memory debug functions are NULL by default, unless the library
16183 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16184 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16185 gives the standard debugging functions that come with OpenSSL) or
16186 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16187 provided by the library user) must be used. When the standard
16188 debugging functions are used, CRYPTO_dbg_set_options can be used to
16189 request additional information:
16190 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16191 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16193 Also, things like CRYPTO_set_mem_functions will always give the
16194 expected result (the new set of functions is used for allocation
16195 and deallocation) at all times, regardless of platform and compiler
16198 To finish it up, some functions that were never use in any other
16199 way than through macros have a new API and new semantic:
16201 CRYPTO_dbg_malloc()
16202 CRYPTO_dbg_realloc()
16205 All macros of value have retained their old syntax.
16207 *Richard Levitte and Bodo Moeller*
16209 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16210 ordering of SMIMECapabilities wasn't in "strength order" and there
16211 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16216 * Some ASN1 types with illegal zero length encoding (INTEGER,
16217 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16219 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16221 * Merge in my S/MIME library for OpenSSL. This provides a simple
16222 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16223 functionality to handle multipart/signed properly) and a utility
16224 called 'smime' to call all this stuff. This is based on code I
16225 originally wrote for Celo who have kindly allowed it to be
16226 included in OpenSSL.
16230 * Add variants des_set_key_checked and des_set_key_unchecked of
16231 des_set_key (aka des_key_sched). Global variable des_check_key
16232 decides which of these is called by des_set_key; this way
16233 des_check_key behaves as it always did, but applications and
16234 the library itself, which was buggy for des_check_key == 1,
16235 have a cleaner way to pick the version they need.
16239 * New function PKCS12_newpass() which changes the password of a
16244 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16245 dynamic mix. In both cases the ids can be used as an index into the
16246 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16247 functions so they accept a list of the field values and the
16248 application doesn't need to directly manipulate the X509_TRUST
16253 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16258 * Modify the way the V3 extension code looks up extensions. This now
16259 works in a similar way to the object code: we have some "standard"
16260 extensions in a static table which is searched with OBJ_bsearch()
16261 and the application can add dynamic ones if needed. The file
16262 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16263 updated whenever a new extension is added to the core code and kept
16264 in ext_nid order. There is a simple program 'tabtest.c' which checks
16265 this. New extensions are not added too often so this file can readily
16266 be maintained manually.
16268 There are two big advantages in doing things this way. The extensions
16269 can be looked up immediately and no longer need to be "added" using
16270 X509V3_add_standard_extensions(): this function now does nothing.
16271 Side note: I get *lots* of email saying the extension code doesn't
16272 work because people forget to call this function.
16273 Also no dynamic allocation is done unless new extensions are added:
16274 so if we don't add custom extensions there is no need to call
16275 X509V3_EXT_cleanup().
16279 * Modify enc utility's salting as follows: make salting the default. Add a
16280 magic header, so unsalted files fail gracefully instead of just decrypting
16281 to garbage. This is because not salting is a big security hole, so people
16282 should be discouraged from doing it.
16286 * Fixes and enhancements to the 'x509' utility. It allowed a message
16287 digest to be passed on the command line but it only used this
16288 parameter when signing a certificate. Modified so all relevant
16289 operations are affected by the digest parameter including the
16290 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16291 DSA key was used because it didn't fix the digest.
16295 * Initial certificate chain verify code. Currently tests the untrusted
16296 certificates for consistency with the verify purpose (which is set
16297 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16299 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16300 this is because it will reject chains with invalid extensions whereas
16301 every previous version of OpenSSL and SSLeay made no checks at all.
16303 Trust code: checks the root CA for the relevant trust settings. Trust
16304 settings have an initial value consistent with the verify purpose: e.g.
16305 if the verify purpose is for SSL client use it expects the CA to be
16306 trusted for SSL client use. However the default value can be changed to
16307 permit custom trust settings: one example of this would be to only trust
16308 certificates from a specific "secure" set of CAs.
16310 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16311 which should be used for version portability: especially since the
16312 verify structure is likely to change more often now.
16314 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16315 to set them. If not set then assume SSL clients will verify SSL servers
16318 Two new options to the verify program: -untrusted allows a set of
16319 untrusted certificates to be passed in and -purpose which sets the
16320 intended purpose of the certificate. If a purpose is set then the
16321 new chain verify code is used to check extension consistency.
16325 * Support for the authority information access extension.
16329 * Modify RSA and DSA PEM read routines to transparently handle
16330 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16331 public keys in a format compatible with certificate
16332 SubjectPublicKeyInfo structures. Unfortunately there were already
16333 functions called *_PublicKey_* which used various odd formats so
16334 these are retained for compatibility: however the DSA variants were
16335 never in a public release so they have been deleted. Changed dsa/rsa
16336 utilities to handle the new format: note no releases ever handled public
16337 keys so we should be OK.
16339 The primary motivation for this change is to avoid the same fiasco
16340 that dogs private keys: there are several incompatible private key
16341 formats some of which are standard and some OpenSSL specific and
16342 require various evil hacks to allow partial transparent handling and
16343 even then it doesn't work with DER formats. Given the option anything
16344 other than PKCS#8 should be dumped: but the other formats have to
16345 stay in the name of compatibility.
16347 With public keys and the benefit of hindsight one standard format
16348 is used which works with EVP_PKEY, RSA or DSA structures: though
16349 it clearly returns an error if you try to read the wrong kind of key.
16351 Added a -pubkey option to the 'x509' utility to output the public key.
16352 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16353 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16354 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16355 that do the same as the `EVP_PKEY_assign_*()` except they up the
16356 reference count of the added key (they don't "swallow" the
16361 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16362 CRLs would fail if the file contained no certificates or no CRLs:
16363 added a new function to read in both types and return the number
16364 read: this means that if none are read it will be an error. The
16365 DER versions of the certificate and CRL reader would always fail
16366 because it isn't possible to mix certificates and CRLs in DER format
16367 without choking one or the other routine. Changed this to just read
16368 a certificate: this is the best we can do. Also modified the code
16369 in `apps/verify.c` to take notice of return codes: it was previously
16370 attempting to read in certificates from NULL pointers and ignoring
16371 any errors: this is one reason why the cert and CRL reader seemed
16372 to work. It doesn't check return codes from the default certificate
16373 routines: these may well fail if the certificates aren't installed.
16377 * Code to support otherName option in GeneralName.
16381 * First update to verify code. Change the verify utility
16382 so it warns if it is passed a self signed certificate:
16383 for consistency with the normal behaviour. X509_verify
16384 has been modified to it will now verify a self signed
16385 certificate if *exactly* the same certificate appears
16386 in the store: it was previously impossible to trust a
16387 single self signed certificate. This means that:
16388 openssl verify ss.pem
16389 now gives a warning about a self signed certificate but
16390 openssl verify -CAfile ss.pem ss.pem
16395 * For servers, store verify_result in SSL_SESSION data structure
16396 (and add it to external session representation).
16397 This is needed when client certificate verifications fails,
16398 but an application-provided verification callback (set by
16399 SSL_CTX_set_cert_verify_callback) allows accepting the session
16400 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16401 but returns 1): When the session is reused, we have to set
16402 ssl->verify_result to the appropriate error code to avoid
16405 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16407 * Fix a bug in the new PKCS#7 code: it didn't consider the
16408 case in PKCS7_dataInit() where the signed PKCS7 structure
16409 didn't contain any existing data because it was being created.
16411 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16413 * Add a salt to the key derivation routines in enc.c. This
16414 forms the first 8 bytes of the encrypted file. Also add a
16415 -S option to allow a salt to be input on the command line.
16419 * New function X509_cmp(). Oddly enough there wasn't a function
16420 to compare two certificates. We do this by working out the SHA1
16421 hash and comparing that. X509_cmp() will be needed by the trust
16426 * SSL_get1_session() is like SSL_get_session(), but increments
16427 the reference count in the SSL_SESSION returned.
