4 This is a high-level summary of the most important changes.
5 For a full list of changes, see the [git commit log][log] and
6 pick the appropriate release branch.
8 [log]: https://github.com/openssl/openssl/commits/
13 - [OpenSSL 3.0](#openssl-30)
14 - [OpenSSL 1.1.1](#openssl-111)
15 - [OpenSSL 1.1.0](#openssl-110)
16 - [OpenSSL 1.0.2](#openssl-102)
17 - [OpenSSL 1.0.1](#openssl-101)
18 - [OpenSSL 1.0.0](#openssl-100)
19 - [OpenSSL 0.9.x](#openssl-09x)
24 ### Changes between 1.1.1 and 3.0 [xx XXX xxxx]
26 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
27 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
28 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
29 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
30 were used to collect all necessary data to form a HTTP request, and to
31 perform the HTTP transfer with that request. With OpenSSL 3.0, the
32 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
33 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
34 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
35 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
36 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
37 OSSL_HTTP_REQ_CTX_set_max_response_length().
39 *Rich Salz and Richard Levitte*
41 * Validation of SM2 keys has been separated from the validation of regular EC
42 keys, allowing to improve the SM2 validation process to reject loaded private
43 keys that are not conforming to the SM2 ISO standard.
44 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
49 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
50 switches: a validation failure triggers an early exit, returning a failure
51 exit status to the parent process.
55 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
56 to ignore unknown ciphers.
60 * The `-cipher-commands` and `-digest-commands` options
61 of the command line utility `list` have been deprecated.
62 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
66 * Deprecated all the libcrypto and libssl error string loading
67 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
68 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
69 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
70 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
71 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
72 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
73 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
74 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
75 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
76 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
78 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
79 now loads error strings automatically.
83 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
84 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
85 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
86 are to be used by servers requiring ephemeral DH keys. Instead applications
87 should consider using the built-in DH parameters that are available by
88 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
89 necessary then applications can use the alternative functions
90 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
91 replacement for the "callback" functions. The callback was originally useful
92 in order to have different parameters for export and non-export ciphersuites.
93 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
94 functions should be replaced by one of the other methods described above.
98 * The `-crypt` option to the `passwd` command line tool has been removed.
102 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
107 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
109 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
110 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
111 The inverse ciphers use AES decryption for wrapping, and
112 AES encryption for unwrapping.
116 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
117 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
118 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
119 OpenSSL 3.0 these are replaced by the more generic functions
120 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
121 The old versions have been converted to deprecated macros that just call the
126 * The security callback, which can be customised by application code, supports
127 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
128 in the "other" parameter. In most places this is what is passed. All these
129 places occur server side. However there was one client side call of this
130 security operation and it passed a DH object instead. This is incorrect
131 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
132 of the other locations. Therefore this client side call has been changed to
133 pass an EVP_PKEY instead.
137 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
138 interface. Their functionality remains unchanged.
142 * Added new option for 'openssl list', '-providers', which will display the
143 list of loaded providers, their names, version and status. It optionally
144 displays their gettable parameters.
148 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
149 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
150 type is internally recognised so the workaround is no longer needed.
152 Functionality is still retained as it is, but will only work with
153 EVP_PKEYs with a legacy internal key.
157 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
158 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
162 * Changed all "STACK" functions to be macros instead of inline functions. Macro
163 parameters are still checked for type safety at compile time via helper
168 * Remove the RAND_DRBG API
170 The RAND_DRBG API did not fit well into the new provider concept as
171 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
172 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
173 and some of its API calls are rather low-level. This holds in particular
174 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
176 Adding a compatibility layer to continue supporting the RAND_DRBG API as
177 a legacy API for a regular deprecation period turned out to come at the
178 price of complicating the new provider API unnecessarily. Since the
179 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
182 *Paul Dale and Matthias St. Pierre*
184 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
185 as well as actual hostnames.
189 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
190 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
191 conversely, silently ignore DTLS protocol version bounds when configuring
192 TLS-based contexts. The commands can be repeated to set bounds of both
193 types. The same applies with the corresponding "min_protocol" and
194 "max_protocol" command-line switches, in case some application uses both TLS
197 SSL_CTX instances that are created for a fixed protocol version (e.g.
198 `TLSv1_server_method()`) also silently ignore version bounds. Previously
199 attempts to apply bounds to these protocol versions would result in an
200 error. Now only the "version-flexible" SSL_CTX instances are subject to
201 limits in configuration files in command-line options.
205 * Deprecated the `ENGINE` API. Engines should be replaced with providers
210 * Reworked the recorded ERR codes to make better space for system errors.
211 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
212 given code is a system error (true) or an OpenSSL error (false).
216 * Reworked the test perl framework to better allow parallel testing.
218 *Nicola Tuveri and David von Oheimb*
220 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
221 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
225 * 'Configure' has been changed to figure out the configuration target if
226 none is given on the command line. Consequently, the 'config' script is
227 now only a mere wrapper. All documentation is changed to only mention
230 *Rich Salz and Richard Levitte*
232 * Added a library context `OSSL_LIB_CTX` that applications as well as
233 other libraries can use to form a separate context within which
234 libcrypto operations are performed.
236 There are two ways this can be used:
238 - Directly, by passing a library context to functions that take
239 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
241 - Indirectly, by creating a new library context and then assigning
242 it as the new default, with `OSSL_LIB_CTX_set0_default`.
244 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
245 apart from the functions directly related to `OSSL_LIB_CTX`, accept
246 NULL to indicate that the default library context should be used.
248 Library code that changes the default library context using
249 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
250 second call before returning to the caller.
252 _(Note: the library context was initially called `OPENSSL_CTX` and
253 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
257 * Handshake now fails if Extended Master Secret extension is dropped
262 * Dropped interactive mode from the `openssl` program. From now on,
263 running it without arguments is equivalent to `openssl help`.
267 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
268 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
269 While the old function names have been retained for backward compatibility
270 they should not be used in new developments
271 because their return values are confusing: Unlike other `_cmp()` functions
272 they do not return 0 in case their arguments are equal.
276 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
277 `EC_GROUP_get_field_type()`.
281 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
282 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
283 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
284 Applications should rely on the library automatically assigning a suitable
285 EC_METHOD internally upon EC_GROUP construction.
289 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
290 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
291 assigned internally without application intervention.
292 Users of EC_GROUP_new() should switch to a different suitable constructor.
296 * Add CAdES-BES signature verification support, mostly derived
297 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
299 *Filipe Raimundo da Silva*
301 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
305 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
306 functions are not widely used and now OpenSSL automatically perform this
307 conversion when needed.
311 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
312 EC_KEY_precompute_mult(). These functions are not widely used and
313 applications should instead switch to named curves which OpenSSL has
314 hardcoded lookup tables for.
318 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
319 should instead use the L<EC_POINT_mul(3)> function.
323 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
324 that are not applicable to the new provider model. Applications should
325 instead use EVP_default_properties_is_fips_enabled() and
326 EVP_default_properties_enable_fips().
330 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
331 is set, an unexpected EOF is ignored, it pretends a close notify was received
332 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
336 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
337 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
338 used and applications should instead use the
339 L<EC_POINT_set_affine_coordinates(3)> and
340 L<EC_POINT_get_affine_coordinates(3)> functions.
344 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
345 arrays to be more easily constructed via a series of utility functions.
346 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
347 the various push functions and finally convert to a passable OSSL_PARAM
348 array using OSSL_PARAM_BLD_to_param().
352 * The security strength of SHA1 and MD5 based signatures in TLS has been
353 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
354 working at the default security level of 1 and instead requires security
355 level 0. The security level can be changed either using the cipher string
356 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
360 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
361 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
362 internal keys, if they correspond to one of those built in types.
366 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
367 contain a provider side internal key.
371 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
372 They are old functions that we don't use, and that you could disable with
373 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
377 * Project text documents not yet having a proper file name extension
378 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
379 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
383 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
384 have been converted to Markdown with the goal to produce documents
385 which not only look pretty when viewed online in the browser, but
386 remain well readable inside a plain text editor.
388 To achieve this goal, a 'minimalistic' Markdown style has been applied
389 which avoids formatting elements that interfere too much with the
390 reading flow in the text file. For example, it
392 * avoids [ATX headings][] and uses [setext headings][] instead
393 (which works for `<h1>` and `<h2>` headings only).
394 * avoids [inline links][] and uses [reference links][] instead.
395 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
397 [ATX headings]: https://github.github.com/gfm/#atx-headings
398 [setext headings]: https://github.github.com/gfm/#setext-headings
399 [inline links]: https://github.github.com/gfm/#inline-link
400 [reference links]: https://github.github.com/gfm/#reference-link
401 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
402 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
404 *Matthias St. Pierre*
406 * The test suite is changed to preserve results of each test recipe.
407 A new directory test-runs/ with subdirectories named like the
408 test recipes are created in the build tree for this purpose.
412 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
413 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
414 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
416 *David von Oheimb, Martin Peylo*
418 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
419 The legacy OCSP-focused and only partly documented API is retained for
420 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
424 * Added `util/check-format.pl`, a tool for checking adherence to the
425 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
426 The checks performed are incomplete and yield some false positives.
427 Still the tool should be useful for detecting most typical glitches.
431 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
432 If domain name resolution yields multiple IP addresses all of them are tried
433 after `connect()` failures.
437 * All of the low level RSA functions have been deprecated including:
439 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
440 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
441 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
442 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
443 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
444 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
445 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
446 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
447 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
448 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
449 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
450 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
451 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
452 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
453 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
454 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
455 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
456 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
457 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
458 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
459 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
460 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
461 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
462 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
463 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
464 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
465 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
466 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
468 Use of these low level functions has been informally discouraged for a long
469 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
470 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
471 L<EVP_PKEY_decrypt(3)>.
475 * X509 certificates signed using SHA1 are no longer allowed at security
477 In TLS/SSL the default security level is 1. It can be set either
478 using the cipher string with `@SECLEVEL`, or calling
479 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
480 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
482 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
483 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
484 options of the commands.
488 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
489 modified to use PKEY APIs. These commands are now in maintenance mode
490 and no new features will be added to them.
494 * The command line utility rsautl has been deprecated.
495 Instead use the pkeyutl program.
499 * The command line utilities genrsa and rsa have been modified to use PKEY
500 APIs. They now write PKCS#8 keys by default. These commands are now in
501 maintenance mode and no new features will be added to them.
505 * All of the low level DH functions have been deprecated including:
507 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
508 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
509 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
510 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
511 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
512 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
513 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
514 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
515 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
516 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
517 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
518 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
519 DH_meth_get_generate_params and DH_meth_set_generate_params.
521 Use of these low level functions has been informally discouraged for a long
522 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
523 and L<EVP_PKEY_derive(3)>.
525 Additionally functions that read and write DH objects such as d2i_DHparams,
526 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
527 functions have also been deprecated. Applications should instead use the
528 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
530 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
531 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
532 `EVP_PKEY_set1_DH()` are also deprecated.
533 Applications should instead either read or write an
534 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
535 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
537 *Paul Dale and Matt Caswell*
539 * All of the low level DSA functions have been deprecated including:
541 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
542 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
543 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
544 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
545 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
546 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
547 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
548 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
549 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
550 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
551 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
552 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
553 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
554 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
555 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
556 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
557 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
559 Use of these low level functions has been informally discouraged for a long
560 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
561 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
565 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
566 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
567 This means that applications don't have to look at the curve NID and
568 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
569 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
570 a no-op when the EVP_PKEY is already of the given type.
572 Parameter and key generation is also reworked to make it possible
573 to generate EVP_PKEY_SM2 parameters and keys without having to go
574 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
575 However, code that does the latter will still work as before.
579 * Deprecated low level ECDH and ECDSA functions. These include:
581 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
582 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
585 Use of these low level functions has been informally discouraged for a long
586 time. Instead applications should use the EVP_PKEY_derive(3),
587 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
591 * Deprecated the EC_KEY_METHOD functions. These include:
593 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
594 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
595 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
596 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
597 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
598 EC_KEY_METHOD_get_verify.
600 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
604 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
605 and EVP_PKEY_decrypt() instead.
606 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
607 and EVP_PKEY_encrypt() instead.
611 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
612 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
613 a new formulation to include all the things it can be used for,
614 as well as words of caution.
618 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
619 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
623 * All of the low level HMAC functions have been deprecated including:
625 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
626 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
629 Use of these low level functions has been informally discouraged for a long
630 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
631 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
632 and L<EVP_MAC_final(3)>.
636 * Over two thousand fixes were made to the documentation, including:
637 - Common options (such as -rand/-writerand, TLS version control, etc)
638 were refactored and point to newly-enhanced descriptions in openssl.pod.
639 - Added style conformance for all options (with help from Richard Levitte),
640 documented all reported missing options, added a CI build to check
641 that all options are documented and that no unimplemented options
643 - Documented some internals, such as all use of environment variables.
644 - Addressed all internal broken L<> references.
648 * All of the low level CMAC functions have been deprecated including:
650 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
651 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
653 Use of these low level functions has been informally discouraged for a long
654 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
655 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
656 and L<EVP_MAC_final(3)>.
660 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
661 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
664 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
665 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
666 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
667 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
668 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
669 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
670 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
671 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
672 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
673 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
675 Use of these low level functions has been informally discouraged
676 for a long time. Applications should use the EVP_DigestInit_ex(3),
677 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
681 * Corrected the documentation of the return values from the `EVP_DigestSign*`
682 set of functions. The documentation mentioned negative values for some
683 errors, but this was never the case, so the mention of negative values
686 Code that followed the documentation and thereby check with something
687 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
691 * All of the low level cipher functions have been deprecated including:
693 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
694 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
695 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
696 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
697 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
698 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
699 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
700 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
701 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
702 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
703 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
704 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
705 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
706 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
707 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
708 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
709 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
710 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
711 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
712 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
713 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
714 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
715 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
716 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
717 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
718 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
719 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
720 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
721 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
723 Use of these low level functions has been informally discouraged for
724 a long time. Applications should use the high level EVP APIs, e.g.
725 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
726 equivalently named decrypt functions instead.
728 *Matt Caswell and Paul Dale*
730 * Removed include/openssl/opensslconf.h.in and replaced it with
731 include/openssl/configuration.h.in, which differs in not including
732 <openssl/macros.h>. A short header include/openssl/opensslconf.h
733 was added to include both.
735 This allows internal hacks where one might need to modify the set
736 of configured macros, for example this if deprecated symbols are
737 still supposed to be available internally:
739 #include <openssl/configuration.h>
741 #undef OPENSSL_NO_DEPRECATED
742 #define OPENSSL_SUPPRESS_DEPRECATED
744 #include <openssl/macros.h>
746 This should not be used by applications that use the exported
747 symbols, as that will lead to linking errors.
751 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
752 used in exponentiation with 512-bit moduli. No EC algorithms are
753 affected. Analysis suggests that attacks against 2-prime RSA1024,
754 3-prime RSA1536, and DSA1024 as a result of this defect would be very
755 difficult to perform and are not believed likely. Attacks against DH512
756 are considered just feasible. However, for an attack the target would
757 have to re-use the DH512 private key, which is not recommended anyway.
758 Also applications directly using the low level API BN_mod_exp may be
759 affected if they use BN_FLG_CONSTTIME.
764 * Most memory-debug features have been deprecated, and the functionality
765 replaced with no-ops.
769 * Added documentation for the STACK API.
773 * Introduced a new method type and API, OSSL_ENCODER, to
774 represent generic encoders. An implementation is expected to
775 be able to encode an object associated with a given name (such
776 as an algorithm name for an asymmetric key) into forms given by
777 implementation properties.
779 Encoders are primarily used from inside libcrypto, through
780 calls to functions like EVP_PKEY_print_private(),
781 PEM_write_bio_PrivateKey() and similar.
783 Encoders are specified in such a way that they can be made to
784 directly handle the provider side portion of an object, if this
785 provider side part comes from the same provider as the encoder
786 itself, but can also be made to handle objects in parametrized
787 form (as an OSSL_PARAM array of data). This allows a provider to
788 offer generic encoders as a service for any other provider.
792 * Added a .pragma directive to the syntax of configuration files, to
793 allow varying behavior in a supported and predictable manner.
794 Currently added pragma:
798 This allows dollar signs to be a keyword character unless it's
799 followed by a opening brace or parenthesis. This is useful for
800 platforms where dollar signs are commonly used in names, such as
801 volume names and system directory names on VMS.
805 * Added functionality to create an EVP_PKEY from user data. This
806 is effectively the same as creating a RSA, DH or DSA object and
807 then assigning them to an EVP_PKEY, but directly using algorithm
808 agnostic EVP functions. A benefit is that this should be future
809 proof for public key algorithms to come.
813 * Change the interpretation of the '--api' configuration option to
814 mean that this is a desired API compatibility level with no
815 further meaning. The previous interpretation, that this would
816 also mean to remove all deprecated symbols up to and including
817 the given version, no requires that 'no-deprecated' is also used
818 in the configuration.
820 When building applications, the desired API compatibility level
821 can be set with the OPENSSL_API_COMPAT macro like before. For
822 API compatibility version below 3.0, the old style numerical
823 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
824 For version 3.0 and on, the value is expected to be the decimal
825 value calculated from the major and minor version like this:
827 MAJOR * 10000 + MINOR * 100
831 -DOPENSSL_API_COMPAT=30000 For 3.0
832 -DOPENSSL_API_COMPAT=30200 For 3.2
834 To hide declarations that are deprecated up to and including the
835 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
836 given when building the application as well.
840 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
841 access to certificate and CRL stores via URIs and OSSL_STORE
844 This adds the following functions:
846 - X509_LOOKUP_store()
847 - X509_STORE_load_file()
848 - X509_STORE_load_path()
849 - X509_STORE_load_store()
850 - SSL_add_store_cert_subjects_to_stack()
851 - SSL_CTX_set_default_verify_store()
852 - SSL_CTX_load_verify_file()
853 - SSL_CTX_load_verify_dir()
854 - SSL_CTX_load_verify_store()
858 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
859 The presence of this system service is determined at run-time.
863 * Added functionality to create an EVP_PKEY context based on data
864 for methods from providers. This takes an algorithm name and a
865 property query string and simply stores them, with the intent
866 that any operation that uses this context will use those strings
867 to fetch the needed methods implicitly, thereby making the port
868 of application written for pre-3.0 OpenSSL easier.
872 * The undocumented function NCONF_WIN32() has been deprecated; for
873 conversion details see the HISTORY section of doc/man5/config.pod
877 * Introduced the new functions EVP_DigestSignInit_ex() and
878 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
879 EVP_DigestVerifyUpdate() have been converted to functions. See the man
880 pages for further details.
884 * Over two thousand fixes were made to the documentation, including:
885 adding missing command flags, better style conformance, documentation
888 *Rich Salz, Richard Levitte*
890 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
891 X25519, X448, Ed25519 and Ed448.
895 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
900 * Deprecated the public definition of `ERR_STATE` as well as the function
901 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
906 * Added ERR functionality to give callers access to the stored function
907 names that have replaced the older function code based functions.
909 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
910 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
911 ERR_peek_error_all() and ERR_peek_last_error_all().
913 These functions have become deprecated: ERR_get_error_line(),
914 ERR_get_error_line_data(), ERR_peek_error_line_data(),
915 ERR_peek_last_error_line_data() and ERR_func_error_string().
917 Users are recommended to use ERR_get_error_all(), or to pick information
918 with ERR_peek functions and finish off with getting the error code by using
923 * Extended testing to be verbose for failing tests only. The make variables
924 VERBOSE_FAILURE or VF can be used to enable this:
926 $ make VF=1 test # Unix
927 $ mms /macro=(VF=1) test ! OpenVMS
928 $ nmake VF=1 test # Windows
932 * Added the `-copy_extensions` option to the `x509` command for use with
933 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
934 all extensions in the request are copied to the certificate or vice versa.
936 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
938 * Added the `-copy_extensions` option to the `req` command for use with
939 `-x509`. When given with the `copy` or `copyall` argument,
940 all extensions in the certification request are copied to the certificate.
944 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
945 they generate are by default RFC 5280 compliant in the following sense:
946 There is a subjectKeyIdentifier extension with a hash value of the public key
947 and for not self-signed certs there is an authorityKeyIdentifier extension
948 with a keyIdentifier field or issuer information identifying the signing key.
949 This is done unless some configuration overrides the new default behavior,
950 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
954 * Added several checks to `X509_verify_cert()` according to requirements in
955 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
956 (which may be done by using the CLI option `-x509_strict`):
957 * The basicConstraints of CA certificates must be marked critical.
958 * CA certificates must explicitly include the keyUsage extension.
959 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
960 * The issuer name of any certificate must not be empty.
961 * The subject name of CA certs, certs with keyUsage crlSign,
962 and certs without subjectAlternativeName must not be empty.
963 * If a subjectAlternativeName extension is given it must not be empty.
964 * The signatureAlgorithm field and the cert signature must be consistent.
965 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
966 must not be marked critical.
967 * The authorityKeyIdentifier must be given for X.509v3 certs
968 unless they are self-signed.
969 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
973 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
974 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
978 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
979 used even when parsing explicit parameters, when loading a encoded key
980 or calling `EC_GROUP_new_from_ecpkparameters()`/
981 `EC_GROUP_new_from_ecparameters()`.
982 This prevents bypass of security hardening and performance gains,
983 especially for curves with specialized EC_METHODs.
984 By default, if a key encoded with explicit parameters is loaded and later
985 encoded, the output is still encoded with explicit parameters, even if
986 internally a "named" EC_GROUP is used for computation.
990 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
991 this change, EC_GROUP_set_generator would accept order and/or cofactor as
992 NULL. After this change, only the cofactor parameter can be NULL. It also
993 does some minimal sanity checks on the passed order.
998 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
999 An attack is simple, if the first CMS_recipientInfo is valid but the
1000 second CMS_recipientInfo is chosen ciphertext. If the second
1001 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1002 encryption key will be replaced by garbage, and the message cannot be
1003 decoded, but if the RSA decryption fails, the correct encryption key is
1004 used and the recipient will not notice the attack.
1005 As a work around for this potential attack the length of the decrypted
1006 key must be equal to the cipher default key length, in case the
1007 certifiate is not given and all recipientInfo are tried out.
1008 The old behaviour can be re-enabled in the CMS code by setting the
1009 CMS_DEBUG_DECRYPT flag.
1013 * Early start up entropy quality from the DEVRANDOM seed source has been
1014 improved for older Linux systems. The RAND subsystem will wait for
1015 /dev/random to be producing output before seeding from /dev/urandom.
1016 The seeded state is stored for future library initialisations using
1017 a system global shared memory segment. The shared memory identifier
1018 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1019 the desired value. The default identifier is 114.
1023 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1024 when primes for RSA keys are computed.
1025 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1026 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1027 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1028 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1029 This avoids possible fingerprinting of newly generated RSA modules.
1033 * Correct the extended master secret constant on EBCDIC systems. Without this
1034 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1035 negotiate EMS will fail. Unfortunately this also means that TLS connections
1036 between EBCDIC systems with this fix, and EBCDIC systems without this
1037 fix will fail if they negotiate EMS.
1041 * Changed the library initialisation so that the config file is now loaded
1042 by default. This was already the case for libssl. It now occurs for both
1043 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1044 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1048 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1049 where the former acts as a replacement for `ERR_put_error()`, and the
1050 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1051 `ERR_raise_data()` adds more flexibility by taking a format string and
1052 an arbitrary number of arguments following it, to be processed with
1057 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1058 to check if a named provider is loaded and available. When called, it
1059 will also activate all fallback providers if such are still present.
1063 * Enforce a minimum DH modulus size of 512 bits.
1067 * Changed DH parameters to generate the order q subgroup instead of 2q.
1068 Previously generated DH parameters are still accepted by DH_check
1069 but DH_generate_key works around that by clearing bit 0 of the
1070 private key for those. This avoids leaking bit 0 of the private key.
1074 * Significantly reduce secure memory usage by the randomness pools.
1078 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1083 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1084 algorithms. An implementation of a key exchange algorithm can be obtained
1085 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1086 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1087 the older EVP_PKEY_derive_init() function. See the man pages for the new
1088 functions for further details.
1092 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1096 * Removed the function names from error messages and deprecated the
1099 * Removed NextStep support and the macro OPENSSL_UNISTD
1103 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1104 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1105 Also removed "export var as function" capability; we do not export
1106 variables, only functions.
1110 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1111 an error and 1 indicating success. In previous versions of OpenSSL this
1112 was a void type. If a key was set longer than the maximum possible this
1117 * Support SM2 signing and verification schemes with X509 certificate.
1121 * Use SHA256 as the default digest for TS query in the `ts` app.
1125 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1126 This checks that the salt length is at least 128 bits, the derived key
1127 length is at least 112 bits, and that the iteration count is at least 1000.
1128 For backwards compatibility these checks are disabled by default in the
1129 default provider, but are enabled by default in the fips provider.
1130 To enable or disable these checks use the control
1131 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1135 * Default cipher lists/suites are now available via a function, the
1136 #defines are deprecated.
1140 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1141 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1142 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1146 * Join the directories crypto/x509 and crypto/x509v3
1150 * Added command 'openssl kdf' that uses the EVP_KDF API.
1154 * Added command 'openssl mac' that uses the EVP_MAC API.
1158 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1159 as default directories. Also added the command 'openssl info'
1160 for scripting purposes.
1164 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1165 deprecated. These undocumented functions were never integrated into the EVP
1166 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1167 Bi-directional IGE mode. These modes were never formally standardised and
1168 usage of these functions is believed to be very small. In particular
1169 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1170 is ever used. The security implications are believed to be minimal, but
1171 this issue was never fixed for backwards compatibility reasons. New code
1172 should not use these modes.
1176 * Add prediction resistance to the DRBG reseeding process.
1180 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1181 mandated by IEEE Std 1619-2018.
1185 * Added newline escaping functionality to a filename when using openssl dgst.
1186 This output format is to replicate the output format found in the `*sum`
1187 checksum programs. This aims to preserve backward compatibility.
1189 *Matt Eaton, Richard Levitte, and Paul Dale*
1191 * Removed the heartbeat message in DTLS feature, as it has very
1192 little usage and doesn't seem to fulfill a valuable purpose.
1193 The configuration option is now deprecated.
1197 * Changed the output of 'openssl {digestname} < file' to display the
1198 digest name in its output.
1202 * Added a new generic trace API which provides support for enabling
1203 instrumentation through trace output. This feature is mainly intended
1204 as an aid for developers and is disabled by default. To utilize it,
1205 OpenSSL needs to be configured with the `enable-trace` option.
1207 If the tracing API is enabled, the application can activate trace output
1208 by registering BIOs as trace channels for a number of tracing and debugging
1211 The `openssl` program has been expanded to enable any of the types
1212 available via environment variables defined by the user, and serves as
1213 one possible example on how to use this functionality.
1215 *Richard Levitte & Matthias St. Pierre*
1217 * Added build tests for C++. These are generated files that only do one
1218 thing, to include one public OpenSSL head file each. This tests that
1219 the public header files can be usefully included in a C++ application.
1221 This test isn't enabled by default. It can be enabled with the option
1222 'enable-buildtest-c++'.
1226 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1230 * Add KMAC to EVP_MAC.
1234 * Added property based algorithm implementation selection framework to
1239 * Added SCA hardening for modular field inversion in EC_GROUP through
1240 a new dedicated field_inv() pointer in EC_METHOD.
1241 This also addresses a leakage affecting conversions from projective
1242 to affine coordinates.
1244 *Billy Bob Brumley, Nicola Tuveri*
1246 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1247 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1248 those algorithms that were already supported through the EVP_PKEY API
1249 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1250 and scrypt are now wrappers that call EVP_KDF.
1254 * Build devcrypto engine as a dynamic engine.
1256 *Eneas U de Queiroz*
1258 * Add keyed BLAKE2 to EVP_MAC.
1262 * Fix a bug in the computation of the endpoint-pair shared secret used
1263 by DTLS over SCTP. This breaks interoperability with older versions
1264 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1265 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1266 interoperability with such broken implementations. However, enabling
1267 this switch breaks interoperability with correct implementations.
1269 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1270 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1274 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1278 * Change the license to the Apache License v2.0.
1282 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1284 - Major releases (indicated by incrementing the MAJOR release number)
1285 may introduce incompatible API/ABI changes.
1286 - Minor releases (indicated by incrementing the MINOR release number)
1287 may introduce new features but retain API/ABI compatibility.
1288 - Patch releases (indicated by incrementing the PATCH number)
1289 are intended for bug fixes and other improvements of existing
1290 features only (like improving performance or adding documentation)
1291 and retain API/ABI compatibility.
1295 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1299 * Remove the 'dist' target and add a tarball building script. The
1300 'dist' target has fallen out of use, and it shouldn't be
1301 necessary to configure just to create a source distribution.
1305 * Recreate the OS390-Unix config target. It no longer relies on a
1306 special script like it did for OpenSSL pre-1.1.0.
1310 * Instead of having the source directories listed in Configure, add
1311 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1316 * Add GMAC to EVP_MAC.
1320 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1324 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1325 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1326 to facilitate the continued use of MACs through raw private keys in
1327 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1331 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1332 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1336 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1337 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1338 are retained for backwards compatibility.
1342 * AES-XTS mode now enforces that its two keys are different to mitigate
1343 the attacked described in "Efficient Instantiations of Tweakable
1344 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1345 Details of this attack can be obtained from:
1346 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1350 * Rename the object files, i.e. give them other names than in previous
1351 versions. Their names now include the name of the final product, as
1352 well as its type mnemonic (bin, lib, shlib).
1356 * Added new option for 'openssl list', '-objects', which will display the
1357 list of built in objects, i.e. OIDs with names.
1361 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1362 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1367 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1368 improves application performance by removing data copies and providing
1369 applications with zero-copy system calls such as sendfile and splice.
1373 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1374 option is set, openssl cleanses (zeroize) plaintext bytes from
1375 internal buffers after delivering them to the application. Note,
1376 the application is still responsible for cleansing other copies
1377 (e.g.: data received by SSL_read(3)).
1381 * `PKCS12_parse` now maintains the order of the parsed certificates
1382 when outputting them via `*ca` (rather than reversing it).
1386 * Deprecated pthread fork support methods. These were unused so no
1387 replacement is required.
1389 - OPENSSL_fork_prepare()
1390 - OPENSSL_fork_parent()
1391 - OPENSSL_fork_child()
1398 ### Changes between 1.1.1i and 1.1.1j [xx XXX xxxx]
1400 * Fixed SRP_Calc_client_key so that it uses constant time. The previous
1401 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1402 could be exploited in a side channel attack to recover the password. Since
1403 the attack is local host only this is outside of the current OpenSSL
1404 threat model and therefore no CVE is assigned.
1406 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1411 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1413 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1414 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1415 If an attacker can control both items being compared then this could lead
1416 to a possible denial of service attack. OpenSSL itself uses the
1417 GENERAL_NAME_cmp function for two purposes:
1418 1) Comparing CRL distribution point names between an available CRL and a
1419 CRL distribution point embedded in an X509 certificate
1420 2) When verifying that a timestamp response token signer matches the
1421 timestamp authority name (exposed via the API functions
1422 TS_RESP_verify_response and TS_RESP_verify_token)
1427 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1429 * Certificates with explicit curve parameters are now disallowed in
1430 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1434 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1435 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1436 conversely, silently ignore DTLS protocol version bounds when configuring
1437 TLS-based contexts. The commands can be repeated to set bounds of both
1438 types. The same applies with the corresponding "min_protocol" and
1439 "max_protocol" command-line switches, in case some application uses both TLS
1442 SSL_CTX instances that are created for a fixed protocol version (e.g.
1443 TLSv1_server_method()) also silently ignore version bounds. Previously
1444 attempts to apply bounds to these protocol versions would result in an
1445 error. Now only the "version-flexible" SSL_CTX instances are subject to
1446 limits in configuration files in command-line options.
1450 * Handshake now fails if Extended Master Secret extension is dropped
1455 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1457 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1459 * Fixed segmentation fault in SSL_check_chain()
1460 Server or client applications that call the SSL_check_chain() function
1461 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1462 dereference as a result of incorrect handling of the
1463 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1464 or unrecognised signature algorithm is received from the peer. This could
1465 be exploited by a malicious peer in a Denial of Service attack.
1470 * Added AES consttime code for no-asm configurations
1471 an optional constant time support for AES was added
1472 when building openssl for no-asm.
1473 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1474 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1475 At this time this feature is by default disabled.
1476 It will be enabled by default in 3.0.
1480 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1482 * Revert the change of EOF detection while reading in libssl to avoid
1483 regressions in applications depending on the current way of reporting
1484 the EOF. As the existing method is not fully accurate the change to
1485 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1486 branch and will be present in the 3.0 release.
1490 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1491 when primes for RSA keys are computed.
1492 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1493 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1494 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1495 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1496 This avoids possible fingerprinting of newly generated RSA modules.
1500 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1502 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1503 while reading in libssl then we would report an error back to the
1504 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1505 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1506 therefore give a hint as to what went wrong.
1510 * Check that ed25519 and ed448 are allowed by the security level. Previously
1511 signature algorithms not using an MD were not being checked that they were
1512 allowed by the security level.
1516 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1517 was not quite right. The behaviour was not consistent between resumption
1518 and normal handshakes, and also not quite consistent with historical
1519 behaviour. The behaviour in various scenarios has been clarified and
1520 it has been updated to make it match historical behaviour as closely as
1525 * *[VMS only]* The header files that the VMS compilers include automatically,
1526 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1527 that the C++ compiler doesn't understand. This is a shortcoming in the
1528 compiler, but can be worked around with `__cplusplus` guards.
1530 C++ applications that use OpenSSL libraries must be compiled using the
1531 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1532 functions. Otherwise, only functions with symbols of less than 31
1533 characters can be used, as the linker will not be able to successfully
1534 resolve symbols with longer names.
1538 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1539 The presence of this system service is determined at run-time.
1543 * Added newline escaping functionality to a filename when using openssl dgst.
1544 This output format is to replicate the output format found in the `*sum`
1545 checksum programs. This aims to preserve backward compatibility.
1547 *Matt Eaton, Richard Levitte, and Paul Dale*
1549 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1554 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1556 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1557 number generator (RNG). This was intended to include protection in the
1558 event of a fork() system call in order to ensure that the parent and child
1559 processes did not share the same RNG state. However this protection was not
1560 being used in the default case.
1562 A partial mitigation for this issue is that the output from a high
1563 precision timer is mixed into the RNG state so the likelihood of a parent
1564 and child process sharing state is significantly reduced.
1566 If an application already calls OPENSSL_init_crypto() explicitly using
1567 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1570 *Matthias St. Pierre*
1572 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1573 used even when parsing explicit parameters, when loading a encoded key
1574 or calling `EC_GROUP_new_from_ecpkparameters()`/
1575 `EC_GROUP_new_from_ecparameters()`.
1576 This prevents bypass of security hardening and performance gains,
1577 especially for curves with specialized EC_METHODs.
1578 By default, if a key encoded with explicit parameters is loaded and later
1579 encoded, the output is still encoded with explicit parameters, even if
1580 internally a "named" EC_GROUP is used for computation.
1584 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1585 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1586 NULL. After this change, only the cofactor parameter can be NULL. It also
1587 does some minimal sanity checks on the passed order.
1592 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1593 An attack is simple, if the first CMS_recipientInfo is valid but the
1594 second CMS_recipientInfo is chosen ciphertext. If the second
1595 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1596 encryption key will be replaced by garbage, and the message cannot be
1597 decoded, but if the RSA decryption fails, the correct encryption key is
1598 used and the recipient will not notice the attack.
1599 As a work around for this potential attack the length of the decrypted
1600 key must be equal to the cipher default key length, in case the
1601 certifiate is not given and all recipientInfo are tried out.
1602 The old behaviour can be re-enabled in the CMS code by setting the
1603 CMS_DEBUG_DECRYPT flag.
1608 * Early start up entropy quality from the DEVRANDOM seed source has been
1609 improved for older Linux systems. The RAND subsystem will wait for
1610 /dev/random to be producing output before seeding from /dev/urandom.
1611 The seeded state is stored for future library initialisations using
1612 a system global shared memory segment. The shared memory identifier
1613 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1614 the desired value. The default identifier is 114.
1618 * Correct the extended master secret constant on EBCDIC systems. Without this
1619 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1620 negotiate EMS will fail. Unfortunately this also means that TLS connections
1621 between EBCDIC systems with this fix, and EBCDIC systems without this
1622 fix will fail if they negotiate EMS.
1626 * Use Windows installation paths in the mingw builds
1628 Mingw isn't a POSIX environment per se, which means that Windows
1629 paths should be used for installation.
1634 * Changed DH_check to accept parameters with order q and 2q subgroups.
1635 With order 2q subgroups the bit 0 of the private key is not secret
1636 but DH_generate_key works around that by clearing bit 0 of the
1637 private key for those. This avoids leaking bit 0 of the private key.
1641 * Significantly reduce secure memory usage by the randomness pools.
1645 * Revert the DEVRANDOM_WAIT feature for Linux systems
1647 The DEVRANDOM_WAIT feature added a select() call to wait for the
1648 /dev/random device to become readable before reading from the
1649 /dev/urandom device.
1651 It turned out that this change had negative side effects on
1652 performance which were not acceptable. After some discussion it
1653 was decided to revert this feature and leave it up to the OS
1654 resp. the platform maintainer to ensure a proper initialization
1655 during early boot time.
1657 *Matthias St. Pierre*
1659 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1661 * Add build tests for C++. These are generated files that only do one
1662 thing, to include one public OpenSSL head file each. This tests that
1663 the public header files can be usefully included in a C++ application.
1665 This test isn't enabled by default. It can be enabled with the option
1666 'enable-buildtest-c++'.
1670 * Enable SHA3 pre-hashing for ECDSA and DSA.
1674 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1675 This changes the size when using the `genpkey` command when no size is given.
1676 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1677 generation commands to use 2048 bits by default.
1681 * Reorganize the manual pages to consistently have RETURN VALUES,
1682 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1683 util/fix-doc-nits accordingly.
1685 *Paul Yang, Joshua Lock*
1687 * Add the missing accessor EVP_PKEY_get0_engine()
1691 * Have commands like `s_client` and `s_server` output the signature scheme
1692 along with other cipher suite parameters when debugging.
1696 * Make OPENSSL_config() error agnostic again.
1700 * Do the error handling in RSA decryption constant time.
1704 * Prevent over long nonces in ChaCha20-Poly1305.
1706 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1707 for every encryption operation. RFC 7539 specifies that the nonce value
1708 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1709 and front pads the nonce with 0 bytes if it is less than 12
1710 bytes. However it also incorrectly allows a nonce to be set of up to 16
1711 bytes. In this case only the last 12 bytes are significant and any
1712 additional leading bytes are ignored.
1714 It is a requirement of using this cipher that nonce values are
1715 unique. Messages encrypted using a reused nonce value are susceptible to
1716 serious confidentiality and integrity attacks. If an application changes
1717 the default nonce length to be longer than 12 bytes and then makes a
1718 change to the leading bytes of the nonce expecting the new value to be a
1719 new unique nonce then such an application could inadvertently encrypt
1720 messages with a reused nonce.
1722 Additionally the ignored bytes in a long nonce are not covered by the
1723 integrity guarantee of this cipher. Any application that relies on the
1724 integrity of these ignored leading bytes of a long nonce may be further
1725 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1726 is safe because no such use sets such a long nonce value. However user
1727 applications that use this cipher directly and set a non-default nonce
1728 length to be longer than 12 bytes may be vulnerable.
1730 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1736 * Add DEVRANDOM_WAIT feature for Linux systems
1738 On older Linux systems where the getrandom() system call is not available,
1739 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1740 Contrary to getrandom(), the /dev/urandom device will not block during
1741 early boot when the kernel CSPRNG has not been seeded yet.
1743 To mitigate this known weakness, use select() to wait for /dev/random to
1744 become readable before reading from /dev/urandom.
1746 * Ensure that SM2 only uses SM3 as digest algorithm
1750 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1752 * Change the info callback signals for the start and end of a post-handshake
1753 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1754 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1755 confused by this and assume that a TLSv1.2 renegotiation has started. This
1756 can break KeyUpdate handling. Instead we no longer signal the start and end
1757 of a post handshake message exchange (although the messages themselves are
1758 still signalled). This could break some applications that were expecting
1759 the old signals. However without this KeyUpdate is not usable for many
1764 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1766 * Timing vulnerability in DSA signature generation
1768 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1769 timing side channel attack. An attacker could use variations in the signing
1770 algorithm to recover the private key.
1772 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1777 * Timing vulnerability in ECDSA signature generation
1779 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1780 timing side channel attack. An attacker could use variations in the signing
1781 algorithm to recover the private key.
1783 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1788 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1789 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1790 of two gigabytes and the error handling improved.
1792 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1793 categorized as a normal bug, not a security issue, because the DRBG reseeds
1794 automatically and is fully functional even without additional randomness
1795 provided by the application.
1797 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1799 * Add a new ClientHello callback. Provides a callback interface that gives
1800 the application the ability to adjust the nascent SSL object at the
1801 earliest stage of ClientHello processing, immediately after extensions have
1802 been collected but before they have been processed. In particular, this
1803 callback can adjust the supported TLS versions in response to the contents
1808 * Add SM2 base algorithm support.
1812 * s390x assembly pack: add (improved) hardware-support for the following
1813 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1814 aes-cfb/cfb8, aes-ecb.
1818 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1819 parameter is no longer accepted, as it leads to a corrupt table. NULL
1820 pem_str is reserved for alias entries only.
1824 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1825 step for prime curves. The new implementation is based on formulae from
1826 differential addition-and-doubling in homogeneous projective coordinates
1827 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1828 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1829 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1830 to work in projective coordinates.
1832 *Billy Bob Brumley, Nicola Tuveri*
1834 * Change generating and checking of primes so that the error rate of not
1835 being prime depends on the intended use based on the size of the input.
1836 For larger primes this will result in more rounds of Miller-Rabin.
1837 The maximal error rate for primes with more than 1080 bits is lowered
1840 *Kurt Roeckx, Annie Yousar*
1842 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1846 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1847 moving between systems, and to avoid confusion when a Windows build is
1848 done with mingw vs with MSVC. For POSIX installs, there's still a
1849 symlink or copy named 'tsget' to avoid that confusion as well.
1853 * Revert blinding in ECDSA sign and instead make problematic addition
1854 length-invariant. Switch even to fixed-length Montgomery multiplication.
1858 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1859 step for binary curves. The new implementation is based on formulae from
1860 differential addition-and-doubling in mixed Lopez-Dahab projective
1861 coordinates, modified to independently blind the operands.
1863 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1865 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1866 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1867 EC_METHODs to implement their own specialized "ladder step", to take
1868 advantage of more favorable coordinate systems or more efficient
1869 differential addition-and-doubling algorithms.
1871 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1873 * Modified the random device based seed sources to keep the relevant
1874 file descriptors open rather than reopening them on each access.
1875 This allows such sources to operate in a chroot() jail without
1876 the associated device nodes being available. This behaviour can be
1877 controlled using RAND_keep_random_devices_open().
1881 * Numerous side-channel attack mitigations have been applied. This may have
1882 performance impacts for some algorithms for the benefit of improved
1883 security. Specific changes are noted in this change log by their respective
1888 * AIX shared library support overhaul. Switch to AIX "natural" way of
1889 handling shared libraries, which means collecting shared objects of
1890 different versions and bitnesses in one common archive. This allows to
1891 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1892 doesn't affect the way 3rd party applications are linked, only how
1893 multi-version installation is managed.
1897 * Make ec_group_do_inverse_ord() more robust and available to other
1898 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1899 mitigations are applied to the fallback BN_mod_inverse().
1900 When using this function rather than BN_mod_inverse() directly, new
1901 EC cryptosystem implementations are then safer-by-default.
1905 * Add coordinate blinding for EC_POINT and implement projective
1906 coordinate blinding for generic prime curves as a countermeasure to
1907 chosen point SCA attacks.
1909 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1911 * Add blinding to ECDSA and DSA signatures to protect against side channel
1912 attacks discovered by Keegan Ryan (NCC Group).
1916 * Enforce checking in the `pkeyutl` command to ensure that the input
1917 length does not exceed the maximum supported digest length when performing
1918 a sign, verify or verifyrecover operation.
1922 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1923 I/O in combination with something like select() or poll() will hang. This
1924 can be turned off again using SSL_CTX_clear_mode().
1925 Many applications do not properly handle non-application data records, and
1926 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1927 around the problems in those applications, but can also break some.
1928 It's recommended to read the manpages about SSL_read(), SSL_write(),
1929 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1930 SSL_CTX_set_read_ahead() again.
1934 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1935 now allow empty (zero character) pass phrases.
1939 * Apply blinding to binary field modular inversion and remove patent
1940 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1944 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1945 binary and prime elliptic curves.
1949 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1950 constant time fixed point multiplication.
1954 * Revise elliptic curve scalar multiplication with timing attack
1955 defenses: ec_wNAF_mul redirects to a constant time implementation
1956 when computing fixed point and variable point multiplication (which
1957 in OpenSSL are mostly used with secret scalars in keygen, sign,
1958 ECDH derive operations).
1959 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1962 * Updated CONTRIBUTING
1966 * Updated DRBG / RAND to request nonce and additional low entropy
1967 randomness from the system.
1969 *Matthias St. Pierre*
1971 * Updated 'openssl rehash' to use OpenSSL consistent default.
1975 * Moved the load of the ssl_conf module to libcrypto, which helps
1976 loading engines that libssl uses before libssl is initialised.
1980 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1984 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1986 *Ingo Schwarze, Rich Salz*
1988 * Added output of accepting IP address and port for 'openssl s_server'
1992 * Added a new API for TLSv1.3 ciphersuites:
1993 SSL_CTX_set_ciphersuites()
1994 SSL_set_ciphersuites()
1998 * Memory allocation failures consistently add an error to the error
2003 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2004 in libcrypto when run as setuid/setgid.
2008 * Load any config file by default when libssl is used.
2012 * Added new public header file <openssl/rand_drbg.h> and documentation
2013 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2015 *Matthias St. Pierre*
2017 * QNX support removed (cannot find contributors to get their approval
2018 for the license change).
2022 * TLSv1.3 replay protection for early data has been implemented. See the
2023 SSL_read_early_data() man page for further details.
2027 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2028 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2029 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2030 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2031 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2032 configuration has been separated out. See the ciphers man page or the
2033 SSL_CTX_set_ciphersuites() man page for more information.
2037 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2038 in responder mode now supports the new "-multi" option, which
2039 spawns the specified number of child processes to handle OCSP
2040 requests. The "-timeout" option now also limits the OCSP
2041 responder's patience to wait to receive the full client request
2042 on a newly accepted connection. Child processes are respawned
2043 as needed, and the CA index file is automatically reloaded
2044 when changed. This makes it possible to run the "ocsp" responder
2045 as a long-running service, making the OpenSSL CA somewhat more
2046 feature-complete. In this mode, most diagnostic messages logged
2047 after entering the event loop are logged via syslog(3) rather than
2052 * Added support for X448 and Ed448. Heavily based on original work by
2057 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2058 objects loaded. This adds the functions OSSL_STORE_expect() and
2059 OSSL_STORE_find() as well as needed tools to construct searches and
2060 get the search data out of them.
2064 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2065 version of OpenSSL should review their configuration settings to ensure
2066 that they are still appropriate for TLSv1.3. For further information see:
2067 <https://wiki.openssl.org/index.php/TLS1.3>
2071 * Grand redesign of the OpenSSL random generator
2073 The default RAND method now utilizes an AES-CTR DRBG according to
2074 NIST standard SP 800-90Ar1. The new random generator is essentially
2075 a port of the default random generator from the OpenSSL FIPS 2.0
2076 object module. It is a hybrid deterministic random bit generator
2077 using an AES-CTR bit stream and which seeds and reseeds itself
2078 automatically using trusted system entropy sources.
2080 Some of its new features are:
2081 - Support for multiple DRBG instances with seed chaining.
2082 - The default RAND method makes use of a DRBG.
2083 - There is a public and private DRBG instance.
2084 - The DRBG instances are fork-safe.
2085 - Keep all global DRBG instances on the secure heap if it is enabled.
2086 - The public and private DRBG instance are per thread for lock free
2089 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2091 * Changed Configure so it only says what it does and doesn't dump
2092 so much data. Instead, ./configdata.pm should be used as a script
2093 to display all sorts of configuration data.
2097 * Added processing of "make variables" to Configure.
2101 * Added SHA512/224 and SHA512/256 algorithm support.
2105 * The last traces of Netware support, first removed in 1.1.0, have
2110 * Get rid of Makefile.shared, and in the process, make the processing
2111 of certain files (rc.obj, or the .def/.map/.opt files produced from
2112 the ordinal files) more visible and hopefully easier to trace and
2113 debug (or make silent).
2117 * Make it possible to have environment variable assignments as
2118 arguments to config / Configure.
2122 * Add multi-prime RSA (RFC 8017) support.
2126 * Add SM3 implemented according to GB/T 32905-2016
2127 *Jack Lloyd <jack.lloyd@ribose.com>,*
2128 *Ronald Tse <ronald.tse@ribose.com>,*
2129 *Erick Borsboom <erick.borsboom@ribose.com>*
2131 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2132 as documented in RFC6066.
2133 Based on a patch from Tomasz Moń
2135 *Filipe Raimundo da Silva*
2137 * Add SM4 implemented according to GB/T 32907-2016.
2138 *Jack Lloyd <jack.lloyd@ribose.com>,*
2139 *Ronald Tse <ronald.tse@ribose.com>,*
2140 *Erick Borsboom <erick.borsboom@ribose.com>*
2142 * Reimplement -newreq-nodes and ERR_error_string_n; the
2143 original author does not agree with the license change.
2147 * Add ARIA AEAD TLS support.
2151 * Some macro definitions to support VS6 have been removed. Visual
2152 Studio 6 has not worked since 1.1.0
2156 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2157 without clearing the errors.
2161 * Add "atfork" functions. If building on a system that without
2162 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2163 requirements. The RAND facility now uses/requires this.
2171 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2172 not possible to disable entirely. However, it's still possible to
2173 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2176 To disable, configure with 'no-ui-console'. 'no-ui' is still
2177 possible to use as an alias. Check at compile time with the
2178 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2179 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2183 * Add a STORE module, which implements a uniform and URI based reader of
2184 stores that can contain keys, certificates, CRLs and numerous other
2185 objects. The main API is loosely based on a few stdio functions,
2186 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2187 OSSL_STORE_error and OSSL_STORE_close.
2188 The implementation uses backends called "loaders" to implement arbitrary
2189 URI schemes. There is one built in "loader" for the 'file' scheme.
2193 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2194 then adjusted to work on FreeBSD 8.4 as well.
2195 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2196 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2200 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2201 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2202 error code calls like this:
2204 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2206 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2207 that can be encoded in C. For the foreseeable future, this will only
2210 *Richard Levitte and Tim Hudson*
2212 * Removed BSD cryptodev engine.
2216 * Add a build target 'build_all_generated', to build all generated files
2217 and only that. This can be used to prepare everything that requires
2218 things like perl for a system that lacks perl and then move everything
2219 to that system and do the rest of the build there.
2223 * In the UI interface, make it possible to duplicate the user data. This
2224 can be used by engines that need to retain the data for a longer time
2225 than just the call where this user data is passed.
2229 * Ignore the '-named_curve auto' value for compatibility of applications
2232 *Tomas Mraz <tmraz@fedoraproject.org>*
2234 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2235 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2236 alerts across multiple records (some of which could be empty). In practice
2237 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2238 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2239 support this at all. Supporting it adds significant complexity to the
2240 record layer, and its removal is unlikely to cause interoperability
2245 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2246 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2247 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2252 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2253 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2255 *Richard Levitte, Andy Polyakov*
2257 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2262 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2263 platform rather than 'mingw'.
2267 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2268 success if they are asked to add an object which already exists
2269 in the store. This change cascades to other functions which load
2270 certificates and CRLs.
2274 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2275 facilitate stack unwinding even from assembly subroutines.
2279 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2280 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2284 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2285 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2286 which is the minimum version we support.
2290 * Certificate time validation (X509_cmp_time) enforces stricter
2291 compliance with RFC 5280. Fractional seconds and timezone offsets
2292 are no longer allowed.
2296 * Add support for ARIA
2300 * s_client will now send the Server Name Indication (SNI) extension by
2301 default unless the new "-noservername" option is used. The server name is
2302 based on the host provided to the "-connect" option unless overridden by
2303 using "-servername".
2307 * Add support for SipHash
2311 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2312 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2313 prevent issues where no progress is being made and the peer continually
2314 sends unrecognised record types, using up resources processing them.
2318 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2319 using the algorithm defined in
2320 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2324 * Heartbeat support has been removed; the ABI is changed for now.
2326 *Richard Levitte, Rich Salz*
2328 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2332 * The RSA "null" method, which was partially supported to avoid patent
2333 issues, has been replaced to always returns NULL.
2340 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2342 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2343 used even when parsing explicit parameters, when loading a encoded key
2344 or calling `EC_GROUP_new_from_ecpkparameters()`/
2345 `EC_GROUP_new_from_ecparameters()`.
2346 This prevents bypass of security hardening and performance gains,
2347 especially for curves with specialized EC_METHODs.
2348 By default, if a key encoded with explicit parameters is loaded and later
2349 encoded, the output is still encoded with explicit parameters, even if
2350 internally a "named" EC_GROUP is used for computation.
2354 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2355 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2356 NULL. After this change, only the cofactor parameter can be NULL. It also
2357 does some minimal sanity checks on the passed order.
2362 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2363 An attack is simple, if the first CMS_recipientInfo is valid but the
2364 second CMS_recipientInfo is chosen ciphertext. If the second
2365 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2366 encryption key will be replaced by garbage, and the message cannot be
2367 decoded, but if the RSA decryption fails, the correct encryption key is
2368 used and the recipient will not notice the attack.
2369 As a work around for this potential attack the length of the decrypted
2370 key must be equal to the cipher default key length, in case the
2371 certifiate is not given and all recipientInfo are tried out.
2372 The old behaviour can be re-enabled in the CMS code by setting the
2373 CMS_DEBUG_DECRYPT flag.
2378 * Use Windows installation paths in the mingw builds
2380 Mingw isn't a POSIX environment per se, which means that Windows
2381 paths should be used for installation.
2386 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2388 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2389 This changes the size when using the `genpkey` command when no size is given.
2390 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2391 generation commands to use 2048 bits by default.
2395 * Prevent over long nonces in ChaCha20-Poly1305.
2397 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2398 for every encryption operation. RFC 7539 specifies that the nonce value
2399 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2400 and front pads the nonce with 0 bytes if it is less than 12
2401 bytes. However it also incorrectly allows a nonce to be set of up to 16
2402 bytes. In this case only the last 12 bytes are significant and any
2403 additional leading bytes are ignored.
2405 It is a requirement of using this cipher that nonce values are
2406 unique. Messages encrypted using a reused nonce value are susceptible to
2407 serious confidentiality and integrity attacks. If an application changes
2408 the default nonce length to be longer than 12 bytes and then makes a
2409 change to the leading bytes of the nonce expecting the new value to be a
2410 new unique nonce then such an application could inadvertently encrypt
2411 messages with a reused nonce.
2413 Additionally the ignored bytes in a long nonce are not covered by the
2414 integrity guarantee of this cipher. Any application that relies on the
2415 integrity of these ignored leading bytes of a long nonce may be further
2416 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2417 is safe because no such use sets such a long nonce value. However user
2418 applications that use this cipher directly and set a non-default nonce
2419 length to be longer than 12 bytes may be vulnerable.
2421 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2427 * Added SCA hardening for modular field inversion in EC_GROUP through
2428 a new dedicated field_inv() pointer in EC_METHOD.
2429 This also addresses a leakage affecting conversions from projective
2430 to affine coordinates.
2432 *Billy Bob Brumley, Nicola Tuveri*
2434 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2435 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2439 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2443 * Remove the 'dist' target and add a tarball building script. The
2444 'dist' target has fallen out of use, and it shouldn't be
2445 necessary to configure just to create a source distribution.
2449 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2451 * Timing vulnerability in DSA signature generation
2453 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2454 timing side channel attack. An attacker could use variations in the signing
2455 algorithm to recover the private key.
2457 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2462 * Timing vulnerability in ECDSA signature generation
2464 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2465 timing side channel attack. An attacker could use variations in the signing
2466 algorithm to recover the private key.
2468 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2473 * Add coordinate blinding for EC_POINT and implement projective
2474 coordinate blinding for generic prime curves as a countermeasure to
2475 chosen point SCA attacks.
2477 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2479 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2481 * Client DoS due to large DH parameter
2483 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2484 malicious server can send a very large prime value to the client. This will
2485 cause the client to spend an unreasonably long period of time generating a
2486 key for this prime resulting in a hang until the client has finished. This
2487 could be exploited in a Denial Of Service attack.
2489 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2494 * Cache timing vulnerability in RSA Key Generation
2496 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2497 a cache timing side channel attack. An attacker with sufficient access to
2498 mount cache timing attacks during the RSA key generation process could
2499 recover the private key.
2501 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2502 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2507 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2508 parameter is no longer accepted, as it leads to a corrupt table. NULL
2509 pem_str is reserved for alias entries only.
2513 * Revert blinding in ECDSA sign and instead make problematic addition
2514 length-invariant. Switch even to fixed-length Montgomery multiplication.
2518 * Change generating and checking of primes so that the error rate of not
2519 being prime depends on the intended use based on the size of the input.
2520 For larger primes this will result in more rounds of Miller-Rabin.
2521 The maximal error rate for primes with more than 1080 bits is lowered
2524 *Kurt Roeckx, Annie Yousar*
2526 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2530 * Add blinding to ECDSA and DSA signatures to protect against side channel
2531 attacks discovered by Keegan Ryan (NCC Group).
2535 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2536 now allow empty (zero character) pass phrases.
2540 * Certificate time validation (X509_cmp_time) enforces stricter
2541 compliance with RFC 5280. Fractional seconds and timezone offsets
2542 are no longer allowed.
2546 * Fixed a text canonicalisation bug in CMS
2548 Where a CMS detached signature is used with text content the text goes
2549 through a canonicalisation process first prior to signing or verifying a
2550 signature. This process strips trailing space at the end of lines, converts
2551 line terminators to CRLF and removes additional trailing line terminators
2552 at the end of a file. A bug in the canonicalisation process meant that
2553 some characters, such as form-feed, were incorrectly treated as whitespace
2554 and removed. This is contrary to the specification (RFC5485). This fix
2555 could mean that detached text data signed with an earlier version of
2556 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2557 signed with a fixed OpenSSL may fail to verify with an earlier version of
2558 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2559 and use the "-binary" flag (for the "cms" command line application) or set
2560 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2564 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2566 * Constructed ASN.1 types with a recursive definition could exceed the stack
2568 Constructed ASN.1 types with a recursive definition (such as can be found
2569 in PKCS7) could eventually exceed the stack given malicious input with
2570 excessive recursion. This could result in a Denial Of Service attack. There
2571 are no such structures used within SSL/TLS that come from untrusted sources
2572 so this is considered safe.
2574 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2580 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2582 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2583 effectively reduced to only comparing the least significant bit of each
2584 byte. This allows an attacker to forge messages that would be considered as
2585 authenticated in an amount of tries lower than that guaranteed by the
2586 security claims of the scheme. The module can only be compiled by the
2587 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2589 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2595 * Add a build target 'build_all_generated', to build all generated files
2596 and only that. This can be used to prepare everything that requires
2597 things like perl for a system that lacks perl and then move everything
2598 to that system and do the rest of the build there.
2602 * Backport SSL_OP_NO_RENGOTIATION
2604 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2605 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2606 changes this is no longer possible in 1.1.0. Therefore the new
2607 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2608 1.1.0 to provide equivalent functionality.
2610 Note that if an application built against 1.1.0h headers (or above) is run
2611 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2612 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2616 * Removed the OS390-Unix config target. It relied on a script that doesn't
2621 * rsaz_1024_mul_avx2 overflow bug on x86_64
2623 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2624 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2625 Analysis suggests that attacks against RSA and DSA as a result of this
2626 defect would be very difficult to perform and are not believed likely.
2627 Attacks against DH1024 are considered just feasible, because most of the
2628 work necessary to deduce information about a private key may be performed
2629 offline. The amount of resources required for such an attack would be
2630 significant. However, for an attack on TLS to be meaningful, the server
2631 would have to share the DH1024 private key among multiple clients, which is
2632 no longer an option since CVE-2016-0701.
2634 This only affects processors that support the AVX2 but not ADX extensions
2635 like Intel Haswell (4th generation).
2637 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2638 was originally found via the OSS-Fuzz project.
2643 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2645 * bn_sqrx8x_internal carry bug on x86_64
2647 There is a carry propagating bug in the x86_64 Montgomery squaring
2648 procedure. No EC algorithms are affected. Analysis suggests that attacks
2649 against RSA and DSA as a result of this defect would be very difficult to
2650 perform and are not believed likely. Attacks against DH are considered just
2651 feasible (although very difficult) because most of the work necessary to
2652 deduce information about a private key may be performed offline. The amount
2653 of resources required for such an attack would be very significant and
2654 likely only accessible to a limited number of attackers. An attacker would
2655 additionally need online access to an unpatched system using the target
2656 private key in a scenario with persistent DH parameters and a private
2657 key that is shared between multiple clients.
2659 This only affects processors that support the BMI1, BMI2 and ADX extensions
2660 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2662 This issue was reported to OpenSSL by the OSS-Fuzz project.
2667 * Malformed X.509 IPAddressFamily could cause OOB read
2669 If an X.509 certificate has a malformed IPAddressFamily extension,
2670 OpenSSL could do a one-byte buffer overread. The most likely result
2671 would be an erroneous display of the certificate in text format.
2673 This issue was reported to OpenSSL by the OSS-Fuzz project.
2678 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2680 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2681 platform rather than 'mingw'.
2685 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2686 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2687 which is the minimum version we support.
2691 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2693 * Encrypt-Then-Mac renegotiation crash
2695 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2696 negotiated where it was not in the original handshake (or vice-versa) then
2697 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2698 and servers are affected.
2700 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2705 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2707 * Truncated packet could crash via OOB read
2709 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2710 cipher is being used, then a truncated packet can cause that host to
2711 perform an out-of-bounds read, usually resulting in a crash.
2713 This issue was reported to OpenSSL by Robert Święcki of Google.
2718 * Bad (EC)DHE parameters cause a client crash
2720 If a malicious server supplies bad parameters for a DHE or ECDHE key
2721 exchange then this can result in the client attempting to dereference a
2722 NULL pointer leading to a client crash. This could be exploited in a Denial
2725 This issue was reported to OpenSSL by Guido Vranken.
2730 * BN_mod_exp may produce incorrect results on x86_64
2732 There is a carry propagating bug in the x86_64 Montgomery squaring
2733 procedure. No EC algorithms are affected. Analysis suggests that attacks
2734 against RSA and DSA as a result of this defect would be very difficult to
2735 perform and are not believed likely. Attacks against DH are considered just
2736 feasible (although very difficult) because most of the work necessary to
2737 deduce information about a private key may be performed offline. The amount
2738 of resources required for such an attack would be very significant and
2739 likely only accessible to a limited number of attackers. An attacker would
2740 additionally need online access to an unpatched system using the target
2741 private key in a scenario with persistent DH parameters and a private
2742 key that is shared between multiple clients. For example this can occur by
2743 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2744 similar to CVE-2015-3193 but must be treated as a separate problem.
2746 This issue was reported to OpenSSL by the OSS-Fuzz project.
2751 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2753 * ChaCha20/Poly1305 heap-buffer-overflow
2755 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2756 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2757 crash. This issue is not considered to be exploitable beyond a DoS.
2759 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2764 * CMS Null dereference
2766 Applications parsing invalid CMS structures can crash with a NULL pointer
2767 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2768 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2769 structure callback if an attempt is made to free certain invalid encodings.
2770 Only CHOICE structures using a callback which do not handle NULL value are
2773 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2778 * Montgomery multiplication may produce incorrect results
2780 There is a carry propagating bug in the Broadwell-specific Montgomery
2781 multiplication procedure that handles input lengths divisible by, but
2782 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2783 and DH private keys are impossible. This is because the subroutine in
2784 question is not used in operations with the private key itself and an input
2785 of the attacker's direct choice. Otherwise the bug can manifest itself as
2786 transient authentication and key negotiation failures or reproducible
2787 erroneous outcome of public-key operations with specially crafted input.
2788 Among EC algorithms only Brainpool P-512 curves are affected and one
2789 presumably can attack ECDH key negotiation. Impact was not analyzed in
2790 detail, because pre-requisites for attack are considered unlikely. Namely
2791 multiple clients have to choose the curve in question and the server has to
2792 share the private key among them, neither of which is default behaviour.
2793 Even then only clients that chose the curve will be affected.
2795 This issue was publicly reported as transient failures and was not
2796 initially recognized as a security issue. Thanks to Richard Morgan for
2797 providing reproducible case.
2802 * Removed automatic addition of RPATH in shared libraries and executables,
2803 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2807 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2809 * Fix Use After Free for large message sizes
2811 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2812 message larger than approx 16k is received then the underlying buffer to
2813 store the incoming message is reallocated and moved. Unfortunately a
2814 dangling pointer to the old location is left which results in an attempt to
2815 write to the previously freed location. This is likely to result in a
2816 crash, however it could potentially lead to execution of arbitrary code.
2818 This issue only affects OpenSSL 1.1.0a.
2820 This issue was reported to OpenSSL by Robert Święcki.
2825 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2827 * OCSP Status Request extension unbounded memory growth
2829 A malicious client can send an excessively large OCSP Status Request
2830 extension. If that client continually requests renegotiation, sending a
2831 large OCSP Status Request extension each time, then there will be unbounded
2832 memory growth on the server. This will eventually lead to a Denial Of
2833 Service attack through memory exhaustion. Servers with a default
2834 configuration are vulnerable even if they do not support OCSP. Builds using
2835 the "no-ocsp" build time option are not affected.
2837 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2842 * SSL_peek() hang on empty record
2844 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2845 sends an empty record. This could be exploited by a malicious peer in a
2846 Denial Of Service attack.
2848 This issue was reported to OpenSSL by Alex Gaynor.
2853 * Excessive allocation of memory in tls_get_message_header() and
2854 dtls1_preprocess_fragment()
2856 A (D)TLS message includes 3 bytes for its length in the header for the
2857 message. This would allow for messages up to 16Mb in length. Messages of
2858 this length are excessive and OpenSSL includes a check to ensure that a
2859 peer is sending reasonably sized messages in order to avoid too much memory
2860 being consumed to service a connection. A flaw in the logic of version
2861 1.1.0 means that memory for the message is allocated too early, prior to
2862 the excessive message length check. Due to way memory is allocated in
2863 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2864 to service a connection. This could lead to a Denial of Service through
2865 memory exhaustion. However, the excessive message length check still takes
2866 place, and this would cause the connection to immediately fail. Assuming
2867 that the application calls SSL_free() on the failed connection in a timely
2868 manner then the 21Mb of allocated memory will then be immediately freed
2869 again. Therefore the excessive memory allocation will be transitory in
2870 nature. This then means that there is only a security impact if:
2872 1) The application does not call SSL_free() in a timely manner in the event
2873 that the connection fails
2875 2) The application is working in a constrained environment where there is
2876 very little free memory
2878 3) The attacker initiates multiple connection attempts such that there are
2879 multiple connections in a state where memory has been allocated for the
2880 connection; SSL_free() has not yet been called; and there is insufficient
2881 memory to service the multiple requests.
2883 Except in the instance of (1) above any Denial Of Service is likely to be
2884 transitory because as soon as the connection fails the memory is
2885 subsequently freed again in the SSL_free() call. However there is an
2886 increased risk during this period of application crashes due to the lack of
2887 memory - which would then mean a more serious Denial of Service.
2889 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2890 (CVE-2016-6307 and CVE-2016-6308)
2894 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2895 had to be removed. Primary reason is that vendor assembler can't
2896 assemble our modules with -KPIC flag. As result it, assembly
2897 support, was not even available as option. But its lack means
2898 lack of side-channel resistant code, which is incompatible with
2899 security by todays standards. Fortunately gcc is readily available
2900 prepackaged option, which we firmly point at...
2904 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2906 * Windows command-line tool supports UTF-8 opt-in option for arguments
2907 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2908 (to any value) allows Windows user to access PKCS#12 file generated
2909 with Windows CryptoAPI and protected with non-ASCII password, as well
2910 as files generated under UTF-8 locale on Linux also protected with
2915 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2916 have been disabled by default and removed from DEFAULT, just like RC4.
2917 See the RC4 item below to re-enable both.
2921 * The method for finding the storage location for the Windows RAND seed file
2922 has changed. First we check %RANDFILE%. If that is not set then we check
2923 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2924 all else fails we fall back to C:\.
2928 * The EVP_EncryptUpdate() function has had its return type changed from void
2929 to int. A return of 0 indicates and error while a return of 1 indicates
2934 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2935 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2936 off the constant time implementation for RSA, DSA and DH have been made
2937 no-ops and deprecated.
2941 * Windows RAND implementation was simplified to only get entropy by
2942 calling CryptGenRandom(). Various other RAND-related tickets
2945 *Joseph Wylie Yandle, Rich Salz*
2947 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2948 and `OPENSSL_LH_`, respectively. The old names are available
2949 with API compatibility. They new names are now completely documented.
2953 * Unify TYPE_up_ref(obj) methods signature.
2954 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2955 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2956 int (instead of void) like all others TYPE_up_ref() methods.
2957 So now these methods also check the return value of CRYPTO_atomic_add(),
2958 and the validity of object reference counter.
2960 *fdasilvayy@gmail.com*
2962 * With Windows Visual Studio builds, the .pdb files are installed
2963 alongside the installed libraries and executables. For a static
2964 library installation, ossl_static.pdb is the associate compiler
2965 generated .pdb file to be used when linking programs.
2969 * Remove openssl.spec. Packaging files belong with the packagers.
2973 * Automatic Darwin/OSX configuration has had a refresh, it will now
2974 recognise x86_64 architectures automatically. You can still decide
2975 to build for a different bitness with the environment variable
2976 KERNEL_BITS (can be 32 or 64), for example:
2978 KERNEL_BITS=32 ./config
2982 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2983 256 bit AES and HMAC with SHA256.
2987 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2991 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2995 * To enable users to have their own config files and build file templates,
2996 Configure looks in the directory indicated by the environment variable
2997 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2998 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2999 name and is used as is.
3003 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3004 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3005 X509_CERT_FILE_CTX was removed.
3009 * "shared" builds are now the default. To create only static libraries use
3010 the "no-shared" Configure option.
3014 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3015 All of these option have not worked for some while and are fundamental
3020 * Make various cleanup routines no-ops and mark them as deprecated. Most
3021 global cleanup functions are no longer required because they are handled
3022 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3023 Explicitly de-initing can cause problems (e.g. where a library that uses
3024 OpenSSL de-inits, but an application is still using it). The affected
3025 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3026 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3027 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3028 COMP_zlib_cleanup().
3032 * --strict-warnings no longer enables runtime debugging options
3033 such as REF_DEBUG. Instead, debug options are automatically
3034 enabled with '--debug' builds.
3036 *Andy Polyakov, Emilia Käsper*
3038 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3039 have been moved out of the public header files. New functions for managing
3040 these have been added.
3044 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3045 objects have been moved out of the public header files. New
3046 functions for managing these have been added.
3050 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3051 have been moved out of the public header files. New functions for managing
3052 these have been added.
3056 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3057 moved out of the public header files. New functions for managing these
3062 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3066 * Removed the mk1mf build scripts.
3070 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3071 it is always safe to #include a header now.
3075 * Removed the aged BC-32 config and all its supporting scripts
3079 * Removed support for Ultrix, Netware, and OS/2.
3083 * Add support for HKDF.
3085 *Alessandro Ghedini*
3087 * Add support for blake2b and blake2s
3091 * Added support for "pipelining". Ciphers that have the
3092 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3093 encryptions/decryptions simultaneously. There are currently no built-in
3094 ciphers with this property but the expectation is that engines will be able
3095 to offer it to significantly improve throughput. Support has been extended
3096 into libssl so that multiple records for a single connection can be
3097 processed in one go (for >=TLS 1.1).
3101 * Added the AFALG engine. This is an async capable engine which is able to
3102 offload work to the Linux kernel. In this initial version it only supports
3103 AES128-CBC. The kernel must be version 4.1.0 or greater.
3107 * OpenSSL now uses a new threading API. It is no longer necessary to
3108 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3109 are two supported threading models: pthreads and windows threads. It is
3110 also possible to configure OpenSSL at compile time for "no-threads". The
3111 old threading API should no longer be used. The functions have been
3112 replaced with "no-op" compatibility macros.
3114 *Alessandro Ghedini, Matt Caswell*
3116 * Modify behavior of ALPN to invoke callback after SNI/servername
3117 callback, such that updates to the SSL_CTX affect ALPN.
3121 * Add SSL_CIPHER queries for authentication and key-exchange.
3125 * Changes to the DEFAULT cipherlist:
3126 - Prefer (EC)DHE handshakes over plain RSA.
3127 - Prefer AEAD ciphers over legacy ciphers.
3128 - Prefer ECDSA over RSA when both certificates are available.
3129 - Prefer TLSv1.2 ciphers/PRF.
3130 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3135 * Change the ECC default curve list to be this, in order: x25519,
3136 secp256r1, secp521r1, secp384r1.
3140 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3141 disabled by default. They can be re-enabled using the
3142 enable-weak-ssl-ciphers option to Configure.
3146 * If the server has ALPN configured, but supports no protocols that the
3147 client advertises, send a fatal "no_application_protocol" alert.
3148 This behaviour is SHALL in RFC 7301, though it isn't universally
3149 implemented by other servers.
3153 * Add X25519 support.
3154 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3155 for public and private key encoding using the format documented in
3156 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3157 key generation and key derivation.
3159 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3164 * Deprecate SRP_VBASE_get_by_user.
3165 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3166 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3167 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3168 seed, even if the seed is configured.
3170 Users should use SRP_VBASE_get1_by_user instead. Note that in
3171 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3172 also that even though configuring the SRP seed attempts to hide
3173 invalid usernames by continuing the handshake with fake
3174 credentials, this behaviour is not constant time and no strong
3175 guarantees are made that the handshake is indistinguishable from
3176 that of a valid user.
3180 * Configuration change; it's now possible to build dynamic engines
3181 without having to build shared libraries and vice versa. This
3182 only applies to the engines in `engines/`, those in `crypto/engine/`
3183 will always be built into libcrypto (i.e. "static").
3185 Building dynamic engines is enabled by default; to disable, use
3186 the configuration option "disable-dynamic-engine".
3188 The only requirements for building dynamic engines are the
3189 presence of the DSO module and building with position independent
3190 code, so they will also automatically be disabled if configuring
3191 with "disable-dso" or "disable-pic".
3193 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3194 are also taken away from openssl/opensslconf.h, as they are
3199 * Configuration change; if there is a known flag to compile
3200 position independent code, it will always be applied on the
3201 libcrypto and libssl object files, and never on the application
3202 object files. This means other libraries that use routines from
3203 libcrypto / libssl can be made into shared libraries regardless
3204 of how OpenSSL was configured.
3206 If this isn't desirable, the configuration options "disable-pic"
3207 or "no-pic" can be used to disable the use of PIC. This will
3208 also disable building shared libraries and dynamic engines.
3212 * Removed JPAKE code. It was experimental and has no wide use.
3216 * The INSTALL_PREFIX Makefile variable has been renamed to
3217 DESTDIR. That makes for less confusion on what this variable
3218 is for. Also, the configuration option --install_prefix is
3223 * Heartbeat for TLS has been removed and is disabled by default
3224 for DTLS; configure with enable-heartbeats. Code that uses the
3225 old #define's might need to be updated.
3227 *Emilia Käsper, Rich Salz*
3229 * Rename REF_CHECK to REF_DEBUG.
3233 * New "unified" build system
3235 The "unified" build system is aimed to be a common system for all
3236 platforms we support. With it comes new support for VMS.
3238 This system builds supports building in a different directory tree
3239 than the source tree. It produces one Makefile (for unix family
3240 or lookalikes), or one descrip.mms (for VMS).
3242 The source of information to make the Makefile / descrip.mms is
3243 small files called 'build.info', holding the necessary
3244 information for each directory with source to compile, and a
3245 template in Configurations, like unix-Makefile.tmpl or
3248 With this change, the library names were also renamed on Windows
3249 and on VMS. They now have names that are closer to the standard
3250 on Unix, and include the major version number, and in certain
3251 cases, the architecture they are built for. See "Notes on shared
3252 libraries" in INSTALL.
3254 We rely heavily on the perl module Text::Template.
3258 * Added support for auto-initialisation and de-initialisation of the library.
3259 OpenSSL no longer requires explicit init or deinit routines to be called,
3260 except in certain circumstances. See the OPENSSL_init_crypto() and
3261 OPENSSL_init_ssl() man pages for further information.
3265 * The arguments to the DTLSv1_listen function have changed. Specifically the
3266 "peer" argument is now expected to be a BIO_ADDR object.
3268 * Rewrite of BIO networking library. The BIO library lacked consistent
3269 support of IPv6, and adding it required some more extensive
3270 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3271 which hold all types of addresses and chains of address information.
3272 It also introduces a new API, with functions like BIO_socket,
3273 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3274 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3275 have been adapted accordingly.
3279 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3284 * CRIME protection: disable compression by default, even if OpenSSL is
3285 compiled with zlib enabled. Applications can still enable compression
3286 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3287 using the SSL_CONF library to configure compression.
3291 * The signature of the session callback configured with
3292 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3293 was explicitly marked as `const unsigned char*` instead of
3298 * Always DPURIFY. Remove the use of uninitialized memory in the
3299 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3303 * Removed many obsolete configuration items, including
3304 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3305 MD2_CHAR, MD2_INT, MD2_LONG
3307 IDEA_SHORT, IDEA_LONG
3308 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3310 *Rich Salz, with advice from Andy Polyakov*
3312 * Many BN internals have been moved to an internal header file.
3314 *Rich Salz with help from Andy Polyakov*
3316 * Configuration and writing out the results from it has changed.
3317 Files such as Makefile include/openssl/opensslconf.h and are now
3318 produced through general templates, such as Makefile.in and
3319 crypto/opensslconf.h.in and some help from the perl module
3322 Also, the center of configuration information is no longer
3323 Makefile. Instead, Configure produces a perl module in
3324 configdata.pm which holds most of the config data (in the hash
3325 table %config), the target data that comes from the target
3326 configuration in one of the `Configurations/*.conf` files (in
3331 * To clarify their intended purposes, the Configure options
3332 --prefix and --openssldir change their semantics, and become more
3333 straightforward and less interdependent.
3335 --prefix shall be used exclusively to give the location INSTALLTOP
3336 where programs, scripts, libraries, include files and manuals are
3337 going to be installed. The default is now /usr/local.
3339 --openssldir shall be used exclusively to give the default
3340 location OPENSSLDIR where certificates, private keys, CRLs are
3341 managed. This is also where the default openssl.cnf gets
3343 If the directory given with this option is a relative path, the
3344 values of both the --prefix value and the --openssldir value will
3345 be combined to become OPENSSLDIR.
3346 The default for --openssldir is INSTALLTOP/ssl.
3348 Anyone who uses --openssldir to specify where OpenSSL is to be
3349 installed MUST change to use --prefix instead.
3353 * The GOST engine was out of date and therefore it has been removed. An up
3354 to date GOST engine is now being maintained in an external repository.
3355 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3356 support for GOST ciphersuites (these are only activated if a GOST engine
3361 * EGD is no longer supported by default; use enable-egd when
3364 *Ben Kaduk and Rich Salz*
3366 * The distribution now has Makefile.in files, which are used to
3367 create Makefile's when Configure is run. *Configure must be run
3368 before trying to build now.*
3372 * The return value for SSL_CIPHER_description() for error conditions
3377 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3379 Obtaining and performing DNSSEC validation of TLSA records is
3380 the application's responsibility. The application provides
3381 the TLSA records of its choice to OpenSSL, and these are then
3382 used to authenticate the peer.
3384 The TLSA records need not even come from DNS. They can, for
3385 example, be used to implement local end-entity certificate or
3386 trust-anchor "pinning", where the "pin" data takes the form
3387 of TLSA records, which can augment or replace verification
3388 based on the usual WebPKI public certification authorities.
3392 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3393 continues to support deprecated interfaces in default builds.
3394 However, applications are strongly advised to compile their
3395 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3396 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3397 or the 1.1.0 releases.
3399 In environments in which all applications have been ported to
3400 not use any deprecated interfaces OpenSSL's Configure script
3401 should be used with the --api=1.1.0 option to entirely remove
3402 support for the deprecated features from the library and
3403 unconditionally disable them in the installed headers.
3404 Essentially the same effect can be achieved with the "no-deprecated"
3405 argument to Configure, except that this will always restrict
3406 the build to just the latest API, rather than a fixed API
3409 As applications are ported to future revisions of the API,
3410 they should update their compile-time OPENSSL_API_COMPAT define
3411 accordingly, but in most cases should be able to continue to
3412 compile with later releases.
3414 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3415 0x10000000L and 0x00908000L, respectively. However those
3416 versions did not support the OPENSSL_API_COMPAT feature, and
3417 so applications are not typically tested for explicit support
3418 of just the undeprecated features of either release.
3422 * Add support for setting the minimum and maximum supported protocol.
3423 It can bet set via the SSL_set_min_proto_version() and
3424 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3425 MaxProtocol. It's recommended to use the new APIs to disable
3426 protocols instead of disabling individual protocols using
3427 SSL_set_options() or SSL_CONF's Protocol. This change also
3428 removes support for disabling TLS 1.2 in the OpenSSL TLS
3429 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3433 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3437 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3438 and integrates ECDSA and ECDH functionality into EC. Implementations can
3439 now redirect key generation and no longer need to convert to or from
3442 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3443 include the ec.h header file instead.
3447 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3448 ciphers who are no longer supported and drops support the ephemeral RSA key
3449 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3453 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3454 opaque. For HMAC_CTX, the following constructors and destructors
3457 HMAC_CTX *HMAC_CTX_new(void);
3458 void HMAC_CTX_free(HMAC_CTX *ctx);
3460 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3461 destroy such methods has been added. See EVP_MD_meth_new(3) and
3462 EVP_CIPHER_meth_new(3) for documentation.
3465 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3466 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3467 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3468 an already created structure.
3469 2) For consistency with the majority of our object creators and
3470 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3471 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3472 for deprecated builds.
3476 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3477 cryptographic operations to be performed asynchronously as long as an
3478 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3479 further details. Libssl has also had this capability integrated with the
3480 introduction of the new mode SSL_MODE_ASYNC and associated error
3481 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3482 pages. This work was developed in partnership with Intel Corp.
3486 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3487 always enabled now. If you want to disable the support you should
3488 exclude it using the list of supported ciphers. This also means that the
3489 "-no_ecdhe" option has been removed from s_server.
3493 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3494 SSL_{CTX_}set1_curves() which can set a list.
3498 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3499 curve you want to support using SSL_{CTX_}set1_curves().
3503 * State machine rewrite. The state machine code has been significantly
3504 refactored in order to remove much duplication of code and solve issues
3505 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3506 further details). This change does have some associated API changes.
3507 Notably the SSL_state() function has been removed and replaced by
3508 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3509 SSL_set_state() has been removed altogether. The previous handshake states
3510 defined in ssl.h and ssl3.h have also been removed.
3514 * All instances of the string "ssleay" in the public API were replaced
3515 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3516 Some error codes related to internal RSA_eay API's were renamed.
3520 * The demo files in crypto/threads were moved to demo/threads.
3524 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3527 *Matt Caswell, Rich Salz*
3529 * New ASN.1 embed macro.
3531 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3532 structure is not allocated: it is part of the parent. That is instead of
3540 This reduces memory fragmentation and make it impossible to accidentally
3541 set a mandatory field to NULL.
3543 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3544 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3545 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3550 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3554 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3555 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3556 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3557 DES and RC4 ciphersuites.
3561 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3562 This changes the decoding behaviour for some invalid messages,
3563 though the change is mostly in the more lenient direction, and
3564 legacy behaviour is preserved as much as possible.
3568 * Fix no-stdio build.
3569 *David Woodhouse <David.Woodhouse@intel.com> and also*
3570 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3572 * New testing framework
3573 The testing framework has been largely rewritten and is now using
3574 perl and the perl modules Test::Harness and an extended variant of
3575 Test::More called OpenSSL::Test to do its work. All test scripts in
3576 test/ have been rewritten into test recipes, and all direct calls to
3577 executables in test/Makefile have become individual recipes using the
3578 simplified testing OpenSSL::Test::Simple.
3580 For documentation on our testing modules, do:
3582 perldoc test/testlib/OpenSSL/Test/Simple.pm
3583 perldoc test/testlib/OpenSSL/Test.pm
3587 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3588 are used; the latter aborts on memory leaks (usually checked on exit).
3589 Some undocumented "set malloc, etc., hooks" functions were removed
3590 and others were changed. All are now documented.
3594 * In DSA_generate_parameters_ex, if the provided seed is too short,
3597 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3599 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3600 from RFC4279, RFC4785, RFC5487, RFC5489.
3602 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3603 original RSA_PSK patch.
3607 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3608 era flag was never set throughout the codebase (only read). Also removed
3609 SSL3_FLAGS_POP_BUFFER which was only used if
3610 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3614 * Changed the default name options in the "ca", "crl", "req" and "x509"
3615 to be "oneline" instead of "compat".
3619 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3620 not aware of clients that still exhibit this bug, and the workaround
3621 hasn't been working properly for a while.
3625 * The return type of BIO_number_read() and BIO_number_written() as well as
3626 the corresponding num_read and num_write members in the BIO structure has
3627 changed from unsigned long to uint64_t. On platforms where an unsigned
3628 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3633 * Given the pervasive nature of TLS extensions it is inadvisable to run
3634 OpenSSL without support for them. It also means that maintaining
3635 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3636 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3640 * Removed support for the two export grade static DH ciphersuites
3641 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3642 were newly added (along with a number of other static DH ciphersuites) to
3643 1.0.2. However the two export ones have *never* worked since they were
3644 introduced. It seems strange in any case to be adding new export
3645 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3649 * Version negotiation has been rewritten. In particular SSLv23_method(),
3650 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3651 and turned into macros which simply call the new preferred function names
3652 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3653 should use the new names instead. Also as part of this change the ssl23.h
3654 header file has been removed.
3658 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3659 code and the associated standard is no longer considered fit-for-purpose.
3663 * RT2547 was closed. When generating a private key, try to make the
3664 output file readable only by the owner. This behavior change might
3665 be noticeable when interacting with other software.
3667 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3672 * Added HTTP GET support to the ocsp command.
3676 * Changed default digest for the dgst and enc commands from MD5 to
3681 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3685 * Added support for TLS extended master secret from
3686 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3687 initial patch which was a great help during development.
3691 * All libssl internal structures have been removed from the public header
3692 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3693 now redundant). Users should not attempt to access internal structures
3694 directly. Instead they should use the provided API functions.
3698 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3699 Access to deprecated functions can be re-enabled by running config with
3700 "enable-deprecated". In addition applications wishing to use deprecated
3701 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3702 will, by default, disable some transitive includes that previously existed
3703 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3707 * Added support for OCB mode. OpenSSL has been granted a patent license
3708 compatible with the OpenSSL license for use of OCB. Details are available
3709 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3710 for OCB can be removed by calling config with no-ocb.
3714 * SSLv2 support has been removed. It still supports receiving a SSLv2
3715 compatible client hello.
3719 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3720 done while fixing the error code for the key-too-small case.
3722 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3724 * CA.sh has been removed; use CA.pl instead.
3728 * Removed old DES API.
3732 * Remove various unsupported platforms:
3738 Sinix/ReliantUNIX RM400
3743 16-bit platforms such as WIN16
3747 * Clean up OPENSSL_NO_xxx #define's
3748 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3749 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3750 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3751 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3752 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3753 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3754 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3755 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3756 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3757 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3761 * Cleaned up dead code
3762 Remove all but one '#ifdef undef' which is to be looked at.
3766 * Clean up calling of xxx_free routines.
3767 Just like free(), fix most of the xxx_free routines to accept
3768 NULL. Remove the non-null checks from callers. Save much code.
3772 * Add secure heap for storage of private keys (when possible).
3773 Add BIO_s_secmem(), CBIGNUM, etc.
3774 Contributed by Akamai Technologies under our Corporate CLA.
3778 * Experimental support for a new, fast, unbiased prime candidate generator,
3779 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3781 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3783 * New output format NSS in the sess_id command line tool. This allows
3784 exporting the session id and the master key in NSS keylog format.
3786 *Martin Kaiser <martin@kaiser.cx>*
3788 * Harmonize version and its documentation. -f flag is used to display
3791 *mancha <mancha1@zoho.com>*
3793 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3794 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3796 *mancha <mancha1@zoho.com>*
3798 * Fix some double frees. These are not thought to be exploitable.
3800 *mancha <mancha1@zoho.com>*
3802 * A missing bounds check in the handling of the TLS heartbeat extension
3803 can be used to reveal up to 64k of memory to a connected client or
3806 Thanks for Neel Mehta of Google Security for discovering this bug and to
3807 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3808 preparing the fix ([CVE-2014-0160])
3810 *Adam Langley, Bodo Moeller*
3812 * Fix for the attack described in the paper "Recovering OpenSSL
3813 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3814 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3815 <http://eprint.iacr.org/2014/140>
3817 Thanks to Yuval Yarom and Naomi Benger for discovering this
3818 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3820 *Yuval Yarom and Naomi Benger*
3822 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3823 this fixes a limitation in previous versions of OpenSSL.
3827 * Experimental encrypt-then-mac support.
3829 Experimental support for encrypt then mac from
3830 draft-gutmann-tls-encrypt-then-mac-02.txt
3832 To enable it set the appropriate extension number (0x42 for the test
3833 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3835 For non-compliant peers (i.e. just about everything) this should have no
3838 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3842 * Add EVP support for key wrapping algorithms, to avoid problems with
3843 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3844 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3845 algorithms and include tests cases.
3849 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3854 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3855 MGF1 digest and OAEP label.
3859 * Make openssl verify return errors.
3861 *Chris Palmer <palmer@google.com> and Ben Laurie*
3863 * New function ASN1_TIME_diff to calculate the difference between two
3864 ASN1_TIME structures or one structure and the current time.
3868 * Update fips_test_suite to support multiple command line options. New
3869 test to induce all self test errors in sequence and check expected
3874 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3875 sign or verify all in one operation.
3879 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3880 test programs and fips_test_suite. Includes functionality to parse
3881 the minimal script output of fipsalgest.pl directly.
3885 * Add authorisation parameter to FIPS_module_mode_set().
3889 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3893 * Use separate DRBG fields for internal and external flags. New function
3894 FIPS_drbg_health_check() to perform on demand health checking. Add
3895 generation tests to fips_test_suite with reduced health check interval to
3896 demonstrate periodic health checking. Add "nodh" option to
3897 fips_test_suite to skip very slow DH test.
3901 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3906 * More extensive health check for DRBG checking many more failure modes.
3907 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3908 combination: call this in fips_test_suite.
3912 * Add support for canonical generation of DSA parameter 'g'. See
3915 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3916 POST to handle HMAC cases.
3920 * Add functions FIPS_module_version() and FIPS_module_version_text()
3921 to return numerical and string versions of the FIPS module number.
3925 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3926 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3927 outside the validated module in the FIPS capable OpenSSL.
3931 * Minor change to DRBG entropy callback semantics. In some cases
3932 there is no multiple of the block length between min_len and
3933 max_len. Allow the callback to return more than max_len bytes
3934 of entropy but discard any extra: it is the callback's responsibility
3935 to ensure that the extra data discarded does not impact the
3936 requested amount of entropy.
3940 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3941 information in FIPS186-3, SP800-57 and SP800-131A.
3945 * CCM support via EVP. Interface is very similar to GCM case except we
3946 must supply all data in one chunk (i.e. no update, final) and the
3947 message length must be supplied if AAD is used. Add algorithm test
3952 * Initial version of POST overhaul. Add POST callback to allow the status
3953 of POST to be monitored and/or failures induced. Modify fips_test_suite
3954 to use callback. Always run all selftests even if one fails.
3958 * XTS support including algorithm test driver in the fips_gcmtest program.
3959 Note: this does increase the maximum key length from 32 to 64 bytes but
3960 there should be no binary compatibility issues as existing applications
3961 will never use XTS mode.
3965 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3966 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3967 performs algorithm blocking for unapproved PRNG types. Also do not
3968 set PRNG type in FIPS_mode_set(): leave this to the application.
3969 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3970 the standard OpenSSL PRNG: set additional data to a date time vector.
3974 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3975 This shouldn't present any incompatibility problems because applications
3976 shouldn't be using these directly and any that are will need to rethink
3977 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3981 * Extensive self tests and health checking required by SP800-90 DRBG.
3982 Remove strength parameter from FIPS_drbg_instantiate and always
3983 instantiate at maximum supported strength.
3987 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3991 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3995 * New function DH_compute_key_padded() to compute a DH key and pad with
3996 leading zeroes if needed: this complies with SP800-56A et al.
4000 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4001 anything, incomplete, subject to change and largely untested at present.
4005 * Modify fipscanisteronly build option to only build the necessary object
4006 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4010 * Add experimental option FIPSSYMS to give all symbols in
4011 fipscanister.o and FIPS or fips prefix. This will avoid
4012 conflicts with future versions of OpenSSL. Add perl script
4013 util/fipsas.pl to preprocess assembly language source files
4014 and rename any affected symbols.
4018 * Add selftest checks and algorithm block of non-fips algorithms in
4019 FIPS mode. Remove DES2 from selftests.
4023 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4024 return internal method without any ENGINE dependencies. Add new
4025 tiny fips sign and verify functions.
4029 * New build option no-ec2m to disable characteristic 2 code.
4033 * New build option "fipscanisteronly". This only builds fipscanister.o
4034 and (currently) associated fips utilities. Uses the file Makefile.fips
4035 instead of Makefile.org as the prototype.
4039 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4040 Update fips_gcmtest to use IV generator.
4044 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4045 setting output buffer to NULL. The `*Final` function must be
4046 called although it will not retrieve any additional data. The tag
4047 can be set or retrieved with a ctrl. The IV length is by default 12
4048 bytes (96 bits) but can be set to an alternative value. If the IV
4049 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4054 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4055 underlying do_cipher function handles all cipher semantics itself
4056 including padding and finalisation. This is useful if (for example)
4057 an ENGINE cipher handles block padding itself. The behaviour of
4058 do_cipher is subtly changed if this flag is set: the return value
4059 is the number of characters written to the output buffer (zero is
4060 no longer an error code) or a negative error code. Also if the
4061 input buffer is NULL and length 0 finalisation should be performed.
4065 * If a candidate issuer certificate is already part of the constructed
4066 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4070 * Improve forward-security support: add functions
4072 void SSL_CTX_set_not_resumable_session_callback(
4073 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4074 void SSL_set_not_resumable_session_callback(
4075 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4077 for use by SSL/TLS servers; the callback function will be called whenever a
4078 new session is created, and gets to decide whether the session may be
4079 cached to make it resumable (return 0) or not (return 1). (As by the
4080 SSL/TLS protocol specifications, the session_id sent by the server will be
4081 empty to indicate that the session is not resumable; also, the server will
4082 not generate RFC 4507 (RFC 5077) session tickets.)
4084 A simple reasonable callback implementation is to return is_forward_secure.
4085 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4086 by the SSL/TLS server library, indicating whether it can provide forward
4089 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4091 * New -verify_name option in command line utilities to set verification
4096 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4097 Add CMAC pkey methods.
4101 * Experimental renegotiation in s_server -www mode. If the client
4102 browses /reneg connection is renegotiated. If /renegcert it is
4103 renegotiated requesting a certificate.
4107 * Add an "external" session cache for debugging purposes to s_server. This
4108 should help trace issues which normally are only apparent in deployed
4109 multi-process servers.
4113 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4114 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4115 BIO_set_cipher() and some obscure PEM functions were changed so they
4116 can now return an error. The RAND changes required a change to the
4117 RAND_METHOD structure.
4121 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4122 a gcc attribute to warn if the result of a function is ignored. This
4123 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4124 whose return value is often ignored.
4128 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4129 These allow SCTs (signed certificate timestamps) to be requested and
4130 validated when establishing a connection.
4132 *Rob Percival <robpercival@google.com>*
4137 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4139 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4140 used even when parsing explicit parameters, when loading a encoded key
4141 or calling `EC_GROUP_new_from_ecpkparameters()`/
4142 `EC_GROUP_new_from_ecparameters()`.
4143 This prevents bypass of security hardening and performance gains,
4144 especially for curves with specialized EC_METHODs.
4145 By default, if a key encoded with explicit parameters is loaded and later
4146 encoded, the output is still encoded with explicit parameters, even if
4147 internally a "named" EC_GROUP is used for computation.
4151 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4152 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4153 NULL. After this change, only the cofactor parameter can be NULL. It also
4154 does some minimal sanity checks on the passed order.
4159 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4160 An attack is simple, if the first CMS_recipientInfo is valid but the
4161 second CMS_recipientInfo is chosen ciphertext. If the second
4162 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4163 encryption key will be replaced by garbage, and the message cannot be
4164 decoded, but if the RSA decryption fails, the correct encryption key is
4165 used and the recipient will not notice the attack.
4166 As a work around for this potential attack the length of the decrypted
4167 key must be equal to the cipher default key length, in case the
4168 certifiate is not given and all recipientInfo are tried out.
4169 The old behaviour can be re-enabled in the CMS code by setting the
4170 CMS_DEBUG_DECRYPT flag.
4175 * Document issue with installation paths in diverse Windows builds
4177 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4178 binaries and run-time config file.
4183 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4185 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4186 This changes the size when using the `genpkey` command when no size is given.
4187 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4188 generation commands to use 2048 bits by default.
4192 * Add FIPS support for Android Arm 64-bit
4194 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4195 Module in Version 2.0.10. For some reason, the corresponding target
4196 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4197 built with FIPS support on Android Arm 64-bit. This omission has been
4200 *Matthias St. Pierre*
4202 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4204 * 0-byte record padding oracle
4206 If an application encounters a fatal protocol error and then calls
4207 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4208 then OpenSSL can respond differently to the calling application if a 0 byte
4209 record is received with invalid padding compared to if a 0 byte record is
4210 received with an invalid MAC. If the application then behaves differently
4211 based on that in a way that is detectable to the remote peer, then this
4212 amounts to a padding oracle that could be used to decrypt data.
4214 In order for this to be exploitable "non-stitched" ciphersuites must be in
4215 use. Stitched ciphersuites are optimised implementations of certain
4216 commonly used ciphersuites. Also the application must call SSL_shutdown()
4217 twice even if a protocol error has occurred (applications should not do
4218 this but some do anyway).
4220 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4221 Aviram, with additional investigation by Steven Collison and Andrew
4222 Hourselt. It was reported to OpenSSL on 10th December 2018.
4227 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4231 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4233 * Microarchitecture timing vulnerability in ECC scalar multiplication
4235 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4236 shown to be vulnerable to a microarchitecture timing side channel attack.
4237 An attacker with sufficient access to mount local timing attacks during
4238 ECDSA signature generation could recover the private key.
4240 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4241 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4247 * Timing vulnerability in DSA signature generation
4249 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4250 timing side channel attack. An attacker could use variations in the signing
4251 algorithm to recover the private key.
4253 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4258 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4259 Module, accidentally introduced while backporting security fixes from the
4260 development branch and hindering the use of ECC in FIPS mode.
4264 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4266 * Client DoS due to large DH parameter
4268 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4269 malicious server can send a very large prime value to the client. This will
4270 cause the client to spend an unreasonably long period of time generating a
4271 key for this prime resulting in a hang until the client has finished. This
4272 could be exploited in a Denial Of Service attack.
4274 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4279 * Cache timing vulnerability in RSA Key Generation
4281 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4282 a cache timing side channel attack. An attacker with sufficient access to
4283 mount cache timing attacks during the RSA key generation process could
4284 recover the private key.
4286 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4287 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4292 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4293 parameter is no longer accepted, as it leads to a corrupt table. NULL
4294 pem_str is reserved for alias entries only.
4298 * Revert blinding in ECDSA sign and instead make problematic addition
4299 length-invariant. Switch even to fixed-length Montgomery multiplication.
4303 * Change generating and checking of primes so that the error rate of not
4304 being prime depends on the intended use based on the size of the input.
4305 For larger primes this will result in more rounds of Miller-Rabin.
4306 The maximal error rate for primes with more than 1080 bits is lowered
4309 *Kurt Roeckx, Annie Yousar*
4311 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4315 * Add blinding to ECDSA and DSA signatures to protect against side channel
4316 attacks discovered by Keegan Ryan (NCC Group).
4320 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4321 now allow empty (zero character) pass phrases.
4325 * Certificate time validation (X509_cmp_time) enforces stricter
4326 compliance with RFC 5280. Fractional seconds and timezone offsets
4327 are no longer allowed.
4331 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4333 * Constructed ASN.1 types with a recursive definition could exceed the stack
4335 Constructed ASN.1 types with a recursive definition (such as can be found
4336 in PKCS7) could eventually exceed the stack given malicious input with
4337 excessive recursion. This could result in a Denial Of Service attack. There
4338 are no such structures used within SSL/TLS that come from untrusted sources
4339 so this is considered safe.
4341 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4347 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4349 * Read/write after SSL object in error state
4351 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4352 mechanism. The intent was that if a fatal error occurred during a handshake
4353 then OpenSSL would move into the error state and would immediately fail if
4354 you attempted to continue the handshake. This works as designed for the
4355 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4356 SSL_connect()), however due to a bug it does not work correctly if
4357 SSL_read() or SSL_write() is called directly. In that scenario, if the
4358 handshake fails then a fatal error will be returned in the initial function
4359 call. If SSL_read()/SSL_write() is subsequently called by the application
4360 for the same SSL object then it will succeed and the data is passed without
4361 being decrypted/encrypted directly from the SSL/TLS record layer.
4363 In order to exploit this issue an application bug would have to be present
4364 that resulted in a call to SSL_read()/SSL_write() being issued after having
4365 already received a fatal error.
4367 This issue was reported to OpenSSL by David Benjamin (Google).
4372 * rsaz_1024_mul_avx2 overflow bug on x86_64
4374 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4375 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4376 Analysis suggests that attacks against RSA and DSA as a result of this
4377 defect would be very difficult to perform and are not believed likely.