16429 *Geoff Thorpe <geoff@eu.c2.net>*
16431 * Fix for 'req': it was adding a null to request attributes.
16432 Also change the X509_LOOKUP and X509_INFO code to handle
16433 certificate auxiliary information.
16437 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16442 * Add the possibility to add extra information to the memory leak
16443 detecting output, to form tracebacks, showing from where each
16444 allocation was originated: CRYPTO_push_info("constant string") adds
16445 the string plus current file name and line number to a per-thread
16446 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16447 is like calling CYRPTO_pop_info() until the stack is empty.
16448 Also updated memory leak detection code to be multi-thread-safe.
16452 * Add options -text and -noout to pkcs7 utility and delete the
16453 encryption options which never did anything. Update docs.
16457 * Add options to some of the utilities to allow the pass phrase
16458 to be included on either the command line (not recommended on
16459 OSes like Unix) or read from the environment. Update the
16460 manpages and fix a few bugs.
16464 * Add a few manpages for some of the openssl commands.
16468 * Fix the -revoke option in ca. It was freeing up memory twice,
16469 leaking and not finding already revoked certificates.
16473 * Extensive changes to support certificate auxiliary information.
16474 This involves the use of X509_CERT_AUX structure and X509_AUX
16475 functions. An X509_AUX function such as PEM_read_X509_AUX()
16476 can still read in a certificate file in the usual way but it
16477 will also read in any additional "auxiliary information". By
16478 doing things this way a fair degree of compatibility can be
16479 retained: existing certificates can have this information added
16480 using the new 'x509' options.
16482 Current auxiliary information includes an "alias" and some trust
16483 settings. The trust settings will ultimately be used in enhanced
16484 certificate chain verification routines: currently a certificate
16485 can only be trusted if it is self signed and then it is trusted
16490 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16491 The problem was that one of the replacement routines had not been working
16492 since SSLeay releases. For now the offending routine has been replaced
16493 with non-optimised assembler. Even so, this now gives around 95%
16494 performance improvement for 1024 bit RSA signs.
16498 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16499 handling. Most clients have the effective key size in bits equal to
16500 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16501 A few however don't do this and instead use the size of the decrypted key
16502 to determine the RC2 key length and the AlgorithmIdentifier to determine
16503 the effective key length. In this case the effective key length can still
16504 be 40 bits but the key length can be 168 bits for example. This is fixed
16505 by manually forcing an RC2 key into the EVP_PKEY structure because the
16506 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16507 the key length and effective key length are equal.
16511 * Add a bunch of functions that should simplify the creation of
16512 X509_NAME structures. Now you should be able to do:
16513 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16514 and have it automatically work out the correct field type and fill in
16515 the structures. The more adventurous can try:
16516 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16517 and it will (hopefully) work out the correct multibyte encoding.
16521 * Change the 'req' utility to use the new field handling and multibyte
16522 copy routines. Before the DN field creation was handled in an ad hoc
16523 way in req, ca, and x509 which was rather broken and didn't support
16524 BMPStrings or UTF8Strings. Since some software doesn't implement
16525 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16526 using the dirstring_type option. See the new comment in the default
16527 openssl.cnf for more info.
16531 * Make crypto/rand/md_rand.c more robust:
16532 - Assure unique random numbers after fork().
16533 - Make sure that concurrent threads access the global counter and
16534 md serializably so that we never lose entropy in them
16535 or use exactly the same state in multiple threads.
16536 Access to the large state is not always serializable because
16537 the additional locking could be a performance killer, and
16538 md should be large enough anyway.
16542 * New file `apps/app_rand.c` with commonly needed functionality
16543 for handling the random seed file.
16545 Use the random seed file in some applications that previously did not:
16547 dsaparam -genkey (which also ignored its '-rand' option),
16550 x509 (when signing).
16551 Except on systems with /dev/urandom, it is crucial to have a random
16552 seed file at least for key creation, DSA signing, and for DH exchanges;
16553 for RSA signatures we could do without one.
16555 gendh and gendsa (unlike genrsa) used to read only the first byte
16556 of each file listed in the '-rand' option. The function as previously
16557 found in genrsa is now in app_rand.c and is used by all programs
16558 that support '-rand'.
16562 * In RAND_write_file, use mode 0600 for creating files;
16563 don't just chmod when it may be too late.
16567 * Report an error from X509_STORE_load_locations
16568 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16572 * New function ASN1_mbstring_copy() this copies a string in either
16573 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16574 into an ASN1_STRING type. A mask of permissible types is passed
16575 and it chooses the "minimal" type to use or an error if not type
16580 * Add function equivalents to the various macros in asn1.h. The old
16581 macros are retained with an `M_` prefix. Code inside the library can
16582 use the `M_` macros. External code (including the openssl utility)
16583 should *NOT* in order to be "shared library friendly".
16587 * Add various functions that can check a certificate's extensions
16588 to see if it usable for various purposes such as SSL client,
16589 server or S/MIME and CAs of these types. This is currently
16590 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16591 verification. Also added a -purpose flag to x509 utility to
16592 print out all the purposes.
16596 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16601 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16602 for, obtain and decode and extension and obtain its critical flag.
16603 This allows all the necessary extension code to be handled in a
16604 single function call.
16608 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16609 platforms. See crypto/rc4/rc4_enc.c for further details.
16613 * New -noout option to asn1parse. This causes no output to be produced
16614 its main use is when combined with -strparse and -out to extract data
16615 from a file (which may not be in ASN.1 format).
16619 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16620 when producing the local key id.
16622 *Richard Levitte <levitte@stacken.kth.se>*
16624 * New option -dhparam in s_server. This allows a DH parameter file to be
16625 stated explicitly. If it is not stated then it tries the first server
16626 certificate file. The previous behaviour hard coded the filename
16631 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16632 a public key to be input or output. For example:
16633 openssl rsa -in key.pem -pubout -out pubkey.pem
16634 Also added necessary DSA public key functions to handle this.
16638 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16639 in the message. This was handled by allowing
16640 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16642 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16644 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16645 to the end of the strings whereas this didn't. This would cause problems
16646 if strings read with d2i_ASN1_bytes() were later modified.
16648 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16650 * Fix for base64 decode bug. When a base64 bio reads only one line of
16651 data and it contains EOF it will end up returning an error. This is
16652 caused by input 46 bytes long. The cause is due to the way base64
16653 BIOs find the start of base64 encoded data. They do this by trying a
16654 trial decode on each line until they find one that works. When they
16655 do a flag is set and it starts again knowing it can pass all the
16656 data directly through the decoder. Unfortunately it doesn't reset
16657 the context it uses. This means that if EOF is reached an attempt
16658 is made to pass two EOFs through the context and this causes the
16659 resulting error. This can also cause other problems as well. As is
16660 usual with these problems it takes *ages* to find and the fix is
16661 trivial: move one line.
16663 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16665 * Ugly workaround to get s_client and s_server working under Windows. The
16666 old code wouldn't work because it needed to select() on sockets and the
16667 tty (for keypresses and to see if data could be written). Win32 only
16668 supports select() on sockets so we select() with a 1s timeout on the
16669 sockets and then see if any characters are waiting to be read, if none
16670 are present then we retry, we also assume we can always write data to
16671 the tty. This isn't nice because the code then blocks until we've
16672 received a complete line of data and it is effectively polling the
16673 keyboard at 1s intervals: however it's quite a bit better than not
16674 working at all :-) A dedicated Windows application might handle this
16675 with an event loop for example.
16679 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16680 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16681 will be called when RSA_sign() and RSA_verify() are used. This is useful
16682 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16683 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16684 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16685 This necessitated the support of an extra signature type NID_md5_sha1
16686 for SSL signatures and modifications to the SSL library to use it instead
16687 of calling RSA_public_decrypt() and RSA_private_encrypt().
16691 * Add new -verify -CAfile and -CApath options to the crl program, these
16692 will lookup a CRL issuers certificate and verify the signature in a
16693 similar way to the verify program. Tidy up the crl program so it
16694 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16695 less strict. It will now permit CRL extensions even if it is not
16696 a V2 CRL: this will allow it to tolerate some broken CRLs.
16700 * Initialize all non-automatic variables each time one of the openssl
16701 sub-programs is started (this is necessary as they may be started
16702 multiple times from the "OpenSSL>" prompt).
16704 *Lennart Bang, Bodo Moeller*
16706 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16707 removing all other RSA functionality (this is what NO_RSA does). This
16708 is so (for example) those in the US can disable those operations covered
16709 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16714 * Non-copying interface to BIO pairs.
16715 (still largely untested)
16719 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16720 ASCII string. This was handled independently in various places before.
16724 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16725 UTF8 strings a character at a time.
16729 * Use client_version from client hello to select the protocol
16730 (s23_srvr.c) and for RSA client key exchange verification
16731 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16735 * Add various utility functions to handle SPKACs, these were previously
16736 handled by poking round in the structure internals. Added new function
16737 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16738 print, verify and generate SPKACs. Based on an original idea from
16739 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16743 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16747 * Allow the config file extension section to be overwritten on the
16748 command line. Based on an original idea from Massimiliano Pala
16749 <madwolf@comune.modena.it>. The new option is called -extensions
16750 and can be applied to ca, req and x509. Also -reqexts to override
16751 the request extensions in req and -crlexts to override the crl extensions
16756 * Add new feature to the SPKAC handling in ca. Now you can include
16757 the same field multiple times by preceding it by "XXXX." for example:
16760 this is the same syntax as used in the req config file.
16764 * Allow certificate extensions to be added to certificate requests. These
16765 are specified in a 'req_extensions' option of the req section of the
16766 config file. They can be printed out with the -text option to req but
16767 are otherwise ignored at present.
16771 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16772 data read consists of only the final block it would not decrypted because
16773 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16774 A misplaced 'break' also meant the decrypted final block might not be
16775 copied until the next read.
16779 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16780 a few extra parameters to the DH structure: these will be useful if
16781 for example we want the value of 'q' or implement X9.42 DH.
16785 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16786 provides hooks that allow the default DSA functions or functions on a
16787 "per key" basis to be replaced. This allows hardware acceleration and
16788 hardware key storage to be handled without major modification to the
16789 library. Also added low level modexp hooks and CRYPTO_EX structure and
16790 associated functions.
16794 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16795 as "read only": it can't be written to and the buffer it points to will
16796 not be freed. Reading from a read only BIO is much more efficient than
16797 a normal memory BIO. This was added because there are several times when
16798 an area of memory needs to be read from a BIO. The previous method was
16799 to create a memory BIO and write the data to it, this results in two
16800 copies of the data and an O(n^2) reading algorithm. There is a new
16801 function BIO_new_mem_buf() which creates a read only memory BIO from
16802 an area of memory. Also modified the PKCS#7 routines to use read only
16807 * Bugfix: ssl23_get_client_hello did not work properly when called in
16808 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16809 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16810 but a retry condition occurred while trying to read the rest.
16814 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16815 NID_pkcs7_encrypted by default: this was wrong since this should almost
16816 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16817 the encrypted data type: this is a more sensible place to put it and it
16818 allows the PKCS#12 code to be tidied up that duplicated this
16823 * Changed obj_dat.pl script so it takes its input and output files on
16824 the command line. This should avoid shell escape redirection problems
16829 * Initial support for certificate extension requests, these are included
16830 in things like Xenroll certificate requests. Included functions to allow
16831 extensions to be obtained and added.
16835 * -crlf option to s_client and s_server for sending newlines as
16836 CRLF (as required by many protocols).
16840 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16842 * Install libRSAglue.a when OpenSSL is built with RSAref.
16844 *Ralf S. Engelschall*
16846 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16848 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16850 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16855 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16856 DH parameters/keys (q is lost during that conversion, but the resulting
16857 DH parameters contain its length).
16859 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16860 much faster than DH_generate_parameters (which creates parameters
16861 where `p = 2*q + 1`), and also the smaller q makes DH computations
16862 much more efficient (160-bit exponentiation instead of 1024-bit
16863 exponentiation); so this provides a convenient way to support DHE
16864 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16865 utter importance to use
16866 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16868 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16869 when such DH parameters are used, because otherwise small subgroup
16870 attacks may become possible!
16874 * Avoid memory leak in i2d_DHparams.
16878 * Allow the -k option to be used more than once in the enc program:
16879 this allows the same encrypted message to be read by multiple recipients.
16883 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16884 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16885 it will always use the numerical form of the OID, even if it has a short
16890 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16891 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16892 otherwise bn_mod_exp was called. In the case of hardware keys for example
16893 no private key components need be present and it might store extra data
16894 in the RSA structure, which cannot be accessed from bn_mod_exp.
16895 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16896 private key operations.
16900 * Added support for SPARC Linux.
16904 * pem_password_cb function type incompatibly changed from
16905 typedef int pem_password_cb(char *buf, int size, int rwflag);
16907 ....(char *buf, int size, int rwflag, void *userdata);
16908 so that applications can pass data to their callbacks:
16909 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16910 additional void * argument, which is just handed through whenever
16911 the password callback is called.
16913 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16915 New function SSL_CTX_set_default_passwd_cb_userdata.
16917 Compatibility note: As many C implementations push function arguments
16918 onto the stack in reverse order, the new library version is likely to
16919 interoperate with programs that have been compiled with the old
16920 pem_password_cb definition (PEM_whatever takes some data that
16921 happens to be on the stack as its last argument, and the callback
16922 just ignores this garbage); but there is no guarantee whatsoever that
16925 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16926 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16927 problems not only on Windows, but also on some Unix platforms.
16928 To avoid problematic command lines, these definitions are now in an
16929 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16930 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16934 * MIPS III/IV assembler module is reimplemented.
16938 * More DES library cleanups: remove references to srand/rand and
16939 delete an unused file.
16943 * Add support for the free Netwide assembler (NASM) under Win32,
16944 since not many people have MASM (ml) and it can be hard to obtain.
16945 This is currently experimental but it seems to work OK and pass all
16946 the tests. Check out INSTALL.W32 for info.
16950 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16951 without temporary keys kept an extra copy of the server key,
16952 and connections with temporary keys did not free everything in case
16957 * New function RSA_check_key and new openssl rsa option -check
16958 for verifying the consistency of RSA keys.
16960 *Ulf Moeller, Bodo Moeller*
16962 * Various changes to make Win32 compile work:
16963 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16964 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16965 comparison" warnings.
16966 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16970 * Add a debugging option to PKCS#5 v2 key generation function: when
16971 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16972 derived keys are printed to stderr.
16976 * Copy the flags in ASN1_STRING_dup().
16978 *Roman E. Pavlov <pre@mo.msk.ru>*
16980 * The x509 application mishandled signing requests containing DSA
16981 keys when the signing key was also DSA and the parameters didn't match.
16983 It was supposed to omit the parameters when they matched the signing key:
16984 the verifying software was then supposed to automatically use the CA's
16985 parameters if they were absent from the end user certificate.
16987 Omitting parameters is no longer recommended. The test was also
16988 the wrong way round! This was probably due to unusual behaviour in
16989 EVP_cmp_parameters() which returns 1 if the parameters match.
16990 This meant that parameters were omitted when they *didn't* match and
16991 the certificate was useless. Certificates signed with 'ca' didn't have
16994 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16996 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16997 The interface is as follows:
16998 Applications can use
16999 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
17000 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
17001 "off" is now the default.
17002 The library internally uses
17003 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
17004 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
17005 to disable memory-checking temporarily.
17007 Some inconsistent states that previously were possible (and were
17008 even the default) are now avoided.
17010 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
17011 with each memory chunk allocated; this is occasionally more helpful
17012 than just having a counter.
17014 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
17016 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
17021 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
17022 which largely parallels "options", but is for changing API behaviour,
17023 whereas "options" are about protocol behaviour.