4378 Attacks against DH1024 are considered just feasible, because most of the
4379 work necessary to deduce information about a private key may be performed
4380 offline. The amount of resources required for such an attack would be
4381 significant. However, for an attack on TLS to be meaningful, the server
4382 would have to share the DH1024 private key among multiple clients, which is
4383 no longer an option since CVE-2016-0701.
4385 This only affects processors that support the AVX2 but not ADX extensions
4386 like Intel Haswell (4th generation).
4388 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4389 was originally found via the OSS-Fuzz project.
4394 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4396 * bn_sqrx8x_internal carry bug on x86_64
4398 There is a carry propagating bug in the x86_64 Montgomery squaring
4399 procedure. No EC algorithms are affected. Analysis suggests that attacks
4400 against RSA and DSA as a result of this defect would be very difficult to
4401 perform and are not believed likely. Attacks against DH are considered just
4402 feasible (although very difficult) because most of the work necessary to
4403 deduce information about a private key may be performed offline. The amount
4404 of resources required for such an attack would be very significant and
4405 likely only accessible to a limited number of attackers. An attacker would
4406 additionally need online access to an unpatched system using the target
4407 private key in a scenario with persistent DH parameters and a private
4408 key that is shared between multiple clients.
4410 This only affects processors that support the BMI1, BMI2 and ADX extensions
4411 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4413 This issue was reported to OpenSSL by the OSS-Fuzz project.
4418 * Malformed X.509 IPAddressFamily could cause OOB read
4420 If an X.509 certificate has a malformed IPAddressFamily extension,
4421 OpenSSL could do a one-byte buffer overread. The most likely result
4422 would be an erroneous display of the certificate in text format.
4424 This issue was reported to OpenSSL by the OSS-Fuzz project.
4429 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4431 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4432 platform rather than 'mingw'.
4436 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4438 * Truncated packet could crash via OOB read
4440 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4441 cipher is being used, then a truncated packet can cause that host to
4442 perform an out-of-bounds read, usually resulting in a crash.
4444 This issue was reported to OpenSSL by Robert Święcki of Google.
4449 * BN_mod_exp may produce incorrect results on x86_64
4451 There is a carry propagating bug in the x86_64 Montgomery squaring
4452 procedure. No EC algorithms are affected. Analysis suggests that attacks
4453 against RSA and DSA as a result of this defect would be very difficult to
4454 perform and are not believed likely. Attacks against DH are considered just
4455 feasible (although very difficult) because most of the work necessary to
4456 deduce information about a private key may be performed offline. The amount
4457 of resources required for such an attack would be very significant and
4458 likely only accessible to a limited number of attackers. An attacker would
4459 additionally need online access to an unpatched system using the target
4460 private key in a scenario with persistent DH parameters and a private
4461 key that is shared between multiple clients. For example this can occur by
4462 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4463 similar to CVE-2015-3193 but must be treated as a separate problem.
4465 This issue was reported to OpenSSL by the OSS-Fuzz project.
4470 * Montgomery multiplication may produce incorrect results
4472 There is a carry propagating bug in the Broadwell-specific Montgomery
4473 multiplication procedure that handles input lengths divisible by, but
4474 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4475 and DH private keys are impossible. This is because the subroutine in
4476 question is not used in operations with the private key itself and an input
4477 of the attacker's direct choice. Otherwise the bug can manifest itself as
4478 transient authentication and key negotiation failures or reproducible
4479 erroneous outcome of public-key operations with specially crafted input.
4480 Among EC algorithms only Brainpool P-512 curves are affected and one
4481 presumably can attack ECDH key negotiation. Impact was not analyzed in
4482 detail, because pre-requisites for attack are considered unlikely. Namely
4483 multiple clients have to choose the curve in question and the server has to
4484 share the private key among them, neither of which is default behaviour.
4485 Even then only clients that chose the curve will be affected.
4487 This issue was publicly reported as transient failures and was not
4488 initially recognized as a security issue. Thanks to Richard Morgan for
4489 providing reproducible case.
4494 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4495 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4496 prevent issues where no progress is being made and the peer continually
4497 sends unrecognised record types, using up resources processing them.
4501 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4503 * Missing CRL sanity check
4505 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4506 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4507 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4509 This issue only affects the OpenSSL 1.0.2i
4514 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4516 * OCSP Status Request extension unbounded memory growth
4518 A malicious client can send an excessively large OCSP Status Request
4519 extension. If that client continually requests renegotiation, sending a
4520 large OCSP Status Request extension each time, then there will be unbounded
4521 memory growth on the server. This will eventually lead to a Denial Of
4522 Service attack through memory exhaustion. Servers with a default
4523 configuration are vulnerable even if they do not support OCSP. Builds using
4524 the "no-ocsp" build time option are not affected.
4526 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4531 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4534 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4540 * OOB write in MDC2_Update()
4542 An overflow can occur in MDC2_Update() either if called directly or
4543 through the EVP_DigestUpdate() function using MDC2. If an attacker
4544 is able to supply very large amounts of input data after a previous
4545 call to EVP_EncryptUpdate() with a partial block then a length check
4546 can overflow resulting in a heap corruption.
4548 The amount of data needed is comparable to SIZE_MAX which is impractical
4551 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4556 * Malformed SHA512 ticket DoS
4558 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4559 DoS attack where a malformed ticket will result in an OOB read which will
4562 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4563 a custom server callback and ticket lookup mechanism.
4565 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4570 * OOB write in BN_bn2dec()
4572 The function BN_bn2dec() does not check the return value of BN_div_word().
4573 This can cause an OOB write if an application uses this function with an
4574 overly large BIGNUM. This could be a problem if an overly large certificate
4575 or CRL is printed out from an untrusted source. TLS is not affected because
4576 record limits will reject an oversized certificate before it is parsed.
4578 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4583 * OOB read in TS_OBJ_print_bio()
4585 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4586 the total length the OID text representation would use and not the amount
4587 of data written. This will result in OOB reads when large OIDs are
4590 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4595 * Pointer arithmetic undefined behaviour
4597 Avoid some undefined pointer arithmetic
4599 A common idiom in the codebase is to check limits in the following manner:
4602 Where "p" points to some malloc'd data of SIZE bytes and
4605 "len" here could be from some externally supplied data (e.g. from a TLS
4608 The rules of C pointer arithmetic are such that "p + len" is only well
4609 defined where len <= SIZE. Therefore the above idiom is actually
4610 undefined behaviour.
4612 For example this could cause problems if some malloc implementation
4613 provides an address for "p" such that "p + len" actually overflows for
4614 values of len that are too big and therefore p + len < limit.
4616 This issue was reported to OpenSSL by Guido Vranken
4621 * Constant time flag not preserved in DSA signing
4623 Operations in the DSA signing algorithm should run in constant time in
4624 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4625 implementation means that a non-constant time codepath is followed for
4626 certain operations. This has been demonstrated through a cache-timing
4627 attack to be sufficient for an attacker to recover the private DSA key.
4629 This issue was reported by César Pereida (Aalto University), Billy Brumley
4630 (Tampere University of Technology), and Yuval Yarom (The University of
4631 Adelaide and NICTA).
4636 * DTLS buffered message DoS
4638 In a DTLS connection where handshake messages are delivered out-of-order
4639 those messages that OpenSSL is not yet ready to process will be buffered
4640 for later use. Under certain circumstances, a flaw in the logic means that
4641 those messages do not get removed from the buffer even though the handshake
4642 has been completed. An attacker could force up to approx. 15 messages to
4643 remain in the buffer when they are no longer required. These messages will
4644 be cleared when the DTLS connection is closed. The default maximum size for
4645 a message is 100k. Therefore the attacker could force an additional 1500k
4646 to be consumed per connection. By opening many simulataneous connections an
4647 attacker could cause a DoS attack through memory exhaustion.
4649 This issue was reported to OpenSSL by Quan Luo.
4654 * DTLS replay protection DoS
4656 A flaw in the DTLS replay attack protection mechanism means that records
4657 that arrive for future epochs update the replay protection "window" before
4658 the MAC for the record has been validated. This could be exploited by an
4659 attacker by sending a record for the next epoch (which does not have to
4660 decrypt or have a valid MAC), with a very large sequence number. This means
4661 that all subsequent legitimate packets are dropped causing a denial of
4662 service for a specific DTLS connection.
4664 This issue was reported to OpenSSL by the OCAP audit team.
4669 * Certificate message OOB reads
4671 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4672 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4673 theoretical DoS risk but this has not been observed in practice on common
4676 The messages affected are client certificate, client certificate request
4677 and server certificate. As a result the attack can only be performed
4678 against a client or a server which enables client authentication.
4680 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4685 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4687 * Prevent padding oracle in AES-NI CBC MAC check
4689 A MITM attacker can use a padding oracle attack to decrypt traffic
4690 when the connection uses an AES CBC cipher and the server support
4693 This issue was introduced as part of the fix for Lucky 13 padding
4694 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4695 constant time by making sure that always the same bytes are read and
4696 compared against either the MAC or padding bytes. But it no longer
4697 checked that there was enough data to have both the MAC and padding
4700 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4705 * Fix EVP_EncodeUpdate overflow
4707 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4708 Base64 encoding of binary data. If an attacker is able to supply very large
4709 amounts of input data then a length check can overflow resulting in a heap
4712 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4713 the `PEM_write_bio*` family of functions. These are mainly used within the
4714 OpenSSL command line applications, so any application which processes data
4715 from an untrusted source and outputs it as a PEM file should be considered
4716 vulnerable to this issue. User applications that call these APIs directly
4717 with large amounts of untrusted data may also be vulnerable.
4719 This issue was reported by Guido Vranken.
4724 * Fix EVP_EncryptUpdate overflow
4726 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4727 is able to supply very large amounts of input data after a previous call to
4728 EVP_EncryptUpdate() with a partial block then a length check can overflow
4729 resulting in a heap corruption. Following an analysis of all OpenSSL
4730 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4731 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4732 the first called function after an EVP_EncryptInit(), and therefore that
4733 specific call must be safe. The second form is where the length passed to
4734 EVP_EncryptUpdate() can be seen from the code to be some small value and
4735 therefore there is no possibility of an overflow. Since all instances are
4736 one of these two forms, it is believed that there can be no overflows in
4737 internal code due to this problem. It should be noted that
4738 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4739 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4740 of these calls have also been analysed too and it is believed there are no
4741 instances in internal usage where an overflow could occur.
4743 This issue was reported by Guido Vranken.
4748 * Prevent ASN.1 BIO excessive memory allocation
4750 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4751 a short invalid encoding can cause allocation of large amounts of memory
4752 potentially consuming excessive resources or exhausting memory.
4754 Any application parsing untrusted data through d2i BIO functions is
4755 affected. The memory based functions such as d2i_X509() are *not* affected.
4756 Since the memory based functions are used by the TLS library, TLS
4757 applications are not affected.
4759 This issue was reported by Brian Carpenter.
4766 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4767 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4768 in arbitrary stack data being returned in the buffer.
4770 This issue was reported by Guido Vranken.
4775 * Modify behavior of ALPN to invoke callback after SNI/servername
4776 callback, such that updates to the SSL_CTX affect ALPN.
4780 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4785 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4786 methods are enabled and ssl2 is disabled the methods return NULL.
4790 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4792 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4793 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4794 provide any "EXPORT" or "LOW" strength ciphers.
4798 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4799 is by default disabled at build-time. Builds that are not configured with
4800 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4801 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4802 will need to explicitly call either of:
4804 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4806 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4808 as appropriate. Even if either of those is used, or the application
4809 explicitly uses the version-specific SSLv2_method() or its client and
4810 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4811 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4812 ciphers, and SSLv2 56-bit DES are no longer available.
4817 * Fix a double-free in DSA code
4819 A double free bug was discovered when OpenSSL parses malformed DSA private
4820 keys and could lead to a DoS attack or memory corruption for applications
4821 that receive DSA private keys from untrusted sources. This scenario is
4824 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4830 * Disable SRP fake user seed to address a server memory leak.
4832 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4834 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4835 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4836 was changed to ignore the "fake user" SRP seed, even if the seed
4839 Users should use SRP_VBASE_get1_by_user instead. Note that in
4840 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4841 also that even though configuring the SRP seed attempts to hide
4842 invalid usernames by continuing the handshake with fake
4843 credentials, this behaviour is not constant time and no strong
4844 guarantees are made that the handshake is indistinguishable from
4845 that of a valid user.
4850 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4852 In the BN_hex2bn function the number of hex digits is calculated using an
4853 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4854 large values of `i` this can result in `bn_expand` not allocating any
4855 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4856 field as NULL leading to a subsequent NULL ptr deref. For very large values
4857 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4858 In this case memory is allocated to the internal BIGNUM data field, but it
4859 is insufficiently sized leading to heap corruption. A similar issue exists
4860 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4861 is ever called by user applications with very large untrusted hex/dec data.
4862 This is anticipated to be a rare occurrence.
4864 All OpenSSL internal usage of these functions use data that is not expected
4865 to be untrusted, e.g. config file data or application command line
4866 arguments. If user developed applications generate config file data based
4867 on untrusted data then it is possible that this could also lead to security
4868 consequences. This is also anticipated to be rare.
4870 This issue was reported to OpenSSL by Guido Vranken.
4875 * Fix memory issues in `BIO_*printf` functions
4877 The internal `fmtstr` function used in processing a "%s" format string in
4878 the `BIO_*printf` functions could overflow while calculating the length of a
4879 string and cause an OOB read when printing very long strings.
4881 Additionally the internal `doapr_outch` function can attempt to write to an
4882 OOB memory location (at an offset from the NULL pointer) in the event of a
4883 memory allocation failure. In 1.0.2 and below this could be caused where
4884 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4885 could be in processing a very long "%s" format string. Memory leaks can
4888 The first issue may mask the second issue dependent on compiler behaviour.
4889 These problems could enable attacks where large amounts of untrusted data
4890 is passed to the `BIO_*printf` functions. If applications use these functions
4891 in this way then they could be vulnerable. OpenSSL itself uses these
4892 functions when printing out human-readable dumps of ASN.1 data. Therefore
4893 applications that print this data could be vulnerable if the data is from
4894 untrusted sources. OpenSSL command line applications could also be
4895 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4896 as command line arguments.
4898 Libssl is not considered directly vulnerable. Additionally certificates etc
4899 received via remote connections via libssl are also unlikely to be able to
4900 trigger these issues because of message size limits enforced within libssl.
4902 This issue was reported to OpenSSL Guido Vranken.
4907 * Side channel attack on modular exponentiation
4909 A side-channel attack was found which makes use of cache-bank conflicts on
4910 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4911 of RSA keys. The ability to exploit this issue is limited as it relies on
4912 an attacker who has control of code in a thread running on the same
4913 hyper-threaded core as the victim thread which is performing decryptions.
4915 This issue was reported to OpenSSL by Yuval Yarom, The University of
4916 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4917 Nadia Heninger, University of Pennsylvania with more information at
4918 <http://cachebleed.info>.
4923 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4924 if no keysize is specified with default_bits. This fixes an
4925 omission in an earlier change that changed all RSA/DSA key generation
4926 commands to use 2048 bits by default.
4930 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4932 * DH small subgroups
4934 Historically OpenSSL only ever generated DH parameters based on "safe"
4935 primes. More recently (in version 1.0.2) support was provided for
4936 generating X9.42 style parameter files such as those required for RFC 5114
4937 support. The primes used in such files may not be "safe". Where an
4938 application is using DH configured with parameters based on primes that are
4939 not "safe" then an attacker could use this fact to find a peer's private
4940 DH exponent. This attack requires that the attacker complete multiple
4941 handshakes in which the peer uses the same private DH exponent. For example
4942 this could be used to discover a TLS server's private DH exponent if it's
4943 reusing the private DH exponent or it's using a static DH ciphersuite.
4945 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4946 TLS. It is not on by default. If the option is not set then the server
4947 reuses the same private DH exponent for the life of the server process and
4948 would be vulnerable to this attack. It is believed that many popular
4949 applications do set this option and would therefore not be at risk.
4951 The fix for this issue adds an additional check where a "q" parameter is
4952 available (as is the case in X9.42 based parameters). This detects the
4953 only known attack, and is the only possible defense for static DH
4954 ciphersuites. This could have some performance impact.
4956 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4957 default and cannot be disabled. This could have some performance impact.
4959 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4964 * SSLv2 doesn't block disabled ciphers
4966 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4967 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4968 been disabled, provided that the SSLv2 protocol was not also disabled via
4971 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4972 and Sebastian Schinzel.
4977 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4979 * BN_mod_exp may produce incorrect results on x86_64
4981 There is a carry propagating bug in the x86_64 Montgomery squaring
4982 procedure. No EC algorithms are affected. Analysis suggests that attacks
4983 against RSA and DSA as a result of this defect would be very difficult to
4984 perform and are not believed likely. Attacks against DH are considered just
4985 feasible (although very difficult) because most of the work necessary to
4986 deduce information about a private key may be performed offline. The amount
4987 of resources required for such an attack would be very significant and
4988 likely only accessible to a limited number of attackers. An attacker would
4989 additionally need online access to an unpatched system using the target
4990 private key in a scenario with persistent DH parameters and a private
4991 key that is shared between multiple clients. For example this can occur by
4992 default in OpenSSL DHE based SSL/TLS ciphersuites.
4994 This issue was reported to OpenSSL by Hanno Böck.
4999 * Certificate verify crash with missing PSS parameter
5001 The signature verification routines will crash with a NULL pointer
5002 dereference if presented with an ASN.1 signature using the RSA PSS
5003 algorithm and absent mask generation function parameter. Since these
5004 routines are used to verify certificate signature algorithms this can be
5005 used to crash any certificate verification operation and exploited in a
5006 DoS attack. Any application which performs certificate verification is
5007 vulnerable including OpenSSL clients and servers which enable client
5010 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5015 * X509_ATTRIBUTE memory leak
5017 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5018 memory. This structure is used by the PKCS#7 and CMS routines so any
5019 application which reads PKCS#7 or CMS data from untrusted sources is
5020 affected. SSL/TLS is not affected.
5022 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5028 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5029 This changes the decoding behaviour for some invalid messages,
5030 though the change is mostly in the more lenient direction, and
5031 legacy behaviour is preserved as much as possible.
5035 * In DSA_generate_parameters_ex, if the provided seed is too short,
5038 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5040 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5042 * Alternate chains certificate forgery
5044 During certificate verification, OpenSSL will attempt to find an
5045 alternative certificate chain if the first attempt to build such a chain
5046 fails. An error in the implementation of this logic can mean that an
5047 attacker could cause certain checks on untrusted certificates to be
5048 bypassed, such as the CA flag, enabling them to use a valid leaf
5049 certificate to act as a CA and "issue" an invalid certificate.
5051 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5056 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5058 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5059 incompatibility in the handling of HMAC. The previous ABI has now been
5064 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5066 * Malformed ECParameters causes infinite loop
5068 When processing an ECParameters structure OpenSSL enters an infinite loop
5069 if the curve specified is over a specially malformed binary polynomial
5072 This can be used to perform denial of service against any
5073 system which processes public keys, certificate requests or
5074 certificates. This includes TLS clients and TLS servers with
5075 client authentication enabled.
5077 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5082 * Exploitable out-of-bounds read in X509_cmp_time
5084 X509_cmp_time does not properly check the length of the ASN1_TIME
5085 string and can read a few bytes out of bounds. In addition,
5086 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5089 An attacker can use this to craft malformed certificates and CRLs of
5090 various sizes and potentially cause a segmentation fault, resulting in
5091 a DoS on applications that verify certificates or CRLs. TLS clients
5092 that verify CRLs are affected. TLS clients and servers with client
5093 authentication enabled may be affected if they use custom verification
5096 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5097 independently by Hanno Böck.
5102 * PKCS7 crash with missing EnvelopedContent
5104 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5105 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5106 with missing content and trigger a NULL pointer dereference on parsing.
5108 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5109 structures from untrusted sources are affected. OpenSSL clients and
5110 servers are not affected.
5112 This issue was reported to OpenSSL by Michal Zalewski (Google).
5117 * CMS verify infinite loop with unknown hash function
5119 When verifying a signedData message the CMS code can enter an infinite loop
5120 if presented with an unknown hash function OID. This can be used to perform
5121 denial of service against any system which verifies signedData messages using
5123 This issue was reported to OpenSSL by Johannes Bauer.
5128 * Race condition handling NewSessionTicket
5130 If a NewSessionTicket is received by a multi-threaded client when attempting to
5131 reuse a previous ticket then a race condition can occur potentially leading to
5132 a double free of the ticket data.
5137 * Only support 256-bit or stronger elliptic curves with the
5138 'ecdh_auto' setting (server) or by default (client). Of supported
5139 curves, prefer P-256 (both).
5143 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5145 * ClientHello sigalgs DoS fix
5147 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5148 invalid signature algorithms extension a NULL pointer dereference will
5149 occur. This can be exploited in a DoS attack against the server.
5151 This issue was was reported to OpenSSL by David Ramos of Stanford
5155 *Stephen Henson and Matt Caswell*
5157 * Multiblock corrupted pointer fix
5159 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5160 feature only applies on 64 bit x86 architecture platforms that support AES
5161 NI instructions. A defect in the implementation of "multiblock" can cause
5162 OpenSSL's internal write buffer to become incorrectly set to NULL when
5163 using non-blocking IO. Typically, when the user application is using a
5164 socket BIO for writing, this will only result in a failed connection.
5165 However if some other BIO is used then it is likely that a segmentation
5166 fault will be triggered, thus enabling a potential DoS attack.
5168 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5173 * Segmentation fault in DTLSv1_listen fix
5175 The DTLSv1_listen function is intended to be stateless and processes the
5176 initial ClientHello from many peers. It is common for user code to loop
5177 over the call to DTLSv1_listen until a valid ClientHello is received with
5178 an associated cookie. A defect in the implementation of DTLSv1_listen means
5179 that state is preserved in the SSL object from one invocation to the next
5180 that can lead to a segmentation fault. Errors processing the initial
5181 ClientHello can trigger this scenario. An example of such an error could be
5182 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5185 This issue was reported to OpenSSL by Per Allansson.
5190 * Segmentation fault in ASN1_TYPE_cmp fix
5192 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5193 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5194 certificate signature algorithm consistency this can be used to crash any
5195 certificate verification operation and exploited in a DoS attack. Any
5196 application which performs certificate verification is vulnerable including
5197 OpenSSL clients and servers which enable client authentication.
5202 * Segmentation fault for invalid PSS parameters fix
5204 The signature verification routines will crash with a NULL pointer
5205 dereference if presented with an ASN.1 signature using the RSA PSS
5206 algorithm and invalid parameters. Since these routines are used to verify
5207 certificate signature algorithms this can be used to crash any
5208 certificate verification operation and exploited in a DoS attack. Any
5209 application which performs certificate verification is vulnerable including
5210 OpenSSL clients and servers which enable client authentication.
5212 This issue was was reported to OpenSSL by Brian Carpenter.
5217 * ASN.1 structure reuse memory corruption fix
5219 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5220 memory corruption via an invalid write. Such reuse is and has been
5221 strongly discouraged and is believed to be rare.
5223 Applications that parse structures containing CHOICE or ANY DEFINED BY
5224 components may be affected. Certificate parsing (d2i_X509 and related
5225 functions) are however not affected. OpenSSL clients and servers are
5231 * PKCS7 NULL pointer dereferences fix
5233 The PKCS#7 parsing code does not handle missing outer ContentInfo
5234 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5235 missing content and trigger a NULL pointer dereference on parsing.
5237 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5238 otherwise parse PKCS#7 structures from untrusted sources are
5239 affected. OpenSSL clients and servers are not affected.
5241 This issue was reported to OpenSSL by Michal Zalewski (Google).
5246 * DoS via reachable assert in SSLv2 servers fix
5248 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5249 servers that both support SSLv2 and enable export cipher suites by sending
5250 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5252 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5253 (OpenSSL development team).
5258 * Empty CKE with client auth and DHE fix
5260 If client auth is used then a server can seg fault in the event of a DHE
5261 ciphersuite being selected and a zero length ClientKeyExchange message
5262 being sent by the client. This could be exploited in a DoS attack.
5267 * Handshake with unseeded PRNG fix
5269 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5270 with an unseeded PRNG. The conditions are:
5271 - The client is on a platform where the PRNG has not been seeded
5272 automatically, and the user has not seeded manually
5273 - A protocol specific client method version has been used (i.e. not
5274 SSL_client_methodv23)
5275 - A ciphersuite is used that does not require additional random data from
5276 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5278 If the handshake succeeds then the client random that has been used will
5279 have been generated from a PRNG with insufficient entropy and therefore the
5280 output may be predictable.
5282 For example using the following command with an unseeded openssl will
5283 succeed on an unpatched platform:
5285 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5290 * Use After Free following d2i_ECPrivatekey error fix
5292 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5293 could cause a use after free condition. This, in turn, could cause a double
5294 free in several private key parsing functions (such as d2i_PrivateKey
5295 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5296 for applications that receive EC private keys from untrusted
5297 sources. This scenario is considered rare.
5299 This issue was discovered by the BoringSSL project and fixed in their
5305 * X509_to_X509_REQ NULL pointer deref fix
5307 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5308 the certificate key is invalid. This function is rarely used in practice.
5310 This issue was discovered by Brian Carpenter.
5315 * Removed the export ciphers from the DEFAULT ciphers
5319 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5321 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5322 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5323 So far those who have to target multiple platforms would compromise
5324 and argue that binary targeting say ARMv5 would still execute on
5325 ARMv8. "Universal" build resolves this compromise by providing
5326 near-optimal performance even on newer platforms.
5330 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5331 (other platforms pending).
5333 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5335 * Add support for the SignedCertificateTimestampList certificate and
5336 OCSP response extensions from RFC6962.
5340 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5341 for corner cases. (Certain input points at infinity could lead to
5342 bogus results, with non-infinity inputs mapped to infinity too.)
5346 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5347 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5348 common cases are optimized and there still is room for further
5349 improvements. Vector Permutation AES for Altivec is also added.
5353 * Add support for little-endian ppc64 Linux target.
5355 *Marcelo Cerri (IBM)*
5357 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5358 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5359 are optimized and there still is room for further improvements.
5360 Both 32- and 64-bit modes are supported.
5362 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5364 * Improved ARMv7 NEON support.
5368 * Support for SPARC Architecture 2011 crypto extensions, first
5369 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5370 SHA256/512, MD5, GHASH and modular exponentiation.
5372 *Andy Polyakov, David Miller*
5374 * Accelerated modular exponentiation for Intel processors, a.k.a.
5377 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5379 * Support for new and upcoming Intel processors, including AVX2,
5380 BMI and SHA ISA extensions. This includes additional "stitched"
5381 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5384 This work was sponsored by Intel Corp.
5388 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5389 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5390 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5394 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5395 this fixes a limitation in previous versions of OpenSSL.
5399 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5400 MGF1 digest and OAEP label.
5404 * Add EVP support for key wrapping algorithms, to avoid problems with
5405 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5406 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5407 algorithms and include tests cases.
5411 * Add functions to allocate and set the fields of an ECDSA_METHOD
5414 *Douglas E. Engert, Steve Henson*
5416 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5417 difference in days and seconds between two tm or ASN1_TIME structures.
5421 * Add -rev test option to s_server to just reverse order of characters
5422 received by client and send back to server. Also prints an abbreviated
5423 summary of the connection parameters.
5427 * New option -brief for s_client and s_server to print out a brief summary
5428 of connection parameters.
5432 * Add callbacks for arbitrary TLS extensions.
5434 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5436 * New option -crl_download in several openssl utilities to download CRLs
5437 from CRLDP extension in certificates.
5441 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5445 * New function X509_CRL_diff to generate a delta CRL from the difference
5446 of two full CRLs. Add support to "crl" utility.
5450 * New functions to set lookup_crls function and to retrieve
5451 X509_STORE from X509_STORE_CTX.
5455 * Print out deprecated issuer and subject unique ID fields in
5460 * Extend OCSP I/O functions so they can be used for simple general purpose
5461 HTTP as well as OCSP. New wrapper function which can be used to download
5462 CRLs using the OCSP API.
5466 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5470 * `SSL_CONF*` functions. These provide a common framework for application
5471 configuration using configuration files or command lines.
5475 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5476 message callback and prints the results. Needs compile time option
5477 "enable-ssl-trace". New options to s_client and s_server to enable
5482 * New ctrl and macro to retrieve supported points extensions.
5483 Print out extension in s_server and s_client.
5487 * New functions to retrieve certificate signature and signature
5492 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5497 * New Suite B modes for TLS code. These use and enforce the requirements
5498 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5499 only use Suite B curves. The Suite B modes can be set by using the
5500 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5504 * New chain verification flags for Suite B levels of security. Check
5505 algorithms are acceptable when flags are set in X509_verify_cert.
5509 * Make tls1_check_chain return a set of flags indicating checks passed
5510 by a certificate chain. Add additional tests to handle client
5511 certificates: checks for matching certificate type and issuer name
5516 * If an attempt is made to use a signature algorithm not in the peer
5517 preference list abort the handshake. If client has no suitable
5518 signature algorithms in response to a certificate request do not
5519 use the certificate.
5523 * If server EC tmp key is not in client preference list abort handshake.
5527 * Add support for certificate stores in CERT structure. This makes it
5528 possible to have different stores per SSL structure or one store in
5529 the parent SSL_CTX. Include distinct stores for certificate chain
5530 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5531 to build and store a certificate chain in CERT structure: returning
5532 an error if the chain cannot be built: this will allow applications
5533 to test if a chain is correctly configured.
5535 Note: if the CERT based stores are not set then the parent SSL_CTX
5536 store is used to retain compatibility with existing behaviour.
5540 * New function ssl_set_client_disabled to set a ciphersuite disabled
5541 mask based on the current session, check mask when sending client
5542 hello and checking the requested ciphersuite.
5546 * New ctrls to retrieve and set certificate types in a certificate
5547 request message. Print out received values in s_client. If certificate
5548 types is not set with custom values set sensible values based on
5549 supported signature algorithms.
5553 * Support for distinct client and server supported signature algorithms.
5557 * Add certificate callback. If set this is called whenever a certificate
5558 is required by client or server. An application can decide which
5559 certificate chain to present based on arbitrary criteria: for example
5560 supported signature algorithms. Add very simple example to s_server.
5561 This fixes many of the problems and restrictions of the existing client
5562 certificate callback: for example you can now clear an existing
5563 certificate and specify the whole chain.
5567 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5568 the certificate can be used for (if anything). Set valid_flags field
5569 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5570 to have similar checks in it.
5572 Add new "cert_flags" field to CERT structure and include a "strict mode".
5573 This enforces some TLS certificate requirements (such as only permitting
5574 certificate signature algorithms contained in the supported algorithms
5575 extension) which some implementations ignore: this option should be used
5576 with caution as it could cause interoperability issues.
5580 * Update and tidy signature algorithm extension processing. Work out
5581 shared signature algorithms based on preferences and peer algorithms
5582 and print them out in s_client and s_server. Abort handshake if no
5583 shared signature algorithms.
5587 * Add new functions to allow customised supported signature algorithms
5588 for SSL and SSL_CTX structures. Add options to s_client and s_server
5593 * New function SSL_certs_clear() to delete all references to certificates
5594 from an SSL structure. Before this once a certificate had been added
5595 it couldn't be removed.
5599 * Integrate hostname, email address and IP address checking with certificate
5600 verification. New verify options supporting checking in openssl utility.
5604 * Fixes and wildcard matching support to hostname and email checking
5605 functions. Add manual page.
5607 *Florian Weimer (Red Hat Product Security Team)*
5609 * New functions to check a hostname email or IP address against a
5610 certificate. Add options x509 utility to print results of checks against
5615 * Fix OCSP checking.
5617 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5619 * Initial experimental support for explicitly trusted non-root CAs.
5620 OpenSSL still tries to build a complete chain to a root but if an
5621 intermediate CA has a trust setting included that is used. The first
5622 setting is used: whether to trust (e.g., -addtrust option to the x509
5627 * Add -trusted_first option which attempts to find certificates in the
5628 trusted store even if an untrusted chain is also supplied.
5632 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5633 platform support for Linux and Android.
5637 * Support for linux-x32, ILP32 environment in x86_64 framework.
5641 * Experimental multi-implementation support for FIPS capable OpenSSL.
5642 When in FIPS mode the approved implementations are used as normal,
5643 when not in FIPS mode the internal unapproved versions are used instead.
5644 This means that the FIPS capable OpenSSL isn't forced to use the
5645 (often lower performance) FIPS implementations outside FIPS mode.
5649 * Transparently support X9.42 DH parameters when calling
5650 PEM_read_bio_DHparameters. This means existing applications can handle
5651 the new parameter format automatically.
5655 * Initial experimental support for X9.42 DH parameter format: mainly
5656 to support use of 'q' parameter for RFC5114 parameters.
5660 * Add DH parameters from RFC5114 including test data to dhtest.
5664 * Support for automatic EC temporary key parameter selection. If enabled
5665 the most preferred EC parameters are automatically used instead of
5666 hardcoded fixed parameters. Now a server just has to call:
5667 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5668 support ECDH and use the most appropriate parameters.
5672 * Enhance and tidy EC curve and point format TLS extension code. Use
5673 static structures instead of allocation if default values are used.
5674 New ctrls to set curves we wish to support and to retrieve shared curves.
5675 Print out shared curves in s_server. New options to s_server and s_client
5676 to set list of supported curves.
5680 * New ctrls to retrieve supported signature algorithms and
5681 supported curve values as an array of NIDs. Extend openssl utility
5682 to print out received values.
5686 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5687 between NIDs and the more common NIST names such as "P-256". Enhance
5688 ecparam utility and ECC method to recognise the NIST names for curves.
5692 * Enhance SSL/TLS certificate chain handling to support different
5693 chains for each certificate instead of one chain in the parent SSL_CTX.
5697 * Support for fixed DH ciphersuite client authentication: where both
5698 server and client use DH certificates with common parameters.
5702 * Support for fixed DH ciphersuites: those requiring DH server
5707 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5709 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5710 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5711 X509_CINF_get_signature were reverted post internal team review.
5716 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5718 * OCSP Status Request extension unbounded memory growth
5720 A malicious client can send an excessively large OCSP Status Request
5721 extension. If that client continually requests renegotiation, sending a
5722 large OCSP Status Request extension each time, then there will be unbounded
5723 memory growth on the server. This will eventually lead to a Denial Of
5724 Service attack through memory exhaustion. Servers with a default
5725 configuration are vulnerable even if they do not support OCSP. Builds using
5726 the "no-ocsp" build time option are not affected.
5728 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5733 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5736 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5742 * OOB write in MDC2_Update()
5744 An overflow can occur in MDC2_Update() either if called directly or
5745 through the EVP_DigestUpdate() function using MDC2. If an attacker
5746 is able to supply very large amounts of input data after a previous
5747 call to EVP_EncryptUpdate() with a partial block then a length check
5748 can overflow resulting in a heap corruption.
5750 The amount of data needed is comparable to SIZE_MAX which is impractical
5753 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5758 * Malformed SHA512 ticket DoS
5760 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5761 DoS attack where a malformed ticket will result in an OOB read which will
5764 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5765 a custom server callback and ticket lookup mechanism.
5767 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5772 * OOB write in BN_bn2dec()
5774 The function BN_bn2dec() does not check the return value of BN_div_word().
5775 This can cause an OOB write if an application uses this function with an
5776 overly large BIGNUM. This could be a problem if an overly large certificate
5777 or CRL is printed out from an untrusted source. TLS is not affected because
5778 record limits will reject an oversized certificate before it is parsed.
5780 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5785 * OOB read in TS_OBJ_print_bio()
5787 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5788 the total length the OID text representation would use and not the amount
5789 of data written. This will result in OOB reads when large OIDs are
5792 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5797 * Pointer arithmetic undefined behaviour
5799 Avoid some undefined pointer arithmetic
5801 A common idiom in the codebase is to check limits in the following manner:
5804 Where "p" points to some malloc'd data of SIZE bytes and
5807 "len" here could be from some externally supplied data (e.g. from a TLS
5810 The rules of C pointer arithmetic are such that "p + len" is only well
5811 defined where len <= SIZE. Therefore the above idiom is actually
5812 undefined behaviour.
5814 For example this could cause problems if some malloc implementation
5815 provides an address for "p" such that "p + len" actually overflows for
5816 values of len that are too big and therefore p + len < limit.
5818 This issue was reported to OpenSSL by Guido Vranken
5823 * Constant time flag not preserved in DSA signing
5825 Operations in the DSA signing algorithm should run in constant time in
5826 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5827 implementation means that a non-constant time codepath is followed for
5828 certain operations. This has been demonstrated through a cache-timing
5829 attack to be sufficient for an attacker to recover the private DSA key.
5831 This issue was reported by César Pereida (Aalto University), Billy Brumley
5832 (Tampere University of Technology), and Yuval Yarom (The University of
5833 Adelaide and NICTA).
5838 * DTLS buffered message DoS
5840 In a DTLS connection where handshake messages are delivered out-of-order
5841 those messages that OpenSSL is not yet ready to process will be buffered
5842 for later use. Under certain circumstances, a flaw in the logic means that
5843 those messages do not get removed from the buffer even though the handshake
5844 has been completed. An attacker could force up to approx. 15 messages to
5845 remain in the buffer when they are no longer required. These messages will
5846 be cleared when the DTLS connection is closed. The default maximum size for
5847 a message is 100k. Therefore the attacker could force an additional 1500k
5848 to be consumed per connection. By opening many simulataneous connections an
5849 attacker could cause a DoS attack through memory exhaustion.
5851 This issue was reported to OpenSSL by Quan Luo.
5856 * DTLS replay protection DoS
5858 A flaw in the DTLS replay attack protection mechanism means that records
5859 that arrive for future epochs update the replay protection "window" before
5860 the MAC for the record has been validated. This could be exploited by an
5861 attacker by sending a record for the next epoch (which does not have to
5862 decrypt or have a valid MAC), with a very large sequence number. This means
5863 that all subsequent legitimate packets are dropped causing a denial of
5864 service for a specific DTLS connection.
5866 This issue was reported to OpenSSL by the OCAP audit team.
5871 * Certificate message OOB reads
5873 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5874 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5875 theoretical DoS risk but this has not been observed in practice on common
5878 The messages affected are client certificate, client certificate request
5879 and server certificate. As a result the attack can only be performed
5880 against a client or a server which enables client authentication.
5882 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5887 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5889 * Prevent padding oracle in AES-NI CBC MAC check
5891 A MITM attacker can use a padding oracle attack to decrypt traffic
5892 when the connection uses an AES CBC cipher and the server support
5895 This issue was introduced as part of the fix for Lucky 13 padding
5896 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5897 constant time by making sure that always the same bytes are read and
5898 compared against either the MAC or padding bytes. But it no longer
5899 checked that there was enough data to have both the MAC and padding
5902 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5907 * Fix EVP_EncodeUpdate overflow
5909 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5910 Base64 encoding of binary data. If an attacker is able to supply very large
5911 amounts of input data then a length check can overflow resulting in a heap
5914 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5915 the `PEM_write_bio*` family of functions. These are mainly used within the
5916 OpenSSL command line applications, so any application which processes data
5917 from an untrusted source and outputs it as a PEM file should be considered
5918 vulnerable to this issue. User applications that call these APIs directly
5919 with large amounts of untrusted data may also be vulnerable.
5921 This issue was reported by Guido Vranken.
5926 * Fix EVP_EncryptUpdate overflow
5928 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5929 is able to supply very large amounts of input data after a previous call to
5930 EVP_EncryptUpdate() with a partial block then a length check can overflow
5931 resulting in a heap corruption. Following an analysis of all OpenSSL
5932 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5933 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5934 the first called function after an EVP_EncryptInit(), and therefore that
5935 specific call must be safe. The second form is where the length passed to
5936 EVP_EncryptUpdate() can be seen from the code to be some small value and
5937 therefore there is no possibility of an overflow. Since all instances are
5938 one of these two forms, it is believed that there can be no overflows in
5939 internal code due to this problem. It should be noted that
5940 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5941 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5942 of these calls have also been analysed too and it is believed there are no
5943 instances in internal usage where an overflow could occur.
5945 This issue was reported by Guido Vranken.
5950 * Prevent ASN.1 BIO excessive memory allocation
5952 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5953 a short invalid encoding can casuse allocation of large amounts of memory
5954 potentially consuming excessive resources or exhausting memory.
5956 Any application parsing untrusted data through d2i BIO functions is
5957 affected. The memory based functions such as d2i_X509() are *not* affected.
5958 Since the memory based functions are used by the TLS library, TLS
5959 applications are not affected.
5961 This issue was reported by Brian Carpenter.
5968 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5969 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5970 in arbitrary stack data being returned in the buffer.
5972 This issue was reported by Guido Vranken.
5977 * Modify behavior of ALPN to invoke callback after SNI/servername
5978 callback, such that updates to the SSL_CTX affect ALPN.
5982 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5987 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5988 methods are enabled and ssl2 is disabled the methods return NULL.
5992 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5994 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5995 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5996 provide any "EXPORT" or "LOW" strength ciphers.
6000 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6001 is by default disabled at build-time. Builds that are not configured with
6002 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6003 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6004 will need to explicitly call either of:
6006 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6008 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6010 as appropriate. Even if either of those is used, or the application
6011 explicitly uses the version-specific SSLv2_method() or its client and
6012 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6013 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6014 ciphers, and SSLv2 56-bit DES are no longer available.
6019 * Fix a double-free in DSA code
6021 A double free bug was discovered when OpenSSL parses malformed DSA private
6022 keys and could lead to a DoS attack or memory corruption for applications
6023 that receive DSA private keys from untrusted sources. This scenario is
6026 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6032 * Disable SRP fake user seed to address a server memory leak.
6034 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6036 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6037 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6038 was changed to ignore the "fake user" SRP seed, even if the seed
6041 Users should use SRP_VBASE_get1_by_user instead. Note that in
6042 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6043 also that even though configuring the SRP seed attempts to hide
6044 invalid usernames by continuing the handshake with fake
6045 credentials, this behaviour is not constant time and no strong
6046 guarantees are made that the handshake is indistinguishable from
6047 that of a valid user.
6052 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6054 In the BN_hex2bn function the number of hex digits is calculated using an
6055 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6056 large values of `i` this can result in `bn_expand` not allocating any
6057 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6058 field as NULL leading to a subsequent NULL ptr deref. For very large values
6059 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6060 In this case memory is allocated to the internal BIGNUM data field, but it
6061 is insufficiently sized leading to heap corruption. A similar issue exists
6062 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6063 is ever called by user applications with very large untrusted hex/dec data.
6064 This is anticipated to be a rare occurrence.
6066 All OpenSSL internal usage of these functions use data that is not expected
6067 to be untrusted, e.g. config file data or application command line
6068 arguments. If user developed applications generate config file data based
6069 on untrusted data then it is possible that this could also lead to security
6070 consequences. This is also anticipated to be rare.
6072 This issue was reported to OpenSSL by Guido Vranken.
6077 * Fix memory issues in `BIO_*printf` functions
6079 The internal `fmtstr` function used in processing a "%s" format string in
6080 the `BIO_*printf` functions could overflow while calculating the length of a
6081 string and cause an OOB read when printing very long strings.
6083 Additionally the internal `doapr_outch` function can attempt to write to an
6084 OOB memory location (at an offset from the NULL pointer) in the event of a
6085 memory allocation failure. In 1.0.2 and below this could be caused where
6086 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6087 could be in processing a very long "%s" format string. Memory leaks can
6090 The first issue may mask the second issue dependent on compiler behaviour.
6091 These problems could enable attacks where large amounts of untrusted data
6092 is passed to the `BIO_*printf` functions. If applications use these functions
6093 in this way then they could be vulnerable. OpenSSL itself uses these
6094 functions when printing out human-readable dumps of ASN.1 data. Therefore
6095 applications that print this data could be vulnerable if the data is from
6096 untrusted sources. OpenSSL command line applications could also be
6097 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6098 as command line arguments.
6100 Libssl is not considered directly vulnerable. Additionally certificates etc
6101 received via remote connections via libssl are also unlikely to be able to
6102 trigger these issues because of message size limits enforced within libssl.
6104 This issue was reported to OpenSSL Guido Vranken.
6109 * Side channel attack on modular exponentiation
6111 A side-channel attack was found which makes use of cache-bank conflicts on
6112 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6113 of RSA keys. The ability to exploit this issue is limited as it relies on
6114 an attacker who has control of code in a thread running on the same
6115 hyper-threaded core as the victim thread which is performing decryptions.
6117 This issue was reported to OpenSSL by Yuval Yarom, The University of
6118 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6119 Nadia Heninger, University of Pennsylvania with more information at
6120 <http://cachebleed.info>.
6125 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6126 if no keysize is specified with default_bits. This fixes an
6127 omission in an earlier change that changed all RSA/DSA key generation
6128 commands to use 2048 bits by default.
6132 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6134 * Protection for DH small subgroup attacks
6136 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6137 switched on by default and cannot be disabled. This could have some
6142 * SSLv2 doesn't block disabled ciphers
6144 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6145 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6146 been disabled, provided that the SSLv2 protocol was not also disabled via
6149 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6150 and Sebastian Schinzel.
6155 * Reject DH handshakes with parameters shorter than 1024 bits.
6159 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6161 * Certificate verify crash with missing PSS parameter
6163 The signature verification routines will crash with a NULL pointer
6164 dereference if presented with an ASN.1 signature using the RSA PSS
6165 algorithm and absent mask generation function parameter. Since these
6166 routines are used to verify certificate signature algorithms this can be
6167 used to crash any certificate verification operation and exploited in a
6168 DoS attack. Any application which performs certificate verification is
6169 vulnerable including OpenSSL clients and servers which enable client
6172 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6177 * X509_ATTRIBUTE memory leak
6179 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6180 memory. This structure is used by the PKCS#7 and CMS routines so any
6181 application which reads PKCS#7 or CMS data from untrusted sources is
6182 affected. SSL/TLS is not affected.
6184 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6190 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6191 This changes the decoding behaviour for some invalid messages,
6192 though the change is mostly in the more lenient direction, and
6193 legacy behaviour is preserved as much as possible.
6197 * In DSA_generate_parameters_ex, if the provided seed is too short,
6198 use a random seed, as already documented.
6200 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6202 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6204 * Alternate chains certificate forgery
6206 During certificate verfification, OpenSSL will attempt to find an
6207 alternative certificate chain if the first attempt to build such a chain
6208 fails. An error in the implementation of this logic can mean that an
6209 attacker could cause certain checks on untrusted certificates to be
6210 bypassed, such as the CA flag, enabling them to use a valid leaf
6211 certificate to act as a CA and "issue" an invalid certificate.
6213 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6219 * Race condition handling PSK identify hint
6221 If PSK identity hints are received by a multi-threaded client then
6222 the values are wrongly updated in the parent SSL_CTX structure. This can
6223 result in a race condition potentially leading to a double free of the
6229 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6231 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6232 incompatibility in the handling of HMAC. The previous ABI has now been
6235 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6237 * Malformed ECParameters causes infinite loop
6239 When processing an ECParameters structure OpenSSL enters an infinite loop
6240 if the curve specified is over a specially malformed binary polynomial
6243 This can be used to perform denial of service against any
6244 system which processes public keys, certificate requests or
6245 certificates. This includes TLS clients and TLS servers with
6246 client authentication enabled.
6248 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6253 * Exploitable out-of-bounds read in X509_cmp_time
6255 X509_cmp_time does not properly check the length of the ASN1_TIME
6256 string and can read a few bytes out of bounds. In addition,
6257 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6260 An attacker can use this to craft malformed certificates and CRLs of
6261 various sizes and potentially cause a segmentation fault, resulting in
6262 a DoS on applications that verify certificates or CRLs. TLS clients
6263 that verify CRLs are affected. TLS clients and servers with client
6264 authentication enabled may be affected if they use custom verification
6267 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6268 independently by Hanno Böck.
6273 * PKCS7 crash with missing EnvelopedContent
6275 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6276 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6277 with missing content and trigger a NULL pointer dereference on parsing.
6279 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6280 structures from untrusted sources are affected. OpenSSL clients and
6281 servers are not affected.
6283 This issue was reported to OpenSSL by Michal Zalewski (Google).
6288 * CMS verify infinite loop with unknown hash function
6290 When verifying a signedData message the CMS code can enter an infinite loop
6291 if presented with an unknown hash function OID. This can be used to perform
6292 denial of service against any system which verifies signedData messages using
6294 This issue was reported to OpenSSL by Johannes Bauer.
6299 * Race condition handling NewSessionTicket
6301 If a NewSessionTicket is received by a multi-threaded client when attempting to
6302 reuse a previous ticket then a race condition can occur potentially leading to
6303 a double free of the ticket data.
6308 * Reject DH handshakes with parameters shorter than 768 bits.
6310 *Kurt Roeckx and Emilia Kasper*
6312 * dhparam: generate 2048-bit parameters by default.
6314 *Kurt Roeckx and Emilia Kasper*
6316 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6318 * Segmentation fault in ASN1_TYPE_cmp fix
6320 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6321 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6322 certificate signature algorithm consistency this can be used to crash any
6323 certificate verification operation and exploited in a DoS attack. Any
6324 application which performs certificate verification is vulnerable including
6325 OpenSSL clients and servers which enable client authentication.
6330 * ASN.1 structure reuse memory corruption fix
6332 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6333 memory corruption via an invalid write. Such reuse is and has been
6334 strongly discouraged and is believed to be rare.
6336 Applications that parse structures containing CHOICE or ANY DEFINED BY
6337 components may be affected. Certificate parsing (d2i_X509 and related
6338 functions) are however not affected. OpenSSL clients and servers are
6344 * PKCS7 NULL pointer dereferences fix
6346 The PKCS#7 parsing code does not handle missing outer ContentInfo
6347 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6348 missing content and trigger a NULL pointer dereference on parsing.
6350 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6351 otherwise parse PKCS#7 structures from untrusted sources are
6352 affected. OpenSSL clients and servers are not affected.
6354 This issue was reported to OpenSSL by Michal Zalewski (Google).
6359 * DoS via reachable assert in SSLv2 servers fix
6361 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6362 servers that both support SSLv2 and enable export cipher suites by sending
6363 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6365 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6366 (OpenSSL development team).
6371 * Use After Free following d2i_ECPrivatekey error fix
6373 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6374 could cause a use after free condition. This, in turn, could cause a double
6375 free in several private key parsing functions (such as d2i_PrivateKey
6376 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6377 for applications that receive EC private keys from untrusted
6378 sources. This scenario is considered rare.
6380 This issue was discovered by the BoringSSL project and fixed in their
6386 * X509_to_X509_REQ NULL pointer deref fix
6388 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6389 the certificate key is invalid. This function is rarely used in practice.
6391 This issue was discovered by Brian Carpenter.
6396 * Removed the export ciphers from the DEFAULT ciphers
6400 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6402 * Build fixes for the Windows and OpenVMS platforms
6404 *Matt Caswell and Richard Levitte*
6406 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6408 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6409 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6410 dereference. This could lead to a Denial Of Service attack. Thanks to
6411 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6416 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6417 dtls1_buffer_record function under certain conditions. In particular this
6418 could occur if an attacker sent repeated DTLS records with the same
6419 sequence number but for the next epoch. The memory leak could be exploited
6420 by an attacker in a Denial of Service attack through memory exhaustion.