17024 Initial "mode" flags are:
17026 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
17027 a single record has been written.
17028 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
17029 retries use the same buffer location.
17030 (But all of the contents must be
17035 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17038 * Fix problems with no-hmac etc.
17040 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17042 * New functions RSA_get_default_method(), RSA_set_method() and
17043 RSA_get_method(). These allows replacement of RSA_METHODs without having
17044 to mess around with the internals of an RSA structure.
17048 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17049 Also really enable memory leak checks in openssl.c and in some
17052 *Chad C. Mulligan, Bodo Moeller*
17054 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17055 up the length of negative integers. This has now been simplified to just
17056 store the length when it is first determined and use it later, rather
17057 than trying to keep track of where data is copied and updating it to
17059 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17061 * Add a new function PKCS7_signatureVerify. This allows the verification
17062 of a PKCS#7 signature but with the signing certificate passed to the
17063 function itself. This contrasts with PKCS7_dataVerify which assumes the
17064 certificate is present in the PKCS#7 structure. This isn't always the
17065 case: certificates can be omitted from a PKCS#7 structure and be
17066 distributed by "out of band" means (such as a certificate database).
17070 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17071 function prototypes in pem.h, also change util/mkdef.pl to add the
17072 necessary function names.
17076 * mk1mf.pl (used by Windows builds) did not properly read the
17077 options set by Configure in the top level Makefile, and Configure
17078 was not even able to write more than one option correctly.
17079 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17083 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17084 file to be loaded from a BIO or FILE pointer. The BIO version will
17085 for example allow memory BIOs to contain config info.
17089 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17090 Whoever hopes to achieve shared-library compatibility across versions
17091 must use this, not the compile-time macro.
17092 (Exercise 0.9.4: Which is the minimum library version required by
17094 Note: All this applies only to multi-threaded programs, others don't
17099 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17100 through a BIO pair triggered the default case, i.e.
17101 SSLerr(...,SSL_R_UNKNOWN_STATE).
17105 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17106 can use the SSL library even if none of the specific BIOs is
17111 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17112 for the encoded length.
17114 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17116 * Add initial documentation of the X509V3 functions.
17120 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17121 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17122 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17123 secure PKCS#8 private key format with a high iteration count.
17127 * Fix determination of Perl interpreter: A perl or perl5
17128 *directory* in $PATH was also accepted as the interpreter.
17130 *Ralf S. Engelschall*
17132 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17133 wrong with it but it was very old and did things like calling
17134 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17135 unusual formatting.
17139 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17140 to use the new extension code.
17144 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17145 with macros. This should make it easier to change their form, add extra
17146 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17151 * Add to configuration table a new entry that can specify an alternative
17152 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17153 according to Mark Crispin <MRC@Panda.COM>.
17157 * DES CBC did not update the IV. Weird.
17161 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17162 Changing the behaviour of the former might break existing programs --
17163 where IV updating is needed, des_ncbc_encrypt can be used.
17166 * When bntest is run from "make test" it drives bc to check its
17167 calculations, as well as internally checking them. If an internal check
17168 fails, it needs to cause bc to give a non-zero result or make test carries
17169 on without noticing the failure. Fixed.
17173 * DES library cleanups.
17177 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17178 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17179 ciphers. NOTE: although the key derivation function has been verified
17180 against some published test vectors it has not been extensively tested
17181 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17186 * Instead of "mkdir -p", which is not fully portable, use new
17187 Perl script "util/mkdir-p.pl".
17191 * Rewrite the way password based encryption (PBE) is handled. It used to
17192 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17193 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17194 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17195 the 'parameter' field of the AlgorithmIdentifier is passed to the
17196 underlying key generation function so it must do its own ASN1 parsing.
17197 This has also changed the EVP_PBE_CipherInit() function which now has a
17198 'parameter' argument instead of literal salt and iteration count values
17199 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17203 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17204 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17205 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17206 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17207 value was just used as a "magic string" and not used directly its
17208 value doesn't matter.
17212 * Introduce some semblance of const correctness to BN. Shame C doesn't
17217 * "linux-sparc64" configuration (ultrapenguin).
17219 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17220 "linux-sparc" configuration.
17222 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17224 * config now generates no-xxx options for missing ciphers.
17228 * Support the EBCDIC character set (work in progress).
17229 File ebcdic.c not yet included because it has a different license.
17231 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17233 * Support BS2000/OSD-POSIX.
17235 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17237 * Make callbacks for key generation use `void *` instead of `char *`.
17241 * Make S/MIME samples compile (not yet tested).
17245 * Additional typesafe stacks.
17249 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17253 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17255 * New configuration variant "sco5-gcc".
17257 * Updated some demos.
17259 *Sean O Riordain, Wade Scholine*
17261 * Add missing BIO_free at exit of pkcs12 application.
17265 * Fix memory leak in conf.c.
17269 * Updates for Win32 to assembler version of MD5.
17273 * Set #! path to perl in `apps/der_chop` to where we found it
17274 instead of using a fixed path.
17278 * SHA library changes for irix64-mips4-cc.
17282 * Improvements for VMS support.
17286 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17288 * Bignum library bug fix. IRIX 6 passes "make test" now!
17289 This also avoids the problems with SC4.2 and unpatched SC5.
17291 *Andy Polyakov <appro@fy.chalmers.se>*
17293 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17294 These are required because of the typesafe stack would otherwise break
17295 existing code. If old code used a structure member which used to be STACK
17296 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17297 sk_num or sk_value it would produce an error because the num, data members
17298 are not present in STACK_OF. Now it just produces a warning. sk_set
17299 replaces the old method of assigning a value to sk_value
17300 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17301 that does this will no longer work (and should use sk_set instead) but
17302 this could be regarded as a "questionable" behaviour anyway.
17306 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17307 correctly handle encrypted S/MIME data.
17311 * Change type of various DES function arguments from des_cblock
17312 (which means, in function argument declarations, pointer to char)
17313 to des_cblock * (meaning pointer to array with 8 char elements),
17314 which allows the compiler to do more typechecking; it was like
17315 that back in SSLeay, but with lots of ugly casts.
17317 Introduce new type const_des_cblock.
17321 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17322 problems: find RecipientInfo structure that matches recipient certificate
17323 and initialise the ASN1 structures properly based on passed cipher.
17327 * Belatedly make the BN tests actually check the results.
17331 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17332 to and from BNs: it was completely broken. New compilation option
17333 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17334 key elements as negative integers.
17338 * Reorganize and speed up MD5.
17340 *Andy Polyakov <appro@fy.chalmers.se>*
17344 *Richard Levitte <richard@levitte.org>*
17346 * New option -out to asn1parse to allow the parsed structure to be
17347 output to a file. This is most useful when combined with the -strparse
17348 option to examine the output of things like OCTET STRINGS.
17352 * Make SSL library a little more fool-proof by not requiring any longer
17353 that `SSL_set_{accept,connect}_state` be called before
17354 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17355 in many applications because usually everything *appeared* to work as
17356 intended anyway -- now it really works as intended).
17360 * Move openssl.cnf out of lib/.
17364 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17365 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17366 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17368 *Ralf S. Engelschall*
17370 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17371 handle PKCS#7 enveloped data properly.
17373 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17375 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17376 copying pointers. The cert_st handling is changed by this in
17377 various ways (and thus what used to be known as ctx->default_cert
17378 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17379 any longer when s->cert does not give us what we need).
17380 ssl_cert_instantiate becomes obsolete by this change.
17381 As soon as we've got the new code right (possibly it already is?),
17382 we have solved a couple of bugs of the earlier code where s->cert
17383 was used as if it could not have been shared with other SSL structures.
17385 Note that using the SSL API in certain dirty ways now will result
17386 in different behaviour than observed with earlier library versions:
17387 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17388 does not influence s as it used to.
17390 In order to clean up things more thoroughly, inside SSL_SESSION
17391 we don't use CERT any longer, but a new structure SESS_CERT
17392 that holds per-session data (if available); currently, this is
17393 the peer's certificate chain and, for clients, the server's certificate
17394 and temporary key. CERT holds only those values that can have
17395 meaningful defaults in an SSL_CTX.
17399 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17400 from the internal representation. Various PKCS#7 fixes: remove some
17401 evil casts and set the enc_dig_alg field properly based on the signing
17406 * Allow PKCS#12 password to be set from the command line or the
17407 environment. Let 'ca' get its config file name from the environment
17408 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17413 * Allow certificate policies extension to use an IA5STRING for the
17414 organization field. This is contrary to the PKIX definition but
17415 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17420 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17421 without disallowing inline assembler and the like for non-pedantic builds.
17425 * Support Borland C++ builder.
17427 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17433 * SHA-1 cleanups and performance enhancements.
17435 *Andy Polyakov <appro@fy.chalmers.se>*
17437 * Sparc v8plus assembler for the bignum library.
17439 *Andy Polyakov <appro@fy.chalmers.se>*
17441 * Accept any -xxx and +xxx compiler options in Configure.
17445 * Update HPUX configuration.
17449 * Add missing `sk_<type>_unshift()` function to safestack.h
17451 *Ralf S. Engelschall*
17453 * New function SSL_CTX_use_certificate_chain_file that sets the
17454 "extra_cert"s in addition to the certificate. (This makes sense
17455 only for "PEM" format files, as chains as a whole are not
17460 * Support verify_depth from the SSL API.
17461 x509_vfy.c had what can be considered an off-by-one-error:
17462 Its depth (which was not part of the external interface)
17463 was actually counting the number of certificates in a chain;
17464 now it really counts the depth.
17468 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17469 instead of X509err, which often resulted in confusing error
17470 messages since the error codes are not globally unique
17471 (e.g. an alleged error in ssl3_accept when a certificate
17472 didn't match the private key).
17474 * New function SSL_CTX_set_session_id_context that allows to set a default
17475 value (so that you don't need SSL_set_session_id_context for each
17476 connection using the SSL_CTX).
17480 * OAEP decoding bug fix.
17484 * Support INSTALL_PREFIX for package builders, as proposed by
17489 * New Configure options "threads" and "no-threads". For systems
17490 where the proper compiler options are known (currently Solaris
17491 and Linux), "threads" is the default.
17495 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17499 * Install various scripts to $(OPENSSLDIR)/misc, not to
17500 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17501 such as /usr/local/bin.
17505 * "make linux-shared" to build shared libraries.
17507 *Niels Poppe <niels@netbox.org>*
17509 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17513 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17514 extension adding in x509 utility.
17518 * Remove NOPROTO sections and error code comments.
17522 * Partial rewrite of the DEF file generator to now parse the ANSI
17527 * New Configure options --prefix=DIR and --openssldir=DIR.
17531 * Complete rewrite of the error code script(s). It is all now handled
17532 by one script at the top level which handles error code gathering,
17533 header rewriting and C source file generation. It should be much better
17534 than the old method: it now uses a modified version of Ulf's parser to
17535 read the ANSI prototypes in all header files (thus the old K&R definitions
17536 aren't needed for error creation any more) and do a better job of
17537 translating function codes into names. The old 'ASN1 error code embedded
17538 in a comment' is no longer necessary and it doesn't use .err files which
17539 have now been deleted. Also the error code call doesn't have to appear all
17540 on one line (which resulted in some large lines...).
17544 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17548 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17549 0 (which usually indicates a closed connection), but continue reading.
17553 * Fix some race conditions.
17557 * Add support for CRL distribution points extension. Add Certificate
17558 Policies and CRL distribution points documentation.
17562 * Move the autogenerated header file parts to crypto/opensslconf.h.
17566 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17567 8 of keying material. Merlin has also confirmed interop with this fix
17568 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17570 *Merlin Hughes <merlin@baltimore.ie>*
17572 * Fix lots of warnings.
17574 *Richard Levitte <levitte@stacken.kth.se>*
17576 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17577 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17579 *Richard Levitte <levitte@stacken.kth.se>*
17581 * Fix problems with sizeof(long) == 8.
17583 *Andy Polyakov <appro@fy.chalmers.se>*
17585 * Change functions to ANSI C.
17589 * Fix typos in error codes.
17591 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17593 * Remove defunct assembler files from Configure.
17597 * SPARC v8 assembler BIGNUM implementation.
17599 *Andy Polyakov <appro@fy.chalmers.se>*
17601 * Support for Certificate Policies extension: both print and set.
17602 Various additions to support the r2i method this uses.
17606 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17607 return a const string when you are expecting an allocated buffer.
17611 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17612 types DirectoryString and DisplayText.
17616 * Add code to allow r2i extensions to access the configuration database,
17617 add an LHASH database driver and add several ctx helper functions.
17621 * Fix an evil bug in bn_expand2() which caused various BN functions to
17622 fail when they extended the size of a BIGNUM.
17626 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17627 support typesafe stack.
17631 * Fix typo in SSL_[gs]et_options().
17633 *Nils Frostberg <nils@medcom.se>*
17635 * Delete various functions and files that belonged to the (now obsolete)
17636 old X509V3 handling code.
17640 * New Configure option "rsaref".
17644 * Don't auto-generate pem.h.
17648 * Introduce type-safe ASN.1 SETs.
17652 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17654 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17656 * Introduce type-safe STACKs. This will almost certainly break lots of code
17657 that links with OpenSSL (well at least cause lots of warnings), but fear
17658 not: the conversion is trivial, and it eliminates loads of evil casts. A
17659 few STACKed things have been converted already. Feel free to convert more.
17660 In the fullness of time, I'll do away with the STACK type altogether.
17664 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17665 specified in `<certfile>` by updating the entry in the index.txt file.
17666 This way one no longer has to edit the index.txt file manually for
17667 revoking a certificate. The -revoke option does the gory details now.
17669 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17671 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17672 `-text` option at all and this way the `-noout -text` combination was
17673 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17675 *Ralf S. Engelschall*
17677 * Make sure a corresponding plain text error message exists for the
17678 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17679 verify callback function determined that a certificate was revoked.
17681 *Ralf S. Engelschall*
17683 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17684 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17685 all available ciphers including rc5, which was forgotten until now.
17686 In order to let the testing shell script know which algorithms
17687 are available, a new (up to now undocumented) command
17688 `openssl list-cipher-commands` is used.
17692 * Bugfix: s_client occasionally would sleep in select() when
17693 it should have checked SSL_pending() first.
17697 * New functions DSA_do_sign and DSA_do_verify to provide access to
17698 the raw DSA values prior to ASN.1 encoding.
17702 * Tweaks to Configure
17704 *Niels Poppe <niels@netbox.org>*
17706 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17711 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17715 * New config option to avoid instructions that are illegal on the 80386.
17716 The default code is faster, but requires at least a 486.
17720 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17721 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17722 same as SSL2_VERSION anyway.
17726 * New "-showcerts" option for s_client.
17730 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17731 application. Various cleanups and fixes.
17735 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17736 modify error routines to work internally. Add error codes and PBE init
17737 to library startup routines.
17741 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17742 packing functions to asn1 and evp. Changed function names and error
17743 codes along the way.
17747 * PKCS12 integration: and so it begins... First of several patches to
17748 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17749 objects to objects.h
17753 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17754 and display support for Thawte strong extranet extension.
17758 * Add LinuxPPC support.
17760 *Jeff Dubrule <igor@pobox.org>*
17762 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17763 bn_div_words in alpha.s.
17765 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17767 * Make sure the RSA OAEP test is skipped under -DRSAref because
17768 OAEP isn't supported when OpenSSL is built with RSAref.
17770 *Ulf Moeller <ulf@fitug.de>*
17772 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17773 so they no longer are missing under -DNOPROTO.
17775 *Soren S. Jorvang <soren@t.dk>*
17777 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17779 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17780 doesn't work when the session is reused. Coming soon!