6421 Thanks to Chris Mueller for reporting this issue.
6426 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6427 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6428 method would be set to NULL which could later result in a NULL pointer
6429 dereference. Thanks to Frank Schmirler for reporting this issue.
6434 * Abort handshake if server key exchange message is omitted for ephemeral
6437 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6438 reporting this issue.
6443 * Remove non-export ephemeral RSA code on client and server. This code
6444 violated the TLS standard by allowing the use of temporary RSA keys in
6445 non-export ciphersuites and could be used by a server to effectively
6446 downgrade the RSA key length used to a value smaller than the server
6447 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6448 INRIA or reporting this issue.
6453 * Fixed issue where DH client certificates are accepted without verification.
6454 An OpenSSL server will accept a DH certificate for client authentication
6455 without the certificate verify message. This effectively allows a client to
6456 authenticate without the use of a private key. This only affects servers
6457 which trust a client certificate authority which issues certificates
6458 containing DH keys: these are extremely rare and hardly ever encountered.
6459 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6465 * Ensure that the session ID context of an SSL is updated when its
6466 SSL_CTX is updated via SSL_set_SSL_CTX.
6468 The session ID context is typically set from the parent SSL_CTX,
6469 and can vary with the CTX.
6473 * Fix various certificate fingerprint issues.
6475 By using non-DER or invalid encodings outside the signed portion of a
6476 certificate the fingerprint can be changed without breaking the signature.
6477 Although no details of the signed portion of the certificate can be changed
6478 this can cause problems with some applications: e.g. those using the
6479 certificate fingerprint for blacklists.
6481 1. Reject signatures with non zero unused bits.
6483 If the BIT STRING containing the signature has non zero unused bits reject
6484 the signature. All current signature algorithms require zero unused bits.
6486 2. Check certificate algorithm consistency.
6488 Check the AlgorithmIdentifier inside TBS matches the one in the
6489 certificate signature. NB: this will result in signature failure
6490 errors for some broken certificates.
6492 Thanks to Konrad Kraszewski from Google for reporting this issue.
6494 3. Check DSA/ECDSA signatures use DER.
6496 Re-encode DSA/ECDSA signatures and compare with the original received
6497 signature. Return an error if there is a mismatch.
6499 This will reject various cases including garbage after signature
6500 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6501 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6502 (negative or with leading zeroes).
6504 Further analysis was conducted and fixes were developed by Stephen Henson
6505 of the OpenSSL core team.
6511 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6512 results on some platforms, including x86_64. This bug occurs at random
6513 with a very low probability, and is not known to be exploitable in any
6514 way, though its exact impact is difficult to determine. Thanks to Pieter
6515 Wuille (Blockstream) who reported this issue and also suggested an initial
6516 fix. Further analysis was conducted by the OpenSSL development team and
6517 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6518 the OpenSSL core team.
6523 * Do not resume sessions on the server if the negotiated protocol
6524 version does not match the session's version. Resuming with a different
6525 version, while not strictly forbidden by the RFC, is of questionable
6526 sanity and breaks all known clients.
6528 *David Benjamin, Emilia Käsper*
6530 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6531 early CCS messages during renegotiation. (Note that because
6532 renegotiation is encrypted, this early CCS was not exploitable.)
6536 * Tighten client-side session ticket handling during renegotiation:
6537 ensure that the client only accepts a session ticket if the server sends
6538 the extension anew in the ServerHello. Previously, a TLS client would
6539 reuse the old extension state and thus accept a session ticket if one was
6540 announced in the initial ServerHello.
6542 Similarly, ensure that the client requires a session ticket if one
6543 was advertised in the ServerHello. Previously, a TLS client would
6544 ignore a missing NewSessionTicket message.
6548 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6552 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6553 sends a carefully crafted handshake message, to cause OpenSSL to fail
6554 to free up to 64k of memory causing a memory leak. This could be
6555 exploited in a Denial Of Service attack. This issue affects OpenSSL
6556 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6557 whether SRTP is used or configured. Implementations of OpenSSL that
6558 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6560 The fix was developed by the OpenSSL team.
6565 * Session Ticket Memory Leak.
6567 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6568 integrity of that ticket is first verified. In the event of a session
6569 ticket integrity check failing, OpenSSL will fail to free memory
6570 causing a memory leak. By sending a large number of invalid session
6571 tickets an attacker could exploit this issue in a Denial Of Service
6577 * Build option no-ssl3 is incomplete.
6579 When OpenSSL is configured with "no-ssl3" as a build option, servers
6580 could accept and complete a SSL 3.0 handshake, and clients could be
6581 configured to send them.
6584 *Akamai and the OpenSSL team*
6586 * Add support for TLS_FALLBACK_SCSV.
6587 Client applications doing fallback retries should call
6588 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6591 *Adam Langley, Bodo Moeller*
6593 * Add additional DigestInfo checks.
6595 Re-encode DigestInto in DER and check against the original when
6596 verifying RSA signature: this will reject any improperly encoded
6597 DigestInfo structures.
6599 Note: this is a precautionary measure and no attacks are currently known.
6603 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6605 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6606 SRP code can be overrun an internal buffer. Add sanity check that
6607 g, A, B < N to SRP code.
6609 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6610 Group for discovering this issue.
6615 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6616 TLS 1.0 instead of higher protocol versions when the ClientHello message
6617 is badly fragmented. This allows a man-in-the-middle attacker to force a
6618 downgrade to TLS 1.0 even if both the server and the client support a
6619 higher protocol version, by modifying the client's TLS records.
6621 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6622 researching this issue.
6627 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6628 to a denial of service attack. A malicious server can crash the client
6629 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6630 ciphersuite and sending carefully crafted handshake messages.
6632 Thanks to Felix Gröbert (Google) for discovering and researching this
6638 * By sending carefully crafted DTLS packets an attacker could cause openssl
6639 to leak memory. This can be exploited through a Denial of Service attack.
6640 Thanks to Adam Langley for discovering and researching this issue.
6645 * An attacker can force openssl to consume large amounts of memory whilst
6646 processing DTLS handshake messages. This can be exploited through a
6647 Denial of Service attack.
6648 Thanks to Adam Langley for discovering and researching this issue.
6653 * An attacker can force an error condition which causes openssl to crash
6654 whilst processing DTLS packets due to memory being freed twice. This
6655 can be exploited through a Denial of Service attack.
6656 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6662 * If a multithreaded client connects to a malicious server using a resumed
6663 session and the server sends an ec point format extension it could write
6664 up to 255 bytes to freed memory.
6666 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6672 * A malicious server can crash an OpenSSL client with a null pointer
6673 dereference (read) by specifying an SRP ciphersuite even though it was not
6674 properly negotiated with the client. This can be exploited through a
6675 Denial of Service attack.
6677 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6678 discovering and researching this issue.
6683 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6684 X509_name_oneline, X509_name_print_ex et al. to leak some information
6685 from the stack. Applications may be affected if they echo pretty printing
6686 output to the attacker.
6688 Thanks to Ivan Fratric (Google) for discovering this issue.
6691 *Emilia Käsper, and Steve Henson*
6693 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6694 for corner cases. (Certain input points at infinity could lead to
6695 bogus results, with non-infinity inputs mapped to infinity too.)
6699 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6701 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6702 handshake can force the use of weak keying material in OpenSSL
6703 SSL/TLS clients and servers.
6705 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6706 researching this issue. ([CVE-2014-0224])
6708 *KIKUCHI Masashi, Steve Henson*
6710 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6711 OpenSSL DTLS client the code can be made to recurse eventually crashing
6714 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6717 *Imre Rad, Steve Henson*
6719 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6720 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6721 client or server. This is potentially exploitable to run arbitrary
6722 code on a vulnerable client or server.
6724 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6726 *Jüri Aedla, Steve Henson*
6728 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6729 are subject to a denial of service attack.
6731 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6732 this issue. ([CVE-2014-3470])
6734 *Felix Gröbert, Ivan Fratric, Steve Henson*
6736 * Harmonize version and its documentation. -f flag is used to display
6739 *mancha <mancha1@zoho.com>*
6741 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6742 in i2d_ECPrivateKey.
6744 *mancha <mancha1@zoho.com>*
6746 * Fix some double frees. These are not thought to be exploitable.
6748 *mancha <mancha1@zoho.com>*
6750 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6752 * A missing bounds check in the handling of the TLS heartbeat extension
6753 can be used to reveal up to 64k of memory to a connected client or
6756 Thanks for Neel Mehta of Google Security for discovering this bug and to
6757 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6758 preparing the fix ([CVE-2014-0160])
6760 *Adam Langley, Bodo Moeller*
6762 * Fix for the attack described in the paper "Recovering OpenSSL
6763 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6764 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6765 <http://eprint.iacr.org/2014/140>
6767 Thanks to Yuval Yarom and Naomi Benger for discovering this
6768 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6770 *Yuval Yarom and Naomi Benger*
6772 * TLS pad extension: draft-agl-tls-padding-03
6774 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6775 TLS client Hello record length value would otherwise be > 255 and
6776 less that 512 pad with a dummy extension containing zeroes so it
6777 is at least 512 bytes long.
6779 *Adam Langley, Steve Henson*
6781 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6783 * Fix for TLS record tampering bug. A carefully crafted invalid
6784 handshake could crash OpenSSL with a NULL pointer exception.
6785 Thanks to Anton Johansson for reporting this issues.
6788 * Keep original DTLS digest and encryption contexts in retransmission
6789 structures so we can use the previous session parameters if they need
6790 to be resent. ([CVE-2013-6450])
6794 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6795 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6796 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6797 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6798 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6799 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6801 *Rob Stradling, Adam Langley*
6803 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6805 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6806 supporting platforms or when small records were transferred.
6808 *Andy Polyakov, Steve Henson*
6810 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6812 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6814 This addresses the flaw in CBC record processing discovered by
6815 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6816 at: <http://www.isg.rhul.ac.uk/tls/>
6818 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6819 Security Group at Royal Holloway, University of London
6820 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6821 Emilia Käsper for the initial patch.
6824 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6826 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6827 ciphersuites which can be exploited in a denial of service attack.
6828 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6829 and detecting this bug and to Wolfgang Ettlinger
6830 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6835 * Return an error when checking OCSP signatures when key is NULL.
6836 This fixes a DoS attack. ([CVE-2013-0166])
6840 * Make openssl verify return errors.
6842 *Chris Palmer <palmer@google.com> and Ben Laurie*
6844 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6845 the right response is stapled. Also change SSL_get_certificate()
6846 so it returns the certificate actually sent.
6847 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6849 *Rob Stradling <rob.stradling@comodo.com>*
6851 * Fix possible deadlock when decoding public keys.
6855 * Don't use TLS 1.0 record version number in initial client hello
6860 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6862 * Sanity check record length before skipping explicit IV in TLS
6863 1.2, 1.1 and DTLS to fix DoS attack.
6865 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6866 fuzzing as a service testing platform.
6871 * Initialise tkeylen properly when encrypting CMS messages.
6872 Thanks to Solar Designer of Openwall for reporting this issue.
6876 * In FIPS mode don't try to use composite ciphers as they are not
6881 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6883 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6884 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6885 mean any application compiled against OpenSSL 1.0.0 headers setting
6886 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6887 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6888 0x10000000L Any application which was previously compiled against
6889 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6890 will need to be recompiled as a result. Letting be results in
6891 inability to disable specifically TLS 1.1 and in client context,
6892 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6896 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6897 disable just protocol X, but all protocols above X *if* there are
6898 protocols *below* X still enabled. In more practical terms it means
6899 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6900 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6901 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6906 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6908 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6909 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6910 in CRYPTO_realloc_clean.
6912 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6913 issue and to Adam Langley <agl@chromium.org> for fixing it.
6916 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6918 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6922 * Workarounds for some broken servers that "hang" if a client hello
6923 record length exceeds 255 bytes.
6925 1. Do not use record version number > TLS 1.0 in initial client
6926 hello: some (but not all) hanging servers will now work.
6927 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6928 the number of ciphers sent in the client hello. This should be
6929 set to an even number, such as 50, for example by passing:
6930 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6931 Most broken servers should now work.
6932 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6933 TLS 1.2 client support entirely.
6937 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6941 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6943 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6944 STRING form instead of a DigestInfo.
6948 * The format used for MDC2 RSA signatures is inconsistent between EVP
6949 and the RSA_sign/RSA_verify functions. This was made more apparent when
6950 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6951 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6952 the correct format in RSA_verify so both forms transparently work.
6956 * Some servers which support TLS 1.0 can choke if we initially indicate
6957 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6958 encrypted premaster secret. As a workaround use the maximum permitted
6959 client version in client hello, this should keep such servers happy
6960 and still work with previous versions of OpenSSL.
6964 * Add support for TLS/DTLS heartbeats.
6966 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6968 * Add support for SCTP.
6970 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6972 * Improved PRNG seeding for VOS.
6974 *Paul Green <Paul.Green@stratus.com>*
6976 * Extensive assembler packs updates, most notably:
6978 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6979 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6980 - x86_64: bit-sliced AES implementation;
6981 - ARM: NEON support, contemporary platforms optimizations;
6982 - s390x: z196 support;
6983 - `*`: GHASH and GF(2^m) multiplication implementations;
6987 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6988 (removal of unnecessary code)
6990 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6992 * Add TLS key material exporter from RFC 5705.
6996 * Add DTLS-SRTP negotiation from RFC 5764.
7000 * Add Next Protocol Negotiation,
7001 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7002 disabled with a no-npn flag to config or Configure. Code donated
7005 *Adam Langley <agl@google.com> and Ben Laurie*
7007 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7008 NIST-P256, NIST-P521, with constant-time single point multiplication on
7009 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7010 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7011 Code made available under Apache License version 2.0.
7013 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7014 line to include this in your build of OpenSSL, and run "make depend" (or
7015 "make update"). This enables the following EC_METHODs:
7017 EC_GFp_nistp224_method()
7018 EC_GFp_nistp256_method()
7019 EC_GFp_nistp521_method()
7021 EC_GROUP_new_by_curve_name() will automatically use these (while
7022 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7025 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7027 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7028 all platforms. Move ssize_t definition from e_os.h to the public
7029 header file e_os2.h as it now appears in public header file cms.h
7033 * New -sigopt option to the ca, req and x509 utilities. Additional
7034 signature parameters can be passed using this option and in
7039 * Add RSA PSS signing function. This will generate and set the
7040 appropriate AlgorithmIdentifiers for PSS based on those in the
7041 corresponding EVP_MD_CTX structure. No application support yet.
7045 * Support for companion algorithm specific ASN1 signing routines.
7046 New function ASN1_item_sign_ctx() signs a pre-initialised
7047 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7048 the appropriate parameters.
7052 * Add new algorithm specific ASN1 verification initialisation function
7053 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7054 handling will be the same no matter what EVP_PKEY_METHOD is used.
7055 Add a PSS handler to support verification of PSS signatures: checked
7056 against a number of sample certificates.
7060 * Add signature printing for PSS. Add PSS OIDs.
7062 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7064 * Add algorithm specific signature printing. An individual ASN1 method
7065 can now print out signatures instead of the standard hex dump.
7067 More complex signatures (e.g. PSS) can print out more meaningful
7068 information. Include DSA version that prints out the signature
7073 * Password based recipient info support for CMS library: implementing
7078 * Split password based encryption into PBES2 and PBKDF2 functions. This
7079 neatly separates the code into cipher and PBE sections and is required
7080 for some algorithms that split PBES2 into separate pieces (such as
7081 password based CMS).
7085 * Session-handling fixes:
7086 - Fix handling of connections that are resuming with a session ID,
7087 but also support Session Tickets.
7088 - Fix a bug that suppressed issuing of a new ticket if the client
7089 presented a ticket with an expired session.
7090 - Try to set the ticket lifetime hint to something reasonable.
7091 - Make tickets shorter by excluding irrelevant information.
7092 - On the client side, don't ignore renewed tickets.
7094 *Adam Langley, Bodo Moeller (Google)*
7096 * Fix PSK session representation.
7100 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7102 This work was sponsored by Intel.
7106 * Add GCM support to TLS library. Some custom code is needed to split
7107 the IV between the fixed (from PRF) and explicit (from TLS record)
7108 portions. This adds all GCM ciphersuites supported by RFC5288 and
7109 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7110 add a special AESGCM string for GCM only.
7114 * Expand range of ctrls for AES GCM. Permit setting invocation
7115 field on decrypt and retrieval of invocation field only on encrypt.
7119 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7120 As required by RFC5289 these ciphersuites cannot be used if for
7121 versions of TLS earlier than 1.2.
7125 * For FIPS capable OpenSSL interpret a NULL default public key method
7126 as unset and return the appropriate default but do *not* set the default.
7127 This means we can return the appropriate method in applications that
7128 switch between FIPS and non-FIPS modes.
7132 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7133 ENGINE is used then we cannot handle that in the FIPS module so we
7134 keep original code iff non-FIPS operations are allowed.
7138 * Add -attime option to openssl utilities.
7140 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7142 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7146 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7147 FIPS EC methods unconditionally for now.
7151 * New build option no-ec2m to disable characteristic 2 code.
7155 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7156 all cases can be covered as some introduce binary incompatibilities.
7160 * Redirect RSA operations to FIPS module including keygen,
7161 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7165 * Add similar low level API blocking to ciphers.
7169 * Low level digest APIs are not approved in FIPS mode: any attempt
7170 to use these will cause a fatal error. Applications that *really* want
7171 to use them can use the `private_*` version instead.
7175 * Redirect cipher operations to FIPS module for FIPS builds.
7179 * Redirect digest operations to FIPS module for FIPS builds.
7183 * Update build system to add "fips" flag which will link in fipscanister.o
7184 for static and shared library builds embedding a signature if needed.
7188 * Output TLS supported curves in preference order instead of numerical
7189 order. This is currently hardcoded for the highest order curves first.
7190 This should be configurable so applications can judge speed vs strength.
7194 * Add TLS v1.2 server support for client authentication.
7198 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7203 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7204 FIPS modules versions.
7208 * Add TLS v1.2 client side support for client authentication. Keep cache
7209 of handshake records longer as we don't know the hash algorithm to use
7210 until after the certificate request message is received.
7214 * Initial TLS v1.2 client support. Add a default signature algorithms
7215 extension including all the algorithms we support. Parse new signature
7216 format in client key exchange. Relax some ECC signing restrictions for
7217 TLS v1.2 as indicated in RFC5246.
7221 * Add server support for TLS v1.2 signature algorithms extension. Switch
7222 to new signature format when needed using client digest preference.
7223 All server ciphersuites should now work correctly in TLS v1.2. No client
7224 support yet and no support for client certificates.
7228 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7229 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7230 ciphersuites. At present only RSA key exchange ciphersuites work with
7231 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7232 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7233 and version checking.
7237 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7238 with this defined it will not be affected by any changes to ssl internal
7239 structures. Add several utility functions to allow openssl application
7240 to work with OPENSSL_NO_SSL_INTERN defined.
7244 * A long standing patch to add support for SRP from EdelWeb (Peter
7245 Sylvester and Christophe Renou) was integrated.
7246 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7247 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7250 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7254 * Permit abbreviated handshakes when renegotiating using the function
7255 SSL_renegotiate_abbreviated().
7257 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7259 * Add call to ENGINE_register_all_complete() to
7260 ENGINE_load_builtin_engines(), so some implementations get used
7261 automatically instead of needing explicit application support.
7265 * Add support for TLS key exporter as described in RFC5705.
7267 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7269 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7270 a few changes are required:
7272 Add SSL_OP_NO_TLSv1_1 flag.
7273 Add TLSv1_1 methods.
7274 Update version checking logic to handle version 1.1.
7275 Add explicit IV handling (ported from DTLS code).
7276 Add command line options to s_client/s_server.
7283 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7285 * X509_ATTRIBUTE memory leak
7287 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7288 memory. This structure is used by the PKCS#7 and CMS routines so any
7289 application which reads PKCS#7 or CMS data from untrusted sources is
7290 affected. SSL/TLS is not affected.
7292 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7298 * Race condition handling PSK identify hint
7300 If PSK identity hints are received by a multi-threaded client then
7301 the values are wrongly updated in the parent SSL_CTX structure. This can
7302 result in a race condition potentially leading to a double free of the
7308 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7310 * Malformed ECParameters causes infinite loop
7312 When processing an ECParameters structure OpenSSL enters an infinite loop
7313 if the curve specified is over a specially malformed binary polynomial
7316 This can be used to perform denial of service against any
7317 system which processes public keys, certificate requests or
7318 certificates. This includes TLS clients and TLS servers with
7319 client authentication enabled.
7321 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7326 * Exploitable out-of-bounds read in X509_cmp_time
7328 X509_cmp_time does not properly check the length of the ASN1_TIME
7329 string and can read a few bytes out of bounds. In addition,
7330 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7333 An attacker can use this to craft malformed certificates and CRLs of
7334 various sizes and potentially cause a segmentation fault, resulting in
7335 a DoS on applications that verify certificates or CRLs. TLS clients
7336 that verify CRLs are affected. TLS clients and servers with client
7337 authentication enabled may be affected if they use custom verification
7340 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7341 independently by Hanno Böck.
7346 * PKCS7 crash with missing EnvelopedContent
7348 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7349 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7350 with missing content and trigger a NULL pointer dereference on parsing.
7352 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7353 structures from untrusted sources are affected. OpenSSL clients and
7354 servers are not affected.
7356 This issue was reported to OpenSSL by Michal Zalewski (Google).
7361 * CMS verify infinite loop with unknown hash function
7363 When verifying a signedData message the CMS code can enter an infinite loop
7364 if presented with an unknown hash function OID. This can be used to perform
7365 denial of service against any system which verifies signedData messages using
7367 This issue was reported to OpenSSL by Johannes Bauer.
7372 * Race condition handling NewSessionTicket
7374 If a NewSessionTicket is received by a multi-threaded client when attempting to
7375 reuse a previous ticket then a race condition can occur potentially leading to
7376 a double free of the ticket data.
7381 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7383 * Segmentation fault in ASN1_TYPE_cmp fix
7385 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7386 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7387 certificate signature algorithm consistency this can be used to crash any
7388 certificate verification operation and exploited in a DoS attack. Any
7389 application which performs certificate verification is vulnerable including
7390 OpenSSL clients and servers which enable client authentication.
7395 * ASN.1 structure reuse memory corruption fix
7397 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7398 memory corruption via an invalid write. Such reuse is and has been
7399 strongly discouraged and is believed to be rare.
7401 Applications that parse structures containing CHOICE or ANY DEFINED BY
7402 components may be affected. Certificate parsing (d2i_X509 and related
7403 functions) are however not affected. OpenSSL clients and servers are
7409 * PKCS7 NULL pointer dereferences fix
7411 The PKCS#7 parsing code does not handle missing outer ContentInfo
7412 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7413 missing content and trigger a NULL pointer dereference on parsing.
7415 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7416 otherwise parse PKCS#7 structures from untrusted sources are
7417 affected. OpenSSL clients and servers are not affected.
7419 This issue was reported to OpenSSL by Michal Zalewski (Google).
7424 * DoS via reachable assert in SSLv2 servers fix
7426 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7427 servers that both support SSLv2 and enable export cipher suites by sending
7428 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7430 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7431 (OpenSSL development team).
7436 * Use After Free following d2i_ECPrivatekey error fix
7438 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7439 could cause a use after free condition. This, in turn, could cause a double
7440 free in several private key parsing functions (such as d2i_PrivateKey
7441 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7442 for applications that receive EC private keys from untrusted
7443 sources. This scenario is considered rare.
7445 This issue was discovered by the BoringSSL project and fixed in their
7451 * X509_to_X509_REQ NULL pointer deref fix
7453 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7454 the certificate key is invalid. This function is rarely used in practice.
7456 This issue was discovered by Brian Carpenter.
7461 * Removed the export ciphers from the DEFAULT ciphers
7465 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7467 * Build fixes for the Windows and OpenVMS platforms
7469 *Matt Caswell and Richard Levitte*
7471 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7473 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7474 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7475 dereference. This could lead to a Denial Of Service attack. Thanks to
7476 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7481 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7482 dtls1_buffer_record function under certain conditions. In particular this
7483 could occur if an attacker sent repeated DTLS records with the same
7484 sequence number but for the next epoch. The memory leak could be exploited
7485 by an attacker in a Denial of Service attack through memory exhaustion.
7486 Thanks to Chris Mueller for reporting this issue.
7491 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7492 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7493 method would be set to NULL which could later result in a NULL pointer
7494 dereference. Thanks to Frank Schmirler for reporting this issue.
7499 * Abort handshake if server key exchange message is omitted for ephemeral
7502 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7503 reporting this issue.
7508 * Remove non-export ephemeral RSA code on client and server. This code
7509 violated the TLS standard by allowing the use of temporary RSA keys in
7510 non-export ciphersuites and could be used by a server to effectively
7511 downgrade the RSA key length used to a value smaller than the server
7512 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7513 INRIA or reporting this issue.
7518 * Fixed issue where DH client certificates are accepted without verification.
7519 An OpenSSL server will accept a DH certificate for client authentication
7520 without the certificate verify message. This effectively allows a client to
7521 authenticate without the use of a private key. This only affects servers
7522 which trust a client certificate authority which issues certificates
7523 containing DH keys: these are extremely rare and hardly ever encountered.
7524 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7530 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7531 results on some platforms, including x86_64. This bug occurs at random
7532 with a very low probability, and is not known to be exploitable in any
7533 way, though its exact impact is difficult to determine. Thanks to Pieter
7534 Wuille (Blockstream) who reported this issue and also suggested an initial
7535 fix. Further analysis was conducted by the OpenSSL development team and
7536 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7537 the OpenSSL core team.
7542 * Fix various certificate fingerprint issues.
7544 By using non-DER or invalid encodings outside the signed portion of a
7545 certificate the fingerprint can be changed without breaking the signature.
7546 Although no details of the signed portion of the certificate can be changed
7547 this can cause problems with some applications: e.g. those using the
7548 certificate fingerprint for blacklists.
7550 1. Reject signatures with non zero unused bits.
7552 If the BIT STRING containing the signature has non zero unused bits reject
7553 the signature. All current signature algorithms require zero unused bits.
7555 2. Check certificate algorithm consistency.
7557 Check the AlgorithmIdentifier inside TBS matches the one in the
7558 certificate signature. NB: this will result in signature failure
7559 errors for some broken certificates.
7561 Thanks to Konrad Kraszewski from Google for reporting this issue.
7563 3. Check DSA/ECDSA signatures use DER.
7565 Reencode DSA/ECDSA signatures and compare with the original received
7566 signature. Return an error if there is a mismatch.
7568 This will reject various cases including garbage after signature
7569 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7570 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7571 (negative or with leading zeroes).
7573 Further analysis was conducted and fixes were developed by Stephen Henson
7574 of the OpenSSL core team.
7580 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7582 * Session Ticket Memory Leak.
7584 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7585 integrity of that ticket is first verified. In the event of a session
7586 ticket integrity check failing, OpenSSL will fail to free memory
7587 causing a memory leak. By sending a large number of invalid session
7588 tickets an attacker could exploit this issue in a Denial Of Service
7594 * Build option no-ssl3 is incomplete.
7596 When OpenSSL is configured with "no-ssl3" as a build option, servers
7597 could accept and complete a SSL 3.0 handshake, and clients could be
7598 configured to send them.
7601 *Akamai and the OpenSSL team*
7603 * Add support for TLS_FALLBACK_SCSV.
7604 Client applications doing fallback retries should call
7605 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7608 *Adam Langley, Bodo Moeller*
7610 * Add additional DigestInfo checks.
7612 Reencode DigestInto in DER and check against the original when
7613 verifying RSA signature: this will reject any improperly encoded
7614 DigestInfo structures.
7616 Note: this is a precautionary measure and no attacks are currently known.
7620 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7622 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7623 to a denial of service attack. A malicious server can crash the client
7624 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7625 ciphersuite and sending carefully crafted handshake messages.
7627 Thanks to Felix Gröbert (Google) for discovering and researching this
7633 * By sending carefully crafted DTLS packets an attacker could cause openssl
7634 to leak memory. This can be exploited through a Denial of Service attack.
7635 Thanks to Adam Langley for discovering and researching this issue.
7640 * An attacker can force openssl to consume large amounts of memory whilst
7641 processing DTLS handshake messages. This can be exploited through a
7642 Denial of Service attack.
7643 Thanks to Adam Langley for discovering and researching this issue.
7648 * An attacker can force an error condition which causes openssl to crash
7649 whilst processing DTLS packets due to memory being freed twice. This
7650 can be exploited through a Denial of Service attack.
7651 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7657 * If a multithreaded client connects to a malicious server using a resumed
7658 session and the server sends an ec point format extension it could write
7659 up to 255 bytes to freed memory.
7661 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7667 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7668 X509_name_oneline, X509_name_print_ex et al. to leak some information
7669 from the stack. Applications may be affected if they echo pretty printing
7670 output to the attacker.
7672 Thanks to Ivan Fratric (Google) for discovering this issue.
7675 *Emilia Käsper, and Steve Henson*
7677 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7678 for corner cases. (Certain input points at infinity could lead to
7679 bogus results, with non-infinity inputs mapped to infinity too.)
7683 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7685 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7686 handshake can force the use of weak keying material in OpenSSL
7687 SSL/TLS clients and servers.
7689 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7690 researching this issue. ([CVE-2014-0224])
7692 *KIKUCHI Masashi, Steve Henson*
7694 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7695 OpenSSL DTLS client the code can be made to recurse eventually crashing
7698 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7701 *Imre Rad, Steve Henson*
7703 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7704 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7705 client or server. This is potentially exploitable to run arbitrary
7706 code on a vulnerable client or server.
7708 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7710 *Jüri Aedla, Steve Henson*
7712 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7713 are subject to a denial of service attack.
7715 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7716 this issue. ([CVE-2014-3470])
7718 *Felix Gröbert, Ivan Fratric, Steve Henson*
7720 * Harmonize version and its documentation. -f flag is used to display
7723 *mancha <mancha1@zoho.com>*
7725 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7726 in i2d_ECPrivateKey.
7728 *mancha <mancha1@zoho.com>*
7730 * Fix some double frees. These are not thought to be exploitable.
7732 *mancha <mancha1@zoho.com>*
7734 * Fix for the attack described in the paper "Recovering OpenSSL
7735 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7736 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7737 <http://eprint.iacr.org/2014/140>
7739 Thanks to Yuval Yarom and Naomi Benger for discovering this
7740 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7742 *Yuval Yarom and Naomi Benger*
7744 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7746 * Keep original DTLS digest and encryption contexts in retransmission
7747 structures so we can use the previous session parameters if they need
7748 to be resent. ([CVE-2013-6450])
7752 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7753 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7754 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7755 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7756 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7757 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7759 *Rob Stradling, Adam Langley*
7761 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7763 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7765 This addresses the flaw in CBC record processing discovered by
7766 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7767 at: <http://www.isg.rhul.ac.uk/tls/>
7769 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7770 Security Group at Royal Holloway, University of London
7771 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7772 Emilia Käsper for the initial patch.
7775 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7777 * Return an error when checking OCSP signatures when key is NULL.
7778 This fixes a DoS attack. ([CVE-2013-0166])
7782 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7783 the right response is stapled. Also change SSL_get_certificate()
7784 so it returns the certificate actually sent.
7785 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7786 (This is a backport)
7788 *Rob Stradling <rob.stradling@comodo.com>*
7790 * Fix possible deadlock when decoding public keys.
7794 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7796 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7799 * Sanity check record length before skipping explicit IV in DTLS
7802 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7803 fuzzing as a service testing platform.
7808 * Initialise tkeylen properly when encrypting CMS messages.
7809 Thanks to Solar Designer of Openwall for reporting this issue.
7813 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7815 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7816 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7817 in CRYPTO_realloc_clean.
7819 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7820 issue and to Adam Langley <agl@chromium.org> for fixing it.
7823 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7825 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7827 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7828 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7829 content decryption and always return the same error. Note: this attack
7830 needs on average 2^20 messages so it only affects automated senders. The
7831 old behaviour can be re-enabled in the CMS code by setting the
7832 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7833 an MMA defence is not necessary.
7834 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7835 this issue. ([CVE-2012-0884])
7839 * Fix CVE-2011-4619: make sure we really are receiving a
7840 client hello before rejecting multiple SGC restarts. Thanks to
7841 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7845 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7847 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7848 Thanks to Antonio Martin, Enterprise Secure Access Research and
7849 Development, Cisco Systems, Inc. for discovering this bug and
7850 preparing a fix. ([CVE-2012-0050])
7854 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7856 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7857 of the Vaudenay padding oracle attack on CBC mode encryption
7858 which enables an efficient plaintext recovery attack against
7859 the OpenSSL implementation of DTLS. Their attack exploits timing
7860 differences arising during decryption processing. A research
7861 paper describing this attack can be found at:
7862 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7863 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7864 Security Group at Royal Holloway, University of London
7865 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7866 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7867 for preparing the fix. ([CVE-2011-4108])
7869 *Robin Seggelmann, Michael Tuexen*
7871 * Clear bytes used for block padding of SSL 3.0 records.
7874 *Adam Langley (Google)*
7876 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7877 Kadianakis <desnacked@gmail.com> for discovering this issue and
7878 Adam Langley for preparing the fix. ([CVE-2011-4619])
7880 *Adam Langley (Google)*
7882 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7884 *Andrey Kulikov <amdeich@gmail.com>*
7886 * Prevent malformed RFC3779 data triggering an assertion failure.
7887 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7888 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7890 *Rob Austein <sra@hactrn.net>*
7892 * Improved PRNG seeding for VOS.
7894 *Paul Green <Paul.Green@stratus.com>*
7896 * Fix ssl_ciph.c set-up race.
7898 *Adam Langley (Google)*
7900 * Fix spurious failures in ecdsatest.c.
7902 *Emilia Käsper (Google)*
7904 * Fix the BIO_f_buffer() implementation (which was mixing different
7905 interpretations of the `..._len` fields).
7907 *Adam Langley (Google)*
7909 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7910 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7911 threads won't reuse the same blinding coefficients.
7913 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7914 lock to call BN_BLINDING_invert_ex, and avoids one use of
7915 BN_BLINDING_update for each BN_BLINDING structure (previously,
7916 the last update always remained unused).
7918 *Emilia Käsper (Google)*
7920 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7922 *Bob Buckholz (Google)*
7924 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7926 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7927 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7929 *Kaspar Brand <ossl@velox.ch>*
7931 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7932 for multi-threaded use of ECDH. ([CVE-2011-3210])
7934 *Adam Langley (Google)*
7936 * Fix x509_name_ex_d2i memory leak on bad inputs.
7940 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7941 signature public key algorithm by using OID xref utilities instead.
7942 Before this you could only use some ECC ciphersuites with SHA1 only.
7946 * Add protection against ECDSA timing attacks as mentioned in the paper
7947 by Billy Bob Brumley and Nicola Tuveri, see:
7948 <http://eprint.iacr.org/2011/232.pdf>
7950 *Billy Bob Brumley and Nicola Tuveri*
7952 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7954 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7956 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7958 * Fix bug in string printing code: if *any* escaping is enabled we must
7959 escape the escape character (backslash) or the resulting string is
7964 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7966 * Disable code workaround for ancient and obsolete Netscape browsers
7967 and servers: an attacker can use it in a ciphersuite downgrade attack.
7968 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7972 * Fixed J-PAKE implementation error, originally discovered by
7973 Sebastien Martini, further info and confirmation from Stefan
7974 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7978 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7980 * Fix extension code to avoid race conditions which can result in a buffer
7981 overrun vulnerability: resumed sessions must not be modified as they can
7982 be shared by multiple threads. CVE-2010-3864
7986 * Fix WIN32 build system to correctly link an ENGINE directory into
7991 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7993 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7996 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7998 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8000 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8001 context. The operation can be customised via the ctrl mechanism in
8002 case ENGINEs want to include additional functionality.
8006 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8010 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8011 output hashes compatible with older versions of OpenSSL.
8013 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8015 * Fix compression algorithm handling: if resuming a session use the
8016 compression algorithm of the resumed session instead of determining
8017 it from client hello again. Don't allow server to change algorithm.
8021 * Add load_crls() function to commands tidying load_certs() too. Add option
8022 to verify utility to allow additional CRLs to be included.
8026 * Update OCSP request code to permit adding custom headers to the request:
8027 some responders need this.
8031 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8034 *Julia Lawall <julia@diku.dk>*
8036 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8037 needlessly dereferenced structures, used obsolete functions and
8038 didn't handle all updated verify codes correctly.
8042 * Disable MD2 in the default configuration.
8046 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8047 indicate the initial BIO being pushed or popped. This makes it possible
8048 to determine whether the BIO is the one explicitly called or as a result
8049 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8050 it handles reference counts correctly and doesn't zero out the I/O bio
8051 when it is not being explicitly popped. WARNING: applications which
8052 included workarounds for the old buggy behaviour will need to be modified
8053 or they could free up already freed BIOs.
8057 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8058 renaming to all platforms (within the 0.9.8 branch, this was
8059 done conditionally on Netware platforms to avoid a name clash).
8061 *Guenter <lists@gknw.net>*
8063 * Add ECDHE and PSK support to DTLS.
8065 *Michael Tuexen <tuexen@fh-muenster.de>*
8067 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8072 * Add "missing" function EVP_MD_flags() (without this the only way to
8073 retrieve a digest flags is by accessing the structure directly. Update
8074 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8075 or cipher is registered as in the "from" argument. Print out all
8076 registered digests in the dgst usage message instead of manually
8077 attempting to work them out.
8081 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8082 this allows the use of compression and extensions. Change default cipher
8083 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8084 by default unless an application cipher string requests it.
8088 * Alter match criteria in PKCS12_parse(). It used to try to use local
8089 key ids to find matching certificates and keys but some PKCS#12 files
8090 don't follow the (somewhat unwritten) rules and this strategy fails.
8091 Now just gather all certificates together and the first private key
8092 then look for the first certificate that matches the key.
8096 * Support use of registered digest and cipher names for dgst and cipher
8097 commands instead of having to add each one as a special case. So now
8104 openssl dgst -sha256 foo
8106 and this works for ENGINE based algorithms too.
8110 * Update Gost ENGINE to support parameter files.
8112 *Victor B. Wagner <vitus@cryptocom.ru>*
8114 * Support GeneralizedTime in ca utility.
8116 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8118 * Enhance the hash format used for certificate directory links. The new
8119 form uses the canonical encoding (meaning equivalent names will work
8120 even if they aren't identical) and uses SHA1 instead of MD5. This form
8121 is incompatible with the older format and as a result c_rehash should
8122 be used to rebuild symbolic links.
8126 * Make PKCS#8 the default write format for private keys, replacing the
8127 traditional format. This form is standardised, more secure and doesn't
8128 include an implicit MD5 dependency.
8132 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8133 committed to OpenSSL should pass this lot as a minimum.
8137 * Add session ticket override functionality for use by EAP-FAST.
8139 *Jouni Malinen <j@w1.fi>*
8141 * Modify HMAC functions to return a value. Since these can be implemented
8142 in an ENGINE errors can occur.
8146 * Type-checked OBJ_bsearch_ex.
8150 * Type-checked OBJ_bsearch. Also some constification necessitated
8151 by type-checking. Still to come: TXT_DB, bsearch(?),
8152 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8157 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8158 seconds to a tm structure directly, instead of going through OS
8159 specific date routines. This avoids any issues with OS routines such
8160 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8161 and X509_time_adj_ex() to cover the extended range. The existing
8162 X509_time_adj() is still usable and will no longer have any date issues.
8166 * Delta CRL support. New use deltas option which will attempt to locate
8167 and search any appropriate delta CRLs available.
8169 This work was sponsored by Google.
8173 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8174 code and add additional score elements. Validate alternate CRL paths
8175 as part of the CRL checking and indicate a new error "CRL path validation
8176 error" in this case. Applications wanting additional details can use
8177 the verify callback and check the new "parent" field. If this is not
8178 NULL CRL path validation is taking place. Existing applications won't
8179 see this because it requires extended CRL support which is off by
8182 This work was sponsored by Google.
8186 * Support for freshest CRL extension.
8188 This work was sponsored by Google.
8192 * Initial indirect CRL support. Currently only supported in the CRLs
8193 passed directly and not via lookup. Process certificate issuer
8194 CRL entry extension and lookup CRL entries by bother issuer name
8195 and serial number. Check and process CRL issuer entry in IDP extension.
8197 This work was sponsored by Google.
8201 * Add support for distinct certificate and CRL paths. The CRL issuer
8202 certificate is validated separately in this case. Only enabled if
8203 an extended CRL support flag is set: this flag will enable additional
8204 CRL functionality in future.
8206 This work was sponsored by Google.
8210 * Add support for policy mappings extension.
8212 This work was sponsored by Google.
8216 * Fixes to pathlength constraint, self issued certificate handling,
8217 policy processing to align with RFC3280 and PKITS tests.
8219 This work was sponsored by Google.
8223 * Support for name constraints certificate extension. DN, email, DNS
8224 and URI types are currently supported.
8226 This work was sponsored by Google.
8230 * To cater for systems that provide a pointer-based thread ID rather
8231 than numeric, deprecate the current numeric thread ID mechanism and
8232 replace it with a structure and associated callback type. This
8233 mechanism allows a numeric "hash" to be extracted from a thread ID in
8234 either case, and on platforms where pointers are larger than 'long',
8235 mixing is done to help ensure the numeric 'hash' is usable even if it
8236 can't be guaranteed unique. The default mechanism is to use "&errno"
8237 as a pointer-based thread ID to distinguish between threads.
8239 Applications that want to provide their own thread IDs should now use
8240 CRYPTO_THREADID_set_callback() to register a callback that will call
8241 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8243 Note that ERR_remove_state() is now deprecated, because it is tied
8244 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8245 to free the current thread's error state should be replaced by
8246 ERR_remove_thread_state(NULL).
8248 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8249 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8250 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8251 application was previously providing a numeric thread callback that
8252 was inappropriate for distinguishing threads, then uniqueness might
8253 have been obtained with &errno that happened immediately in the
8254 intermediate development versions of OpenSSL; this is no longer the
8255 case, the numeric thread callback will now override the automatic use
8258 *Geoff Thorpe, with help from Bodo Moeller*
8260 * Initial support for different CRL issuing certificates. This covers a
8261 simple case where the self issued certificates in the chain exist and
8262 the real CRL issuer is higher in the existing chain.
8264 This work was sponsored by Google.
8268 * Removed effectively defunct crypto/store from the build.
8272 * Revamp of STACK to provide stronger type-checking. Still to come:
8273 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8274 ASN1_STRING, CONF_VALUE.
8278 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8279 RAM on SSL connections. This option can save about 34k per idle SSL.
8283 * Revamp of LHASH to provide stronger type-checking. Still to come:
8284 STACK, TXT_DB, bsearch, qsort.
8288 * Initial support for Cryptographic Message Syntax (aka CMS) based
8289 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8290 support for data, signedData, compressedData, digestedData and
8291 encryptedData, envelopedData types included. Scripts to check against
8292 RFC4134 examples draft and interop and consistency checks of many
8293 content types and variants.
8297 * Add options to enc utility to support use of zlib compression BIO.
8301 * Extend mk1mf to support importing of options and assembly language
8302 files from Configure script, currently only included in VC-WIN32.
8303 The assembly language rules can now optionally generate the source
8304 files from the associated perl scripts.
8308 * Implement remaining functionality needed to support GOST ciphersuites.
8309 Interop testing has been performed using CryptoPro implementations.
8311 *Victor B. Wagner <vitus@cryptocom.ru>*
8313 * s390x assembler pack.
8317 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8322 * Implement Opaque PRF Input TLS extension as specified in
8323 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8324 official specification yet and no extension type assignment by
8325 IANA exists, this extension (for now) will have to be explicitly
8326 enabled when building OpenSSL by providing the extension number
8327 to use. For example, specify an option
8329 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8331 to the "config" or "Configure" script to enable the extension,
8332 assuming extension number 0x9527 (which is a completely arbitrary
8333 and unofficial assignment based on the MD5 hash of the Internet
8334 Draft). Note that by doing so, you potentially lose
8335 interoperability with other TLS implementations since these might
8336 be using the same extension number for other purposes.
8338 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8339 opaque PRF input value to use in the handshake. This will create
8340 an internal copy of the length-'len' string at 'src', and will
8341 return non-zero for success.
8343 To get more control and flexibility, provide a callback function
8346 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8347 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8351 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8354 Callback function 'cb' will be called in handshakes, and is
8355 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8356 Argument 'arg' is for application purposes (the value as given to
8357 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8358 be provided to the callback function). The callback function
8359 has to return non-zero to report success: usually 1 to use opaque
8360 PRF input just if possible, or 2 to enforce use of the opaque PRF
8361 input. In the latter case, the library will abort the handshake
8362 if opaque PRF input is not successfully negotiated.
8364 Arguments 'peerinput' and 'len' given to the callback function
8365 will always be NULL and 0 in the case of a client. A server will
8366 see the client's opaque PRF input through these variables if
8367 available (NULL and 0 otherwise). Note that if the server
8368 provides an opaque PRF input, the length must be the same as the
8369 length of the client's opaque PRF input.
8371 Note that the callback function will only be called when creating
8372 a new session (session resumption can resume whatever was
8373 previously negotiated), and will not be called in SSL 2.0
8374 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8375 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8376 for applications that need to enforce opaque PRF input.
8380 * Update ssl code to support digests other than SHA1+MD5 for handshake
8383 *Victor B. Wagner <vitus@cryptocom.ru>*
8385 * Add RFC4507 support to OpenSSL. This includes the corrections in
8386 RFC4507bis. The encrypted ticket format is an encrypted encoded
8387 SSL_SESSION structure, that way new session features are automatically
8390 If a client application caches session in an SSL_SESSION structure
8391 support is transparent because tickets are now stored in the encoded
8394 The SSL_CTX structure automatically generates keys for ticket
8395 protection in servers so again support should be possible
8396 with no application modification.
8398 If a client or server wishes to disable RFC4507 support then the option
8399 SSL_OP_NO_TICKET can be set.
8401 Add a TLS extension debugging callback to allow the contents of any client
8402 or server extensions to be examined.
8404 This work was sponsored by Google.
8408 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8409 OpenSSL should now compile cleanly on gcc 4.2
8411 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8413 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8414 support including streaming MAC support: this is required for GOST
8415 ciphersuite support.
8417 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8419 * Add option -stream to use PKCS#7 streaming in smime utility. New
8420 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8421 to output in BER and PEM format.
8425 * Experimental support for use of HMAC via EVP_PKEY interface. This
8426 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8427 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8428 ENGINE support for HMAC keys which are unextractable. New -mac and
8429 -macopt options to dgst utility.
8433 * New option -sigopt to dgst utility. Update dgst to use
8434 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8435 alternative signing parameters such as X9.31 or PSS in the dgst
8440 * Change ssl_cipher_apply_rule(), the internal function that does
8441 the work each time a ciphersuite string requests enabling
8442 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8443 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8444 the order of disabled ciphersuites such that those ciphersuites
8445 that most recently went from enabled to disabled not only stay
8446 in order with respect to each other, but also have higher priority
8447 than other disabled ciphersuites the next time ciphersuites are
8450 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8451 the same ciphersuites as with "HIGH" alone, but in a specific
8452 order where the PSK ciphersuites come first (since they are the
8453 most recently disabled ciphersuites when "HIGH" is parsed).
8455 Also, change ssl_create_cipher_list() (using this new
8456 functionality) such that between otherwise identical
8457 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8462 * Change ssl_create_cipher_list() so that it automatically
8463 arranges the ciphersuites in reasonable order before starting
8464 to process the rule string. Thus, the definition for "DEFAULT"
8465 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8466 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8467 This makes it much easier to arrive at a reasonable default order
8468 in applications for which anonymous ciphers are OK (meaning
8469 that you can't actually use DEFAULT).
8471 *Bodo Moeller; suggested by Victor Duchovni*
8473 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8474 processing) into multiple integers instead of setting
8475 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8476 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8477 (These masks as well as the individual bit definitions are hidden
8478 away into the non-exported interface ssl/ssl_locl.h, so this
8479 change to the definition of the SSL_CIPHER structure shouldn't
8480 affect applications.) This give us more bits for each of these
8481 categories, so there is no longer a need to coagulate AES128 and
8482 AES256 into a single algorithm bit, and to coagulate Camellia128
8483 and Camellia256 into a single algorithm bit, which has led to all
8486 Thus, among other things, the kludge introduced in 0.9.7m and
8487 0.9.8e for masking out AES256 independently of AES128 or masking
8488 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8490 With the change, we also introduce new ciphersuite aliases that
8491 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8496 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8497 Use the leftmost N bytes of the signature input if the input is
8498 larger than the prime q (with N being the size in bytes of q).
8502 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8503 it yet and it is largely untested.
8507 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8511 * Initial incomplete changes to avoid need for function casts in OpenSSL
8512 some compilers (gcc 4.2 and later) reject their use. Safestack is
8513 reimplemented. Update ASN1 to avoid use of legacy functions.
8517 * Win32/64 targets are linked with Winsock2.
8521 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8522 to external functions. This can be used to increase CRL handling
8523 efficiency especially when CRLs are very large by (for example) storing
8524 the CRL revoked certificates in a database.
8528 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8529 new CRLs added to a directory can be used. New command line option
8530 -verify_return_error to s_client and s_server. This causes real errors
8531 to be returned by the verify callback instead of carrying on no matter
8532 what. This reflects the way a "real world" verify callback would behave.
8536 * GOST engine, supporting several GOST algorithms and public key formats.
8537 Kindly donated by Cryptocom.
8541 * Partial support for Issuing Distribution Point CRL extension. CRLs
8542 partitioned by DP are handled but no indirect CRL or reason partitioning
8543 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8544 selected via a scoring technique which handles IDP and AKID in CRLs.
8548 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8549 will ultimately be used for all verify operations: this will remove the
8550 X509_STORE dependency on certificate verification and allow alternative
8551 lookup methods. X509_STORE based implementations of these two callbacks.
8555 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8556 Modify get_crl() to find a valid (unexpired) CRL if possible.
8560 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8561 this would be called X509_CRL_cmp() but that name is already used by
8562 a function that just compares CRL issuer names. Cache several CRL
8563 extensions in X509_CRL structure and cache CRLDP in X509.
8567 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8568 this maps equivalent X509_NAME structures into a consistent structure.
8569 Name comparison can then be performed rapidly using memcmp().
8573 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8578 * Allow digests to supply their own micalg string for S/MIME type using
8579 the ctrl EVP_MD_CTRL_MICALG.
8583 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8584 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8585 ctrl. It can then customise the structure before and/or after signing
8590 * New function OBJ_add_sigid() to allow application defined signature OIDs
8591 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8592 to free up any added signature OIDs.
8596 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8597 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8598 digest and cipher tables. New options added to openssl utility:
8599 list-message-digest-algorithms and list-cipher-algorithms.
8603 * Change the array representation of binary polynomials: the list
8604 of degrees of non-zero coefficients is now terminated with -1.
8605 Previously it was terminated with 0, which was also part of the
8606 value; thus, the array representation was not applicable to
8607 polynomials where t^0 has coefficient zero. This change makes
8608 the array representation useful in a more general context.
8612 * Various modifications and fixes to SSL/TLS cipher string
8613 handling. For ECC, the code now distinguishes between fixed ECDH
8614 with RSA certificates on the one hand and with ECDSA certificates
8615 on the other hand, since these are separate ciphersuites. The
8616 unused code for Fortezza ciphersuites has been removed.