17784 * Fix a security hole, that allows sessions to be reused in the wrong
17785 context thus bypassing client cert protection! All software that uses
17786 client certs and session caches in multiple contexts NEEDS PATCHING to
17787 allow session reuse! A fuller solution is in the works.
17789 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17791 * Some more source tree cleanups (removed obsolete files
17792 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17793 permission on "config" script to be executable) and a fix for the INSTALL
17796 *Ulf Moeller <ulf@fitug.de>*
17798 * Remove some legacy and erroneous uses of malloc, free instead of
17801 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17803 * Make rsa_oaep_test return non-zero on error.
17805 *Ulf Moeller <ulf@fitug.de>*
17807 * Add support for native Solaris shared libraries. Configure
17808 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17809 if someone would make that last step automatic.
17811 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17813 * ctx_size was not built with the right compiler during "make links". Fixed.
17817 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17818 except NULL ciphers". This means the default cipher list will no longer
17819 enable NULL ciphers. They need to be specifically enabled e.g. with
17820 the string "DEFAULT:eNULL".
17824 * Fix to RSA private encryption routines: if p < q then it would
17825 occasionally produce an invalid result. This will only happen with
17826 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17830 * Be less restrictive and allow also `perl util/perlpath.pl
17831 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17832 because this way one can also use an interpreter named `perl5` (which is
17833 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17834 installed as `perl`).
17836 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17838 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17840 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17842 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17843 advapi32.lib to Win32 build and change the pem test comparison
17844 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17845 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17846 and crypto/des/ede_cbcm_enc.c.
17850 * DES quad checksum was broken on big-endian architectures. Fixed.
17854 * Comment out two functions in bio.h that aren't implemented. Fix up the
17855 Win32 test batch file so it (might) work again. The Win32 test batch file
17856 is horrible: I feel ill....
17860 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17861 in e_os.h. Audit of header files to check ANSI and non ANSI
17862 sections: 10 functions were absent from non ANSI section and not exported
17863 from Windows DLLs. Fixed up libeay.num for new functions.
17867 * Make `openssl version` output lines consistent.
17869 *Ralf S. Engelschall*
17871 * Fix Win32 symbol export lists for BIO functions: Added
17872 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17873 to ms/libeay{16,32}.def.
17875 *Ralf S. Engelschall*
17877 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17878 fine under Unix and passes some trivial tests I've now added. But the
17879 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17880 added to make sure no one expects that this stuff really works in the
17881 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17882 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17885 *Ralf S. Engelschall*
17887 * Fix the generation of two part addresses in perl.
17889 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17891 * Add config entry for Linux on MIPS.
17893 *John Tobey <jtobey@channel1.com>*
17895 * Make links whenever Configure is run, unless we are on Windoze.
17899 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17900 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17905 * Add a useful kludge to allow package maintainers to specify compiler and
17906 other platforms details on the command line without having to patch the
17907 Configure script every time: One now can use
17908 `perl Configure <id>:<details>`,
17909 i.e. platform ids are allowed to have details appended
17910 to them (separated by colons). This is treated as there would be a static
17911 pre-configured entry in Configure's %table under key `<id>` with value
17912 `<details>` and `perl Configure <id>` is called. So, when you want to
17913 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17914 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17915 now, which overrides the FreeBSD-elf entry on-the-fly.
17917 *Ralf S. Engelschall*
17919 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17923 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17924 on the `perl Configure ...` command line. This way one can compile
17925 OpenSSL libraries with Position Independent Code (PIC) which is needed
17926 for linking it into DSOs.
17928 *Ralf S. Engelschall*
17930 * Remarkably, export ciphers were totally broken and no-one had noticed!
17935 * Cleaned up the LICENSE document: The official contact for any license
17936 questions now is the OpenSSL core team under openssl-core@openssl.org.
17937 And add a paragraph about the dual-license situation to make sure people
17938 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17939 to the OpenSSL toolkit.
17941 *Ralf S. Engelschall*
17943 * General source tree makefile cleanups: Made `making xxx in yyy...`
17944 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17945 Additionally cleaned up the `make links` target: Remove unnecessary
17946 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17947 to speed processing and no longer clutter the display with confusing
17948 stuff. Instead only the actually done links are displayed.
17950 *Ralf S. Engelschall*
17952 * Permit null encryption ciphersuites, used for authentication only. It used
17953 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17954 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17959 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17960 signed attributes when verifying signatures (this would break them),
17961 the detached data encoding was wrong and public keys obtained using
17962 X509_get_pubkey() weren't freed.
17966 * Add text documentation for the BUFFER functions. Also added a work around
17967 to a Win95 console bug. This was triggered by the password read stuff: the
17968 last character typed gets carried over to the next fread(). If you were
17969 generating a new cert request using 'req' for example then the last
17970 character of the passphrase would be CR which would then enter the first
17975 * Added the new 'Includes OpenSSL Cryptography Software' button as
17976 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17977 button and can be used by applications based on OpenSSL to show the
17978 relationship to the OpenSSL project.
17980 *Ralf S. Engelschall*
17982 * Remove confusing variables in function signatures in files
17983 ssl/ssl_lib.c and ssl/ssl.h.
17985 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17987 * Don't install bss_file.c under PREFIX/include/
17989 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17991 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17992 functions that return function pointers and has support for NT specific
17993 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17994 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17995 unsigned to signed types: this was killing the Win32 compile.
17999 * Add new certificate file to stack functions,
18000 SSL_add_dir_cert_subjects_to_stack() and
18001 SSL_add_file_cert_subjects_to_stack(). These largely supplant
18002 SSL_load_client_CA_file(), and can be used to add multiple certs easily
18003 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
18004 This means that Apache-SSL and similar packages don't have to mess around
18005 to add as many CAs as they want to the preferred list.
18009 * Experiment with doxygen documentation. Currently only partially applied to
18011 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
18012 openssl.doxy as the configuration file.
18016 * Get rid of remaining C++-style comments which strict C compilers hate.
18018 *Ralf S. Engelschall, pointed out by Carlos Amengual*
18020 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
18021 compiled in by default: it has problems with large keys.
18025 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
18026 DH private keys and/or callback functions which directly correspond to
18027 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
18028 is needed for applications which have to configure certificates on a
18029 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18031 For the RSA certificate situation is makes no difference, but
18032 for the DSA certificate situation this fixes the "no shared cipher"
18033 problem where the OpenSSL cipher selection procedure failed because the
18034 temporary keys were not overtaken from the context and the API provided
18035 no way to reconfigure them.
18036 The new functions now let applications reconfigure the stuff and they
18037 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18038 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18039 non-public-API function ssl_cert_instantiate() is used as a helper
18040 function and also to reduce code redundancy inside ssl_rsa.c.
18042 *Ralf S. Engelschall*
18044 * Move s_server -dcert and -dkey options out of the undocumented feature
18045 area because they are useful for the DSA situation and should be
18046 recognized by the users.
18048 *Ralf S. Engelschall*
18050 * Fix the cipher decision scheme for export ciphers: the export bits are
18051 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18052 SSL_EXP_MASK. So, the original variable has to be used instead of the
18053 already masked variable.
18055 *Richard Levitte <levitte@stacken.kth.se>*
18057 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18059 *Richard Levitte <levitte@stacken.kth.se>*
18061 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18062 from `int` to `unsigned int` because it is a length and initialized by
18063 EVP_DigestFinal() which expects an `unsigned int *`.
18065 *Richard Levitte <levitte@stacken.kth.se>*
18067 * Don't hard-code path to Perl interpreter on shebang line of Configure
18068 script. Instead use the usual Shell->Perl transition trick.
18070 *Ralf S. Engelschall*
18072 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18073 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18074 -noout -modulus` as it's already the case for `openssl rsa -noout
18075 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18076 currently the public key is printed (a decision which was already done by
18077 `openssl dsa -modulus` in the past) which serves a similar purpose.
18078 Additionally the NO_RSA no longer completely removes the whole -modulus
18079 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18082 *Ralf S. Engelschall*
18084 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18085 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18087 *Arne Ansper <arne@ats.cyber.ee>*
18089 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18090 to be added. Now both 'req' and 'ca' can use new objects defined in the
18095 * Add cool BIO that does syslog (or event log on NT).
18097 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18099 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18100 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18101 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18102 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18106 * Add preliminary config info for new extension code.
18110 * Make RSA_NO_PADDING really use no padding.
18112 *Ulf Moeller <ulf@fitug.de>*
18114 * Generate errors when private/public key check is done.
18118 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18119 for some CRL extensions and new objects added.
18123 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18124 key usage extension and fuller support for authority key id.
18128 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18129 padding method for RSA, which is recommended for new applications in PKCS
18130 #1 v2.0 (RFC 2437, October 1998).
18131 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18132 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18133 against Bleichbacher's attack on RSA.
18134 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18137 * Updates to the new SSL compression code
18139 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18141 * Fix so that the version number in the master secret, when passed
18142 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18143 (because the server will not accept higher), that the version number
18144 is 0x03,0x01, not 0x03,0x00
18146 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18148 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18149 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18150 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18154 * Support for RAW extensions where an arbitrary extension can be
18155 created by including its DER encoding. See `apps/openssl.cnf` for
18160 * Make sure latest Perl versions don't interpret some generated C array
18161 code as Perl array code in the crypto/err/err_genc.pl script.
18163 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18165 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18166 not many people have the assembler. Various Win32 compilation fixes and
18167 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18168 build instructions.
18172 * Modify configure script 'Configure' to automatically create crypto/date.h
18173 file under Win32 and also build pem.h from pem.org. New script
18174 util/mkfiles.pl to create the MINFO file on environments that can't do a
18175 'make files': perl util/mkfiles.pl >MINFO should work.
18179 * Major rework of DES function declarations, in the pursuit of correctness
18180 and purity. As a result, many evil casts evaporated, and some weirdness,
18181 too. You may find this causes warnings in your code. Zapping your evil
18182 casts will probably fix them. Mostly.
18186 * Fix for a typo in asn1.h. Bug fix to object creation script
18187 obj_dat.pl. It considered a zero in an object definition to mean
18188 "end of object": none of the objects in objects.h have any zeros
18189 so it wasn't spotted.
18191 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18193 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18194 Masking (CBCM). In the absence of test vectors, the best I have been able
18195 to do is check that the decrypt undoes the encrypt, so far. Send me test
18196 vectors if you have them.
18200 * Correct calculation of key length for export ciphers (too much space was
18201 allocated for null ciphers). This has not been tested!
18205 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18206 message is now correct (it understands "crypto" and "ssl" on its
18207 command line). There is also now an "update" option. This will update
18208 the util/ssleay.num and util/libeay.num files with any new functions.
18210 perl util/mkdef.pl crypto ssl update
18211 it will update them.
18215 * Overhauled the Perl interface:
18216 - ported BN stuff to OpenSSL's different BN library
18217 - made the perl/ source tree CVS-aware
18218 - renamed the package from SSLeay to OpenSSL (the files still contain
18219 their history because I've copied them in the repository)
18220 - removed obsolete files (the test scripts will be replaced
18221 by better Test::Harness variants in the future)
18223 *Ralf S. Engelschall*
18225 * First cut for a very conservative source tree cleanup:
18226 1. merge various obsolete readme texts into doc/ssleay.txt
18227 where we collect the old documents and readme texts.
18228 2. remove the first part of files where I'm already sure that we no
18229 longer need them because of three reasons: either they are just temporary
18230 files which were left by Eric or they are preserved original files where
18231 I've verified that the diff is also available in the CVS via "cvs diff
18232 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18233 the crypto/md/ stuff).
18235 *Ralf S. Engelschall*
18237 * More extension code. Incomplete support for subject and issuer alt
18238 name, issuer and authority key id. Change the i2v function parameters
18239 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18240 what that's for :-) Fix to ASN1 macro which messed up
18241 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18245 * Preliminary support for ENUMERATED type. This is largely copied from the
18250 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18252 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18254 * Make sure `make rehash` target really finds the `openssl` program.
18256 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18258 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18259 like to hear about it if this slows down other processors.
18263 * Add CygWin32 platform information to Configure script.
18265 *Alan Batie <batie@aahz.jf.intel.com>*
18267 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18269 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18271 * New program nseq to manipulate netscape certificate sequences
18275 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18280 * Fixes to BN code. Previously the default was to define BN_RECURSION
18281 but the BN code had some problems that would cause failures when
18282 doing certificate verification and some other functions.
18284 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18286 * Add ASN1 and PEM code to support netscape certificate sequences.
18290 * Add ASN1 and PEM code to support netscape certificate sequences.
18294 * Add several PKIX and private extended key usage OIDs.
18298 * Modify the 'ca' program to handle the new extension code. Modify
18299 openssl.cnf for new extension format, add comments.
18303 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18304 and add a sample to openssl.cnf so req -x509 now adds appropriate
18309 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18310 error code, add initial support to X509_print() and x509 application.
18314 * Takes a deep breath and start adding X509 V3 extension support code. Add
18315 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18316 stuff is currently isolated and isn't even compiled yet.
18320 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18321 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18322 Removed the versions check from X509 routines when loading extensions:
18323 this allows certain broken certificates that don't set the version
18324 properly to be processed.
18328 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18329 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18330 can still be regenerated with "make depend".
18334 * Spelling mistake in C version of CAST-128.
18336 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18338 * Changes to the error generation code. The perl script err-code.pl
18339 now reads in the old error codes and retains the old numbers, only
18340 adding new ones if necessary. It also only changes the .err files if new
18341 codes are added. The makefiles have been modified to only insert errors
18342 when needed (to avoid needlessly modifying header files). This is done
18343 by only inserting errors if the .err file is newer than the auto generated
18344 C file. To rebuild all the error codes from scratch (the old behaviour)
18345 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18346 or delete all the .err files.
18350 * CAST-128 was incorrectly implemented for short keys. The C version has
18351 been fixed, but is untested. The assembler versions are also fixed, but
18352 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18353 to regenerate it if needed.
18354 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18355 Hagino <itojun@kame.net>*
18357 * File was opened incorrectly in randfile.c.
18359 *Ulf Möller <ulf@fitug.de>*
18361 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18362 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18363 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18364 al: it's just almost always a UTCTime. Note this patch adds new error
18365 codes so do a "make errors" if there are problems.
18369 * Correct Linux 1 recognition in config.
18371 *Ulf Möller <ulf@fitug.de>*
18373 * Remove pointless MD5 hash when using DSA keys in ca.
18375 *Anonymous <nobody@replay.com>*
18377 * Generate an error if given an empty string as a cert directory. Also
18378 generate an error if handed NULL (previously returned 0 to indicate an
18379 error, but didn't set one).
18381 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18383 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18387 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18388 parameters. This was causing a warning which killed off the Win32 compile.
18392 * Remove C++ style comments from crypto/bn/bn_local.h.
18394 *Neil Costigan <neil.costigan@celocom.com>*
18396 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18397 based on a text string, looking up short and long names and finally
18398 "dot" format. The "dot" format stuff didn't work. Added new function
18399 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18400 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18401 OID is not part of the table.
18405 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18406 X509_LOOKUP_by_alias().
18410 * Sort openssl functions by name.
18414 * Get the `gendsa` command working and add it to the `list` command. Remove
18415 encryption from sample DSA keys (in case anyone is interested the password
18420 * Make *all* `*_free` functions accept a NULL pointer.
18422 *Frans Heymans <fheymans@isaserver.be>*
18424 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18427 *Anonymous <nobody@replay.com>*
18429 * s_server should send the CAfile as acceptable CAs, not its own cert.
18431 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18433 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18435 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18437 * Temp key "for export" tests were wrong in s3_srvr.c.