8618 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8619 (not "ECDHE"). For consistency with the code for DH
8620 certificates, use of ECDH certificates is now considered ECDH
8621 authentication, not RSA or ECDSA authentication (the latter is
8622 merely the CA's signing algorithm and not actively used in the
8625 The temporary ciphersuite alias "ECCdraft" is no longer
8626 available, and ECC ciphersuites are no longer excluded from "ALL"
8627 and "DEFAULT". The following aliases now exist for RFC 4492
8628 ciphersuites, most of these by analogy with the DH case:
8630 kECDHr - ECDH cert, signed with RSA
8631 kECDHe - ECDH cert, signed with ECDSA
8632 kECDH - ECDH cert (signed with either RSA or ECDSA)
8633 kEECDH - ephemeral ECDH
8634 ECDH - ECDH cert or ephemeral ECDH
8640 AECDH - anonymous ECDH
8641 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8645 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8646 Use correct micalg parameters depending on digest(s) in signed message.
8650 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8651 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8655 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8656 an engine to register a method. Add ENGINE lookups for methods and
8657 functional reference processing.
8661 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8662 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8667 * New -resign option to smime utility. This adds one or more signers
8668 to an existing PKCS#7 signedData structure. Also -md option to use an
8669 alternative message digest algorithm for signing.
8673 * Tidy up PKCS#7 routines and add new functions to make it easier to
8674 create PKCS7 structures containing multiple signers. Update smime
8675 application to support multiple signers.
8679 * New -macalg option to pkcs12 utility to allow setting of an alternative
8684 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8685 Reorganize PBE internals to lookup from a static table using NIDs,
8686 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8687 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8688 PRF which will be automatically used with PBES2.
8692 * Replace the algorithm specific calls to generate keys in "req" with the
8697 * Update PKCS#7 enveloped data routines to use new API. This is now
8698 supported by any public key method supporting the encrypt operation. A
8699 ctrl is added to allow the public key algorithm to examine or modify
8700 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8705 * Add a ctrl to asn1 method to allow a public key algorithm to express
8706 a default digest type to use. In most cases this will be SHA1 but some
8707 algorithms (such as GOST) need to specify an alternative digest. The
8708 return value indicates how strong the preference is 1 means optional and
8709 2 is mandatory (that is it is the only supported type). Modify
8710 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8711 use the default md. Update openssl utilities to use the default digest
8712 type for signing if it is not explicitly indicated.
8716 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8717 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8718 signing method from the key type. This effectively removes the link
8719 between digests and public key types.
8723 * Add an OID cross reference table and utility functions. Its purpose is to
8724 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8725 rsaEncryption. This will allow some of the algorithm specific hackery
8726 needed to use the correct OID to be removed.
8730 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8731 structures for PKCS7_sign(). They are now set up by the relevant public
8736 * Add provisional EC pkey method with support for ECDSA and ECDH.
8740 * Add support for key derivation (agreement) in the API, DH method and
8745 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8746 public and private key formats. As a side effect these add additional
8747 command line functionality not previously available: DSA signatures can be
8748 generated and verified using pkeyutl and DH key support and generation in
8755 *Oliver Tappe <zooey@hirschkaefer.de>*
8757 * New make target "install_html_docs" installs HTML renditions of the
8760 *Oliver Tappe <zooey@hirschkaefer.de>*
8762 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8763 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8764 support key and parameter generation and add initial key generation
8765 functionality for RSA.
8769 * Add functions for main EVP_PKEY_method operations. The undocumented
8770 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8771 `EVP_PKEY_{encrypt,decrypt}_old`.
8775 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8776 key API, doesn't do much yet.
8780 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8781 public key algorithms. New option to openssl utility:
8782 "list-public-key-algorithms" to print out info.
8786 * Implement the Supported Elliptic Curves Extension for
8787 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8791 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8792 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8796 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8797 utilities such as rsa, dsa, dsaparam etc except they process any key
8802 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8803 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8804 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8809 * Initial support for pluggable public key ASN1.
8810 De-spaghettify the public key ASN1 handling. Move public and private
8811 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8812 algorithm specific handling to a single module within the relevant
8813 algorithm directory. Add functions to allow (near) opaque processing
8814 of public and private key structures.
8818 * Implement the Supported Point Formats Extension for
8819 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8823 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8824 for the psk identity [hint] and the psk callback functions to the
8825 SSL_SESSION, SSL and SSL_CTX structure.
8828 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8832 SSL_CTX_use_psk_identity_hint
8833 SSL_get_psk_identity_hint
8834 SSL_get_psk_identity
8835 SSL_use_psk_identity_hint
8837 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8839 * Add RFC 3161 compliant time stamp request creation, response generation
8840 and response verification functionality.
8842 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8844 * Add initial support for TLS extensions, specifically for the server_name
8845 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8846 have new members for a host name. The SSL data structure has an
8847 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8848 stored in that context to allow for session resumption, even after the
8849 SSL has been switched to a new SSL_CTX in reaction to a client's
8850 server_name extension.
8852 New functions (subject to change):
8854 SSL_get_servername()
8855 SSL_get_servername_type()
8858 New CTRL codes and macros (subject to change):
8860 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8861 - SSL_CTX_set_tlsext_servername_callback()
8862 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8863 - SSL_CTX_set_tlsext_servername_arg()
8864 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8866 openssl s_client has a new '-servername ...' option.
8868 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8869 '-key2 ...', '-servername_fatal' (subject to change). This allows
8870 testing the HostName extension for a specific single host name ('-cert'
8871 and '-key' remain fallbacks for handshakes without HostName
8872 negotiation). If the unrecognized_name alert has to be sent, this by
8873 default is a warning; it becomes fatal with the '-servername_fatal'
8876 *Peter Sylvester, Remy Allais, Christophe Renou*
8878 * Whirlpool hash implementation is added.
8882 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8883 bn(64,32). Because of instruction set limitations it doesn't have
8884 any negative impact on performance. This was done mostly in order
8885 to make it possible to share assembler modules, such as bn_mul_mont
8886 implementations, between 32- and 64-bit builds without hassle.
8890 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8891 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8896 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8897 dedicated Montgomery multiplication procedure, is introduced.
8898 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8899 "64-bit" performance on certain 32-bit targets.
8903 * New option SSL_OP_NO_COMP to disable use of compression selectively
8904 in SSL structures. New SSL ctrl to set maximum send fragment size.
8905 Save memory by setting the I/O buffer sizes dynamically instead of
8906 using the maximum available value.
8910 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8911 in addition to the text details.
8915 * Very, very preliminary EXPERIMENTAL support for printing of general
8916 ASN1 structures. This currently produces rather ugly output and doesn't
8917 handle several customised structures at all.
8921 * Integrated support for PVK file format and some related formats such
8922 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8923 these in the 'rsa' and 'dsa' utilities.
8927 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8931 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8932 place for the (very old) "NETSCAPE" format certificates which are now
8933 handled using new ASN1 code equivalents.
8937 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8938 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8939 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8943 * Modify CRL distribution points extension code to print out previously
8944 unsupported fields. Enhance extension setting code to allow setting of
8949 * Add print and set support for Issuing Distribution Point CRL extension.
8953 * Change 'Configure' script to enable Camellia by default.
8960 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8962 * When rejecting SSL/TLS records due to an incorrect version number, never
8963 update s->server with a new major version number. As of
8964 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8965 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8966 the previous behavior could result in a read attempt at NULL when
8967 receiving specific incorrect SSL/TLS records once record payload
8968 protection is active. ([CVE-2010-0740])
8970 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8972 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8973 could be crashed if the relevant tables were not present (e.g. chrooted).
8975 *Tomas Hoger <thoger@redhat.com>*
8977 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8979 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8981 *Martin Olsson, Neel Mehta*
8983 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8984 accommodate for stack sorting, always a write lock!).
8988 * On some versions of WIN32 Heap32Next is very slow. This can cause
8989 excessive delays in the RAND_poll(): over a minute. As a workaround
8990 include a time check in the inner Heap32Next loop too.
8994 * The code that handled flushing of data in SSL/TLS originally used the
8995 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8996 the problem outlined in PR#1949. The fix suggested there however can
8997 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8998 of Apache). So instead simplify the code to flush unconditionally.
8999 This should be fine since flushing with no data to flush is a no op.
9003 * Handle TLS versions 2.0 and later properly and correctly use the
9004 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9005 off ancient servers have a habit of sticking around for a while...
9009 * Modify compression code so it frees up structures without using the
9010 ex_data callbacks. This works around a problem where some applications
9011 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9012 restarting) then use compression (e.g. SSL with compression) later.
9013 This results in significant per-connection memory leaks and
9014 has caused some security issues including CVE-2008-1678 and
9019 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9020 change when encrypting or decrypting.
9024 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9025 connect and renegotiate with servers which do not support RI.
9026 Until RI is more widely deployed this option is enabled by default.
9030 * Add "missing" ssl ctrls to clear options and mode.
9034 * If client attempts to renegotiate and doesn't support RI respond with
9035 a no_renegotiation alert as required by RFC5746. Some renegotiating
9036 TLS clients will continue a connection gracefully when they receive
9037 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9038 waiting for a server hello which it will never receive. Now we treat a
9039 received no_renegotiation alert as a fatal error. This is because
9040 applications requesting a renegotiation might well expect it to succeed
9041 and would have no code in place to handle the server denying it so the
9042 only safe thing to do is to terminate the connection.
9046 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9047 peer supports secure renegotiation and 0 otherwise. Print out peer
9048 renegotiation support in s_client/s_server.
9052 * Replace the highly broken and deprecated SPKAC certification method with
9053 the updated NID creation version. This should correctly handle UTF8.
9057 * Implement RFC5746. Re-enable renegotiation but require the extension
9058 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9059 turns out to be a bad idea. It has been replaced by
9060 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9061 SSL_CTX_set_options(). This is really not recommended unless you
9062 know what you are doing.
9064 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9066 * Fixes to stateless session resumption handling. Use initial_ctx when
9067 issuing and attempting to decrypt tickets in case it has changed during
9068 servername handling. Use a non-zero length session ID when attempting
9069 stateless session resumption: this makes it possible to determine if
9070 a resumption has occurred immediately after receiving server hello
9071 (several places in OpenSSL subtly assume this) instead of later in
9076 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9077 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9078 fixes for a few places where the return code is not checked
9081 *Julia Lawall <julia@diku.dk>*
9083 * Add --strict-warnings option to Configure script to include devteam
9084 warnings in other configurations.
9088 * Add support for --libdir option and LIBDIR variable in makefiles. This
9089 makes it possible to install openssl libraries in locations which
9090 have names other than "lib", for example "/usr/lib64" which some
9093 *Steve Henson, based on patch from Jeremy Utley*
9095 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9096 X690 8.9.12 and can produce some misleading textual output of OIDs.
9098 *Steve Henson, reported by Dan Kaminsky*
9100 * Delete MD2 from algorithm tables. This follows the recommendation in
9101 several standards that it is not used in new applications due to
9102 several cryptographic weaknesses. For binary compatibility reasons
9103 the MD2 API is still compiled in by default.
9107 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9112 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9113 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9116 *Guenter <lists@gknw.net>*
9118 * Fix the server certificate chain building code to use X509_verify_cert(),
9119 it used to have an ad-hoc builder which was unable to cope with anything
9120 other than a simple chain.
9122 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9124 * Don't check self signed certificate signatures in X509_verify_cert()
9125 by default (a flag can override this): it just wastes time without
9126 adding any security. As a useful side effect self signed root CAs
9127 with non-FIPS digests are now usable in FIPS mode.
9131 * In dtls1_process_out_of_seq_message() the check if the current message
9132 is already buffered was missing. For every new message was memory
9133 allocated, allowing an attacker to perform an denial of service attack
9134 with sending out of seq handshake messages until there is no memory
9135 left. Additionally every future message was buffered, even if the
9136 sequence number made no sense and would be part of another handshake.
9137 So only messages with sequence numbers less than 10 in advance will be
9138 buffered. ([CVE-2009-1378])
9140 *Robin Seggelmann, discovered by Daniel Mentz*
9142 * Records are buffered if they arrive with a future epoch to be
9143 processed after finishing the corresponding handshake. There is
9144 currently no limitation to this buffer allowing an attacker to perform
9145 a DOS attack with sending records with future epochs until there is no
9146 memory left. This patch adds the pqueue_size() function to determine
9147 the size of a buffer and limits the record buffer to 100 entries.
9150 *Robin Seggelmann, discovered by Daniel Mentz*
9152 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9153 parent structure is freed. ([CVE-2009-1379])
9157 * Handle non-blocking I/O properly in SSL_shutdown() call.
9159 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9161 * Add `2.5.4.*` OIDs
9163 *Ilya O. <vrghost@gmail.com>*
9165 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9167 * Disable renegotiation completely - this fixes a severe security
9168 problem ([CVE-2009-3555]) at the cost of breaking all
9169 renegotiation. Renegotiation can be re-enabled by setting
9170 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9171 run-time. This is really not recommended unless you know what
9176 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9178 * Don't set val to NULL when freeing up structures, it is freed up by
9179 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9180 zeroing past the valid field. ([CVE-2009-0789])
9182 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9184 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9185 checked correctly. This would allow some invalid signed attributes to
9186 appear to verify correctly. ([CVE-2009-0591])
9188 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9190 * Reject UniversalString and BMPString types with invalid lengths. This
9191 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9192 a legal length. ([CVE-2009-0590])
9196 * Set S/MIME signing as the default purpose rather than setting it
9197 unconditionally. This allows applications to override it at the store
9202 * Permit restricted recursion of ASN1 strings. This is needed in practice
9203 to handle some structures.
9207 * Improve efficiency of mem_gets: don't search whole buffer each time
9210 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9212 * New -hex option for openssl rand.
9216 * Print out UTF8String and NumericString when parsing ASN1.
9220 * Support NumericString type for name components.
9224 * Allow CC in the environment to override the automatically chosen
9225 compiler. Note that nothing is done to ensure flags work with the
9230 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9232 * Properly check EVP_VerifyFinal() and similar return values
9235 *Ben Laurie, Bodo Moeller, Google Security Team*
9237 * Enable TLS extensions by default.
9241 * Allow the CHIL engine to be loaded, whether the application is
9242 multithreaded or not. (This does not release the developer from the
9243 obligation to set up the dynamic locking callbacks.)
9245 *Sander Temme <sander@temme.net>*
9247 * Use correct exit code if there is an error in dgst command.
9249 *Steve Henson; problem pointed out by Roland Dirlewanger*
9251 * Tweak Configure so that you need to say "experimental-jpake" to enable
9252 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9256 * Add experimental JPAKE support, including demo authentication in
9257 s_client and s_server.
9261 * Set the comparison function in v3_addr_canonize().
9263 *Rob Austein <sra@hactrn.net>*
9265 * Add support for XMPP STARTTLS in s_client.
9267 *Philip Paeps <philip@freebsd.org>*
9269 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9270 to ensure that even with this option, only ciphersuites in the
9271 server's preference list will be accepted. (Note that the option
9272 applies only when resuming a session, so the earlier behavior was
9273 just about the algorithm choice for symmetric cryptography.)
9277 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9279 * Fix NULL pointer dereference if a DTLS server received
9280 ChangeCipherSpec as first record ([CVE-2009-1386]).
9284 * Fix a state transition in s3_srvr.c and d1_srvr.c
9285 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9289 * The fix in 0.9.8c that supposedly got rid of unsafe
9290 double-checked locking was incomplete for RSA blinding,
9291 addressing just one layer of what turns out to have been
9292 doubly unsafe triple-checked locking.
9294 So now fix this for real by retiring the MONT_HELPER macro
9295 in crypto/rsa/rsa_eay.c.
9297 *Bodo Moeller; problem pointed out by Marius Schilder*
9299 * Various precautionary measures:
9301 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9303 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9304 (NB: This would require knowledge of the secret session ticket key
9305 to exploit, in which case you'd be SOL either way.)
9307 - Change bn_nist.c so that it will properly handle input BIGNUMs
9308 outside the expected range.
9310 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9313 *Neel Mehta, Bodo Moeller*
9315 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9316 the load fails. Useful for distros.
9318 *Ben Laurie and the FreeBSD team*
9320 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9324 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9328 * Expand ENGINE to support engine supplied SSL client certificate functions.
9330 This work was sponsored by Logica.
9334 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9335 keystores. Support for SSL/TLS client authentication too.
9336 Not compiled unless enable-capieng specified to Configure.
9338 This work was sponsored by Logica.
9342 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9343 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9344 attribute creation routines such as certificate requests and PKCS#12
9349 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9351 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9352 handshake which could lead to a client crash as found using the
9353 Codenomicon TLS test suite ([CVE-2008-1672])
9355 *Steve Henson, Mark Cox*
9357 * Fix double free in TLS server name extensions which could lead to
9358 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9362 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9364 Clear the error queue to ensure that error entries left from
9365 older function calls do not interfere with the correct operation.
9367 *Lutz Jaenicke, Erik de Castro Lopo*
9369 * Remove root CA certificates of commercial CAs:
9371 The OpenSSL project does not recommend any specific CA and does not
9372 have any policy with respect to including or excluding any CA.
9373 Therefore it does not make any sense to ship an arbitrary selection
9374 of root CA certificates with the OpenSSL software.
9378 * RSA OAEP patches to fix two separate invalid memory reads.
9379 The first one involves inputs when 'lzero' is greater than
9380 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9381 before the beginning of from). The second one involves inputs where
9382 the 'db' section contains nothing but zeroes (there is a one-byte
9383 invalid read after the end of 'db').
9385 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9387 * Partial backport from 0.9.9-dev:
9389 Introduce bn_mul_mont (dedicated Montgomery multiplication
9390 procedure) as a candidate for BIGNUM assembler implementation.
9391 While 0.9.9-dev uses assembler for various architectures, only
9392 x86_64 is available by default here in the 0.9.8 branch, and
9393 32-bit x86 is available through a compile-time setting.
9395 To try the 32-bit x86 assembler implementation, use Configure
9396 option "enable-montasm" (which exists only for this backport).
9398 As "enable-montasm" for 32-bit x86 disclaims code stability
9399 anyway, in this constellation we activate additional code
9400 backported from 0.9.9-dev for further performance improvements,
9401 namely BN_from_montgomery_word. (To enable this otherwise,
9402 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9404 *Andy Polyakov (backport partially by Bodo Moeller)*
9406 * Add TLS session ticket callback. This allows an application to set
9407 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9408 values. This is useful for key rollover for example where several key
9409 sets may exist with different names.
9413 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9414 This was broken until now in 0.9.8 releases, such that the only way
9415 a registered ENGINE could be used (assuming it initialises
9416 successfully on the host) was to explicitly set it as the default
9417 for the relevant algorithms. This is in contradiction with 0.9.7
9418 behaviour and the documentation. With this fix, when an ENGINE is
9419 registered into a given algorithm's table of implementations, the
9420 'uptodate' flag is reset so that auto-discovery will be used next
9421 time a new context for that algorithm attempts to select an
9424 *Ian Lister (tweaked by Geoff Thorpe)*
9426 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9427 implementation in the following ways:
9429 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9432 Lack of BER streaming support means one pass streaming processing is
9433 only supported if data is detached: setting the streaming flag is
9434 ignored for embedded content.
9436 CMS support is disabled by default and must be explicitly enabled
9437 with the enable-cms configuration option.
9441 * Update the GMP engine glue to do direct copies between BIGNUM and
9442 mpz_t when openssl and GMP use the same limb size. Otherwise the
9443 existing "conversion via a text string export" trick is still used.
9445 *Paul Sheer <paulsheer@gmail.com>*
9447 * Zlib compression BIO. This is a filter BIO which compressed and
9448 uncompresses any data passed through it.
9452 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9453 RFC3394 compatible AES key wrapping.
9457 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9458 sets string data without copying. X509_ALGOR_set0() and
9459 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9460 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9461 from an X509_ATTRIBUTE structure optionally checking it occurs only
9462 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9467 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9468 to get the expected BN_FLG_CONSTTIME behavior.
9470 *Bodo Moeller (Google)*
9474 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9475 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9476 - added some more tests to do_tests.pl
9477 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9478 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9479 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9480 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9481 - various changes to netware.pl to enable gcc-cross builds on Win32
9483 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9484 - various changes to fix missing prototype warnings
9485 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9486 - added AES, WHIRLPOOL and CPUID assembler code to build files
9487 - added missing AES assembler make rules to mk1mf.pl
9488 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9490 *Guenter Knauf <eflash@gmx.net>*
9492 * Implement certificate status request TLS extension defined in RFC3546.
9493 A client can set the appropriate parameters and receive the encoded
9494 OCSP response via a callback. A server can query the supplied parameters
9495 and set the encoded OCSP response in the callback. Add simplified examples
9496 to s_client and s_server.
9500 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9503 + Binary incompatibility of ssl_ctx_st structure
9504 + DTLS interoperation with non-compliant servers
9505 + Don't call get_session_cb() without proposed session
9506 + Fix ia64 assembler code
9508 *Andy Polyakov, Steve Henson*
9510 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9512 * DTLS Handshake overhaul. There were longstanding issues with
9513 OpenSSL DTLS implementation, which were making it impossible for
9514 RFC 4347 compliant client to communicate with OpenSSL server.
9515 Unfortunately just fixing these incompatibilities would "cut off"
9516 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9517 server keeps tolerating non RFC compliant syntax. The opposite is
9518 not true, 0.9.8f client can not communicate with earlier server.
9519 This update even addresses CVE-2007-4995.
9523 * Changes to avoid need for function casts in OpenSSL: some compilers
9524 (gcc 4.2 and later) reject their use.
9525 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9528 * Add RFC4507 support to OpenSSL. This includes the corrections in
9529 RFC4507bis. The encrypted ticket format is an encrypted encoded
9530 SSL_SESSION structure, that way new session features are automatically
9533 If a client application caches session in an SSL_SESSION structure
9534 support is transparent because tickets are now stored in the encoded
9537 The SSL_CTX structure automatically generates keys for ticket
9538 protection in servers so again support should be possible
9539 with no application modification.
9541 If a client or server wishes to disable RFC4507 support then the option
9542 SSL_OP_NO_TICKET can be set.
9544 Add a TLS extension debugging callback to allow the contents of any client
9545 or server extensions to be examined.
9547 This work was sponsored by Google.
9551 * Add initial support for TLS extensions, specifically for the server_name
9552 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9553 have new members for a host name. The SSL data structure has an
9554 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9555 stored in that context to allow for session resumption, even after the
9556 SSL has been switched to a new SSL_CTX in reaction to a client's
9557 server_name extension.
9559 New functions (subject to change):
9561 SSL_get_servername()
9562 SSL_get_servername_type()
9565 New CTRL codes and macros (subject to change):
9567 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9568 - SSL_CTX_set_tlsext_servername_callback()
9569 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9570 - SSL_CTX_set_tlsext_servername_arg()
9571 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9573 openssl s_client has a new '-servername ...' option.
9575 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9576 '-key2 ...', '-servername_fatal' (subject to change). This allows
9577 testing the HostName extension for a specific single host name ('-cert'
9578 and '-key' remain fallbacks for handshakes without HostName
9579 negotiation). If the unrecognized_name alert has to be sent, this by
9580 default is a warning; it becomes fatal with the '-servername_fatal'
9583 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9585 * Add AES and SSE2 assembly language support to VC++ build.
9589 * Mitigate attack on final subtraction in Montgomery reduction.
9593 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9594 (which previously caused an internal error).
9598 * Squeeze another 10% out of IGE mode when in != out.
9602 * AES IGE mode speedup.
9604 *Dean Gaudet (Google)*
9606 * Add the Korean symmetric 128-bit cipher SEED (see
9607 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9608 add SEED ciphersuites from RFC 4162:
9610 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9611 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9612 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9613 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9615 To minimize changes between patchlevels in the OpenSSL 0.9.8
9616 series, SEED remains excluded from compilation unless OpenSSL
9617 is configured with 'enable-seed'.
9619 *KISA, Bodo Moeller*
9621 * Mitigate branch prediction attacks, which can be practical if a
9622 single processor is shared, allowing a spy process to extract
9623 information. For detailed background information, see
9624 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9625 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9626 and Necessary Software Countermeasures"). The core of the change
9627 are new versions BN_div_no_branch() and
9628 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9629 respectively, which are slower, but avoid the security-relevant
9630 conditional branches. These are automatically called by BN_div()
9631 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9632 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9633 remove a conditional branch.
9635 BN_FLG_CONSTTIME is the new name for the previous
9636 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9637 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9638 in the exponent causes BN_mod_exp_mont() to use the alternative
9639 implementation in BN_mod_exp_mont_consttime().) The old name
9640 remains as a deprecated alias.
9642 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9643 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9644 constant-time implementations for more than just exponentiation.
9645 Here too the old name is kept as a deprecated alias.
9647 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9648 the BN_BLINDING structure gets an independent copy of the
9649 modulus. This means that the previous `BIGNUM *m` argument to
9650 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9651 essentially becomes `const BIGNUM *m`, although we can't actually
9652 change this in the header file before 0.9.9. It allows
9653 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9654 enable BN_FLG_CONSTTIME.
9656 *Matthew D Wood (Intel Corp)*
9658 * In the SSL/TLS server implementation, be strict about session ID
9659 context matching (which matters if an application uses a single
9660 external cache for different purposes). Previously,
9661 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9662 set. This did ensure strict client verification, but meant that,
9663 with applications using a single external cache for quite
9664 different requirements, clients could circumvent ciphersuite
9665 restrictions for a given session ID context by starting a session
9666 in a different context.
9670 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9671 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9672 authentication-only ciphersuites.
9676 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9677 not complete and could lead to a possible single byte overflow
9678 ([CVE-2007-5135]) [Ben Laurie]
9680 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9682 * Since AES128 and AES256 (and similarly Camellia128 and
9683 Camellia256) share a single mask bit in the logic of
9684 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9685 kludge to work properly if AES128 is available and AES256 isn't
9686 (or if Camellia128 is available and Camellia256 isn't).
9690 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9691 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9692 When a point or a seed is encoded in a BIT STRING, we need to
9693 prevent the removal of trailing zero bits to get the proper DER
9694 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9695 of a NamedBitList, for which trailing 0 bits need to be removed.)
9699 * Have SSL/TLS server implementation tolerate "mismatched" record
9700 protocol version while receiving ClientHello even if the
9701 ClientHello is fragmented. (The server can't insist on the
9702 particular protocol version it has chosen before the ServerHello
9703 message has informed the client about his choice.)
9707 * Add RFC 3779 support.
9709 *Rob Austein for ARIN, Ben Laurie*
9711 * Load error codes if they are not already present instead of using a
9712 static variable. This allows them to be cleanly unloaded and reloaded.
9713 Improve header file function name parsing.
9717 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9718 or CAPABILITY handshake as required by RFCs.
9722 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9724 * Introduce limits to prevent malicious keys being able to
9725 cause a denial of service. ([CVE-2006-2940])
9727 *Steve Henson, Bodo Moeller*
9729 * Fix ASN.1 parsing of certain invalid structures that can result
9730 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9732 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9733 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9735 * Fix SSL client code which could crash if connecting to a
9736 malicious SSLv2 server. ([CVE-2006-4343])
9738 *Tavis Ormandy and Will Drewry, Google Security Team*
9740 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9741 match only those. Before that, "AES256-SHA" would be interpreted
9742 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9743 the same strength classification in 0.9.7h) as we currently only
9744 have a single AES bit in the ciphersuite description bitmap.
9745 That change, however, also applied to ciphersuite strings such as
9746 "RC4-MD5" that intentionally matched multiple ciphersuites --
9747 namely, SSL 2.0 ciphersuites in addition to the more common ones
9748 from SSL 3.0/TLS 1.0.
9750 So we change the selection algorithm again: Naming an explicit
9751 ciphersuite selects this one ciphersuite, and any other similar
9752 ciphersuite (same bitmap) from *other* protocol versions.
9753 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9754 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9756 Since SSL 2.0 does not have any ciphersuites for which the
9757 128/256 bit distinction would be relevant, this works for now.
9758 The proper fix will be to use different bits for AES128 and
9759 AES256, which would have avoided the problems from the beginning;
9760 however, bits are scarce, so we can only do this in a new release
9761 (not just a patchlevel) when we can change the SSL_CIPHER
9762 definition to split the single 'unsigned long mask' bitmap into
9763 multiple values to extend the available space.
9767 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9769 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9770 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9772 * Add AES IGE and biIGE modes.
9776 * Change the Unix randomness entropy gathering to use poll() when
9777 possible instead of select(), since the latter has some
9778 undesirable limitations.
9780 *Darryl Miles via Richard Levitte and Bodo Moeller*
9782 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9783 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9784 cannot be implicitly activated as part of, e.g., the "AES" alias.
9785 However, please upgrade to OpenSSL 0.9.9[-dev] for
9786 non-experimental use of the ECC ciphersuites to get TLS extension
9787 support, which is required for curve and point format negotiation
9788 to avoid potential handshake problems.
9792 * Disable rogue ciphersuites:
9794 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9795 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9796 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9798 The latter two were purportedly from
9799 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9802 Also deactivate the remaining ciphersuites from
9803 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9804 unofficial, and the ID has long expired.
9808 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9809 dual-core machines) and other potential thread-safety issues.
9813 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9814 versions), which is now available for royalty-free use
9815 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9816 Also, add Camellia TLS ciphersuites from RFC 4132.
9818 To minimize changes between patchlevels in the OpenSSL 0.9.8
9819 series, Camellia remains excluded from compilation unless OpenSSL
9820 is configured with 'enable-camellia'.
9824 * Disable the padding bug check when compression is in use. The padding
9825 bug check assumes the first packet is of even length, this is not
9826 necessarily true if compression is enabled and can result in false
9827 positives causing handshake failure. The actual bug test is ancient
9828 code so it is hoped that implementations will either have fixed it by
9829 now or any which still have the bug do not support compression.
9833 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9835 * When applying a cipher rule check to see if string match is an explicit
9836 cipher suite and only match that one cipher suite if it is.
9840 * Link in manifests for VC++ if needed.
9842 *Austin Ziegler <halostatue@gmail.com>*
9844 * Update support for ECC-based TLS ciphersuites according to
9845 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9846 TLS extensions, which are supported starting with the 0.9.9
9847 branch, not in the OpenSSL 0.9.8 branch).
9851 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9852 opaque EVP_CIPHER_CTX handling.
9856 * Fixes and enhancements to zlib compression code. We now only use
9857 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9858 to conform with the standards mentioned here:
9859 <http://www.zlib.net/DLL_FAQ.txt>
9860 Static zlib linking now works on Windows and the new --with-zlib-include
9861 --with-zlib-lib options to Configure can be used to supply the location
9862 of the headers and library. Gracefully handle case where zlib library
9867 * Several fixes and enhancements to the OID generation code. The old code
9868 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9869 handle numbers larger than ULONG_MAX, truncated printing and had a
9870 non standard OBJ_obj2txt() behaviour.
9874 * Add support for building of engines under engine/ as shared libraries
9875 under VC++ build system.
9879 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9880 Hopefully, we will not see any false combination of paths any more.
9884 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9886 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9887 (part of SSL_OP_ALL). This option used to disable the
9888 countermeasure against man-in-the-middle protocol-version
9889 rollback in the SSL 2.0 server implementation, which is a bad
9890 idea. ([CVE-2005-2969])
9892 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9893 for Information Security, National Institute of Advanced Industrial
9894 Science and Technology [AIST], Japan)*
9896 * Add two function to clear and return the verify parameter flags.
9900 * Keep cipherlists sorted in the source instead of sorting them at
9901 runtime, thus removing the need for a lock.
9905 * Avoid some small subgroup attacks in Diffie-Hellman.
9907 *Nick Mathewson and Ben Laurie*
9909 * Add functions for well-known primes.
9913 * Extended Windows CE support.
9915 *Satoshi Nakamura and Andy Polyakov*
9917 * Initialize SSL_METHOD structures at compile time instead of during
9918 runtime, thus removing the need for a lock.
9922 * Make PKCS7_decrypt() work even if no certificate is supplied by
9923 attempting to decrypt each encrypted key in turn. Add support to
9928 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9930 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9933 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9937 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9938 key into the same file any more.
9942 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9946 * Add -utf8 command line and config file option to 'ca'.
9948 *Stefan <stf@udoma.org*
9950 * Removed the macro des_crypt(), as it seems to conflict with some
9951 libraries. Use DES_crypt().
9955 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9956 involves renaming the source and generated shared-libs for
9957 both. The engines will accept the corrected or legacy ids
9958 ('ncipher' and '4758_cca' respectively) when binding. NB,
9959 this only applies when building 'shared'.
9961 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9963 * Add attribute functions to EVP_PKEY structure. Modify
9964 PKCS12_create() to recognize a CSP name attribute and
9965 use it. Make -CSP option work again in pkcs12 utility.
9969 * Add new functionality to the bn blinding code:
9970 - automatic re-creation of the BN_BLINDING parameters after
9971 a fixed number of uses (currently 32)
9972 - add new function for parameter creation
9973 - introduce flags to control the update behaviour of the
9974 BN_BLINDING parameters
9975 - hide BN_BLINDING structure
9976 Add a second BN_BLINDING slot to the RSA structure to improve
9977 performance when a single RSA object is shared among several
9982 * Add support for DTLS.
9984 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9986 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9987 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9991 * Remove buggy and incomplete DH cert support from
9992 ssl/ssl_rsa.c and ssl/s3_both.c
9996 * Use SHA-1 instead of MD5 as the default digest algorithm for
9997 the `apps/openssl` commands.
10001 * Compile clean with "-Wall -Wmissing-prototypes
10002 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10003 DEBUG_SAFESTACK must also be set.
10007 * Change ./Configure so that certain algorithms can be disabled by default.
10008 The new counterpiece to "no-xxx" is "enable-xxx".
10010 The patented RC5 and MDC2 algorithms will now be disabled unless
10011 "enable-rc5" and "enable-mdc2", respectively, are specified.
10013 (IDEA remains enabled despite being patented. This is because IDEA
10014 is frequently required for interoperability, and there is no license
10015 fee for non-commercial use. As before, "no-idea" can be used to
10016 avoid this algorithm.)
10020 * Add processing of proxy certificates (see RFC 3820). This work was
10021 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10022 EGEE (Enabling Grids for E-science in Europe).
10026 * RC4 performance overhaul on modern architectures/implementations, such
10027 as Intel P4, IA-64 and AMD64.
10031 * New utility extract-section.pl. This can be used specify an alternative
10032 section number in a pod file instead of having to treat each file as
10033 a separate case in Makefile. This can be done by adding two lines to the
10036 =for comment openssl_section:XXX
10038 The blank line is mandatory.
10042 * New arguments -certform, -keyform and -pass for s_client and s_server
10043 to allow alternative format key and certificate files and passphrase
10048 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10049 update associated structures and add various utility functions.
10051 Add new policy related verify parameters, include policy checking in
10052 standard verify code. Enhance 'smime' application with extra parameters
10053 to support policy checking and print out.
10057 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10058 Nehemiah processors. These extensions support AES encryption in hardware
10059 as well as RNG (though RNG support is currently disabled).
10061 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10063 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10067 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10069 *Andy Polyakov and a number of other people*
10071 * Improved PowerPC platform support. Most notably BIGNUM assembler
10072 implementation contributed by IBM.
10074 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10076 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10077 exponent rather than 'unsigned long'. There is a corresponding change to
10078 the new 'rsa_keygen' element of the RSA_METHOD structure.
10080 *Jelte Jansen, Geoff Thorpe*
10082 * Functionality for creating the initial serial number file is now
10083 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10085 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10086 number file to 1, which is bound to cause problems. To avoid
10087 the problems while respecting compatibility between different 0.9.7
10088 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10089 CA.pl for serial number initialization. With the new release 0.9.8,
10090 we can fix the problem directly in the 'ca' utility.)
10094 * Reduced header interdependencies by declaring more opaque objects in
10095 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10096 give fewer recursive includes, which could break lazy source code - so
10097 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10098 developers should define this symbol when building and using openssl to
10099 ensure they track the recommended behaviour, interfaces, [etc], but
10100 backwards-compatible behaviour prevails when this isn't defined.
10104 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10108 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10109 This will generate a random key of the appropriate length based on the
10110 cipher context. The EVP_CIPHER can provide its own random key generation
10111 routine to support keys of a specific form. This is used in the des and
10112 3des routines to generate a key of the correct parity. Update S/MIME
10113 code to use new functions and hence generate correct parity DES keys.
10114 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10115 valid (weak or incorrect parity).
10119 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10120 as looking them up. This is useful when the verified structure may contain
10121 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10122 present unless the new PKCS7_NO_CRL flag is asserted.
10126 * Extend ASN1 oid configuration module. It now additionally accepts the
10129 shortName = some long name, 1.2.3.4
10133 * Reimplemented the BN_CTX implementation. There is now no more static
10134 limitation on the number of variables it can handle nor the depth of the
10135 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10136 information can now expand as required, and rather than having a single
10137 static array of bignums, BN_CTX now uses a linked-list of such arrays
10138 allowing it to expand on demand whilst maintaining the usefulness of
10139 BN_CTX's "bundling".
10143 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10144 to allow all RSA operations to function using a single BN_CTX.
10148 * Preliminary support for certificate policy evaluation and checking. This
10149 is initially intended to pass the tests outlined in "Conformance Testing
10150 of Relying Party Client Certificate Path Processing Logic" v1.07.
10154 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10155 remained unused and not that useful. A variety of other little bignum
10156 tweaks and fixes have also been made continuing on from the audit (see
10161 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10162 associated ASN1, EVP and SSL functions and old ASN1 macros.
10166 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10167 and this should never fail. So the return value from the use of
10168 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10169 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10173 * BN_CTX_get() should return zero-valued bignums, providing the same
10174 initialised value as BN_new().
10176 *Geoff Thorpe, suggested by Ulf Möller*
10178 * Support for inhibitAnyPolicy certificate extension.
10182 * An audit of the BIGNUM code is underway, for which debugging code is
10183 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10184 is considered valid when processing BIGNUMs, and causes execution to
10185 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10186 further steps are taken to deliberately pollute unused data in BIGNUM
10187 structures to try and expose faulty code further on. For now, openssl will
10188 (in its default mode of operation) continue to tolerate the inconsistent
10189 forms that it has tolerated in the past, but authors and packagers should
10190 consider trying openssl and their own applications when compiled with
10191 these debugging symbols defined. It will help highlight potential bugs in
10192 their own code, and will improve the test coverage for OpenSSL itself. At
10193 some point, these tighter rules will become openssl's default to improve
10194 maintainability, though the assert()s and other overheads will remain only
10195 in debugging configurations. See bn.h for more details.
10197 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10199 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10200 that can only be obtained through BN_CTX_new() (which implicitly
10201 initialises it). The presence of this function only made it possible
10202 to overwrite an existing structure (and cause memory leaks).
10206 * Because of the callback-based approach for implementing LHASH as a
10207 template type, lh_insert() adds opaque objects to hash-tables and
10208 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10209 to clean up those corresponding objects before destroying the hash table
10210 (and losing the object pointers). So some over-zealous constifications in
10211 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10212 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10213 prototyped to have "const" restrictions on the object pointers they are
10214 given (and so aren't required to cast them away any more).
10218 * The tmdiff.h API was so ugly and minimal that our own timing utility
10219 (speed) prefers to use its own implementation. The two implementations
10220 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10221 its object type properly exposed (MS_TM) instead of casting to/from
10222 `char *`. This may still change yet if someone realises MS_TM and
10224 aren't necessarily the greatest nomenclatures - but this is what was used
10225 internally to the implementation so I've used that for now.
10229 * Ensure that deprecated functions do not get compiled when
10230 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10231 the self-tests were still using deprecated key-generation functions so
10232 these have been updated also.
10236 * Reorganise PKCS#7 code to separate the digest location functionality
10237 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10238 New function PKCS7_set_digest() to set the digest type for PKCS#7
10239 digestedData type. Add additional code to correctly generate the
10240 digestedData type and add support for this type in PKCS7 initialization
10245 * New function PKCS7_set0_type_other() this initializes a PKCS7
10246 structure of type "other".
10250 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10251 sure the loop does correctly stop and breaking ("division by zero")
10252 modulus operations are not performed. The (pre-generated) prime
10253 table crypto/bn/bn_prime.h was already correct, but it could not be
10254 re-generated on some platforms because of the "division by zero"
10255 situation in the script.
10257 *Ralf S. Engelschall*
10259 * Update support for ECC-based TLS ciphersuites according to
10260 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10261 SHA-1 now is only used for "small" curves (where the
10262 representation of a field element takes up to 24 bytes); for
10263 larger curves, the field element resulting from ECDH is directly
10264 used as premaster secret.
10266 *Douglas Stebila (Sun Microsystems Laboratories)*
10268 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10269 curve secp160r1 to the tests.
10271 *Douglas Stebila (Sun Microsystems Laboratories)*
10273 * Add the possibility to load symbols globally with DSO.
10275 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10277 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10278 control of the error stack.
10282 * Add support for STORE in ENGINE.
10286 * Add the STORE type. The intention is to provide a common interface
10287 to certificate and key stores, be they simple file-based stores, or
10288 HSM-type store, or LDAP stores, or...
10289 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10293 * Add a generic structure called OPENSSL_ITEM. This can be used to
10294 pass a list of arguments to any function as well as provide a way
10295 for a function to pass data back to the caller.
10299 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10300 works like BUF_strdup() but can be used to duplicate a portion of
10301 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10306 * Add the function sk_find_ex() which works like sk_find(), but will
10307 return an index to an element even if an exact match couldn't be
10308 found. The index is guaranteed to point at the element where the
10309 searched-for key would be inserted to preserve sorting order.
10313 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10314 takes an extra flags argument for optional functionality. Currently,
10315 the following flags are defined:
10317 OBJ_BSEARCH_VALUE_ON_NOMATCH
10318 This one gets OBJ_bsearch_ex() to return a pointer to the first
10319 element where the comparing function returns a negative or zero
10322 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10323 This one gets OBJ_bsearch_ex() to return a pointer to the first
10324 element where the comparing function returns zero. This is useful
10325 if there are more than one element where the comparing function
10330 * Make it possible to create self-signed certificates with 'openssl ca'
10331 in such a way that the self-signed certificate becomes part of the
10332 CA database and uses the same mechanisms for serial number generation
10333 as all other certificate signing. The new flag '-selfsign' enables
10334 this functionality. Adapt CA.sh and CA.pl.in.
10338 * Add functionality to check the public key of a certificate request
10339 against a given private. This is useful to check that a certificate
10340 request can be signed by that key (self-signing).
10344 * Make it possible to have multiple active certificates with the same
10345 subject in the CA index file. This is done only if the keyword
10346 'unique_subject' is set to 'no' in the main CA section (default
10347 if 'CA_default') of the configuration file. The value is saved
10348 with the database itself in a separate index attribute file,
10349 named like the index file with '.attr' appended to the name.
10353 * Generate multi-valued AVAs using '+' notation in config files for
10358 * Support for nameConstraints certificate extension.
10362 * Support for policyConstraints certificate extension.
10366 * Support for policyMappings certificate extension.
10370 * Make sure the default DSA_METHOD implementation only uses its
10371 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10372 and change its own handlers to be NULL so as to remove unnecessary
10373 indirection. This lets alternative implementations fallback to the
10374 default implementation more easily.
10378 * Support for directoryName in GeneralName related extensions
10383 * Make it possible to link applications using Makefile.shared.
10384 Make that possible even when linking against static libraries!
10388 * Support for single pass processing for S/MIME signing. This now
10389 means that S/MIME signing can be done from a pipe, in addition
10390 cleartext signing (multipart/signed type) is effectively streaming
10391 and the signed data does not need to be all held in memory.
10393 This is done with a new flag PKCS7_STREAM. When this flag is set
10394 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10395 is done after the data is output (and digests calculated) in
10396 SMIME_write_PKCS7().
10400 * Add full support for -rpath/-R, both in shared libraries and
10401 applications, at least on the platforms where it's known how
10406 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10407 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10408 will now compute a table of multiples of the generator that
10409 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10410 faster (notably in the case of a single point multiplication,
10411 scalar * generator).
10413 *Nils Larsch, Bodo Moeller*
10415 * IPv6 support for certificate extensions. The various extensions
10416 which use the IP:a.b.c.d can now take IPv6 addresses using the
10417 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10422 * Added an ENGINE that implements RSA by performing private key
10423 exponentiations with the GMP library. The conversions to and from
10424 GMP's mpz_t format aren't optimised nor are any montgomery forms
10425 cached, and on x86 it appears OpenSSL's own performance has caught up.
10426 However there are likely to be other architectures where GMP could
10427 provide a boost. This ENGINE is not built in by default, but it can be
10428 specified at Configure time and should be accompanied by the necessary
10429 linker additions, eg;
10430 ./config -DOPENSSL_USE_GMP -lgmp
10434 * "openssl engine" will not display ENGINE/DSO load failure errors when
10435 testing availability of engines with "-t" - the old behaviour is
10436 produced by increasing the feature's verbosity with "-tt".
10440 * ECDSA routines: under certain error conditions uninitialized BN objects
10441 could be freed. Solution: make sure initialization is performed early
10442 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10447 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10448 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10449 software implementations. For DSA and DH, parameter generation can
10450 also be overridden by providing the appropriate method callbacks.
10454 * Change the "progress" mechanism used in key-generation and
10455 primality testing to functions that take a new BN_GENCB pointer in
10456 place of callback/argument pairs. The new API functions have `_ex`
10457 postfixes and the older functions are reimplemented as wrappers for
10458 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10459 declarations of the old functions to help (graceful) attempts to
10460 migrate to the new functions. Also, the new key-generation API
10461 functions operate on a caller-supplied key-structure and return
10462 success/failure rather than returning a key or NULL - this is to
10463 help make "keygen" another member function of RSA_METHOD etc.
10465 Example for using the new callback interface:
10467 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10468 void *my_arg = ...;
10471 BN_GENCB_set(&my_cb, my_callback, my_arg);
10473 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10474 /* For the meaning of a, b in calls to my_callback(), see the
10475 * documentation of the function that calls the callback.
10476 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10477 * my_callback should return 1 if it wants BN_is_prime_ex()
10478 * to continue, or 0 to stop.
10483 * Change the ZLIB compression method to be stateful, and make it
10484 available to TLS with the number defined in
10485 draft-ietf-tls-compression-04.txt.
10489 * Add the ASN.1 structures and functions for CertificatePair, which
10490 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10492 CertificatePair ::= SEQUENCE {
10493 forward [0] Certificate OPTIONAL,
10494 reverse [1] Certificate OPTIONAL,
10495 -- at least one of the pair shall be present -- }
10497 Also implement the PEM functions to read and write certificate
10498 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10500 This needed to be defined, mostly for the sake of the LDAP
10501 attribute crossCertificatePair, but may prove useful elsewhere as
10506 * Make it possible to inhibit symlinking of shared libraries in
10507 Makefile.shared, for Cygwin's sake.
10511 * Extend the BIGNUM API by creating a function
10512 void BN_set_negative(BIGNUM *a, int neg);
10513 and a macro that behave like
10514 int BN_is_negative(const BIGNUM *a);
10516 to avoid the need to access 'a->neg' directly in applications.
10520 * Implement fast modular reduction for pseudo-Mersenne primes
10521 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10522 EC_GROUP_new_curve_GFp() will now automatically use this
10525 *Nils Larsch <nla@trustcenter.de>*
10527 * Add new lock type (CRYPTO_LOCK_BN).
10531 * Change the ENGINE framework to automatically load engines
10532 dynamically from specific directories unless they could be
10533 found to already be built in or loaded. Move all the
10534 current engines except for the cryptodev one to a new
10535 directory engines/.
10536 The engines in engines/ are built as shared libraries if
10537 the "shared" options was given to ./Configure or ./config.
10538 Otherwise, they are inserted in libcrypto.a.
10539 /usr/local/ssl/engines is the default directory for dynamic
10540 engines, but that can be overridden at configure time through
10541 the usual use of --prefix and/or --openssldir, and at run
10542 time with the environment variable OPENSSL_ENGINES.
10544 *Geoff Thorpe and Richard Levitte*
10546 * Add Makefile.shared, a helper makefile to build shared
10547 libraries. Adapt Makefile.org.
10551 * Add version info to Win32 DLLs.
10553 *Peter 'Luna' Runestig" <peter@runestig.com>*
10555 * Add new 'medium level' PKCS#12 API. Certificates and keys
10556 can be added using this API to created arbitrary PKCS#12
10557 files while avoiding the low level API.
10559 New options to PKCS12_create(), key or cert can be NULL and
10560 will then be omitted from the output file. The encryption
10561 algorithm NIDs can be set to -1 for no encryption, the mac
10562 iteration count can be set to 0 to omit the mac.
10564 Enhance pkcs12 utility by making the -nokeys and -nocerts
10565 options work when creating a PKCS#12 file. New option -nomac
10566 to omit the mac, NONE can be set for an encryption algorithm.
10567 New code is modified to use the enhanced PKCS12_create()
10568 instead of the low level API.
10572 * Extend ASN1 encoder to support indefinite length constructed
10573 encoding. This can output sequences tags and octet strings in
10574 this form. Modify pk7_asn1.c to support indefinite length
10575 encoding. This is experimental and needs additional code to
10576 be useful, such as an ASN1 bio and some enhanced streaming
10579 Extend template encode functionality so that tagging is passed
10580 down to the template encoder.
10584 * Let 'openssl req' fail if an argument to '-newkey' is not
10585 recognized instead of using RSA as a default.
10589 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10590 As these are not official, they are not included in "ALL";
10591 the "ECCdraft" ciphersuite group alias can be used to select them.
10593 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10595 * Add ECDH engine support.
10597 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10599 * Add ECDH in new directory crypto/ecdh/.
10601 *Douglas Stebila (Sun Microsystems Laboratories)*
10603 * Let BN_rand_range() abort with an error after 100 iterations
10604 without success (which indicates a broken PRNG).
10608 * Change BN_mod_sqrt() so that it verifies that the input value
10609 is really the square of the return value. (Previously,
10610 BN_mod_sqrt would show GIGO behaviour.)
10614 * Add named elliptic curves over binary fields from X9.62, SECG,
10615 and WAP/WTLS; add OIDs that were still missing.
10617 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10619 * Extend the EC library for elliptic curves over binary fields
10620 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10623 EC_GF2m_simple_method
10627 EC_GROUP_new_curve_GF2m
10628 EC_GROUP_set_curve_GF2m
10629 EC_GROUP_get_curve_GF2m
10630 EC_POINT_set_affine_coordinates_GF2m
10631 EC_POINT_get_affine_coordinates_GF2m
10632 EC_POINT_set_compressed_coordinates_GF2m
10634 Point compression for binary fields is disabled by default for
10635 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10638 As binary polynomials are represented as BIGNUMs, various members
10639 of the EC_GROUP and EC_POINT data structures can be shared
10640 between the implementations for prime fields and binary fields;
10641 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10642 are essentially identical to their `..._GFp` counterparts.
10643 (For simplicity, the `..._GFp` prefix has been dropped from
10644 various internal method names.)
10646 An internal 'field_div' method (similar to 'field_mul' and
10647 'field_sqr') has been added; this is used only for binary fields.
10649 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10651 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10652 through methods ('mul', 'precompute_mult').
10654 The generic implementations (now internally called 'ec_wNAF_mul'
10655 and 'ec_wNAF_precomputed_mult') remain the default if these
10656 methods are undefined.
10658 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10660 * New function EC_GROUP_get_degree, which is defined through
10661 EC_METHOD. For curves over prime fields, this returns the bit
10662 length of the modulus.
10664 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10666 * New functions EC_GROUP_dup, EC_POINT_dup.
10667 (These simply call ..._new and ..._copy).
10669 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10671 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10672 Polynomials are represented as BIGNUMs (where the sign bit is not
10673 used) in the following functions [macros]:
10676 BN_GF2m_sub [= BN_GF2m_add]
10677 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10678 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10679 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10681 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10682 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10683 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10684 BN_GF2m_cmp [= BN_ucmp]
10686 (Note that only the 'mod' functions are actually for fields GF(2^m).
10687 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10689 For some functions, an the irreducible polynomial defining a
10690 field can be given as an 'unsigned int[]' with strictly
10691 decreasing elements giving the indices of those bits that are set;
10692 i.e., p[] represents the polynomial
10693 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10695 p[0] > p[1] > ... > p[k] = 0.
10696 This applies to the following functions:
10699 BN_GF2m_mod_mul_arr
10700 BN_GF2m_mod_sqr_arr
10701 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10702 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10703 BN_GF2m_mod_exp_arr
10704 BN_GF2m_mod_sqrt_arr
10705 BN_GF2m_mod_solve_quad_arr
10709 Conversion can be performed by the following functions:
10714 bntest.c has additional tests for binary polynomial arithmetic.
10716 Two implementations for BN_GF2m_mod_div() are available.
10717 The default algorithm simply uses BN_GF2m_mod_inv() and
10718 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10719 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10720 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10722 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10724 * Add new error code 'ERR_R_DISABLED' that can be used when some
10725 functionality is disabled at compile-time.
10727 *Douglas Stebila <douglas.stebila@sun.com>*
10729 * Change default behaviour of 'openssl asn1parse' so that more
10730 information is visible when viewing, e.g., a certificate:
10732 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10733 mode the content of non-printable OCTET STRINGs is output in a
10734 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10735 avoid the appearance of a printable string.
10737 *Nils Larsch <nla@trustcenter.de>*
10739 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10741 EC_GROUP_set_asn1_flag()
10742 EC_GROUP_get_asn1_flag()
10743 EC_GROUP_set_point_conversion_form()
10744 EC_GROUP_get_point_conversion_form()
10745 These control ASN1 encoding details:
10746 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10747 has been set to OPENSSL_EC_NAMED_CURVE.
10748 - Points are encoded in uncompressed form by default; options for
10749 asn1_for are as for point2oct, namely
10750 POINT_CONVERSION_COMPRESSED
10751 POINT_CONVERSION_UNCOMPRESSED
10752 POINT_CONVERSION_HYBRID
10754 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10756 EC_GROUP_set_seed()
10757 EC_GROUP_get0_seed()
10758 EC_GROUP_get_seed_len()
10759 This is used only for ASN1 purposes (so far).
10761 *Nils Larsch <nla@trustcenter.de>*
10763 * Add 'field_type' member to EC_METHOD, which holds the NID
10764 of the appropriate field type OID. The new function
10765 EC_METHOD_get_field_type() returns this value.
10767 *Nils Larsch <nla@trustcenter.de>*
10770 EC_POINT_point2bn()
10771 EC_POINT_bn2point()
10772 EC_POINT_point2hex()
10773 EC_POINT_hex2point()
10774 providing useful interfaces to EC_POINT_point2oct() and
10775 EC_POINT_oct2point().
10777 *Nils Larsch <nla@trustcenter.de>*
10779 * Change internals of the EC library so that the functions
10780 EC_GROUP_set_generator()
10781 EC_GROUP_get_generator()
10782 EC_GROUP_get_order()
10783 EC_GROUP_get_cofactor()
10784 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10785 to methods, which would lead to unnecessary code duplication when
10786 adding different types of curves.
10788 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10790 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10791 arithmetic, and such that modified wNAFs are generated
10792 (which avoid length expansion in many cases).
10796 * Add a function EC_GROUP_check_discriminant() (defined via
10797 EC_METHOD) that verifies that the curve discriminant is non-zero.
10799 Add a function EC_GROUP_check() that makes some sanity tests
10800 on a EC_GROUP, its generator and order. This includes
10801 EC_GROUP_check_discriminant().
10803 *Nils Larsch <nla@trustcenter.de>*
10805 * Add ECDSA in new directory crypto/ecdsa/.
10807 Add applications 'openssl ecparam' and 'openssl ecdsa'
10808 (these are based on 'openssl dsaparam' and 'openssl dsa').
10810 ECDSA support is also included in various other files across the
10811 library. Most notably,
10812 - 'openssl req' now has a '-newkey ecdsa:file' option;
10813 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10814 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10815 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10816 them suitable for ECDSA where domain parameters must be
10817 extracted before the specific public key;
10818 - ECDSA engine support has been added.
10820 *Nils Larsch <nla@trustcenter.de>*
10822 * Include some named elliptic curves, and add OIDs from X9.62,
10823 SECG, and WAP/WTLS. Each curve can be obtained from the new
10825 EC_GROUP_new_by_curve_name(),
10826 and the list of available named curves can be obtained with
10827 EC_get_builtin_curves().
10828 Also add a 'curve_name' member to EC_GROUP objects, which can be
10830 EC_GROUP_set_curve_name()
10831 EC_GROUP_get_curve_name()
10833 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10835 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10836 was actually never needed) and in BN_mul(). The removal in BN_mul()
10837 required a small change in bn_mul_part_recursive() and the addition
10838 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10839 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10840 bn_sub_words() and bn_add_words() except they take arrays with
10845 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10847 * Cleanse PEM buffers before freeing them since they may contain
10850 *Benjamin Bennett <ben@psc.edu>*
10852 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10853 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10854 authentication-only ciphersuites.
10858 * Since AES128 and AES256 share a single mask bit in the logic of
10859 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10860 kludge to work properly if AES128 is available and AES256 isn't.
10864 * Expand security boundary to match 1.1.1 module.
10868 * Remove redundant features: hash file source, editing of test vectors
10869 modify fipsld to use external fips_premain.c signature.
10873 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10874 run algorithm test programs.
10878 * Make algorithm test programs more tolerant of whitespace.
10882 * Have SSL/TLS server implementation tolerate "mismatched" record
10883 protocol version while receiving ClientHello even if the
10884 ClientHello is fragmented. (The server can't insist on the
10885 particular protocol version it has chosen before the ServerHello
10886 message has informed the client about his choice.)
10890 * Load error codes if they are not already present instead of using a
10891 static variable. This allows them to be cleanly unloaded and reloaded.
10895 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10897 * Introduce limits to prevent malicious keys being able to
10898 cause a denial of service. ([CVE-2006-2940])
10900 *Steve Henson, Bodo Moeller*
10902 * Fix ASN.1 parsing of certain invalid structures that can result
10903 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10905 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10906 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10908 * Fix SSL client code which could crash if connecting to a
10909 malicious SSLv2 server. ([CVE-2006-4343])
10911 *Tavis Ormandy and Will Drewry, Google Security Team*
10913 * Change ciphersuite string processing so that an explicit
10914 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10915 will no longer include "AES128-SHA"), and any other similar
10916 ciphersuite (same bitmap) from *other* protocol versions (so that
10917 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10918 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10919 changes from 0.9.8b and 0.9.8d.
10923 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10925 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10926 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10928 * Change the Unix randomness entropy gathering to use poll() when
10929 possible instead of select(), since the latter has some
10930 undesirable limitations.
10932 *Darryl Miles via Richard Levitte and Bodo Moeller*
10934 * Disable rogue ciphersuites:
10936 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10937 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10938 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10940 The latter two were purportedly from
10941 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10944 Also deactivate the remaining ciphersuites from
10945 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10946 unofficial, and the ID has long expired.
10950 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10951 dual-core machines) and other potential thread-safety issues.
10955 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10957 * Adapt fipsld and the build system to link against the validated FIPS
10958 module in FIPS mode.
10962 * Fixes for VC++ 2005 build under Windows.
10966 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10967 from a Windows bash shell such as MSYS. It is autodetected from the
10968 "config" script when run from a VC++ environment. Modify standard VC++
10969 build to use fipscanister.o from the GNU make build.
10973 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10975 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10976 The value now differs depending on if you build for FIPS or not.
10977 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10978 safely run with a non-FIPSed libcrypto, as it may crash because of
10979 the difference induced by this change.
10983 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10985 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10986 (part of SSL_OP_ALL). This option used to disable the
10987 countermeasure against man-in-the-middle protocol-version
10988 rollback in the SSL 2.0 server implementation, which is a bad
10989 idea. ([CVE-2005-2969])
10991 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10992 for Information Security, National Institute of Advanced Industrial
10993 Science and Technology [AIST, Japan)]*
10995 * Minimal support for X9.31 signatures and PSS padding modes. This is
10996 mainly for FIPS compliance and not fully integrated at this stage.
11000 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11001 the exponentiation using a fixed-length exponent. (Otherwise,
11002 the information leaked through timing could expose the secret key
11003 after many signatures; cf. Bleichenbacher's attack on DSA with
11008 * Make a new fixed-window mod_exp implementation the default for
11009 RSA, DSA, and DH private-key operations so that the sequence of
11010 squares and multiplies and the memory access pattern are
11011 independent of the particular secret key. This will mitigate
11012 cache-timing and potential related attacks.
11014 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11015 and this is automatically used by BN_mod_exp_mont() if the new flag
11016 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11017 will use this BN flag for private exponents unless the flag
11018 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11019 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11021 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11023 * Change the client implementation for SSLv23_method() and
11024 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11025 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11026 (Previously, the SSL 2.0 backwards compatible Client Hello
11027 message format would be used even with SSL_OP_NO_SSLv2.)
11031 * Add support for smime-type MIME parameter in S/MIME messages which some
11036 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11037 a threadsafe manner. Modify rsa code to use new function and add calls
11038 to dsa and dh code (which had race conditions before).
11042 * Include the fixed error library code in the C error file definitions
11043 instead of fixing them up at runtime. This keeps the error code
11044 structures constant.
11048 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11050 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11053 * Fixes for newer kerberos headers. NB: the casts are needed because
11054 the 'length' field is signed on one version and unsigned on another
11055 with no (?) obvious way to tell the difference, without these VC++
11056 complains. Also the "definition" of FAR (blank) is no longer included
11057 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11058 some needed definitions.
11062 * Undo Cygwin change.
11066 * Added support for proxy certificates according to RFC 3820.
11067 Because they may be a security thread to unaware applications,
11068 they must be explicitly allowed in run-time. See
11069 docs/HOWTO/proxy_certificates.txt for further information.
11073 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11075 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11076 server and client random values. Previously
11077 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11078 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11080 This change has negligible security impact because:
11082 1. Server and client random values still have 24 bytes of pseudo random
11085 2. Server and client random values are sent in the clear in the initial
11088 3. The master secret is derived using the premaster secret (48 bytes in
11089 size for static RSA ciphersuites) as well as client server and random
11092 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11095 *Stephen Henson, reported by UK NISCC*
11097 * Use Windows randomness collection on Cygwin.
11101 * Fix hang in EGD/PRNGD query when communication socket is closed
11102 prematurely by EGD/PRNGD.
11104 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11106 * Prompt for pass phrases when appropriate for PKCS12 input format.
11110 * Back-port of selected performance improvements from development
11111 branch, as well as improved support for PowerPC platforms.
11115 * Add lots of checks for memory allocation failure, error codes to indicate
11116 failure and freeing up memory if a failure occurs.
11118 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11120 * Add new -passin argument to dgst.
11124 * Perform some character comparisons of different types in X509_NAME_cmp:
11125 this is needed for some certificates that re-encode DNs into UTF8Strings
11126 (in violation of RFC3280) and can't or won't issue name rollover
11131 * Make an explicit check during certificate validation to see that
11132 the CA setting in each certificate on the chain is correct. As a
11133 side effect always do the following basic checks on extensions,
11134 not just when there's an associated purpose to the check:
11136 - if there is an unhandled critical extension (unless the user
11137 has chosen to ignore this fault)
11138 - if the path length has been exceeded (if one is set at all)
11139 - that certain extensions fit the associated purpose (if one has
11144 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11146 * Avoid a race condition when CRLs are checked in a multi threaded
11147 environment. This would happen due to the reordering of the revoked
11148 entries during signature checking and serial number lookup. Now the
11149 encoding is cached and the serial number sort performed under a lock.
11150 Add new STACK function sk_is_sorted().
11154 * Add Delta CRL to the extension code.
11158 * Various fixes to s3_pkt.c so alerts are sent properly.
11160 *David Holmes <d.holmes@f5.com>*
11162 * Reduce the chances of duplicate issuer name and serial numbers (in
11163 violation of RFC3280) using the OpenSSL certificate creation utilities.
11164 This is done by creating a random 64 bit value for the initial serial
11165 number when a serial number file is created or when a self signed
11166 certificate is created using 'openssl req -x509'. The initial serial
11167 number file is created using 'openssl x509 -next_serial' in CA.pl
11168 rather than being initialized to 1.
11172 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11174 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11175 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11177 *Joe Orton, Steve Henson*
11179 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11182 *Joe Orton, Steve Henson*
11184 * Make it possible to have multiple active certificates with the same
11185 subject in the CA index file. This is done only if the keyword
11186 'unique_subject' is set to 'no' in the main CA section (default
11187 if 'CA_default') of the configuration file. The value is saved
11188 with the database itself in a separate index attribute file,
11189 named like the index file with '.attr' appended to the name.
11193 * X509 verify fixes. Disable broken certificate workarounds when
11194 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11195 keyUsage extension present. Don't accept CRLs with unhandled critical
11196 extensions: since verify currently doesn't process CRL extensions this
11197 rejects a CRL with *any* critical extensions. Add new verify error codes
11202 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11203 A clarification of RFC2560 will require the use of OCTET STRINGs and
11204 some implementations cannot handle the current raw format. Since OpenSSL
11205 copies and compares OCSP nonces as opaque blobs without any attempt at
11206 parsing them this should not create any compatibility issues.
11210 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11211 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11212 this HMAC (and other) operations are several times slower than OpenSSL
11217 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11219 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11221 * Use the correct content when signing type "other".
11225 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11227 * Fix various bugs revealed by running the NISCC test suite:
11229 Stop out of bounds reads in the ASN1 code when presented with
11230 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11232 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11234 If verify callback ignores invalid public key errors don't try to check
11235 certificate signature with the NULL public key.
11239 * New -ignore_err option in ocsp application to stop the server
11240 exiting on the first error in a request.
11244 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11245 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11250 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11251 extra data after the compression methods not only for TLS 1.0
11252 but also for SSL 3.0 (as required by the specification).
11254 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11256 * Change X509_certificate_type() to mark the key as exported/exportable
11257 when it's 512 *bits* long, not 512 bytes.
11261 * Change AES_cbc_encrypt() so it outputs exact multiple of
11262 blocks during encryption.
11266 * Various fixes to base64 BIO and non blocking I/O. On write
11267 flushes were not handled properly if the BIO retried. On read
11268 data was not being buffered properly and had various logic bugs.
11269 This also affects blocking I/O when the data being decoded is a
11274 * Various S/MIME bugfixes and compatibility changes:
11275 output correct application/pkcs7 MIME type if
11276 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11277 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11278 of files as .eml work). Correctly handle very long lines in MIME
11283 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11285 * Countermeasure against the Klima-Pokorny-Rosa extension of
11286 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11287 a protocol version number mismatch like a decryption error
11288 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11292 * Turn on RSA blinding by default in the default implementation
11293 to avoid a timing attack. Applications that don't want it can call
11294 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11295 They would be ill-advised to do so in most cases.
11297 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11299 * Change RSA blinding code so that it works when the PRNG is not
11300 seeded (in this case, the secret RSA exponent is abused as
11301 an unpredictable seed -- if it is not unpredictable, there
11302 is no point in blinding anyway). Make RSA blinding thread-safe
11303 by remembering the creator's thread ID in rsa->blinding and
11304 having all other threads use local one-time blinding factors
11305 (this requires more computation than sharing rsa->blinding, but
11306 avoids excessive locking; and if an RSA object is not shared
11307 between threads, blinding will still be very fast).
11311 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11312 ENGINE as defaults for all supported algorithms irrespective of
11313 the 'flags' parameter. 'flags' is now honoured, so applications
11314 should make sure they are passing it correctly.
11318 * Target "mingw" now allows native Windows code to be generated in
11319 the Cygwin environment as well as with the MinGW compiler.
11323 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11325 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11326 via timing by performing a MAC computation even if incorrect
11327 block cipher padding has been found. This is a countermeasure
11328 against active attacks where the attacker has to distinguish
11329 between bad padding and a MAC verification error. ([CVE-2003-0078])
11331 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11332 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11333 Martin Vuagnoux (EPFL, Ilion)*
11335 * Make the no-err option work as intended. The intention with no-err
11336 is not to have the whole error stack handling routines removed from
11337 libcrypto, it's only intended to remove all the function name and
11338 reason texts, thereby removing some of the footprint that may not
11339 be interesting if those errors aren't displayed anyway.
11341 NOTE: it's still possible for any application or module to have its
11342 own set of error texts inserted. The routines are there, just not
11343 used by default when no-err is given.
11347 * Add support for FreeBSD on IA64.
11349 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11351 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11352 Kerberos function mit_des_cbc_cksum(). Before this change,
11353 the value returned by DES_cbc_cksum() was like the one from
11354 mit_des_cbc_cksum(), except the bytes were swapped.
11356 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11358 * Allow an application to disable the automatic SSL chain building.
11359 Before this a rather primitive chain build was always performed in
11360 ssl3_output_cert_chain(): an application had no way to send the
11361 correct chain if the automatic operation produced an incorrect result.
11363 Now the chain builder is disabled if either:
11365 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11367 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11369 The reasoning behind this is that an application would not want the
11370 auto chain building to take place if extra chain certificates are
11371 present and it might also want a means of sending no additional
11372 certificates (for example the chain has two certificates and the
11377 * Add the possibility to build without the ENGINE framework.
11379 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11381 * Under Win32 gmtime() can return NULL: check return value in
11382 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11386 * DSA routines: under certain error conditions uninitialized BN objects
11387 could be freed. Solution: make sure initialization is performed early
11388 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11389 Nils Larsch <nla@trustcenter.de> via PR#459)
11393 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11394 checked on reconnect on the client side, therefore session resumption
11395 could still fail with a "ssl session id is different" error. This
11396 behaviour is masked when SSL_OP_ALL is used due to
11397 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11398 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11399 followup to PR #377.
11403 * IA-32 assembler support enhancements: unified ELF targets, support
11404 for SCO/Caldera platforms, fix for Cygwin shared build.
11408 * Add support for FreeBSD on sparc64. As a consequence, support for
11409 FreeBSD on non-x86 processors is separate from x86 processors on
11410 the config script, much like the NetBSD support.
11412 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11414 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11416 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11419 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11420 code (06) was taken as the first octet of the session ID and the last
11421 octet was ignored consequently. As a result SSLv2 client side session
11422 caching could not have worked due to the session ID mismatch between
11424 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11429 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11430 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11435 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11436 seems that in spite of existing for more than a year, many application
11437 author have done nothing to provide the necessary callbacks, which
11438 means that this particular engine will not work properly anywhere.
11439 This is a very unfortunate situation which forces us, in the name
11440 of usability, to give the hw_ncipher.c a static lock, which is part
11442 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11443 appear in 0.9.8 or later. We EXPECT application authors to have
11444 dealt properly with this when 0.9.8 is released (unless we actually
11445 make such changes in the libcrypto locking code that changes will
11446 have to be made anyway).
11450 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11451 octets have been read, EOF or an error occurs. Without this change
11452 some truncated ASN1 structures will not produce an error.
11456 * Disable Heimdal support, since it hasn't been fully implemented.
11457 Still give the possibility to force the use of Heimdal, but with
11458 warnings and a request that patches get sent to openssl-dev.
11462 * Add the VC-CE target, introduce the WINCE sysname, and add
11463 INSTALL.WCE and appropriate conditionals to make it build.
11465 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11467 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11468 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11469 edit numbers of the version.
11471 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11473 * Introduce safe string copy and catenation functions
11474 (BUF_strlcpy() and BUF_strlcat()).
11476 *Ben Laurie (CHATS) and Richard Levitte*
11478 * Avoid using fixed-size buffers for one-line DNs.
11480 *Ben Laurie (CHATS)*
11482 * Add BUF_MEM_grow_clean() to avoid information leakage when
11483 resizing buffers containing secrets, and use where appropriate.
11485 *Ben Laurie (CHATS)*
11487 * Avoid using fixed size buffers for configuration file location.
11489 *Ben Laurie (CHATS)*
11491 * Avoid filename truncation for various CA files.
11493 *Ben Laurie (CHATS)*
11495 * Use sizeof in preference to magic numbers.
11497 *Ben Laurie (CHATS)*
11499 * Avoid filename truncation in cert requests.
11501 *Ben Laurie (CHATS)*
11503 * Add assertions to check for (supposedly impossible) buffer
11506 *Ben Laurie (CHATS)*
11508 * Don't cache truncated DNS entries in the local cache (this could
11509 potentially lead to a spoofing attack).
11511 *Ben Laurie (CHATS)*
11513 * Fix various buffers to be large enough for hex/decimal
11514 representations in a platform independent manner.
11516 *Ben Laurie (CHATS)*
11518 * Add CRYPTO_realloc_clean() to avoid information leakage when
11519 resizing buffers containing secrets, and use where appropriate.
11521 *Ben Laurie (CHATS)*
11523 * Add BIO_indent() to avoid much slightly worrying code to do
11526 *Ben Laurie (CHATS)*
11528 * Convert sprintf()/BIO_puts() to BIO_printf().
11530 *Ben Laurie (CHATS)*
11532 * buffer_gets() could terminate with the buffer only half
11535 *Ben Laurie (CHATS)*
11537 * Add assertions to prevent user-supplied crypto functions from
11538 overflowing internal buffers by having large block sizes, etc.
11540 *Ben Laurie (CHATS)*
11542 * New OPENSSL_assert() macro (similar to assert(), but enabled
11545 *Ben Laurie (CHATS)*
11547 * Eliminate unused copy of key in RC4.
11549 *Ben Laurie (CHATS)*
11551 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11553 *Ben Laurie (CHATS)*
11555 * Fix off-by-one error in EGD path.
11557 *Ben Laurie (CHATS)*
11559 * If RANDFILE path is too long, ignore instead of truncating.
11561 *Ben Laurie (CHATS)*
11563 * Eliminate unused and incorrectly sized X.509 structure
11566 *Ben Laurie (CHATS)*
11568 * Eliminate unused and dangerous function knumber().
11570 *Ben Laurie (CHATS)*
11572 * Eliminate unused and dangerous structure, KSSL_ERR.
11574 *Ben Laurie (CHATS)*
11576 * Protect against overlong session ID context length in an encoded
11577 session object. Since these are local, this does not appear to be
11580 *Ben Laurie (CHATS)*
11582 * Change from security patch (see 0.9.6e below) that did not affect
11583 the 0.9.6 release series:
11585 Remote buffer overflow in SSL3 protocol - an attacker could
11586 supply an oversized master key in Kerberos-enabled versions.
11589 *Ben Laurie (CHATS)*
11591 * Change the SSL kerb5 codes to match RFC 2712.
11595 * Make -nameopt work fully for req and add -reqopt switch.
11597 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11599 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11601 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11603 * Make sure tests can be performed even if the corresponding algorithms
11604 have been removed entirely. This was also the last step to make
11605 OpenSSL compilable with DJGPP under all reasonable conditions.
11607 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11609 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11610 to allow version independent disabling of normally unselected ciphers,
11611 which may be activated as a side-effect of selecting a single cipher.
11613 (E.g., cipher list string "RSA" enables ciphersuites that are left
11614 out of "ALL" because they do not provide symmetric encryption.
11615 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11617 *Lutz Jaenicke, Bodo Moeller*
11619 * Add appropriate support for separate platform-dependent build
11620 directories. The recommended way to make a platform-dependent
11621 build directory is the following (tested on Linux), maybe with
11624 # Place yourself outside of the OpenSSL source tree. In
11625 # this example, the environment variable OPENSSL_SOURCE
11626 # is assumed to contain the absolute OpenSSL source directory.
11627 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11628 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11629 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11630 mkdir -p `dirname $F`
11631 ln -s $OPENSSL_SOURCE/$F $F
11634 To be absolutely sure not to disturb the source tree, a "make clean"
11635 is a good thing. If it isn't successful, don't worry about it,
11636 it probably means the source directory is very clean.
11640 * Make sure any ENGINE control commands make local copies of string
11641 pointers passed to them whenever necessary. Otherwise it is possible
11642 the caller may have overwritten (or deallocated) the original string
11643 data when a later ENGINE operation tries to use the stored values.
11645 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11647 * Improve diagnostics in file reading and command-line digests.
11649 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11651 * Add AES modes CFB and OFB to the object database. Correct an
11652 error in AES-CFB decryption.
11656 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11657 allows existing EVP_CIPHER_CTX structures to be reused after
11658 calling `EVP_*Final()`. This behaviour is used by encryption
11659 BIOs and some applications. This has the side effect that
11660 applications must explicitly clean up cipher contexts with
11661 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11665 * Check the values of dna and dnb in bn_mul_recursive before calling
11666 bn_mul_comba (a non zero value means the a or b arrays do not contain
11667 n2 elements) and fallback to bn_mul_normal if either is not zero.
11671 * Fix escaping of non-ASCII characters when using the -subj option
11672 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11676 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11677 form for "surname", serialNumber has no short form.
11678 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11679 therefore remove "mail" short name for "internet 7".
11680 The OID for unique identifiers in X509 certificates is
11681 x500UniqueIdentifier, not uniqueIdentifier.
11682 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11686 * Add an "init" command to the ENGINE config module and auto initialize
11687 ENGINEs. Without any "init" command the ENGINE will be initialized
11688 after all ctrl commands have been executed on it. If init=1 the
11689 ENGINE is initialized at that point (ctrls before that point are run
11690 on the uninitialized ENGINE and after on the initialized one). If
11691 init=0 then the ENGINE will not be initialized at all.
11695 * Fix the 'app_verify_callback' interface so that the user-defined
11696 argument is actually passed to the callback: In the
11697 SSL_CTX_set_cert_verify_callback() prototype, the callback
11698 declaration has been changed from
11701 int (*cb)(X509_STORE_CTX *,void *);
11702 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11703 i=s->ctx->app_verify_callback(&ctx)
11704 has been changed into
11705 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11707 To update applications using SSL_CTX_set_cert_verify_callback(),
11708 a dummy argument can be added to their callback functions.
11710 *D. K. Smetters <smetters@parc.xerox.com>*
11712 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11714 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11716 * Add and OPENSSL_LOAD_CONF define which will cause
11717 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11718 This allows older applications to transparently support certain
11719 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11720 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11721 load the config file and OPENSSL_add_all_algorithms_conf() which will
11722 always load it have also been added.
11726 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11727 Adjust NIDs and EVP layer.
11729 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11731 * Config modules support in openssl utility.
11733 Most commands now load modules from the config file,
11734 though in a few (such as version) this isn't done
11735 because it couldn't be used for anything.
11737 In the case of ca and req the config file used is
11738 the same as the utility itself: that is the -config
11739 command line option can be used to specify an
11744 * Move default behaviour from OPENSSL_config(). If appname is NULL
11745 use "openssl_conf" if filename is NULL use default openssl config file.
11749 * Add an argument to OPENSSL_config() to allow the use of an alternative
11750 config section name. Add a new flag to tolerate a missing config file
11751 and move code to CONF_modules_load_file().
11755 * Support for crypto accelerator cards from Accelerated Encryption
11756 Processing, www.aep.ie. (Use engine 'aep')
11757 The support was copied from 0.9.6c [engine] and adapted/corrected
11758 to work with the new engine framework.
11760 *AEP Inc. and Richard Levitte*
11762 * Support for SureWare crypto accelerator cards from Baltimore
11763 Technologies. (Use engine 'sureware')
11764 The support was copied from 0.9.6c [engine] and adapted
11765 to work with the new engine framework.
11769 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11770 make the newer ENGINE framework commands for the CHIL engine work.
11772 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11774 * Make it possible to produce shared libraries on ReliantUNIX.
11776 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11778 * Add the configuration target debug-linux-ppro.
11779 Make 'openssl rsa' use the general key loading routines
11780 implemented in `apps.c`, and make those routines able to
11781 handle the key format FORMAT_NETSCAPE and the variant
11784 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11786 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11788 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11790 * Add -keyform to rsautl, and document -engine.
11792 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11794 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11795 BIO_R_NO_SUCH_FILE error code rather than the generic
11796 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11800 * Add new functions
11801 ERR_peek_last_error
11802 ERR_peek_last_error_line
11803 ERR_peek_last_error_line_data.
11804 These are similar to
11806 ERR_peek_error_line
11807 ERR_peek_error_line_data,
11808 but report on the latest error recorded rather than the first one
11809 still in the error queue.
11811 *Ben Laurie, Bodo Moeller*
11813 * default_algorithms option in ENGINE config module. This allows things
11815 default_algorithms = ALL
11816 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11820 * Preliminary ENGINE config module.
11824 * New experimental application configuration code.
11828 * Change the AES code to follow the same name structure as all other
11829 symmetric ciphers, and behave the same way. Move everything to
11830 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11832 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11834 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11836 *Ben Laurie and Theo de Raadt*
11838 * Add option to output public keys in req command.
11840 *Massimiliano Pala madwolf@openca.org*
11842 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11843 (up to about 10% better than before for P-192 and P-224).
11847 * New functions/macros
11849 SSL_CTX_set_msg_callback(ctx, cb)
11850 SSL_CTX_set_msg_callback_arg(ctx, arg)
11851 SSL_set_msg_callback(ssl, cb)
11852 SSL_set_msg_callback_arg(ssl, arg)
11854 to request calling a callback function
11856 void cb(int write_p, int version, int content_type,
11857 const void *buf, size_t len, SSL *ssl, void *arg)
11859 whenever a protocol message has been completely received
11860 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11861 protocol version according to which the SSL library interprets
11862 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11863 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11864 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11865 specification (change_cipher_spec(20), alert(21), handshake(22)).
11866 'buf' and 'len' point to the actual message, 'ssl' to the
11867 SSL object, and 'arg' is the application-defined value set by
11868 SSL[_CTX]_set_msg_callback_arg().
11870 'openssl s_client' and 'openssl s_server' have new '-msg' options
11871 to enable a callback that displays all protocol messages.
11875 * Change the shared library support so shared libraries are built as
11876 soon as the corresponding static library is finished, and thereby get
11877 openssl and the test programs linked against the shared library.
11878 This still only happens when the keyword "shard" has been given to
11879 the configuration scripts.
11881 NOTE: shared library support is still an experimental thing, and
11882 backward binary compatibility is still not guaranteed.
11884 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11886 * Add support for Subject Information Access extension.
11888 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11890 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11891 additional bytes when new memory had to be allocated, not just
11892 when reusing an existing buffer.
11896 * New command line and configuration option 'utf8' for the req command.
11897 This allows field values to be specified as UTF8 strings.
11901 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11902 runs for the former and machine-readable output for the latter.
11906 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11907 of the e-mail address in the DN (i.e., it will go into a certificate
11908 extension only). The new configuration file option 'email_in_dn = no'
11909 has the same effect.
11911 *Massimiliano Pala madwolf@openca.org*
11913 * Change all functions with names starting with `des_` to be starting
11914 with `DES_` instead. Add wrappers that are compatible with libdes,
11915 but are named `_ossl_old_des_*`. Finally, add macros that map the
11916 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11917 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11918 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11921 Since we provide two compatibility mappings, the user needs to
11922 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11923 compatibility is desired. The default (i.e., when that macro
11924 isn't defined) is OpenSSL 0.9.6c compatibility.
11926 There are also macros that enable and disable the support of old
11927 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11928 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11929 are defined, the default will apply: to support the old des routines.
11931 In either case, one must include openssl/des.h to get the correct
11932 definitions. Do not try to just include openssl/des_old.h, that
11935 NOTE: This is a major break of an old API into a new one. Software
11936 authors are encouraged to switch to the `DES_` style functions. Some
11937 time in the future, des_old.h and the libdes compatibility functions
11938 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11939 default), and then completely removed.
11943 * Test for certificates which contain unsupported critical extensions.
11944 If such a certificate is found during a verify operation it is
11945 rejected by default: this behaviour can be overridden by either
11946 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11947 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11948 X509_supported_extension() has also been added which returns 1 if a
11949 particular extension is supported.
11953 * Modify the behaviour of EVP cipher functions in similar way to digests
11954 to retain compatibility with existing code.
11958 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11959 compatibility with existing code. In particular the 'ctx' parameter does
11960 not have to be to be initialized before the call to EVP_DigestInit() and
11961 it is tidied up after a call to EVP_DigestFinal(). New function
11962 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11963 EVP_MD_CTX_copy() changed to not require the destination to be
11964 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11965 requires the destination to be valid.
11967 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11968 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11972 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11973 so that complete 'Handshake' protocol structures are kept in memory
11974 instead of overwriting 'msg_type' and 'length' with 'body' data.
11978 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11980 *Massimo Santin via Richard Levitte*
11982 * Major restructuring to the underlying ENGINE code. This includes
11983 reduction of linker bloat, separation of pure "ENGINE" manipulation
11984 (initialisation, etc) from functionality dealing with implementations
11985 of specific crypto interfaces. This change also introduces integrated
11986 support for symmetric ciphers and digest implementations - so ENGINEs
11987 can now accelerate these by providing EVP_CIPHER and EVP_MD
11988 implementations of their own. This is detailed in
11989 [crypto/engine/README.md](crypto/engine/README.md)
11990 as it couldn't be adequately described here. However, there are a few
11991 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11992 were changed in the original introduction of ENGINE code have now
11993 reverted back - the hooking from this code to ENGINE is now a good
11994 deal more passive and at run-time, operations deal directly with
11995 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11996 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11997 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11998 they were not being used by the framework as there is no concept of a
11999 BIGNUM_METHOD and they could not be generalised to the new
12000 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12001 ENGINE_cpy() has been removed as it cannot be consistently defined in
12006 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12010 * Change mkdef.pl to sort symbols that get the same entry number,
12011 and make sure the automatically generated functions `ERR_load_*`
12012 become part of libeay.num as well.
12016 * New function SSL_renegotiate_pending(). This returns true once
12017 renegotiation has been requested (either SSL_renegotiate() call
12018 or HelloRequest/ClientHello received from the peer) and becomes
12019 false once a handshake has been completed.
12020 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12021 sends a HelloRequest, but does not ensure that a handshake takes
12022 place. SSL_renegotiate_pending() is useful for checking if the
12023 client has followed the request.)
12027 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12028 By default, clients may request session resumption even during
12029 renegotiation (if session ID contexts permit); with this option,
12030 session resumption is possible only in the first handshake.
12032 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12033 more bits available for options that should not be part of
12034 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12038 * Add some demos for certificate and certificate request creation.
12042 * Make maximum certificate chain size accepted from the peer application
12043 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12044 "Douglas E. Engert" <deengert@anl.gov>.
12048 * Add support for shared libraries for Unixware-7
12049 (Boyd Lynn Gerber <gerberb@zenez.com>).
12053 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12054 be done prior to destruction. Use this to unload error strings from
12055 ENGINEs that load their own error strings. NB: This adds two new API
12056 functions to "get" and "set" this destroy handler in an ENGINE.
12060 * Alter all existing ENGINE implementations (except "openssl" and
12061 "openbsd") to dynamically instantiate their own error strings. This
12062 makes them more flexible to be built both as statically-linked ENGINEs
12063 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12064 Also, add stub code to each that makes building them as self-contained
12065 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12069 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12070 implementations into applications that are completely implemented in
12071 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12072 commands that can be used to configure what shared-library to load and
12073 to control aspects of the way it is handled. Also, made an update to
12074 the [README-Engine.md](README-Engine.md) file
12075 that brings its information up-to-date and
12076 provides some information and instructions on the "dynamic" ENGINE
12077 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12081 * Make it possible to unload ranges of ERR strings with a new
12082 "ERR_unload_strings" function.
12086 * Add a copy() function to EVP_MD.
12090 * Make EVP_MD routines take a context pointer instead of just the
12091 md_data void pointer.
12095 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12096 that the digest can only process a single chunk of data
12097 (typically because it is provided by a piece of
12098 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12099 is only going to provide a single chunk of data, and hence the
12100 framework needn't accumulate the data for oneshot drivers.
12104 * As with "ERR", make it possible to replace the underlying "ex_data"
12105 functions. This change also alters the storage and management of global
12106 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12107 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12108 index counters. The API functions that use this state have been changed
12109 to take a "class_index" rather than pointers to the class's local STACK
12110 and counter, and there is now an API function to dynamically create new
12111 classes. This centralisation allows us to (a) plug a lot of the
12112 thread-safety problems that existed, and (b) makes it possible to clean
12113 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12114 such data would previously have always leaked in application code and
12115 workarounds were in place to make the memory debugging turn a blind eye
12116 to it. Application code that doesn't use this new function will still
12117 leak as before, but their memory debugging output will announce it now
12118 rather than letting it slide.
12120 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12121 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12122 has a return value to indicate success or failure.
12126 * Make it possible to replace the underlying "ERR" functions such that the
12127 global state (2 LHASH tables and 2 locks) is only used by the "default"
12128 implementation. This change also adds two functions to "get" and "set"
12129 the implementation prior to it being automatically set the first time
12130 any other ERR function takes place. Ie. an application can call "get",
12131 pass the return value to a module it has just loaded, and that module
12132 can call its own "set" function using that value. This means the
12133 module's "ERR" operations will use (and modify) the error state in the
12134 application and not in its own statically linked copy of OpenSSL code.
12138 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12139 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12140 the operation, and provides a more encapsulated way for external code
12141 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12142 to use these functions rather than manually incrementing the counts.
12144 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12148 * Add EVP test program.
12152 * Add symmetric cipher support to ENGINE. Expect the API to change!
12156 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12157 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12158 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12159 These allow a CRL to be built without having to access X509_CRL fields
12160 directly. Modify 'ca' application to use new functions.
12164 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12165 bug workarounds. Rollback attack detection is a security feature.
12166 The problem will only arise on OpenSSL servers when TLSv1 is not
12167 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12168 Software authors not wanting to support TLSv1 will have special reasons
12169 for their choice and can explicitly enable this option.
12171 *Bodo Moeller, Lutz Jaenicke*
12173 * Rationalise EVP so it can be extended: don't include a union of
12174 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12175 (similar to those existing for EVP_CIPHER_CTX).
12180 EVP_MD_CTX_init(&md); /* new function call */
12181 EVP_DigestInit(&md, EVP_sha1());
12182 EVP_DigestUpdate(&md, in, len);
12183 EVP_DigestFinal(&md, out, NULL);
12184 EVP_MD_CTX_cleanup(&md); /* new function call */
12188 * Make DES key schedule conform to the usual scheme, as well as
12189 correcting its structure. This means that calls to DES functions
12190 now have to pass a pointer to a des_key_schedule instead of a
12191 plain des_key_schedule (which was actually always a pointer
12194 des_key_schedule ks;
12196 des_set_key_checked(..., &ks);
12197 des_ncbc_encrypt(..., &ks, ...);
12199 (Note that a later change renames 'des_...' into 'DES_...'.)
12203 * Initial reduction of linker bloat: the use of some functions, such as
12204 PEM causes large amounts of unused functions to be linked in due to
12205 poor organisation. For example pem_all.c contains every PEM function
12206 which has a knock on effect of linking in large amounts of (unused)
12207 ASN1 code. Grouping together similar functions and splitting unrelated
12208 functions prevents this.
12212 * Cleanup of EVP macros.
12216 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12217 correct `_ecb suffix`.
12221 * Add initial OCSP responder support to ocsp application. The
12222 revocation information is handled using the text based index
12223 use by the ca application. The responder can either handle
12224 requests generated internally, supplied in files (for example
12225 via a CGI script) or using an internal minimal server.
12229 * Add configuration choices to get zlib compression for TLS.
12233 * Changes to Kerberos SSL for RFC 2712 compliance:
12234 1. Implemented real KerberosWrapper, instead of just using
12235 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12236 2. Implemented optional authenticator field of KerberosWrapper.
12238 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12239 and authenticator structs; see crypto/krb5/.
12241 Generalized Kerberos calls to support multiple Kerberos libraries.
12242 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12243 via Richard Levitte*
12245 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12246 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12247 values for each of the key sizes rather than having just
12248 parameters (and 'speed' generating keys each time).
12252 * Speed up EVP routines.
12255 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12256 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12257 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12258 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12260 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12261 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12262 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12265 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12267 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12271 * Added the OS2-EMX target.
12273 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12275 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12276 New functions to support `NCONF `routines in extension code.
12277 New function `CONF_set_nconf()`
12278 to allow functions which take an `NCONF` to also handle the old `LHASH`
12279 structure: this means that the old `CONF` compatible routines can be
12280 retained (in particular w.rt. extensions) without having to duplicate the
12281 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12285 * Enhance the general user interface with mechanisms for inner control
12286 and with possibilities to have yes/no kind of prompts.
12290 * Change all calls to low level digest routines in the library and
12291 applications to use EVP. Add missing calls to HMAC_cleanup() and
12292 don't assume HMAC_CTX can be copied using memcpy().
12294 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12296 * Add the possibility to control engines through control names but with
12297 arbitrary arguments instead of just a string.
12298 Change the key loaders to take a UI_METHOD instead of a callback
12299 function pointer. NOTE: this breaks binary compatibility with earlier
12300 versions of OpenSSL [engine].
12301 Adapt the nCipher code for these new conditions and add a card insertion
12306 * Enhance the general user interface with mechanisms to better support
12307 dialog box interfaces, application-defined prompts, the possibility
12308 to use defaults (for example default passwords from somewhere else)
12309 and interrupts/cancellations.
12313 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12314 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12318 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12319 tidy up some unnecessarily weird code in 'sk_new()').
12321 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12323 * Change the key loading routines for ENGINEs to use the same kind
12324 callback (pem_password_cb) as all other routines that need this
12329 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12330 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12331 than this minimum value is recommended.
12335 * New random seeder for OpenVMS, using the system process statistics
12336 that are easily reachable.
12340 * Windows apparently can't transparently handle global
12341 variables defined in DLLs. Initialisations such as:
12343 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12345 won't compile. This is used by the any applications that need to
12346 declare their own ASN1 modules. This was fixed by adding the option
12347 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12348 needed for static libraries under Win32.
12352 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12353 setting of purpose and trust fields. New X509_STORE trust and
12354 purpose functions and tidy up setting in other SSL functions.
12358 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12359 structure. These are inherited by X509_STORE_CTX when it is
12360 initialised. This allows various defaults to be set in the
12361 X509_STORE structure (such as flags for CRL checking and custom
12362 purpose or trust settings) for functions which only use X509_STORE_CTX
12363 internally such as S/MIME.
12365 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12366 trust settings if they are not set in X509_STORE. This allows X509_STORE
12367 purposes and trust (in S/MIME for example) to override any set by default.
12369 Add command line options for CRL checking to smime, s_client and s_server
12374 * Initial CRL based revocation checking. If the CRL checking flag(s)
12375 are set then the CRL is looked up in the X509_STORE structure and
12376 its validity and signature checked, then if the certificate is found
12377 in the CRL the verify fails with a revoked error.
12379 Various new CRL related callbacks added to X509_STORE_CTX structure.
12381 Command line options added to 'verify' application to support this.
12383 This needs some additional work, such as being able to handle multiple
12384 CRLs with different times, extension based lookup (rather than just
12385 by subject name) and ultimately more complete V2 CRL extension
12390 * Add a general user interface API (crypto/ui/). This is designed
12391 to replace things like des_read_password and friends (backward
12392 compatibility functions using this new API are provided).
12393 The purpose is to remove prompting functions from the DES code
12394 section as well as provide for prompting through dialog boxes in
12395 a window system and the like.
12399 * Add "ex_data" support to ENGINE so implementations can add state at a
12400 per-structure level rather than having to store it globally.
12404 * Make it possible for ENGINE structures to be copied when retrieved by
12405 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12406 This causes the "original" ENGINE structure to act like a template,
12407 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12408 operational state can be localised to each ENGINE structure, despite the
12409 fact they all share the same "methods". New ENGINE structures returned in
12410 this case have no functional references and the return value is the single
12411 structural reference. This matches the single structural reference returned
12412 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12417 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12418 needs to match any other type at all we need to manually clear the
12423 * Changes to the "openssl engine" utility to include;
12424 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12425 about an ENGINE's available control commands.
12426 - executing control commands from command line arguments using the
12427 '-pre' and '-post' switches. '-post' is only used if '-t' is
12428 specified and the ENGINE is successfully initialised. The syntax for
12429 the individual commands are colon-separated, for example;
12430 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12434 * New dynamic control command support for ENGINEs. ENGINEs can now
12435 declare their own commands (numbers), names (strings), descriptions,
12436 and input types for run-time discovery by calling applications. A
12437 subset of these commands are implicitly classed as "executable"
12438 depending on their input type, and only these can be invoked through
12439 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12440 can be based on user input, config files, etc). The distinction is
12441 that "executable" commands cannot return anything other than a boolean
12442 result and can only support numeric or string input, whereas some
12443 discoverable commands may only be for direct use through
12444 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12445 pointers, or other custom uses. The "executable" commands are to
12446 support parameterisations of ENGINE behaviour that can be
12447 unambiguously defined by ENGINEs and used consistently across any
12448 OpenSSL-based application. Commands have been added to all the
12449 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12450 control over shared-library paths without source code alterations.
12454 * Changed all ENGINE implementations to dynamically allocate their
12455 ENGINEs rather than declaring them statically. Apart from this being
12456 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12457 this also allows the implementations to compile without using the
12458 internal engine_int.h header.
12462 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12463 'const' value. Any code that should be able to modify a RAND_METHOD
12464 should already have non-const pointers to it (ie. they should only
12465 modify their own ones).
12469 * Made a variety of little tweaks to the ENGINE code.
12470 - "atalla" and "ubsec" string definitions were moved from header files
12471 to C code. "nuron" string definitions were placed in variables
12472 rather than hard-coded - allowing parameterisation of these values
12473 later on via ctrl() commands.
12474 - Removed unused "#if 0"'d code.
12475 - Fixed engine list iteration code so it uses ENGINE_free() to release
12476 structural references.
12477 - Constified the RAND_METHOD element of ENGINE structures.
12478 - Constified various get/set functions as appropriate and added
12479 missing functions (including a catch-all ENGINE_cpy that duplicates
12480 all ENGINE values onto a new ENGINE except reference counts/state).
12481 - Removed NULL parameter checks in get/set functions. Setting a method
12482 or function to NULL is a way of cancelling out a previously set
12483 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12484 and doesn't justify the extra error symbols and code.
12485 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12486 flags from engine_int.h to engine.h.
12487 - Changed prototypes for ENGINE handler functions (init(), finish(),
12488 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12492 * Implement binary inversion algorithm for BN_mod_inverse in addition
12493 to the algorithm using long division. The binary algorithm can be
12494 used only if the modulus is odd. On 32-bit systems, it is faster
12495 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12496 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12497 up to 450 bits. In 64-bit environments, the binary algorithm
12498 appears to be advantageous for much longer moduli; here we use it
12499 for moduli up to 2048 bits.
12503 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12504 could not support the combine flag in choice fields.
12508 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12509 extensions from a certificate request to the certificate.
12513 * Allow multiple 'certopt' and 'nameopt' options to be separated
12514 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12515 file: this allows the display of the certificate about to be
12516 signed to be customised, to allow certain fields to be included
12517 or excluded and extension details. The old system didn't display
12518 multicharacter strings properly, omitted fields not in the policy
12519 and couldn't display additional details such as extensions.
12523 * Function EC_POINTs_mul for multiple scalar multiplication
12524 of an arbitrary number of elliptic curve points
12525 \sum scalars[i]*points[i],
12526 optionally including the generator defined for the EC_GROUP:
12527 scalar*generator + \sum scalars[i]*points[i].
12529 EC_POINT_mul is a simple wrapper function for the typical case
12530 that the point list has just one item (besides the optional
12535 * First EC_METHODs for curves over GF(p):
12537 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12538 operations and provides various method functions that can also
12539 operate with faster implementations of modular arithmetic.
12541 EC_GFp_mont_method() reuses most functions that are part of
12542 EC_GFp_simple_method, but uses Montgomery arithmetic.
12544 *Bodo Moeller; point addition and point doubling
12545 implementation directly derived from source code provided by
12546 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12548 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12549 crypto/ec/ec_lib.c):
12551 Curves are EC_GROUP objects (with an optional group generator)
12552 based on EC_METHODs that are built into the library.
12554 Points are EC_POINT objects based on EC_GROUP objects.
12556 Most of the framework would be able to handle curves over arbitrary
12557 finite fields, but as there are no obvious types for fields other
12558 than GF(p), some functions are limited to that for now.
12562 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12563 that the file contains a complete HTTP response.
12567 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12568 change the def and num file printf format specifier from "%-40sXXX"
12569 to "%-39s XXX". The latter will always guarantee a space after the
12570 field while the former will cause them to run together if the field
12571 is 40 of more characters long.
12575 * Constify the cipher and digest 'method' functions and structures
12576 and modify related functions to take constant EVP_MD and EVP_CIPHER
12581 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12582 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12586 * Modify `EVP_Digest*()` routines so they now return values. Although the
12587 internal software routines can never fail additional hardware versions
12592 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12594 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12595 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12598 ERR_R_NESTED_ASN1_ERROR
12600 ERR_R_MISSING_ASN1_EOS
12601 were 4 .. 9, conflicting with
12602 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12604 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12605 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12607 Add new error code 'ERR_R_INTERNAL_ERROR'.
12611 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12616 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12617 sets the subject name for a new request or supersedes the
12618 subject name in a given request. Formats that can be parsed are
12619 'CN=Some Name, OU=myOU, C=IT'
12621 'CN=Some Name/OU=myOU/C=IT'.
12623 Add options '-batch' and '-verbose' to 'openssl req'.
12625 *Massimiliano Pala <madwolf@hackmasters.net>*
12627 * Introduce the possibility to access global variables through
12628 functions on platform were that's the best way to handle exporting
12629 global variables in shared libraries. To enable this functionality,
12630 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12631 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12632 is normally done by Configure or something similar).
12634 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12635 in the source file (foo.c) like this:
12637 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12638 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12640 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12641 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12643 OPENSSL_DECLARE_GLOBAL(int,foo);
12644 #define foo OPENSSL_GLOBAL_REF(foo)
12645 OPENSSL_DECLARE_GLOBAL(double,bar);
12646 #define bar OPENSSL_GLOBAL_REF(bar)
12648 The #defines are very important, and therefore so is including the
12649 header file everywhere where the defined globals are used.
12651 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12652 of ASN.1 items, but that structure is a bit different.
12654 The largest change is in util/mkdef.pl which has been enhanced with
12655 better and easier to understand logic to choose which symbols should
12656 go into the Windows .def files as well as a number of fixes and code
12657 cleanup (among others, algorithm keywords are now sorted
12658 lexicographically to avoid constant rewrites).
12662 * In BN_div() keep a copy of the sign of 'num' before writing the
12663 result to 'rm' because if rm==num the value will be overwritten
12664 and produce the wrong result if 'num' is negative: this caused
12665 problems with BN_mod() and BN_nnmod().