18439 *Anonymous <nobody@replay.com>*
18441 * Add prototype for temp key callback functions
18442 SSL_CTX_set_tmp_{rsa,dh}_callback().
18446 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18447 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18451 * X509_name_add_entry() freed the wrong thing after an error.
18453 *Arne Ansper <arne@ats.cyber.ee>*
18455 * rsa_eay.c would attempt to free a NULL context.
18457 *Arne Ansper <arne@ats.cyber.ee>*
18459 * BIO_s_socket() had a broken should_retry() on Windoze.
18461 *Arne Ansper <arne@ats.cyber.ee>*
18463 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18465 *Arne Ansper <arne@ats.cyber.ee>*
18467 * Make sure the already existing X509_STORE->depth variable is initialized
18468 in X509_STORE_new(), but document the fact that this variable is still
18469 unused in the certificate verification process.
18471 *Ralf S. Engelschall*
18473 * Fix the various library and `apps/` files to free up pkeys obtained from
18474 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18478 * Fix reference counting in X509_PUBKEY_get(). This makes
18479 demos/maurice/example2.c work, amongst others, probably.
18481 *Steve Henson and Ben Laurie*
18483 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18484 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18485 are no longer created. This way we have a single and consistent command
18486 line interface `openssl <command>`, similar to `cvs <command>`.
18488 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18490 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18491 BIT STRING wrapper always have zero unused bits.
18495 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18499 * Make the top-level INSTALL documentation easier to understand.
18503 * Makefiles updated to exit if an error occurs in a sub-directory
18504 make (including if user presses ^C) [Paul Sutton]
18506 * Make Montgomery context stuff explicit in RSA data structure.
18510 * Fix build order of pem and err to allow for generated pem.h.
18514 * Fix renumbering bug in X509_NAME_delete_entry().
18518 * Enhanced the err-ins.pl script so it makes the error library number
18519 global and can add a library name. This is needed for external ASN1 and
18520 other error libraries.
18524 * Fixed sk_insert which never worked properly.
18528 * Fix ASN1 macros so they can handle indefinite length constructed
18529 EXPLICIT tags. Some non standard certificates use these: they can now
18534 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18535 into a single doc/ssleay.txt bundle. This way the information is still
18536 preserved but no longer messes up this directory. Now it's new room for
18537 the new set of documentation files.
18539 *Ralf S. Engelschall*
18541 * SETs were incorrectly DER encoded. This was a major pain, because they
18542 shared code with SEQUENCEs, which aren't coded the same. This means that
18543 almost everything to do with SETs or SEQUENCEs has either changed name or
18544 number of arguments.
18546 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18548 * Fix test data to work with the above.
18552 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18553 was already fixed by Eric for 0.9.1 it seems.
18555 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18557 * Autodetect FreeBSD3.
18561 * Fix various bugs in Configure. This affects the following platforms:
18565 unixware-2.0-pentium
18570 * Eliminate generated files from CVS. Reorder tests to regenerate files
18571 before they are needed.
18575 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18579 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18581 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18582 changed SSLeay to OpenSSL in version strings.
18584 *Ralf S. Engelschall*
18586 * Some fixups to the top-level documents.
18590 * Fixed the nasty bug where rsaref.h was not found under compile-time
18591 because the symlink to include/ was missing.
18593 *Ralf S. Engelschall*
18595 * Incorporated the popular no-RSA/DSA-only patches
18596 which allow to compile a RSA-free SSLeay.
18598 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18600 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18601 when "ssleay" is still not found.
18603 *Ralf S. Engelschall*
18605 * Added more platforms to Configure: Cray T3E, HPUX 11,
18607 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18609 * Updated the README file.
18611 *Ralf S. Engelschall*
18613 * Added various .cvsignore files in the CVS repository subdirs
18614 to make a "cvs update" really silent.
18616 *Ralf S. Engelschall*
18618 * Recompiled the error-definition header files and added
18619 missing symbols to the Win32 linker tables.
18621 *Ralf S. Engelschall*
18623 * Cleaned up the top-level documents;
18624 o new files: CHANGES and LICENSE
18625 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18626 o merged COPYRIGHT into LICENSE
18627 o removed obsolete TODO file
18628 o renamed MICROSOFT to INSTALL.W32
18630 *Ralf S. Engelschall*
18632 * Removed dummy files from the 0.9.1b source tree:
18634 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18635 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18636 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18637 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18638 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18641 *Ralf S. Engelschall*
18643 * Added various platform portability fixes.
18647 * The Genesis of the OpenSSL rpject:
18648 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18649 Young and Tim J. Hudson created while they were working for C2Net until
18652 *The OpenSSL Project*
18654 ### Changes between 0.9.0b and 0.9.1b [not released]
18656 * Updated a few CA certificates under certs/
18660 * Changed some BIGNUM api stuff.
18664 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18665 DGUX x86, Linux Alpha, etc.
18669 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18670 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18675 * Add -strparse option to asn1pars program which parses nested
18678 *Dr Stephen Henson <shenson@bigfoot.com>*
18680 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18684 * DSA fix for "ca" program.
18688 * Added "-genkey" option to "dsaparam" program.
18692 * Added RIPE MD160 (rmd160) message digest.
18696 * Added -a (all) option to "ssleay version" command.
18700 * Added PLATFORM define which is the id given to Configure.
18704 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18708 * Extended the ASN.1 parser routines.
18712 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18716 * Added a BN_CTX to the BN library.
18720 * Fixed the weak key values in DES library
18724 * Changed API in EVP library for cipher aliases.
18728 * Added support for RC2/64bit cipher.
18732 * Converted the lhash library to the crypto/mem.c functions.
18736 * Added more recognized ASN.1 object ids.
18740 * Added more RSA padding checks for SSL/TLS.
18744 * Added BIO proxy/filter functionality.
18748 * Added extra_certs to SSL_CTX which can be used
18749 send extra CA certificates to the client in the CA cert chain sending
18750 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18754 * Now Fortezza is denied in the authentication phase because
18755 this is key exchange mechanism is not supported by SSLeay at all.
18759 * Additional PKCS1 checks.
18763 * Support the string "TLSv1" for all TLS v1 ciphers.
18767 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18768 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18772 * Fixed a few memory leaks.
18776 * Fixed various code and comment typos.
18780 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18781 bytes sent in the client random.
18783 *Edward Bishop <ebishop@spyglass.com>*
18787 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18788 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18789 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18790 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18791 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18792 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18793 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18794 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18795 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18796 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18797 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18798 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18799 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18800 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18801 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18802 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18803 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18804 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18805 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18806 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18807 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18808 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18809 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18810 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18811 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18812 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18813 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18814 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18815 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18816 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18817 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18818 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18819 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18820 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18821 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18822 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18823 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18824 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18825 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18826 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18827 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18828 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18829 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18830 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18831 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18832 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18833 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18834 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18835 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18836 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18837 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18838 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18839 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18840 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18841 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18842 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18843 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18844 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18845 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18846 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18847 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18848 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18849 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18850 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18851 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18852 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18853 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18854 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18855 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18856 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18857 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18858 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18859 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18860 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18861 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18862 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18863 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18864 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18865 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18866 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18867 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18868 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18869 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18870 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18871 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18872 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18873 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18874 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18875 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18876 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18877 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18878 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18879 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18880 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18881 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18882 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18883 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18884 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18885 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18886 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18887 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18888 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18889 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18890 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18891 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18892 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18893 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18894 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18895 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18896 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18897 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18898 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18899 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18900 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18901 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18902 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18903 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18904 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18905 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18906 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18907 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18908 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18909 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18910 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18911 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18912 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18913 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18914 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18915 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18916 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18917 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18918 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18919 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18920 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18921 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18922 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18923 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18924 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18925 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18926 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18927 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18928 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18929 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18930 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18931 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18932 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18933 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18934 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18935 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18936 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18937 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18938 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18939 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18940 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18941 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18942 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18943 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18944 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18945 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18946 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18947 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18948 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655