12669 * Function OCSP_request_verify(). This checks the signature on an
12670 OCSP request and verifies the signer certificate. The signer
12671 certificate is just checked for a generic purpose and OCSP request
12676 * Add OCSP_check_validity() function to check the validity of OCSP
12677 responses. OCSP responses are prepared in real time and may only
12678 be a few seconds old. Simply checking that the current time lies
12679 between thisUpdate and nextUpdate max reject otherwise valid responses
12680 caused by either OCSP responder or client clock inaccuracy. Instead
12681 we allow thisUpdate and nextUpdate to fall within a certain period of
12682 the current time. The age of the response can also optionally be
12683 checked. Two new options -validity_period and -status_age added to
12688 * If signature or public key algorithm is unrecognized print out its
12689 OID rather that just UNKNOWN.
12693 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12694 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12695 ID to be generated from the issuer certificate alone which can then be
12696 passed to OCSP_id_issuer_cmp().
12700 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12701 ASN1 modules to export functions returning ASN1_ITEM pointers
12702 instead of the ASN1_ITEM structures themselves. This adds several
12703 new macros which allow the underlying ASN1 function/structure to
12704 be accessed transparently. As a result code should not use ASN1_ITEM
12705 references directly (such as &X509_it) but instead use the relevant
12706 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12707 use of the new ASN1 code on platforms where exporting structures
12708 is problematical (for example in shared libraries) but exporting
12709 functions returning pointers to structures is not.
12713 * Add support for overriding the generation of SSL/TLS session IDs.
12714 These callbacks can be registered either in an SSL_CTX or per SSL.
12715 The purpose of this is to allow applications to control, if they wish,
12716 the arbitrary values chosen for use as session IDs, particularly as it
12717 can be useful for session caching in multiple-server environments. A
12718 command-line switch for testing this (and any client code that wishes
12719 to use such a feature) has been added to "s_server".
12721 *Geoff Thorpe, Lutz Jaenicke*
12723 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12724 of the form `#if defined(...) || defined(...) || ...` and
12725 `#if !defined(...) && !defined(...) && ...`. This also avoids
12726 the growing number of special cases it was previously handling.
12730 * Make all configuration macros available for application by making
12731 sure they are available in opensslconf.h, by giving them names starting
12732 with `OPENSSL_` to avoid conflicts with other packages and by making
12733 sure e_os2.h will cover all platform-specific cases together with
12735 Additionally, it is now possible to define configuration/platform-
12736 specific names (called "system identities"). In the C code, these
12737 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12738 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12739 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12744 * New option -set_serial to 'req' and 'x509' this allows the serial
12745 number to use to be specified on the command line. Previously self
12746 signed certificates were hard coded with serial number 0 and the
12747 CA options of 'x509' had to use a serial number in a file which was
12752 * New options to 'ca' utility to support V2 CRL entry extensions.
12753 Currently CRL reason, invalidity date and hold instruction are
12754 supported. Add new CRL extensions to V3 code and some new objects.
12758 * New function EVP_CIPHER_CTX_set_padding() this is used to
12759 disable standard block padding (aka PKCS#5 padding) in the EVP
12760 API, which was previously mandatory. This means that the data is
12761 not padded in any way and so the total length much be a multiple
12762 of the block size, otherwise an error occurs.
12766 * Initial (incomplete) OCSP SSL support.
12770 * New function OCSP_parse_url(). This splits up a URL into its host,
12771 port and path components: primarily to parse OCSP URLs. New -url
12772 option to ocsp utility.
12776 * New nonce behavior. The return value of OCSP_check_nonce() now
12777 reflects the various checks performed. Applications can decide
12778 whether to tolerate certain situations such as an absent nonce
12779 in a response when one was present in a request: the ocsp application
12780 just prints out a warning. New function OCSP_add1_basic_nonce()
12781 this is to allow responders to include a nonce in a response even if
12782 the request is nonce-less.
12786 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12787 skipped when using openssl x509 multiple times on a single input file,
12788 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12792 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12793 set string type: to handle setting ASN1_TIME structures. Fix ca
12794 utility to correctly initialize revocation date of CRLs.
12798 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12799 the clients preferred ciphersuites and rather use its own preferences.
12800 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12801 Internet Explorer by ensuring unchanged hash method during stepup.
12802 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12806 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12807 to aes and add a new 'exist' option to print out symbols that don't
12812 * Additional options to ocsp utility to allow flags to be set and
12813 additional certificates supplied.
12817 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12818 OCSP client a number of certificate to only verify the response
12823 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12824 handle the new API. Currently only ECB, CBC modes supported. Add new
12827 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12828 Encryption Standard (AES) Ciphersuites for Transport Layer
12829 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12830 not enabled by default and were not part of the "ALL" ciphersuite
12831 alias because they were not yet official; they could be
12832 explicitly requested by specifying the "AESdraft" ciphersuite
12833 group alias. In the final release of OpenSSL 0.9.7, the group
12834 alias is called "AES" and is part of "ALL".)
12836 *Ben Laurie, Steve Henson, Bodo Moeller*
12838 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12839 request to response.
12843 * Functions for OCSP responders. OCSP_request_onereq_count(),
12844 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12845 extract information from a certificate request. OCSP_response_create()
12846 creates a response and optionally adds a basic response structure.
12847 OCSP_basic_add1_status() adds a complete single response to a basic
12848 response and returns the OCSP_SINGLERESP structure just added (to allow
12849 extensions to be included for example). OCSP_basic_add1_cert() adds a
12850 certificate to a basic response and OCSP_basic_sign() signs a basic
12851 response with various flags. New helper functions ASN1_TIME_check()
12852 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12853 (converts ASN1_TIME to GeneralizedTime).
12857 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12858 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12859 structure from a certificate. X509_pubkey_digest() digests the public_key
12860 contents: this is used in various key identifiers.
12864 * Make sk_sort() tolerate a NULL argument.
12866 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12868 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12869 passed by the function are trusted implicitly. If any of them signed the
12870 response then it is assumed to be valid and is not verified.
12874 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12875 to data. This was previously part of the PKCS7 ASN1 code. This
12876 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12877 *Steve Henson, reported by Kenneth R. Robinette
12878 <support@securenetterm.com>*
12880 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12881 routines: without these tracing memory leaks is very painful.
12882 Fix leaks in PKCS12 and PKCS7 routines.
12886 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12887 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12888 effectively meant GeneralizedTime would never be used. Now it
12889 is initialised to -1 but X509_time_adj() now has to check the value
12890 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12891 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12892 *Steve Henson, reported by Kenneth R. Robinette
12893 <support@securenetterm.com>*
12895 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12896 result in a zero length in the ASN1_INTEGER structure which was
12897 not consistent with the structure when d2i_ASN1_INTEGER() was used
12898 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12899 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12900 where it did not print out a minus for negative ASN1_INTEGER.
12904 * Add summary printout to ocsp utility. The various functions which
12905 convert status values to strings have been renamed to:
12906 OCSP_response_status_str(), OCSP_cert_status_str() and
12907 OCSP_crl_reason_str() and are no longer static. New options
12908 to verify nonce values and to disable verification. OCSP response
12909 printout format cleaned up.
12913 * Add additional OCSP certificate checks. These are those specified
12914 in RFC2560. This consists of two separate checks: the CA of the
12915 certificate being checked must either be the OCSP signer certificate
12916 or the issuer of the OCSP signer certificate. In the latter case the
12917 OCSP signer certificate must contain the OCSP signing extended key
12918 usage. This check is performed by attempting to match the OCSP
12919 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12920 in the OCSP_CERTID structures of the response.
12924 * Initial OCSP certificate verification added to OCSP_basic_verify()
12925 and related routines. This uses the standard OpenSSL certificate
12926 verify routines to perform initial checks (just CA validity) and
12927 to obtain the certificate chain. Then additional checks will be
12928 performed on the chain. Currently the root CA is checked to see
12929 if it is explicitly trusted for OCSP signing. This is used to set
12930 a root CA as a global signing root: that is any certificate that
12931 chains to that CA is an acceptable OCSP signing certificate.
12935 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12936 extensions from a separate configuration file.
12937 As when reading extensions from the main configuration file,
12938 the '-extensions ...' option may be used for specifying the
12941 *Massimiliano Pala <madwolf@comune.modena.it>*
12943 * New OCSP utility. Allows OCSP requests to be generated or
12944 read. The request can be sent to a responder and the output
12945 parsed, outputted or printed in text form. Not complete yet:
12946 still needs to check the OCSP response validity.
12950 * New subcommands for 'openssl ca':
12951 `openssl ca -status <serial>` prints the status of the cert with
12952 the given serial number (according to the index file).
12953 `openssl ca -updatedb` updates the expiry status of certificates
12956 *Massimiliano Pala <madwolf@comune.modena.it>*
12958 * New '-newreq-nodes' command option to CA.pl. This is like
12959 '-newreq', but calls 'openssl req' with the '-nodes' option
12960 so that the resulting key is not encrypted.
12962 *Damien Miller <djm@mindrot.org>*
12964 * New configuration for the GNU Hurd.
12966 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12968 * Initial code to implement OCSP basic response verify. This
12969 is currently incomplete. Currently just finds the signer's
12970 certificate and verifies the signature on the response.
12974 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12975 value of OPENSSLDIR. This is available via the new '-d' option
12976 to 'openssl version', and is also included in 'openssl version -a'.
12980 * Allowing defining memory allocation callbacks that will be given
12981 file name and line number information in additional arguments
12982 (a `const char*` and an int). The basic functionality remains, as
12983 well as the original possibility to just replace malloc(),
12984 realloc() and free() by functions that do not know about these
12985 additional arguments. To register and find out the current
12986 settings for extended allocation functions, the following
12987 functions are provided:
12989 CRYPTO_set_mem_ex_functions
12990 CRYPTO_set_locked_mem_ex_functions
12991 CRYPTO_get_mem_ex_functions
12992 CRYPTO_get_locked_mem_ex_functions
12994 These work the same way as CRYPTO_set_mem_functions and friends.
12995 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12996 extended allocation function is enabled.
12997 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12998 a conventional allocation function is enabled.
13000 *Richard Levitte, Bodo Moeller*
13002 * Finish off removing the remaining LHASH function pointer casts.
13003 There should no longer be any prototype-casting required when using
13004 the LHASH abstraction, and any casts that remain are "bugs". See
13005 the callback types and macros at the head of lhash.h for details
13006 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13010 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13011 If /dev/[u]random devices are not available or do not return enough
13012 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13014 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13015 /etc/entropy will be queried once each in this sequence, querying stops
13016 when enough entropy was collected without querying more sockets.
13020 * Change the Unix RAND_poll() variant to be able to poll several
13021 random devices, as specified by DEVRANDOM, until a sufficient amount
13022 of data has been collected. We spend at most 10 ms on each file
13023 (select timeout) and read in non-blocking mode. DEVRANDOM now
13024 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13025 (previously it was just the string "/dev/urandom"), so on typical
13026 platforms the 10 ms delay will never occur.
13027 Also separate out the Unix variant to its own file, rand_unix.c.
13028 For VMS, there's a currently-empty rand_vms.c.
13032 * Move OCSP client related routines to ocsp_cl.c. These
13033 provide utility functions which an application needing
13034 to issue a request to an OCSP responder and analyse the
13035 response will typically need: as opposed to those which an
13036 OCSP responder itself would need which will be added later.
13038 OCSP_request_sign() signs an OCSP request with an API similar
13039 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13040 response. OCSP_response_get1_basic() extracts basic response
13041 from response. OCSP_resp_find_status(): finds and extracts status
13042 information from an OCSP_CERTID structure (which will be created
13043 when the request structure is built). These are built from lower
13044 level functions which work on OCSP_SINGLERESP structures but
13045 won't normally be used unless the application wishes to examine
13046 extensions in the OCSP response for example.
13048 Replace nonce routines with a pair of functions.
13049 OCSP_request_add1_nonce() adds a nonce value and optionally
13050 generates a random value. OCSP_check_nonce() checks the
13051 validity of the nonce in an OCSP response.
13055 * Change function OCSP_request_add() to OCSP_request_add0_id().
13056 This doesn't copy the supplied OCSP_CERTID and avoids the
13057 need to free up the newly created id. Change return type
13058 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13059 This can then be used to add extensions to the request.
13060 Deleted OCSP_request_new(), since most of its functionality
13061 is now in OCSP_REQUEST_new() (and the case insensitive name
13062 clash) apart from the ability to set the request name which
13063 will be added elsewhere.
13067 * Update OCSP API. Remove obsolete extensions argument from
13068 various functions. Extensions are now handled using the new
13069 OCSP extension code. New simple OCSP HTTP function which
13070 can be used to send requests and parse the response.
13074 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13075 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13076 uses the special reorder version of SET OF to sort the attributes
13077 and reorder them to match the encoded order. This resolves a long
13078 standing problem: a verify on a PKCS7 structure just after signing
13079 it used to fail because the attribute order did not match the
13080 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13081 it uses the received order. This is necessary to tolerate some broken
13082 software that does not order SET OF. This is handled by encoding
13083 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13084 to produce the required SET OF.
13088 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13089 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13090 files to get correct declarations of the ASN.1 item variables.
13094 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13095 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13096 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13097 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13098 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13099 ASN1_ITEM and no wrapper functions.
13103 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13104 replace the old function pointer based I/O routines. Change most of
13105 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13109 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13110 lines, recognize more "algorithms" that can be deselected, and make
13111 it complain about algorithm deselection that isn't recognised.
13115 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13116 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13117 to use new functions. Add NO_ASN1_OLD which can be set to remove
13118 some old style ASN1 functions: this can be used to determine if old
13119 code will still work when these eventually go away.
13123 * New extension functions for OCSP structures, these follow the
13124 same conventions as certificates and CRLs.
13128 * New function X509V3_add1_i2d(). This automatically encodes and
13129 adds an extension. Its behaviour can be customised with various
13130 flags to append, replace or delete. Various wrappers added for
13131 certificates and CRLs.
13135 * Fix to avoid calling the underlying ASN1 print routine when
13136 an extension cannot be parsed. Correct a typo in the
13137 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13141 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13142 entries for variables.
13146 * Add functionality to `apps/openssl.c` for detecting locking
13147 problems: As the program is single-threaded, all we have
13148 to do is register a locking callback using an array for
13149 storing which locks are currently held by the program.
13153 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13154 SSL_get_ex_data_X509_STORE_idx(), which is used in
13155 ssl_verify_cert_chain() and thus can be called at any time
13156 during TLS/SSL handshakes so that thread-safety is essential.
13157 Unfortunately, the ex_data design is not at all suited
13158 for multi-threaded use, so it probably should be abolished.
13162 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13164 *Broadcom, tweaked and integrated by Geoff Thorpe*
13166 * Move common extension printing code to new function
13167 X509V3_print_extensions(). Reorganise OCSP print routines and
13168 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13172 * New function X509_signature_print() to remove duplication in some
13177 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13178 set (this was treated exactly the same as SET OF previously). This
13179 is used to reorder the STACK representing the structure to match the
13180 encoding. This will be used to get round a problem where a PKCS7
13181 structure which was signed could not be verified because the STACK
13182 order did not reflect the encoded order.
13186 * Reimplement the OCSP ASN1 module using the new code.
13190 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13191 for its ASN1 operations. The old style function pointers still exist
13192 for now but they will eventually go away.
13196 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13197 completely replaces the old ASN1 functionality with a table driven
13198 encoder and decoder which interprets an ASN1_ITEM structure describing
13199 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13200 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13201 has also been converted to the new form.
13205 * Change BN_mod_exp_recp so that negative moduli are tolerated
13206 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13207 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13208 for negative moduli.
13212 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13213 of not touching the result's sign bit.
13217 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13222 * Changed the LHASH code to use prototypes for callbacks, and created
13223 macros to declare and implement thin (optionally static) functions
13224 that provide type-safety and avoid function pointer casting for the
13225 type-specific callbacks.
13229 * Added Kerberos Cipher Suites to be used with TLS, as written in
13231 *Veers Staats <staatsvr@asc.hpc.mil>,
13232 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13234 * Reformat the FAQ so the different questions and answers can be divided
13235 in sections depending on the subject.
13239 * Have the zlib compression code load ZLIB.DLL dynamically under
13244 * New function BN_mod_sqrt for computing square roots modulo a prime
13245 (using the probabilistic Tonelli-Shanks algorithm unless
13246 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13247 be handled deterministically).
13249 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13251 * Make BN_mod_inverse faster by explicitly handling small quotients
13252 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13253 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13257 * New function BN_kronecker.
13261 * Fix BN_gcd so that it works on negative inputs; the result is
13262 positive unless both parameters are zero.
13263 Previously something reasonably close to an infinite loop was
13264 possible because numbers could be growing instead of shrinking
13265 in the implementation of Euclid's algorithm.
13269 * Fix BN_is_word() and BN_is_one() macros to take into account the
13270 sign of the number in question.
13272 Fix BN_is_word(a,w) to work correctly for w == 0.
13274 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13275 because its test if the absolute value of 'a' equals 'w'.
13276 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13277 it exists mostly for use in the implementations of BN_is_zero(),
13278 BN_is_one(), and BN_is_word().
13282 * New function BN_swap.
13286 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13287 the exponentiation functions are more likely to produce reasonable
13288 results on negative inputs.
13292 * Change BN_mod_mul so that the result is always non-negative.
13293 Previously, it could be negative if one of the factors was negative;
13294 I don't think anyone really wanted that behaviour.
13298 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13299 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13300 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13301 and add new functions:
13310 BN_mod_lshift1_quick
13312 BN_mod_lshift_quick
13314 These functions always generate non-negative results.
13316 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13317 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13319 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13320 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13321 be reduced modulo `m`.
13323 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13326 The following entry accidentally appeared in the CHANGES file
13327 distributed with OpenSSL 0.9.7. The modifications described in
13328 it do *not* apply to OpenSSL 0.9.7.
13330 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13331 was actually never needed) and in BN_mul(). The removal in BN_mul()
13332 required a small change in bn_mul_part_recursive() and the addition
13333 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13334 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13335 bn_sub_words() and bn_add_words() except they take arrays with
13341 * In 'openssl passwd', verify passwords read from the terminal
13342 unless the '-salt' option is used (which usually means that
13343 verification would just waste user's time since the resulting
13344 hash is going to be compared with some given password hash)
13345 or the new '-noverify' option is used.
13347 This is an incompatible change, but it does not affect
13348 non-interactive use of 'openssl passwd' (passwords on the command
13349 line, '-stdin' option, '-in ...' option) and thus should not
13350 cause any problems.
13354 * Remove all references to RSAref, since there's no more need for it.
13358 * Make DSO load along a path given through an environment variable
13359 (SHLIB_PATH) with shl_load().
13363 * Constify the ENGINE code as a result of BIGNUM constification.
13364 Also constify the RSA code and most things related to it. In a
13365 few places, most notable in the depth of the ASN.1 code, ugly
13366 casts back to non-const were required (to be solved at a later
13371 * Make it so the openssl application has all engines loaded by default.
13375 * Constify the BIGNUM routines a little more.
13379 * Add the following functions:
13381 ENGINE_load_cswift()
13383 ENGINE_load_atalla()
13384 ENGINE_load_nuron()
13385 ENGINE_load_builtin_engines()
13387 That way, an application can itself choose if external engines that
13388 are built-in in OpenSSL shall ever be used or not. The benefit is
13389 that applications won't have to be linked with libdl or other dso
13390 libraries unless it's really needed.
13392 Changed 'openssl engine' to load all engines on demand.
13393 Changed the engine header files to avoid the duplication of some
13394 declarations (they differed!).
13398 * 'openssl engine' can now list capabilities.
13402 * Better error reporting in 'openssl engine'.
13406 * Never call load_dh_param(NULL) in s_server.
13410 * Add engine application. It can currently list engines by name and
13411 identity, and test if they are actually available.
13415 * Improve RPM specification file by forcing symbolic linking and making
13416 sure the installed documentation is also owned by root.root.
13418 *Damien Miller <djm@mindrot.org>*
13420 * Give the OpenSSL applications more possibilities to make use of
13421 keys (public as well as private) handled by engines.
13425 * Add OCSP code that comes from CertCo.
13429 * Add VMS support for the Rijndael code.
13433 * Added untested support for Nuron crypto accelerator.
13437 * Add support for external cryptographic devices. This code was
13438 previously distributed separately as the "engine" branch.
13440 *Geoff Thorpe, Richard Levitte*
13442 * Rework the filename-translation in the DSO code. It is now possible to
13443 have far greater control over how a "name" is turned into a filename
13444 depending on the operating environment and any oddities about the
13445 different shared library filenames on each system.
13449 * Support threads on FreeBSD-elf in Configure.
13453 * Fix for SHA1 assembly problem with MASM: it produces
13454 warnings about corrupt line number information when assembling
13455 with debugging information. This is caused by the overlapping
13458 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13461 NCONF_get_number() has no error checking at all. As a replacement,
13462 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13463 promoted strongly. The old NCONF_get_number is kept around for
13464 binary backward compatibility.
13465 Make it possible for methods to load from something other than a BIO,
13466 by providing a function pointer that is given a name instead of a BIO.
13467 For example, this could be used to load configuration data from an
13472 * Fix for non blocking accept BIOs. Added new I/O special reason
13473 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13474 with non blocking I/O was not possible because no retry code was
13475 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13480 * Added the beginnings of Rijndael support.
13484 * Fix for bug in DirectoryString mask setting. Add support for
13485 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13486 to allow certificate printing to more controllable, additional
13487 'certopt' option to 'x509' to allow new printing options to be
13492 * Clean old EAY MD5 hack from e_os.h.
13496 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13498 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13499 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13501 *Joe Orton, Steve Henson*
13503 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13505 * Fix additional bug revealed by the NISCC test suite:
13507 Stop bug triggering large recursion when presented with
13508 certain ASN.1 tags ([CVE-2003-0851])
13512 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13514 * Fix various bugs revealed by running the NISCC test suite:
13516 Stop out of bounds reads in the ASN1 code when presented with
13517 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13519 If verify callback ignores invalid public key errors don't try to check
13520 certificate signature with the NULL public key.
13524 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13525 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13530 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13531 extra data after the compression methods not only for TLS 1.0
13532 but also for SSL 3.0 (as required by the specification).
13534 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13536 * Change X509_certificate_type() to mark the key as exported/exportable
13537 when it's 512 *bits* long, not 512 bytes.
13541 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13543 * Countermeasure against the Klima-Pokorny-Rosa extension of
13544 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13545 a protocol version number mismatch like a decryption error
13546 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13550 * Turn on RSA blinding by default in the default implementation
13551 to avoid a timing attack. Applications that don't want it can call
13552 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13553 They would be ill-advised to do so in most cases.
13555 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13557 * Change RSA blinding code so that it works when the PRNG is not
13558 seeded (in this case, the secret RSA exponent is abused as
13559 an unpredictable seed -- if it is not unpredictable, there
13560 is no point in blinding anyway). Make RSA blinding thread-safe
13561 by remembering the creator's thread ID in rsa->blinding and
13562 having all other threads use local one-time blinding factors
13563 (this requires more computation than sharing rsa->blinding, but
13564 avoids excessive locking; and if an RSA object is not shared
13565 between threads, blinding will still be very fast).
13569 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13571 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13572 via timing by performing a MAC computation even if incorrect
13573 block cipher padding has been found. This is a countermeasure
13574 against active attacks where the attacker has to distinguish
13575 between bad padding and a MAC verification error. ([CVE-2003-0078])
13577 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13578 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13579 Martin Vuagnoux (EPFL, Ilion)*
13581 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13583 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13584 memory from its contents. This is done with a counter that will
13585 place alternating values in each byte. This can be used to solve
13586 two issues: 1) the removal of calls to memset() by highly optimizing
13587 compilers, and 2) cleansing with other values than 0, since those can
13588 be read through on certain media, for example a swap space on disk.
13592 * Bugfix: client side session caching did not work with external caching,
13593 because the session->cipher setting was not restored when reloading
13594 from the external cache. This problem was masked, when
13595 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13596 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13600 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13601 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13603 *Zeev Lieber <zeev-l@yahoo.com>*
13605 * Undo an undocumented change introduced in 0.9.6e which caused
13606 repeated calls to OpenSSL_add_all_ciphers() and
13607 OpenSSL_add_all_digests() to be ignored, even after calling
13612 * Change the default configuration reader to deal with last line not
13613 being properly terminated.
13617 * Change X509_NAME_cmp() so it applies the special rules on handling
13618 DN values that are of type PrintableString, as well as RDNs of type
13619 emailAddress where the value has the type ia5String.
13621 *stefank@valicert.com via Richard Levitte*
13623 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13624 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13625 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13626 the bitwise-OR of the two for use by the majority of applications
13627 wanting this behaviour, and update the docs. The documented
13628 behaviour and actual behaviour were inconsistent and had been
13629 changing anyway, so this is more a bug-fix than a behavioural
13632 *Geoff Thorpe, diagnosed by Nadav Har'El*
13634 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13635 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13639 * Fix initialization code race conditions in
13640 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13641 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13642 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13643 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13644 ssl2_get_cipher_by_char(),
13645 ssl3_get_cipher_by_char().
13647 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13649 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13650 the cached sessions are flushed, as the remove_cb() might use ex_data
13651 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13652 (see [openssl.org #212]).
13654 *Geoff Thorpe, Lutz Jaenicke*
13656 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13657 length, instead of the encoding length to d2i_ASN1_OBJECT.
13661 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13663 * [In 0.9.6g-engine release:]
13664 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13666 *Lynn Gazis <lgazis@rainbow.com>*
13668 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13670 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13671 and get fix the header length calculation.
13672 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13673 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13675 * Use proper error handling instead of 'assertions' in buffer
13676 overflow checks added in 0.9.6e. This prevents DoS (the
13677 assertions could call abort()).
13679 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13681 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13683 * Add various sanity checks to asn1_get_length() to reject
13684 the ASN1 length bytes if they exceed sizeof(long), will appear
13685 negative or the content length exceeds the length of the
13688 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13690 * Fix cipher selection routines: ciphers without encryption had no flags
13691 for the cipher strength set and where therefore not handled correctly
13692 by the selection routines (PR #130).
13696 * Fix EVP_dsa_sha macro.
13701 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13702 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13703 that was added in OpenSSL 0.9.6d.
13705 As the countermeasure turned out to be incompatible with some
13706 broken SSL implementations, the new option is part of SSL_OP_ALL.
13707 SSL_OP_ALL is usually employed when compatibility with weird SSL
13708 implementations is desired (e.g. '-bugs' option to 's_client' and
13709 's_server'), so the new option is automatically set in many
13714 * Changes in security patch:
13716 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13717 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13718 Air Force Materiel Command, USAF, under agreement number
13721 * Add various sanity checks to asn1_get_length() to reject
13722 the ASN1 length bytes if they exceed sizeof(long), will appear
13723 negative or the content length exceeds the length of the
13724 supplied buffer. ([CVE-2002-0659])
13726 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13728 * Assertions for various potential buffer overflows, not known to
13729 happen in practice.
13731 *Ben Laurie (CHATS)*
13733 * Various temporary buffers to hold ASCII versions of integers were
13734 too small for 64 bit platforms. ([CVE-2002-0655])
13735 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13737 * Remote buffer overflow in SSL3 protocol - an attacker could
13738 supply an oversized session ID to a client. ([CVE-2002-0656])
13740 *Ben Laurie (CHATS)*
13742 * Remote buffer overflow in SSL2 protocol - an attacker could
13743 supply an oversized client master key. ([CVE-2002-0656])
13745 *Ben Laurie (CHATS)*
13747 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13749 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13750 encoded as NULL) with id-dsa-with-sha1.
13752 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13754 * Check various `X509_...()` return values in `apps/req.c`.
13756 *Nils Larsch <nla@trustcenter.de>*
13758 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13759 an end-of-file condition would erroneously be flagged, when the CRLF
13760 was just at the end of a processed block. The bug was discovered when
13761 processing data through a buffering memory BIO handing the data to a
13762 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13763 <ptsekov@syntrex.com> and Nedelcho Stanev.
13767 * Implement a countermeasure against a vulnerability recently found
13768 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13769 before application data chunks to avoid the use of known IVs
13770 with data potentially chosen by the attacker.
13774 * Fix length checks in ssl3_get_client_hello().
13778 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13779 to prevent ssl3_read_internal() from incorrectly assuming that
13780 ssl3_read_bytes() found application data while handshake
13781 processing was enabled when in fact s->s3->in_read_app_data was
13782 merely automatically cleared during the initial handshake.
13784 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13786 * Fix object definitions for Private and Enterprise: they were not
13787 recognized in their shortname (=lowercase) representation. Extend
13788 obj_dat.pl to issue an error when using undefined keywords instead
13789 of silently ignoring the problem (Svenning Sorensen
13790 <sss@sss.dnsalias.net>).
13794 * Fix DH_generate_parameters() so that it works for 'non-standard'
13795 generators, i.e. generators other than 2 and 5. (Previously, the
13796 code did not properly initialise the 'add' and 'rem' values to
13797 BN_generate_prime().)
13799 In the new general case, we do not insist that 'generator' is
13800 actually a primitive root: This requirement is rather pointless;
13801 a generator of the order-q subgroup is just as good, if not
13806 * Map new X509 verification errors to alerts. Discovered and submitted by
13807 Tom Wu <tom@arcot.com>.
13811 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13812 returning non-zero before the data has been completely received
13813 when using non-blocking I/O.
13815 *Bodo Moeller; problem pointed out by John Hughes*
13817 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13819 *Ben Laurie, Lutz Jaenicke*
13821 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13822 Yoram Zahavi <YoramZ@gilian.com>).
13826 * Add information about CygWin 1.3 and on, and preserve proper
13827 configuration for the versions before that.
13829 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13831 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13832 check whether we deal with a copy of a session and do not delete from
13833 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13834 <izhar@checkpoint.com>.
13838 * Do not store session data into the internal session cache, if it
13839 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13840 flag is set). Proposed by Aslam <aslam@funk.com>.
13844 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13849 * [In 0.9.6d-engine release:]
13850 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13852 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13854 * Add the configuration target linux-s390x.
13856 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13858 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13859 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13860 variable as an indication that a ClientHello message has been
13861 received. As the flag value will be lost between multiple
13862 invocations of ssl3_accept when using non-blocking I/O, the
13863 function may not be aware that a handshake has actually taken
13864 place, thus preventing a new session from being added to the
13867 To avoid this problem, we now set s->new_session to 2 instead of
13868 using a local variable.
13870 *Lutz Jaenicke, Bodo Moeller*
13872 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13873 if the SSL_R_LENGTH_MISMATCH error is detected.
13875 *Geoff Thorpe, Bodo Moeller*
13877 * New 'shared_ldflag' column in Configure platform table.
13881 * Fix EVP_CIPHER_mode macro.
13883 *"Dan S. Camper" <dan@bti.net>*
13885 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13886 type, we must throw them away by setting rr->length to 0.
13888 *D P Chang <dpc@qualys.com>*
13890 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13892 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13893 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13894 worked incorrectly for those cases where range = `10..._2` and
13895 `3*range` is two bits longer than range.)
13899 * Only add signing time to PKCS7 structures if it is not already
13904 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13905 OBJ_ld_ce should be OBJ_id_ce.
13906 Also some ip-pda OIDs in crypto/objects/objects.txt were
13907 incorrect (cf. RFC 3039).
13909 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13911 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13912 returns early because it has nothing to do.
13914 *Andy Schneider <andy.schneider@bjss.co.uk>*
13916 * [In 0.9.6c-engine release:]
13917 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13919 *Andy Schneider <andy.schneider@bjss.co.uk>*
13921 * [In 0.9.6c-engine release:]
13922 Add support for Cryptographic Appliance's keyserver technology.
13923 (Use engine 'keyclient')
13925 *Cryptographic Appliances and Geoff Thorpe*
13927 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13928 is called via tools/c89.sh because arguments have to be
13929 rearranged (all '-L' options must appear before the first object
13932 *Richard Shapiro <rshapiro@abinitio.com>*
13934 * [In 0.9.6c-engine release:]
13935 Add support for Broadcom crypto accelerator cards, backported
13938 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13940 * [In 0.9.6c-engine release:]
13941 Add support for SureWare crypto accelerator cards from
13942 Baltimore Technologies. (Use engine 'sureware')
13944 *Baltimore Technologies and Mark Cox*
13946 * [In 0.9.6c-engine release:]
13947 Add support for crypto accelerator cards from Accelerated
13948 Encryption Processing, www.aep.ie. (Use engine 'aep')
13950 *AEP Inc. and Mark Cox*
13952 * Add a configuration entry for gcc on UnixWare.
13954 *Gary Benson <gbenson@redhat.com>*
13956 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13957 messages are stored in a single piece (fixed-length part and
13958 variable-length part combined) and fix various bugs found on the way.
13962 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13963 instead. BIO_gethostbyname() does not know what timeouts are
13964 appropriate, so entries would stay in cache even when they have
13966 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13968 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13969 faced with a pathologically small ClientHello fragment that does
13970 not contain client_version: Instead of aborting with an error,
13971 simply choose the highest available protocol version (i.e.,
13972 TLS 1.0 unless it is disabled). In practice, ClientHello
13973 messages are never sent like this, but this change gives us
13974 strictly correct behaviour at least for TLS.
13978 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13979 never resets s->method to s->ctx->method when called from within
13980 one of the SSL handshake functions.
13982 *Bodo Moeller; problem pointed out by Niko Baric*
13984 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13985 (sent using the client's version number) if client_version is
13986 smaller than the protocol version in use. Also change
13987 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13988 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13989 the client will at least see that alert.
13993 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13998 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13999 client receives HelloRequest while in a handshake.
14001 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14003 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14004 should end in 'break', not 'goto end' which circumvents various
14005 cleanups done in state SSL_ST_OK. But session related stuff
14006 must be disabled for SSL_ST_OK in the case that we just sent a
14009 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14010 before just sending a HelloRequest.
14012 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14014 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14015 reveal whether illegal block cipher padding was found or a MAC
14016 verification error occurred. (Neither SSLerr() codes nor alerts
14017 are directly visible to potential attackers, but the information
14018 may leak via logfiles.)
14020 Similar changes are not required for the SSL 2.0 implementation
14021 because the number of padding bytes is sent in clear for SSL 2.0,
14022 and the extra bytes are just ignored. However ssl/s2_pkt.c
14023 failed to verify that the purported number of padding bytes is in
14028 * Add OpenUNIX-8 support including shared libraries
14029 (Boyd Lynn Gerber <gerberb@zenez.com>).
14033 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14034 'wristwatch attack' using huge encoding parameters (cf.
14035 James H. Manger's CRYPTO 2001 paper). Note that the
14036 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14037 encoding parameters and hence was not vulnerable.
14041 * BN_sqr() bug fix.
14043 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14045 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14046 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14047 followed by modular reduction.
14049 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14051 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14052 equivalent based on BN_pseudo_rand() instead of BN_rand().
14056 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14057 This function was broken, as the check for a new client hello message
14058 to handle SGC did not allow these large messages.
14059 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14063 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14067 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14068 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14072 * Rework the configuration and shared library support for Tru64 Unix.
14073 The configuration part makes use of modern compiler features and
14074 still retains old compiler behavior for those that run older versions
14075 of the OS. The shared library support part includes a variant that
14076 uses the RPATH feature, and is available through the special
14077 configuration target "alpha-cc-rpath", which will never be selected
14080 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14082 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14083 with the same message size as in ssl3_get_certificate_request().
14084 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14085 messages might inadvertently be reject as too long.
14087 *Petr Lampa <lampa@fee.vutbr.cz>*
14089 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14093 * Modified SSL library such that the verify_callback that has been set
14094 specifically for an SSL object with SSL_set_verify() is actually being
14095 used. Before the change, a verify_callback set with this function was
14096 ignored and the verify_callback() set in the SSL_CTX at the time of
14097 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14098 to allow the necessary settings.
14102 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14103 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14104 done automatically (in contradiction to the requirements of the C
14105 standard). This made problems when used from OpenSSH.
14109 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14110 dh->length and always used
14112 BN_rand_range(priv_key, dh->p).
14114 BN_rand_range() is not necessary for Diffie-Hellman, and this
14115 specific range makes Diffie-Hellman unnecessarily inefficient if
14116 dh->length (recommended exponent length) is much smaller than the
14117 length of dh->p. We could use BN_rand_range() if the order of
14118 the subgroup was stored in the DH structure, but we only have
14123 BN_rand(priv_key, l, ...)
14125 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14132 RSA_eay_public_encrypt
14133 RSA_eay_private_decrypt
14134 RSA_eay_private_encrypt (signing)
14135 RSA_eay_public_decrypt (signature verification)
14137 (default implementations for RSA_public_encrypt,
14138 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14139 always reject numbers >= n.
14143 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14144 to synchronize access to 'locking_thread'. This is necessary on
14145 systems where access to 'locking_thread' (an 'unsigned long'
14146 variable) is not atomic.
14150 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14151 *before* setting the 'crypto_lock_rand' flag. The previous code had
14152 a race condition if 0 is a valid thread ID.
14154 *Travis Vitek <vitek@roguewave.com>*
14156 * Add support for shared libraries under Irix.
14158 *Albert Chin-A-Young <china@thewrittenword.com>*
14160 * Add configuration option to build on Linux on both big-endian and
14161 little-endian MIPS.
14163 *Ralf Baechle <ralf@uni-koblenz.de>*
14165 * Add the possibility to create shared libraries on HP-UX.
14169 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14171 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14172 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14173 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14174 PRNG state recovery was possible based on the output of
14175 one PRNG request appropriately sized to gain knowledge on
14176 'md' followed by enough consecutive 1-byte PRNG requests
14177 to traverse all of 'state'.
14179 1. When updating 'md_local' (the current thread's copy of 'md')
14180 during PRNG output generation, hash all of the previous
14181 'md_local' value, not just the half used for PRNG output.
14183 2. Make the number of bytes from 'state' included into the hash
14184 independent from the number of PRNG bytes requested.
14186 The first measure alone would be sufficient to avoid
14187 Markku-Juhani's attack. (Actually it had never occurred
14188 to me that the half of 'md_local' used for chaining was the
14189 half from which PRNG output bytes were taken -- I had always
14190 assumed that the secret half would be used.) The second
14191 measure makes sure that additional data from 'state' is never
14192 mixed into 'md_local' in small portions; this heuristically
14193 further strengthens the PRNG.
14197 * Fix crypto/bn/asm/mips3.s.
14201 * When only the key is given to "enc", the IV is undefined. Print out
14202 an error message in this case.
14206 * Handle special case when X509_NAME is empty in X509 printing routines.
14210 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14211 positive and less than q.
14215 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14216 used: it isn't thread safe and the add_lock_callback should handle
14219 *Paul Rose <Paul.Rose@bridge.com>*
14221 * Verify that incoming data obeys the block size in
14222 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14228 *Ulf Möller, Bodo Möller*
14230 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14231 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14232 when fixing the server behaviour for backwards-compatible 'client
14233 hello' messages. (Note that the attack is impractical against
14234 SSL 3.0 and TLS 1.0 anyway because length and version checking
14235 means that the probability of guessing a valid ciphertext is
14236 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14239 Before 0.9.5, the countermeasure (hide the error by generating a
14240 random 'decryption result') did not work properly because
14241 ERR_clear_error() was missing, meaning that SSL_get_error() would
14242 detect the supposedly ignored error.
14244 Both problems are now fixed.
14248 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14249 (previously it was 1024).
14253 * Fix for compatibility mode trust settings: ignore trust settings
14254 unless some valid trust or reject settings are present.
14258 * Fix for blowfish EVP: its a variable length cipher.
14262 * Fix various bugs related to DSA S/MIME verification. Handle missing
14263 parameters in DSA public key structures and return an error in the
14264 DSA routines if parameters are absent.
14268 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14269 in the current directory if neither $RANDFILE nor $HOME was set.
14270 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14271 caused some confusion to Windows users who haven't defined $HOME.
14272 Thus RAND_file_name() is changed again: e_os.h can define a
14273 DEFAULT_HOME, which will be used if $HOME is not set.
14274 For Windows, we use "C:"; on other platforms, we still require
14275 environment variables.
14277 * Move 'if (!initialized) RAND_poll()' into regions protected by
14278 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14279 having multiple threads call RAND_poll() concurrently.
14283 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14284 combination of a flag and a thread ID variable.
14285 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14286 flag), *other* threads can enter ssleay_add_bytes without obeying
14287 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14288 that they do not hold after the first thread unsets add_do_not_lock).
14292 * Change bctest again: '-x' expressions are not available in all
14293 versions of 'test'.
14297 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14299 * Fix a couple of memory leaks in PKCS7_dataDecode()
14301 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14303 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14304 the default extension for executables, if any. Also, make the perl
14305 scripts that use symlink() to test if it really exists and use "cp"
14306 if it doesn't. All this made OpenSSL compilable and installable in
14311 * Fix for asn1_GetSequence() for indefinite length constructed data.
14312 If SEQUENCE is length is indefinite just set c->slen to the total
14313 amount of data available.
14315 *Steve Henson, reported by shige@FreeBSD.org*
14317 *This change does not apply to 0.9.7.*
14319 * Change bctest to avoid here-documents inside command substitution
14320 (workaround for FreeBSD /bin/sh bug).
14321 For compatibility with Ultrix, avoid shell functions (introduced
14322 in the bctest version that searches along $PATH).
14326 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14327 with des_encrypt() defined on some operating systems, like Solaris
14332 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14333 On the Importance of Eliminating Errors in Cryptographic
14334 Computations, J. Cryptology 14 (2001) 2, 101-119,
14335 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14339 * MIPS assembler BIGNUM division bug fix.
14343 * Disabled incorrect Alpha assembler code.
14347 * Fix PKCS#7 decode routines so they correctly update the length
14348 after reading an EOC for the EXPLICIT tag.
14352 *This change does not apply to 0.9.7.*
14354 * Fix bug in PKCS#12 key generation routines. This was triggered
14355 if a 3DES key was generated with a 0 initial byte. Include
14356 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14357 (but broken) behaviour.
14361 * Enhance bctest to search for a working bc along $PATH and print
14364 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14366 * Fix memory leaks in err.c: free err_data string if necessary;
14367 don't write to the wrong index in ERR_set_error_data.
14371 * Implement ssl23_peek (analogous to ssl23_read), which previously
14376 * Replace rdtsc with `_emit` statements for VC++ version 5.
14378 *Jeremy Cooper <jeremy@baymoo.org>*
14380 * Make it possible to reuse SSLv2 sessions.
14384 * In copy_email() check for >= 0 as a return value for
14385 X509_NAME_get_index_by_NID() since 0 is a valid index.
14387 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14389 * Avoid coredump with unsupported or invalid public keys by checking if
14390 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14391 PKCS7_verify() fails with non detached data.
14395 * Don't use getenv in library functions when run as setuid/setgid.
14396 New function OPENSSL_issetugid().
14400 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14401 due to incorrect handling of multi-threading:
14403 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14405 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14407 3. Count how many times MemCheck_off() has been called so that
14408 nested use can be treated correctly. This also avoids
14409 inband-signalling in the previous code (which relied on the
14410 assumption that thread ID 0 is impossible).
14414 * Add "-rand" option also to s_client and s_server.
14418 * Fix CPU detection on Irix 6.x.
14419 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14420 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14422 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14427 *This change does not apply to 0.9.7.*
14429 * Use the cached encoding of an X509_NAME structure rather than
14430 copying it. This is apparently the reason for the libsafe "errors"
14431 but the code is actually correct.
14435 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14436 Bleichenbacher's DSA attack.
14437 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14438 to be set and top=0 forces the highest bit to be set; top=-1 is new
14439 and leaves the highest bit random.
14441 *Ulf Moeller, Bodo Moeller*
14443 * In the `NCONF_...`-based implementations for `CONF_...` queries
14444 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14445 a temporary CONF structure with the data component set to NULL
14446 (which gives segmentation faults in lh_retrieve).
14447 Instead, use NULL for the CONF pointer in CONF_get_string and
14448 CONF_get_number (which may use environment variables) and directly
14449 return NULL from CONF_get_section.
14453 * Fix potential buffer overrun for EBCDIC.
14457 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14458 keyUsage if basicConstraints absent for a CA.
14462 * Make SMIME_write_PKCS7() write mail header values with a format that
14463 is more generally accepted (no spaces before the semicolon), since
14464 some programs can't parse those values properly otherwise. Also make
14465 sure BIO's that break lines after each write do not create invalid
14470 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14471 macros previously used would not encode an empty SEQUENCE OF
14472 and break the signature.
14476 *This change does not apply to 0.9.7.*
14478 * Zero the premaster secret after deriving the master secret in
14483 * Add some EVP_add_digest_alias registrations (as found in
14484 OpenSSL_add_all_digests()) to SSL_library_init()
14485 aka OpenSSL_add_ssl_algorithms(). This provides improved
14486 compatibility with peers using X.509 certificates
14487 with unconventional AlgorithmIdentifier OIDs.
14491 * Fix for Irix with NO_ASM.
14493 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14495 * ./config script fixes.
14497 *Ulf Moeller, Richard Levitte*
14499 * Fix 'openssl passwd -1'.
14503 * Change PKCS12_key_gen_asc() so it can cope with non null
14504 terminated strings whose length is passed in the passlen
14505 parameter, for example from PEM callbacks. This was done
14506 by adding an extra length parameter to asc2uni().
14508 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14510 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14511 call failed, free the DSA structure.
14515 * Fix to uni2asc() to cope with zero length Unicode strings.
14516 These are present in some PKCS#12 files.
14520 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14521 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14522 when writing a 32767 byte record.
14524 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14526 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14527 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14529 (RSA objects have a reference count access to which is protected
14530 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14531 so they are meant to be shared between threads.)
14532 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14533 "Reddie, Steven" <Steven.Reddie@ca.com>*
14535 * Fix a deadlock in CRYPTO_mem_leaks().
14539 * Use better test patterns in bntest.
14543 * rand_win.c fix for Borland C.
14547 * BN_rshift bugfix for n == 0.
14551 * Add a 'bctest' script that checks for some known 'bc' bugs
14552 so that 'make test' does not abort just because 'bc' is broken.
14556 * Store verify_result within SSL_SESSION also for client side to
14557 avoid potential security hole. (Re-used sessions on the client side
14558 always resulted in verify_result==X509_V_OK, not using the original
14559 result of the server certificate verification.)
14563 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14564 SSL3_RT_APPLICATION_DATA, return 0.
14565 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14570 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14571 releases, have been re-implemented by renaming the previous
14572 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14573 and ssl3_read_internal, respectively, and adding 'peek' parameters
14574 to them. The new ssl[23]_{read,peek} functions are calls to
14575 ssl[23]_read_internal with the 'peek' flag set appropriately.
14576 A 'peek' parameter has also been added to ssl3_read_bytes, which
14577 does the actual work for ssl3_read_internal.
14581 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14582 the method-specific "init()" handler. Also clean up ex_data after
14583 calling the method-specific "finish()" handler. Previously, this was
14584 happening the other way round.
14588 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14589 The previous value, 12, was not always sufficient for BN_mod_exp().
14593 * Make sure that shared libraries get the internal name engine with
14594 the full version number and not just 0. This should mark the
14595 shared libraries as not backward compatible. Of course, this should
14596 be changed again when we can guarantee backward binary compatibility.
14600 * Fix typo in get_cert_by_subject() in by_dir.c
14602 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14604 * Rework the system to generate shared libraries:
14606 - Make note of the expected extension for the shared libraries and
14607 if there is a need for symbolic links from for example libcrypto.so.0
14608 to libcrypto.so.0.9.7. There is extended info in Configure for
14611 - Make as few rebuilds of the shared libraries as possible.
14613 - Still avoid linking the OpenSSL programs with the shared libraries.
14615 - When installing, install the shared libraries separately from the
14620 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14622 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14623 and not in SSL_clear because the latter is also used by the
14624 accept/connect functions; previously, the settings made by
14625 SSL_set_read_ahead would be lost during the handshake.
14627 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14629 * Correct util/mkdef.pl to be selective about disabled algorithms.
14630 Previously, it would create entries for disabled algorithms no
14635 * Added several new manual pages for SSL_* function.
14639 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14641 * In ssl23_get_client_hello, generate an error message when faced
14642 with an initial SSL 3.0/TLS record that is too small to contain the
14643 first two bytes of the ClientHello message, i.e. client_version.
14644 (Note that this is a pathologic case that probably has never happened
14645 in real life.) The previous approach was to use the version number
14646 from the record header as a substitute; but our protocol choice
14647 should not depend on that one because it is not authenticated
14648 by the Finished messages.
14652 * More robust randomness gathering functions for Windows.
14654 *Jeffrey Altman <jaltman@columbia.edu>*
14656 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14657 not set then we don't setup the error code for issuer check errors
14658 to avoid possibly overwriting other errors which the callback does
14659 handle. If an application does set the flag then we assume it knows
14660 what it is doing and can handle the new informational codes
14665 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14666 a general "ANY" type, as such it should be able to decode anything
14667 including tagged types. However it didn't check the class so it would
14668 wrongly interpret tagged types in the same way as their universal
14669 counterpart and unknown types were just rejected. Changed so that the
14670 tagged and unknown types are handled in the same way as a SEQUENCE:
14671 that is the encoding is stored intact. There is also a new type
14672 "V_ASN1_OTHER" which is used when the class is not universal, in this
14673 case we have no idea what the actual type is so we just lump them all
14678 * On VMS, stdout may very well lead to a file that is written to
14679 in a record-oriented fashion. That means that every write() will
14680 write a separate record, which will be read separately by the
14681 programs trying to read from it. This can be very confusing.
14683 The solution is to put a BIO filter in the way that will buffer
14684 text until a linefeed is reached, and then write everything a
14685 line at a time, so every record written will be an actual line,
14686 not chunks of lines and not (usually doesn't happen, but I've
14687 seen it once) several lines in one record. BIO_f_linebuffer() is
14690 Currently, it's a VMS-only method, because that's where it has
14691 been tested well enough.
14695 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14696 it can return incorrect results.
14697 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14698 but it was in 0.9.6-beta[12].)
14702 * Disable the check for content being present when verifying detached
14703 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14704 include zero length content when signing messages.
14708 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14709 BIO_ctrl (for BIO pairs).
14713 * Add DSO method for VMS.
14717 * Bug fix: Montgomery multiplication could produce results with the
14722 * Add RPM specification openssl.spec and modify it to build three
14723 packages. The default package contains applications, application
14724 documentation and run-time libraries. The devel package contains
14725 include files, static libraries and function documentation. The
14726 doc package contains the contents of the doc directory. The original
14727 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14731 * Add a large number of documentation files for many SSL routines.
14733 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14735 * Add a configuration entry for Sony News 4.
14737 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14739 * Don't set the two most significant bits to one when generating a
14740 random number < q in the DSA library.
14744 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14745 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14746 the underlying transport is blocking) if a handshake took place.
14747 (The default behaviour is needed by applications such as s_client
14748 and s_server that use select() to determine when to use SSL_read;
14749 but for applications that know in advance when to expect data, it
14750 just makes things more complicated.)
14754 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14759 * Add a few more EBCDIC conditionals that make `req` and `x509`
14760 work better on such systems.
14762 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14764 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14765 Update PKCS12_parse() so it copies the friendlyName and the
14766 keyid to the certificates aux info.
14770 * Fix bug in PKCS7_verify() which caused an infinite loop
14771 if there was more than one signature.
14773 *Sven Uszpelkat <su@celocom.de>*
14775 * Major change in util/mkdef.pl to include extra information
14776 about each symbol, as well as presenting variables as well
14777 as functions. This change means that there's n more need
14778 to rebuild the .num files when some algorithms are excluded.
14782 * Allow the verify time to be set by an application,
14783 rather than always using the current time.
14787 * Phase 2 verify code reorganisation. The certificate
14788 verify code now looks up an issuer certificate by a
14789 number of criteria: subject name, authority key id
14790 and key usage. It also verifies self signed certificates
14791 by the same criteria. The main comparison function is
14792 X509_check_issued() which performs these checks.
14794 Lot of changes were necessary in order to support this
14795 without completely rewriting the lookup code.
14797 Authority and subject key identifier are now cached.
14799 The LHASH 'certs' is X509_STORE has now been replaced
14800 by a STACK_OF(X509_OBJECT). This is mainly because an
14801 LHASH can't store or retrieve multiple objects with
14802 the same hash value.
14804 As a result various functions (which were all internal
14805 use only) have changed to handle the new X509_STORE
14806 structure. This will break anything that messed round
14807 with X509_STORE internally.
14809 The functions X509_STORE_add_cert() now checks for an
14810 exact match, rather than just subject name.
14812 The X509_STORE API doesn't directly support the retrieval
14813 of multiple certificates matching a given criteria, however
14814 this can be worked round by performing a lookup first
14815 (which will fill the cache with candidate certificates)
14816 and then examining the cache for matches. This is probably
14817 the best we can do without throwing out X509_LOOKUP
14818 entirely (maybe later...).
14820 The X509_VERIFY_CTX structure has been enhanced considerably.
14822 All certificate lookup operations now go via a get_issuer()
14823 callback. Although this currently uses an X509_STORE it
14824 can be replaced by custom lookups. This is a simple way
14825 to bypass the X509_STORE hackery necessary to make this
14826 work and makes it possible to use more efficient techniques
14827 in future. A very simple version which uses a simple
14828 STACK for its trusted certificate store is also provided
14829 using X509_STORE_CTX_trusted_stack().
14831 The verify_cb() and verify() callbacks now have equivalents
14832 in the X509_STORE_CTX structure.
14834 X509_STORE_CTX also has a 'flags' field which can be used
14835 to customise the verify behaviour.
14839 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14840 excludes S/MIME capabilities.
14844 * When a certificate request is read in keep a copy of the
14845 original encoding of the signed data and use it when outputting
14846 again. Signatures then use the original encoding rather than
14847 a decoded, encoded version which may cause problems if the
14848 request is improperly encoded.
14852 * For consistency with other BIO_puts implementations, call
14853 buffer_write(b, ...) directly in buffer_puts instead of calling
14856 In BIO_puts, increment b->num_write as in BIO_write.
14858 *Peter.Sylvester@EdelWeb.fr*
14860 * Fix BN_mul_word for the case where the word is 0. (We have to use
14861 BN_zero, we may not return a BIGNUM with an array consisting of
14862 words set to zero.)
14866 * Avoid calling abort() from within the library when problems are
14867 detected, except if preprocessor symbols have been defined
14868 (such as REF_CHECK, BN_DEBUG etc.).
14872 * New openssl application 'rsautl'. This utility can be
14873 used for low level RSA operations. DER public key
14874 BIO/fp routines also added.
14878 * New Configure entry and patches for compiling on QNX 4.
14880 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14882 * A demo state-machine implementation was sponsored by
14883 Nuron (<http://www.nuron.com/>) and is now available in
14884 demos/state_machine.
14888 * New options added to the 'dgst' utility for signature
14889 generation and verification.
14893 * Unrecognized PKCS#7 content types are now handled via a
14894 catch all ASN1_TYPE structure. This allows unsupported
14895 types to be stored as a "blob" and an application can
14896 encode and decode it manually.
14900 * Fix various signed/unsigned issues to make a_strex.c
14901 compile under VC++.
14903 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14905 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14906 length if passed a buffer. ASN1_INTEGER_to_BN failed
14907 if passed a NULL BN and its argument was negative.
14909 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14911 * Modification to PKCS#7 encoding routines to output definite
14912 length encoding. Since currently the whole structures are in
14913 memory there's not real point in using indefinite length
14914 constructed encoding. However if OpenSSL is compiled with
14915 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14919 * Added BIO_vprintf() and BIO_vsnprintf().
14923 * Added more prefixes to parse for in the strings written
14924 through a logging bio, to cover all the levels that are available
14925 through syslog. The prefixes are now:
14927 PANIC, EMERG, EMR => LOG_EMERG
14928 ALERT, ALR => LOG_ALERT
14929 CRIT, CRI => LOG_CRIT
14930 ERROR, ERR => LOG_ERR
14931 WARNING, WARN, WAR => LOG_WARNING
14932 NOTICE, NOTE, NOT => LOG_NOTICE
14933 INFO, INF => LOG_INFO
14934 DEBUG, DBG => LOG_DEBUG
14936 and as before, if none of those prefixes are present at the
14937 beginning of the string, LOG_ERR is chosen.
14939 On Win32, the `LOG_*` levels are mapped according to this:
14941 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14942 LOG_WARNING => EVENTLOG_WARNING_TYPE
14943 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14947 * Made it possible to reconfigure with just the configuration
14948 argument "reconf" or "reconfigure". The command line arguments
14949 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14950 and are retrieved from there when reconfiguring.
14956 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14958 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14962 * The obj_dat.pl script was messing up the sorting of object
14963 names. The reason was that it compared the quoted version
14964 of strings as a result "OCSP" > "OCSP Signing" because
14965 " > SPACE. Changed script to store unquoted versions of
14966 names and add quotes on output. It was also omitting some
14967 names from the lookup table if they were given a default
14968 value (that is if SN is missing it is given the same
14969 value as LN and vice versa), these are now added on the
14970 grounds that if an object has a name we should be able to
14971 look it up. Finally added warning output when duplicate
14972 short or long names are found.
14976 * Changes needed for Tandem NSK.
14978 *Scott Uroff <scott@xypro.com>*
14980 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14981 RSA_padding_check_SSLv23(), special padding was never detected
14982 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14983 version rollback attacks was not effective.
14985 In s23_clnt.c, don't use special rollback-attack detection padding
14986 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14987 client; similarly, in s23_srvr.c, don't do the rollback check if
14988 SSL 2.0 is the only protocol enabled in the server.
14992 * Make it possible to get hexdumps of unprintable data with 'openssl
14993 asn1parse'. By implication, the functions ASN1_parse_dump() and
14994 BIO_dump_indent() are added.
14998 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14999 these print out strings and name structures based on various
15000 flags including RFC2253 support and proper handling of
15001 multibyte characters. Added options to the 'x509' utility
15002 to allow the various flags to be set.
15006 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15007 Also change the functions X509_cmp_current_time() and
15008 X509_gmtime_adj() work with an ASN1_TIME structure,
15009 this will enable certificates using GeneralizedTime in validity
15010 dates to be checked.
15014 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15015 negative public key encodings) on by default,
15016 NO_NEG_PUBKEY_BUG can be set to disable it.
15020 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15021 content octets. An i2c_ASN1_OBJECT is unnecessary because
15022 the encoding can be trivially obtained from the structure.
15026 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15027 not read locks (`CRYPTO_r_[un]lock`).
15031 * A first attempt at creating official support for shared
15032 libraries through configuration. I've kept it so the
15033 default is static libraries only, and the OpenSSL programs
15034 are always statically linked for now, but there are
15035 preparations for dynamic linking in place.
15036 This has been tested on Linux and Tru64.
15040 * Randomness polling function for Win9x, as described in:
15041 Peter Gutmann, Software Generation of Practically Strong
15046 * Fix so PRNG is seeded in req if using an already existing
15051 * New options to smime application. -inform and -outform
15052 allow alternative formats for the S/MIME message including
15053 PEM and DER. The -content option allows the content to be
15054 specified separately. This should allow things like Netscape
15055 form signing output easier to verify.
15059 * Fix the ASN1 encoding of tags using the 'long form'.
15063 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15064 STRING types. These convert content octets to and from the
15065 underlying type. The actual tag and length octets are
15066 already assumed to have been read in and checked. These
15067 are needed because all other string types have virtually
15068 identical handling apart from the tag. By having versions
15069 of the ASN1 functions that just operate on content octets
15070 IMPLICIT tagging can be handled properly. It also allows
15071 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15072 and ASN1_INTEGER are identical apart from the tag.
15076 * Change the handling of OID objects as follows:
15078 - New object identifiers are inserted in objects.txt, following
15079 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15080 - objects.pl is used to process obj_mac.num and create a new
15082 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15085 This is currently kind of a hack, and the perl code in objects.pl
15086 isn't very elegant, but it works as I intended. The simplest way
15087 to check that it worked correctly is to look in obj_dat.h and
15088 check the array nid_objs and make sure the objects haven't moved
15089 around (this is important!). Additions are OK, as well as
15090 consistent name changes.
15094 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15098 * Addition of the command line parameter '-rand file' to 'openssl req'.
15099 The given file adds to whatever has already been seeded into the
15100 random pool through the RANDFILE configuration file option or
15101 environment variable, or the default random state file.
15105 * mkstack.pl now sorts each macro group into lexical order.
15106 Previously the output order depended on the order the files
15107 appeared in the directory, resulting in needless rewriting
15112 * Patches to make OpenSSL compile under Win32 again. Mostly
15113 work arounds for the VC++ problem that it treats func() as
15114 func(void). Also stripped out the parts of mkdef.pl that
15115 added extra typesafe functions: these no longer exist.
15119 * Reorganisation of the stack code. The macros are now all
15120 collected in safestack.h . Each macro is defined in terms of
15121 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15122 DEBUG_SAFESTACK is now handled in terms of function casts,
15123 this has the advantage of retaining type safety without the
15124 use of additional functions. If DEBUG_SAFESTACK is not defined
15125 then the non typesafe macros are used instead. Also modified the
15126 mkstack.pl script to handle the new form. Needs testing to see
15127 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15128 the default if no major problems. Similar behaviour for ASN1_SET_OF
15129 and PKCS12_STACK_OF.
15133 * When some versions of IIS use the 'NET' form of private key the
15134 key derivation algorithm is different. Normally MD5(password) is
15135 used as a 128 bit RC4 key. In the modified case
15136 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15137 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15138 as the old Netscape_RSA functions except they have an additional
15139 'sgckey' parameter which uses the modified algorithm. Also added
15140 an -sgckey command line option to the rsa utility. Thanks to
15141 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15142 algorithm to openssl-dev.
15146 * The evp_local.h macros were using 'c.##kname' which resulted in
15147 invalid expansion on some systems (SCO 5.0.5 for example).
15148 Corrected to 'c.kname'.
15150 *Phillip Porch <root@theporch.com>*
15152 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15153 a STACK of email addresses from a certificate or request, these look
15154 in the subject name and the subject alternative name extensions and
15155 omit any duplicate addresses.
15159 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15160 This makes DSA verification about 2 % faster.
15164 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15165 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15166 plus overhead for 1024 bit moduli).
15167 This makes exponentiations about 0.5 % faster for 1024 bit
15168 exponents (as measured by "openssl speed rsa2048").
15172 * Rename memory handling macros to avoid conflicts with other
15174 Malloc => OPENSSL_malloc
15175 Malloc_locked => OPENSSL_malloc_locked
15176 Realloc => OPENSSL_realloc
15177 Free => OPENSSL_free
15181 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15182 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15186 * CygWin32 support.
15188 *John Jarvie <jjarvie@newsguy.com>*
15190 * The type-safe stack code has been rejigged. It is now only compiled
15191 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15192 by default all type-specific stack functions are "#define"d back to
15193 standard stack functions. This results in more streamlined output
15194 but retains the type-safety checking possibilities of the original
15199 * The STACK code has been cleaned up, and certain type declarations
15200 that didn't make a lot of sense have been brought in line. This has
15201 also involved a cleanup of sorts in safestack.h to more correctly
15202 map type-safe stack functions onto their plain stack counterparts.
15203 This work has also resulted in a variety of "const"ifications of
15204 lots of the code, especially `_cmp` operations which should normally
15205 be prototyped with "const" parameters anyway.
15209 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15210 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15211 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15212 where all of 'md' is used each time the PRNG is used, but 'state'
15213 is used only indexed by a cyclic counter. As entropy may not be
15214 well distributed from the beginning, 'md' is important as a
15215 chaining variable. However, the output function chains only half
15216 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15217 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15218 in all of 'state' being rewritten, with the new values depending
15219 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15223 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15224 the handshake is continued after ssl_verify_cert_chain();
15225 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15226 can lead to 'unexplainable' connection aborts later.
15228 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15230 * Major EVP API cipher revision.
15231 Add hooks for extra EVP features. This allows various cipher
15232 parameters to be set in the EVP interface. Support added for variable
15233 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15234 setting of RC2 and RC5 parameters.
15236 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15239 Remove lots of duplicated code from the EVP library. For example *every*
15240 cipher init() function handles the 'iv' in the same way according to the
15241 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15242 for CFB and OFB modes they zero ctx->num.
15244 New functionality allows removal of S/MIME code RC2 hack.
15246 Most of the routines have the same form and so can be declared in terms
15249 By shifting this to the top level EVP_CipherInit() it can be removed from
15250 all individual ciphers. If the cipher wants to handle IVs or keys
15251 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15254 Change lots of functions like EVP_EncryptUpdate() to now return a
15255 value: although software versions of the algorithms cannot fail
15256 any installed hardware versions can.
15260 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15261 this option is set, tolerate broken clients that send the negotiated
15262 protocol version number instead of the requested protocol version
15267 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15268 i.e. non-zero for export ciphersuites, zero otherwise.
15269 Previous versions had this flag inverted, inconsistent with
15270 rsa_tmp_cb (..._TMP_RSA_CB).
15272 *Bodo Moeller; problem reported by Amit Chopra*
15274 * Add missing DSA library text string. Work around for some IIS
15275 key files with invalid SEQUENCE encoding.
15279 * Add a document (doc/standards.txt) that list all kinds of standards
15280 and so on that are implemented in OpenSSL.
15284 * Enhance c_rehash script. Old version would mishandle certificates
15285 with the same subject name hash and wouldn't handle CRLs at all.
15286 Added -fingerprint option to crl utility, to support new c_rehash
15291 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15295 * Fix for SSL server purpose checking. Server checking was
15296 rejecting certificates which had extended key usage present
15297 but no ssl client purpose.
15299 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15301 * Make PKCS#12 code work with no password. The PKCS#12 spec
15302 is a little unclear about how a blank password is handled.
15303 Since the password in encoded as a BMPString with terminating
15304 double NULL a zero length password would end up as just the
15305 double NULL. However no password at all is different and is
15306 handled differently in the PKCS#12 key generation code. NS
15307 treats a blank password as zero length. MSIE treats it as no
15308 password on export: but it will try both on import. We now do
15309 the same: PKCS12_parse() tries zero length and no password if
15310 the password is set to "" or NULL (NULL is now a valid password:
15311 it wasn't before) as does the pkcs12 application.
15315 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15316 perror when PEM_read_bio_X509_REQ fails, the error message must
15317 be obtained from the error queue.
15321 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15322 it in ERR_remove_state if appropriate, and change ERR_get_state
15323 accordingly to avoid race conditions (this is necessary because
15324 thread_hash is no longer constant once set).
15328 * Bugfix for linux-elf makefile.one.
15332 * RSA_get_default_method() will now cause a default
15333 RSA_METHOD to be chosen if one doesn't exist already.
15334 Previously this was only set during a call to RSA_new()
15335 or RSA_new_method(NULL) meaning it was possible for
15336 RSA_get_default_method() to return NULL.
15340 * Added native name translation to the existing DSO code
15341 that will convert (if the flag to do so is set) filenames
15342 that are sufficiently small and have no path information
15343 into a canonical native form. Eg. "blah" converted to
15344 "libblah.so" or "blah.dll" etc.
15348 * New function ERR_error_string_n(e, buf, len) which is like
15349 ERR_error_string(e, buf), but writes at most 'len' bytes
15350 including the 0 terminator. For ERR_error_string_n, 'buf'
15353 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15355 * CONF library reworked to become more general. A new CONF
15356 configuration file reader "class" is implemented as well as a
15357 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15358 old `CONF_*` functions are still there, but are reimplemented to
15359 work in terms of the new functions. Also, a set of functions
15360 to handle the internal storage of the configuration data is
15361 provided to make it easier to write new configuration file
15362 reader "classes" (I can definitely see something reading a
15363 configuration file in XML format, for example), called `_CONF_*`,
15364 or "the configuration storage API"...
15366 The new configuration file reading functions are:
15368 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15369 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15371 NCONF_default, NCONF_WIN32
15373 NCONF_dump_fp, NCONF_dump_bio
15375 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15376 NCONF_new creates a new CONF object. This works in the same way
15377 as other interfaces in OpenSSL, like the BIO interface.
15378 `NCONF_dump_*` dump the internal storage of the configuration file,
15379 which is useful for debugging. All other functions take the same
15380 arguments as the old `CONF_*` functions with the exception of the
15381 first that must be a `CONF *` instead of a `LHASH *`.
15383 To make it easier to use the new classes with the old `CONF_*` functions,
15384 the function CONF_set_default_method is provided.
15388 * Add '-tls1' option to 'openssl ciphers', which was already
15389 mentioned in the documentation but had not been implemented.
15390 (This option is not yet really useful because even the additional
15391 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15395 * Initial DSO code added into libcrypto for letting OpenSSL (and
15396 OpenSSL-based applications) load shared libraries and bind to
15397 them in a portable way.
15399 *Geoff Thorpe, with contributions from Richard Levitte*
15401 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15403 * Make sure _lrotl and _lrotr are only used with MSVC.
15405 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15406 (the default implementation of RAND_status).
15408 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15409 to '-clrext' (= clear extensions), as intended and documented.
15410 *Bodo Moeller; inconsistency pointed out by Michael Attili
15411 <attili@amaxo.com>*
15413 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15414 was larger than the MD block size.
15416 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15418 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15419 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15420 using the passed key: if the passed key was a private key the result
15421 of X509_print(), for example, would be to print out all the private key
15426 * des_quad_cksum() byte order bug fix.
15427 *Ulf Möller, using the problem description in krb4-0.9.7, where
15428 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15430 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15433 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15435 * For easily testing in shell scripts whether some command
15436 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15437 returns with exit code 0 iff no command of the given name is available.
15438 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15439 the output goes to stdout and nothing is printed to stderr.
15440 Additional arguments are always ignored.
15442 Since for each cipher there is a command of the same name,
15443 the 'no-cipher' compilation switches can be tested this way.
15445 ('openssl no-XXX' is not able to detect pseudo-commands such
15446 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15450 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15454 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15455 is set; it will be thrown away anyway because each handshake creates
15457 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15458 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15459 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15460 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15464 * New s_client option -ign_eof: EOF at stdin is ignored, and
15465 'Q' and 'R' lose their special meanings (quit/renegotiate).
15466 This is part of what -quiet does; unlike -quiet, -ign_eof
15467 does not suppress any output.
15471 * Add compatibility options to the purpose and trust code. The
15472 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15473 accepts a certificate or CA, this was the previous behaviour,
15474 with all the associated security issues.
15476 X509_TRUST_COMPAT is the old trust behaviour: only and
15477 automatically trust self signed roots in certificate store. A
15478 new trust setting X509_TRUST_DEFAULT is used to specify that
15479 a purpose has no associated trust setting and it should instead
15480 use the value in the default purpose.
15484 * Fix the PKCS#8 DSA private key code so it decodes keys again
15485 and fix a memory leak.
15489 * In util/mkerr.pl (which implements 'make errors'), preserve
15490 reason strings from the previous version of the .c file, as
15491 the default to have only downcase letters (and digits) in
15492 automatically generated reasons codes is not always appropriate.
15496 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15497 using strerror. Previously, ERR_reason_error_string() returned
15498 library names as reason strings for SYSerr; but SYSerr is a special
15499 case where small numbers are errno values, not library numbers.
15503 * Add '-dsaparam' option to 'openssl dhparam' application. This
15504 converts DSA parameters into DH parameters. (When creating parameters,
15505 DSA_generate_parameters is used.)
15509 * Include 'length' (recommended exponent length) in C code generated
15510 by 'openssl dhparam -C'.
15514 * The second argument to set_label in perlasm was already being used
15515 so couldn't be used as a "file scope" flag. Moved to third argument
15520 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15521 instead of RAND_bytes for encryption IVs and salts.
15525 * Include RAND_status() into RAND_METHOD instead of implementing
15526 it only for md_rand.c Otherwise replacing the PRNG by calling
15527 RAND_set_rand_method would be impossible.
15531 * Don't let DSA_generate_key() enter an infinite loop if the random
15532 number generation fails.
15536 * New 'rand' application for creating pseudo-random output.
15540 * Added configuration support for Linux/IA64
15542 *Rolf Haberrecker <rolf@suse.de>*
15544 * Assembler module support for Mingw32.
15548 * Shared library support for HPUX (in shlib/).
15550 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15552 * Shared library support for Solaris gcc.
15554 *Lutz Behnke <behnke@trustcenter.de>*
15556 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15558 * PKCS7_encrypt() was adding text MIME headers twice because they
15559 were added manually and by SMIME_crlf_copy().
15563 * In bntest.c don't call BN_rand with zero bits argument.
15565 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15567 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15568 case was implemented. This caused BN_div_recp() to fail occasionally.
15572 * Add an optional second argument to the set_label() in the perl
15573 assembly language builder. If this argument exists and is set
15574 to 1 it signals that the assembler should use a symbol whose
15575 scope is the entire file, not just the current function. This
15576 is needed with MASM which uses the format label:: for this scope.
15578 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15580 * Change the ASN1 types so they are typedefs by default. Before
15581 almost all types were #define'd to ASN1_STRING which was causing
15582 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15587 * Change names of new functions to the new get1/get0 naming
15588 convention: After 'get1', the caller owns a reference count
15589 and has to call `..._free`; 'get0' returns a pointer to some
15590 data structure without incrementing reference counters.
15591 (Some of the existing 'get' functions increment a reference
15592 counter, some don't.)
15593 Similarly, 'set1' and 'add1' functions increase reference
15594 counters or duplicate objects.
15598 * Allow for the possibility of temp RSA key generation failure:
15599 the code used to assume it always worked and crashed on failure.
15603 * Fix potential buffer overrun problem in BIO_printf().
15604 *Ulf Möller, using public domain code by Patrick Powell; problem
15605 pointed out by David Sacerdote <das33@cornell.edu>*
15607 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15608 RAND_egd() and RAND_status(). In the command line application,
15609 the EGD socket can be specified like a seed file using RANDFILE
15614 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15615 Some CAs (e.g. Verisign) distribute certificates in this form.
15619 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15620 list to exclude them. This means that no special compilation option
15621 is needed to use anonymous DH: it just needs to be included in the
15626 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15627 EVP_MD_type. The old functionality is available in a new macro called
15628 EVP_MD_md(). Change code that uses it and update docs.
15632 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15633 where the `void *` argument is replaced by a function pointer argument.
15634 Previously `void *` was abused to point to functions, which works on
15635 many platforms, but is not correct. As these functions are usually
15636 called by macros defined in OpenSSL header files, most source code
15637 should work without changes.
15641 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15642 sections with information on -D... compiler switches used for
15643 compiling the library so that applications can see them. To enable
15644 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15645 must be defined. E.g.,
15646 #define OPENSSL_ALGORITHM_DEFINES
15647 #include <openssl/opensslconf.h>
15648 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15650 *Richard Levitte, Ulf and Bodo Möller*
15652 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15657 * Change the 'other' type in certificate aux info to a STACK_OF
15658 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15659 the required ASN1 format: arbitrary types determined by an OID.
15663 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15664 argument to 'req'. This is not because the function is newer or
15665 better than others it just uses the work 'NEW' in the certificate
15666 request header lines. Some software needs this.
15670 * Reorganise password command line arguments: now passwords can be
15671 obtained from various sources. Delete the PEM_cb function and make
15672 it the default behaviour: i.e. if the callback is NULL and the
15673 usrdata argument is not NULL interpret it as a null terminated pass
15674 phrase. If usrdata and the callback are NULL then the pass phrase
15675 is prompted for as usual.
15679 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15680 the support is automatically enabled. The resulting binaries will
15681 autodetect the card and use it if present.
15683 *Ben Laurie and Compaq Inc.*
15685 * Work around for Netscape hang bug. This sends certificate request
15686 and server done in one record. Since this is perfectly legal in the
15687 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15688 the bugs/SSLv3 entry for more info.
15692 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15696 * Add -rand argument to smime and pkcs12 applications and read/write
15701 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15705 * Add command line password options to the remaining applications.
15709 * Bug fix for BN_div_recp() for numerators with an even number of
15714 * More tests in bntest.c, and changed test_bn output.
15718 * ./config recognizes MacOS X now.
15722 * Bug fix for BN_div() when the first words of num and divisor are
15723 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15727 * Add support for various broken PKCS#8 formats, and command line
15728 options to produce them.
15732 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15733 get temporary BIGNUMs from a BN_CTX.
15737 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15742 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15743 include a #define from the old name to the new. The original intent
15744 was that statically linked binaries could for example just call
15745 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15746 link with digests. This never worked because SSLeay_add_all_digests()
15747 and SSLeay_add_all_ciphers() were in the same source file so calling
15748 one would link with the other. They are now in separate source files.
15752 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15756 * Use a less unusual form of the Miller-Rabin primality test (it used
15757 a binary algorithm for exponentiation integrated into the Miller-Rabin
15758 loop, our standard modexp algorithms are faster).
15762 * Support for the EBCDIC character set completed.
15764 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15766 * Source code cleanups: use const where appropriate, eliminate casts,
15767 use `void *` instead of `char *` in lhash.
15771 * Bugfix: ssl3_send_server_key_exchange was not restartable
15772 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15773 this the server could overwrite ephemeral keys that the client
15778 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15779 using 50 iterations of the Rabin-Miller test.
15781 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15782 iterations of the Rabin-Miller test as required by the appendix
15783 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15784 As BN_is_prime_fasttest includes trial division, DSA parameter
15785 generation becomes much faster.
15787 This implies a change for the callback functions in DSA_is_prime
15788 and DSA_generate_parameters: The callback function is called once
15789 for each positive witness in the Rabin-Miller test, not just
15790 occasionally in the inner loop; and the parameters to the
15791 callback function now provide an iteration count for the outer
15792 loop rather than for the current invocation of the inner loop.
15793 DSA_generate_parameters additionally can call the callback
15794 function with an 'iteration count' of -1, meaning that a
15795 candidate has passed the trial division test (when q is generated
15796 from an application-provided seed, trial division is skipped).
15800 * New function BN_is_prime_fasttest that optionally does trial
15801 division before starting the Rabin-Miller test and has
15802 an additional BN_CTX * argument (whereas BN_is_prime always
15803 has to allocate at least one BN_CTX).
15804 'callback(1, -1, cb_arg)' is called when a number has passed the
15805 trial division stage.
15809 * Fix for bug in CRL encoding. The validity dates weren't being handled
15814 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15818 * New function BN_pseudo_rand().
15822 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15823 bignum version of BN_from_montgomery() with the working code from
15824 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15829 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15830 made it impossible to use the same SSL_SESSION data structure in
15831 SSL2 clients in multiple threads.
15835 * The return value of RAND_load_file() no longer counts bytes obtained
15836 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15837 to seed the PRNG (previously an explicit byte count was required).
15839 *Ulf Möller, Bodo Möller*
15841 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15842 used `char *` instead of `void *` and had casts all over the place.
15846 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15850 * Retain source code compatibility for BN_prime_checks macro:
15851 BN_is_prime(..., BN_prime_checks, ...) now uses
15852 BN_prime_checks_for_size to determine the appropriate number of
15853 Rabin-Miller iterations.
15857 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15858 DH_CHECK_P_NOT_SAFE_PRIME.
15859 (Check if this is true? OpenPGP calls them "strong".)
15863 * Merge the functionality of "dh" and "gendh" programs into a new program
15864 "dhparam". The old programs are retained for now but will handle DH keys
15865 (instead of parameters) in future.
15869 * Make the ciphers, s_server and s_client programs check the return values
15870 when a new cipher list is set.
15874 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15875 ciphers. Before when the 56bit ciphers were enabled the sorting was
15878 The syntax for the cipher sorting has been extended to support sorting by
15879 cipher-strength (using the strength_bits hard coded in the tables).
15880 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15882 Fix a bug in the cipher-command parser: when supplying a cipher command
15883 string with an "undefined" symbol (neither command nor alphanumeric
15884 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15885 an error is flagged.
15887 Due to the strength-sorting extension, the code of the
15888 ssl_create_cipher_list() function was completely rearranged. I hope that
15889 the readability was also increased :-)
15891 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15893 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15894 for the first serial number and places 2 in the serial number file. This
15895 avoids problems when the root CA is created with serial number zero and
15896 the first user certificate has the same issuer name and serial number
15901 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15902 the new code. Add documentation for this stuff.
15906 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15907 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15908 structures and behave in an analogous way to the X509v3 functions:
15909 they shouldn't be called directly but wrapper functions should be used
15912 So we also now have some wrapper functions that call the X509at functions
15913 when passed certificate requests. (TO DO: similar things can be done with
15914 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15915 things. Some of these need some d2i or i2d and print functionality
15916 because they handle more complex structures.)
15920 * Add missing #ifndefs that caused missing symbols when building libssl
15921 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15922 NO_RSA in `ssl/s2*.c`.
15924 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15926 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15927 has a return value which indicates the quality of the random data
15928 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15929 error queue. New function RAND_pseudo_bytes() generates output that is
15930 guaranteed to be unique but not unpredictable. RAND_add is like
15931 RAND_seed, but takes an extra argument for an entropy estimate
15932 (RAND_seed always assumes full entropy).
15936 * Do more iterations of Rabin-Miller probable prime test (specifically,
15937 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15938 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15939 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15940 false-positive rate of at most 2^-80 for random input.
15944 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15948 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15949 in the 0.9.5 release), this returns the chain
15950 from an X509_CTX structure with a dup of the stack and all
15951 the X509 reference counts upped: so the stack will exist
15952 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15955 Also make SSL_SESSION_print() print out the verify return
15960 * Add manpage for the pkcs12 command. Also change the default
15961 behaviour so MAC iteration counts are used unless the new
15962 -nomaciter option is used. This improves file security and
15963 only older versions of MSIE (4.0 for example) need it.
15967 * Honor the no-xxx Configure options when creating .DEF files.
15971 * Add PKCS#10 attributes to field table: challengePassword,
15972 unstructuredName and unstructuredAddress. These are taken from
15973 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15974 international characters are used.
15976 More changes to X509_ATTRIBUTE code: allow the setting of types
15977 based on strings. Remove the 'loc' parameter when adding
15978 attributes because these will be a SET OF encoding which is sorted
15983 * Initial changes to the 'req' utility to allow request generation
15984 automation. This will allow an application to just generate a template
15985 file containing all the field values and have req construct the
15988 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15989 used all over the place including certificate requests and PKCS#7
15990 structures. They are currently handled manually where necessary with
15991 some primitive wrappers for PKCS#7. The new functions behave in a
15992 manner analogous to the X509 extension functions: they allow
15993 attributes to be looked up by NID and added.
15995 Later something similar to the X509V3 code would be desirable to
15996 automatically handle the encoding, decoding and printing of the
15997 more complex types. The string types like challengePassword can
15998 be handled by the string table functions.
16000 Also modified the multi byte string table handling. Now there is
16001 a 'global mask' which masks out certain types. The table itself
16002 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16003 is useful when for example there is only one permissible type
16004 (as in countryName) and using the mask might result in no valid
16009 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16010 SSL_get_peer_finished to allow applications to obtain the latest
16011 Finished messages sent to the peer or expected from the peer,
16012 respectively. (SSL_get_peer_finished is usually the Finished message
16013 actually received from the peer, otherwise the protocol will be aborted.)
16015 As the Finished message are message digests of the complete handshake
16016 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16017 be used for external authentication procedures when the authentication
16018 provided by SSL/TLS is not desired or is not enough.
16022 * Enhanced support for Alpha Linux is added. Now ./config checks if
16023 the host supports BWX extension and if Compaq C is present on the
16024 $PATH. Just exploiting of the BWX extension results in 20-30%
16025 performance kick for some algorithms, e.g. DES and RC4 to mention
16026 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16031 * Add support for MS "fast SGC". This is arguably a violation of the
16032 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16033 weak crypto and after checking the certificate is SGC a second one
16034 with strong crypto. MS SGC stops the first handshake after receiving
16035 the server certificate message and sends a second client hello. Since
16036 a server will typically do all the time consuming operations before
16037 expecting any further messages from the client (server key exchange
16038 is the most expensive) there is little difference between the two.
16040 To get OpenSSL to support MS SGC we have to permit a second client
16041 hello message after we have sent server done. In addition we have to
16042 reset the MAC if we do get this second client hello.
16046 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16047 if a DER encoded private key is RSA or DSA traditional format. Changed
16048 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16049 format DER encoded private key. Newer code should use PKCS#8 format which
16050 has the key type encoded in the ASN1 structure. Added DER private key
16051 support to pkcs8 application.
16055 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16056 ciphersuites has been selected (as required by the SSL 3/TLS 1
16057 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16058 is set, we interpret this as a request to violate the specification
16059 (the worst that can happen is a handshake failure, and 'correct'
16060 behaviour would result in a handshake failure anyway).
16064 * In SSL_CTX_add_session, take into account that there might be multiple
16065 SSL_SESSION structures with the same session ID (e.g. when two threads
16066 concurrently obtain them from an external cache).
16067 The internal cache can handle only one SSL_SESSION with a given ID,
16068 so if there's a conflict, we now throw out the old one to achieve
16073 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16074 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16075 some routines that use cipher OIDs: some ciphers do not have OIDs
16076 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16081 * Simplify the trust setting structure and code. Now we just have
16082 two sequences of OIDs for trusted and rejected settings. These will
16083 typically have values the same as the extended key usage extension
16084 and any application specific purposes.
16086 The trust checking code now has a default behaviour: it will just
16087 check for an object with the same NID as the passed id. Functions can
16088 be provided to override either the default behaviour or the behaviour
16089 for a given id. SSL client, server and email already have functions
16090 in place for compatibility: they check the NID and also return "trusted"
16091 if the certificate is self signed.
16095 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16096 traditional format into an EVP_PKEY structure.
16100 * Add a password callback function PEM_cb() which either prompts for
16101 a password if usr_data is NULL or otherwise assumes it is a null
16102 terminated password. Allow passwords to be passed on command line
16103 environment or config files in a few more utilities.
16107 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16108 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16109 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16110 Update documentation.
16114 * Support for ASN1 "NULL" type. This could be handled before by using
16115 ASN1_TYPE but there wasn't any function that would try to read a NULL
16116 and produce an error if it couldn't. For compatibility we also have
16117 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16118 don't allocate anything because they don't need to.
16122 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16125 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16127 * Rebuild of the memory allocation routines used by OpenSSL code and
16128 possibly others as well. The purpose is to make an interface that
16129 provide hooks so anyone can build a separate set of allocation and
16130 deallocation routines to be used by OpenSSL, for example memory
16131 pool implementations, or something else, which was previously hard
16132 since Malloc(), Realloc() and Free() were defined as macros having
16133 the values malloc, realloc and free, respectively (except for Win32
16134 compilations). The same is provided for memory debugging code.
16135 OpenSSL already comes with functionality to find memory leaks, but
16136 this gives people a chance to debug other memory problems.
16138 With these changes, a new set of functions and macros have appeared:
16140 CRYPTO_set_mem_debug_functions() [F]
16141 CRYPTO_get_mem_debug_functions() [F]
16142 CRYPTO_dbg_set_options() [F]
16143 CRYPTO_dbg_get_options() [F]
16144 CRYPTO_malloc_debug_init() [M]
16146 The memory debug functions are NULL by default, unless the library
16147 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16148 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16149 gives the standard debugging functions that come with OpenSSL) or
16150 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16151 provided by the library user) must be used. When the standard
16152 debugging functions are used, CRYPTO_dbg_set_options can be used to
16153 request additional information:
16154 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16155 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16157 Also, things like CRYPTO_set_mem_functions will always give the
16158 expected result (the new set of functions is used for allocation
16159 and deallocation) at all times, regardless of platform and compiler
16162 To finish it up, some functions that were never use in any other
16163 way than through macros have a new API and new semantic:
16165 CRYPTO_dbg_malloc()
16166 CRYPTO_dbg_realloc()
16169 All macros of value have retained their old syntax.
16171 *Richard Levitte and Bodo Moeller*
16173 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16174 ordering of SMIMECapabilities wasn't in "strength order" and there
16175 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16180 * Some ASN1 types with illegal zero length encoding (INTEGER,
16181 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16183 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16185 * Merge in my S/MIME library for OpenSSL. This provides a simple
16186 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16187 functionality to handle multipart/signed properly) and a utility
16188 called 'smime' to call all this stuff. This is based on code I
16189 originally wrote for Celo who have kindly allowed it to be
16190 included in OpenSSL.
16194 * Add variants des_set_key_checked and des_set_key_unchecked of
16195 des_set_key (aka des_key_sched). Global variable des_check_key
16196 decides which of these is called by des_set_key; this way
16197 des_check_key behaves as it always did, but applications and
16198 the library itself, which was buggy for des_check_key == 1,
16199 have a cleaner way to pick the version they need.
16203 * New function PKCS12_newpass() which changes the password of a
16208 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16209 dynamic mix. In both cases the ids can be used as an index into the
16210 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16211 functions so they accept a list of the field values and the
16212 application doesn't need to directly manipulate the X509_TRUST
16217 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16222 * Modify the way the V3 extension code looks up extensions. This now
16223 works in a similar way to the object code: we have some "standard"
16224 extensions in a static table which is searched with OBJ_bsearch()
16225 and the application can add dynamic ones if needed. The file
16226 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16227 updated whenever a new extension is added to the core code and kept
16228 in ext_nid order. There is a simple program 'tabtest.c' which checks
16229 this. New extensions are not added too often so this file can readily
16230 be maintained manually.
16232 There are two big advantages in doing things this way. The extensions
16233 can be looked up immediately and no longer need to be "added" using
16234 X509V3_add_standard_extensions(): this function now does nothing.
16235 Side note: I get *lots* of email saying the extension code doesn't
16236 work because people forget to call this function.
16237 Also no dynamic allocation is done unless new extensions are added:
16238 so if we don't add custom extensions there is no need to call
16239 X509V3_EXT_cleanup().
16243 * Modify enc utility's salting as follows: make salting the default. Add a
16244 magic header, so unsalted files fail gracefully instead of just decrypting
16245 to garbage. This is because not salting is a big security hole, so people
16246 should be discouraged from doing it.
16250 * Fixes and enhancements to the 'x509' utility. It allowed a message
16251 digest to be passed on the command line but it only used this
16252 parameter when signing a certificate. Modified so all relevant
16253 operations are affected by the digest parameter including the
16254 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16255 DSA key was used because it didn't fix the digest.
16259 * Initial certificate chain verify code. Currently tests the untrusted
16260 certificates for consistency with the verify purpose (which is set
16261 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16263 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16264 this is because it will reject chains with invalid extensions whereas
16265 every previous version of OpenSSL and SSLeay made no checks at all.
16267 Trust code: checks the root CA for the relevant trust settings. Trust
16268 settings have an initial value consistent with the verify purpose: e.g.
16269 if the verify purpose is for SSL client use it expects the CA to be
16270 trusted for SSL client use. However the default value can be changed to
16271 permit custom trust settings: one example of this would be to only trust
16272 certificates from a specific "secure" set of CAs.
16274 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16275 which should be used for version portability: especially since the
16276 verify structure is likely to change more often now.
16278 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16279 to set them. If not set then assume SSL clients will verify SSL servers
16282 Two new options to the verify program: -untrusted allows a set of
16283 untrusted certificates to be passed in and -purpose which sets the
16284 intended purpose of the certificate. If a purpose is set then the
16285 new chain verify code is used to check extension consistency.
16289 * Support for the authority information access extension.
16293 * Modify RSA and DSA PEM read routines to transparently handle
16294 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16295 public keys in a format compatible with certificate
16296 SubjectPublicKeyInfo structures. Unfortunately there were already
16297 functions called *_PublicKey_* which used various odd formats so
16298 these are retained for compatibility: however the DSA variants were
16299 never in a public release so they have been deleted. Changed dsa/rsa
16300 utilities to handle the new format: note no releases ever handled public
16301 keys so we should be OK.
16303 The primary motivation for this change is to avoid the same fiasco
16304 that dogs private keys: there are several incompatible private key
16305 formats some of which are standard and some OpenSSL specific and
16306 require various evil hacks to allow partial transparent handling and
16307 even then it doesn't work with DER formats. Given the option anything
16308 other than PKCS#8 should be dumped: but the other formats have to
16309 stay in the name of compatibility.
16311 With public keys and the benefit of hindsight one standard format
16312 is used which works with EVP_PKEY, RSA or DSA structures: though
16313 it clearly returns an error if you try to read the wrong kind of key.
16315 Added a -pubkey option to the 'x509' utility to output the public key.
16316 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16317 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16318 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16319 that do the same as the `EVP_PKEY_assign_*()` except they up the
16320 reference count of the added key (they don't "swallow" the
16325 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16326 CRLs would fail if the file contained no certificates or no CRLs:
16327 added a new function to read in both types and return the number
16328 read: this means that if none are read it will be an error. The
16329 DER versions of the certificate and CRL reader would always fail
16330 because it isn't possible to mix certificates and CRLs in DER format
16331 without choking one or the other routine. Changed this to just read
16332 a certificate: this is the best we can do. Also modified the code
16333 in `apps/verify.c` to take notice of return codes: it was previously
16334 attempting to read in certificates from NULL pointers and ignoring
16335 any errors: this is one reason why the cert and CRL reader seemed
16336 to work. It doesn't check return codes from the default certificate
16337 routines: these may well fail if the certificates aren't installed.
16341 * Code to support otherName option in GeneralName.
16345 * First update to verify code. Change the verify utility
16346 so it warns if it is passed a self signed certificate:
16347 for consistency with the normal behaviour. X509_verify
16348 has been modified to it will now verify a self signed
16349 certificate if *exactly* the same certificate appears
16350 in the store: it was previously impossible to trust a
16351 single self signed certificate. This means that:
16352 openssl verify ss.pem
16353 now gives a warning about a self signed certificate but
16354 openssl verify -CAfile ss.pem ss.pem
16359 * For servers, store verify_result in SSL_SESSION data structure
16360 (and add it to external session representation).
16361 This is needed when client certificate verifications fails,
16362 but an application-provided verification callback (set by
16363 SSL_CTX_set_cert_verify_callback) allows accepting the session
16364 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16365 but returns 1): When the session is reused, we have to set
16366 ssl->verify_result to the appropriate error code to avoid
16369 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16371 * Fix a bug in the new PKCS#7 code: it didn't consider the
16372 case in PKCS7_dataInit() where the signed PKCS7 structure
16373 didn't contain any existing data because it was being created.
16375 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16377 * Add a salt to the key derivation routines in enc.c. This
16378 forms the first 8 bytes of the encrypted file. Also add a
16379 -S option to allow a salt to be input on the command line.
16383 * New function X509_cmp(). Oddly enough there wasn't a function
16384 to compare two certificates. We do this by working out the SHA1
16385 hash and comparing that. X509_cmp() will be needed by the trust
16390 * SSL_get1_session() is like SSL_get_session(), but increments
16391 the reference count in the SSL_SESSION returned.
16393 *Geoff Thorpe <geoff@eu.c2.net>*
16395 * Fix for 'req': it was adding a null to request attributes.
16396 Also change the X509_LOOKUP and X509_INFO code to handle
16397 certificate auxiliary information.
16401 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16406 * Add the possibility to add extra information to the memory leak
16407 detecting output, to form tracebacks, showing from where each
16408 allocation was originated: CRYPTO_push_info("constant string") adds
16409 the string plus current file name and line number to a per-thread
16410 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16411 is like calling CYRPTO_pop_info() until the stack is empty.
16412 Also updated memory leak detection code to be multi-thread-safe.
16416 * Add options -text and -noout to pkcs7 utility and delete the
16417 encryption options which never did anything. Update docs.
16421 * Add options to some of the utilities to allow the pass phrase
16422 to be included on either the command line (not recommended on
16423 OSes like Unix) or read from the environment. Update the
16424 manpages and fix a few bugs.
16428 * Add a few manpages for some of the openssl commands.
16432 * Fix the -revoke option in ca. It was freeing up memory twice,
16433 leaking and not finding already revoked certificates.
16437 * Extensive changes to support certificate auxiliary information.
16438 This involves the use of X509_CERT_AUX structure and X509_AUX
16439 functions. An X509_AUX function such as PEM_read_X509_AUX()
16440 can still read in a certificate file in the usual way but it
16441 will also read in any additional "auxiliary information". By
16442 doing things this way a fair degree of compatibility can be
16443 retained: existing certificates can have this information added
16444 using the new 'x509' options.
16446 Current auxiliary information includes an "alias" and some trust
16447 settings. The trust settings will ultimately be used in enhanced
16448 certificate chain verification routines: currently a certificate
16449 can only be trusted if it is self signed and then it is trusted
16454 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16455 The problem was that one of the replacement routines had not been working
16456 since SSLeay releases. For now the offending routine has been replaced
16457 with non-optimised assembler. Even so, this now gives around 95%
16458 performance improvement for 1024 bit RSA signs.
16462 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16463 handling. Most clients have the effective key size in bits equal to
16464 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16465 A few however don't do this and instead use the size of the decrypted key
16466 to determine the RC2 key length and the AlgorithmIdentifier to determine
16467 the effective key length. In this case the effective key length can still
16468 be 40 bits but the key length can be 168 bits for example. This is fixed
16469 by manually forcing an RC2 key into the EVP_PKEY structure because the
16470 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16471 the key length and effective key length are equal.
16475 * Add a bunch of functions that should simplify the creation of
16476 X509_NAME structures. Now you should be able to do:
16477 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16478 and have it automatically work out the correct field type and fill in
16479 the structures. The more adventurous can try:
16480 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16481 and it will (hopefully) work out the correct multibyte encoding.
16485 * Change the 'req' utility to use the new field handling and multibyte
16486 copy routines. Before the DN field creation was handled in an ad hoc
16487 way in req, ca, and x509 which was rather broken and didn't support
16488 BMPStrings or UTF8Strings. Since some software doesn't implement
16489 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16490 using the dirstring_type option. See the new comment in the default
16491 openssl.cnf for more info.
16495 * Make crypto/rand/md_rand.c more robust:
16496 - Assure unique random numbers after fork().
16497 - Make sure that concurrent threads access the global counter and
16498 md serializably so that we never lose entropy in them
16499 or use exactly the same state in multiple threads.
16500 Access to the large state is not always serializable because
16501 the additional locking could be a performance killer, and
16502 md should be large enough anyway.
16506 * New file `apps/app_rand.c` with commonly needed functionality
16507 for handling the random seed file.
16509 Use the random seed file in some applications that previously did not:
16511 dsaparam -genkey (which also ignored its '-rand' option),
16514 x509 (when signing).
16515 Except on systems with /dev/urandom, it is crucial to have a random
16516 seed file at least for key creation, DSA signing, and for DH exchanges;
16517 for RSA signatures we could do without one.
16519 gendh and gendsa (unlike genrsa) used to read only the first byte
16520 of each file listed in the '-rand' option. The function as previously
16521 found in genrsa is now in app_rand.c and is used by all programs
16522 that support '-rand'.
16526 * In RAND_write_file, use mode 0600 for creating files;
16527 don't just chmod when it may be too late.
16531 * Report an error from X509_STORE_load_locations
16532 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16536 * New function ASN1_mbstring_copy() this copies a string in either
16537 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16538 into an ASN1_STRING type. A mask of permissible types is passed
16539 and it chooses the "minimal" type to use or an error if not type
16544 * Add function equivalents to the various macros in asn1.h. The old
16545 macros are retained with an `M_` prefix. Code inside the library can
16546 use the `M_` macros. External code (including the openssl utility)
16547 should *NOT* in order to be "shared library friendly".
16551 * Add various functions that can check a certificate's extensions
16552 to see if it usable for various purposes such as SSL client,
16553 server or S/MIME and CAs of these types. This is currently
16554 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16555 verification. Also added a -purpose flag to x509 utility to
16556 print out all the purposes.
16560 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16565 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16566 for, obtain and decode and extension and obtain its critical flag.
16567 This allows all the necessary extension code to be handled in a
16568 single function call.
16572 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16573 platforms. See crypto/rc4/rc4_enc.c for further details.
16577 * New -noout option to asn1parse. This causes no output to be produced
16578 its main use is when combined with -strparse and -out to extract data
16579 from a file (which may not be in ASN.1 format).
16583 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16584 when producing the local key id.
16586 *Richard Levitte <levitte@stacken.kth.se>*
16588 * New option -dhparam in s_server. This allows a DH parameter file to be
16589 stated explicitly. If it is not stated then it tries the first server
16590 certificate file. The previous behaviour hard coded the filename
16595 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16596 a public key to be input or output. For example:
16597 openssl rsa -in key.pem -pubout -out pubkey.pem
16598 Also added necessary DSA public key functions to handle this.
16602 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16603 in the message. This was handled by allowing
16604 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16606 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16608 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16609 to the end of the strings whereas this didn't. This would cause problems
16610 if strings read with d2i_ASN1_bytes() were later modified.
16612 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16614 * Fix for base64 decode bug. When a base64 bio reads only one line of
16615 data and it contains EOF it will end up returning an error. This is
16616 caused by input 46 bytes long. The cause is due to the way base64
16617 BIOs find the start of base64 encoded data. They do this by trying a
16618 trial decode on each line until they find one that works. When they
16619 do a flag is set and it starts again knowing it can pass all the
16620 data directly through the decoder. Unfortunately it doesn't reset
16621 the context it uses. This means that if EOF is reached an attempt
16622 is made to pass two EOFs through the context and this causes the
16623 resulting error. This can also cause other problems as well. As is
16624 usual with these problems it takes *ages* to find and the fix is
16625 trivial: move one line.
16627 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16629 * Ugly workaround to get s_client and s_server working under Windows. The
16630 old code wouldn't work because it needed to select() on sockets and the
16631 tty (for keypresses and to see if data could be written). Win32 only
16632 supports select() on sockets so we select() with a 1s timeout on the
16633 sockets and then see if any characters are waiting to be read, if none
16634 are present then we retry, we also assume we can always write data to
16635 the tty. This isn't nice because the code then blocks until we've
16636 received a complete line of data and it is effectively polling the
16637 keyboard at 1s intervals: however it's quite a bit better than not
16638 working at all :-) A dedicated Windows application might handle this
16639 with an event loop for example.
16643 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16644 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16645 will be called when RSA_sign() and RSA_verify() are used. This is useful
16646 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16647 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16648 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16649 This necessitated the support of an extra signature type NID_md5_sha1
16650 for SSL signatures and modifications to the SSL library to use it instead
16651 of calling RSA_public_decrypt() and RSA_private_encrypt().
16655 * Add new -verify -CAfile and -CApath options to the crl program, these
16656 will lookup a CRL issuers certificate and verify the signature in a
16657 similar way to the verify program. Tidy up the crl program so it
16658 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16659 less strict. It will now permit CRL extensions even if it is not
16660 a V2 CRL: this will allow it to tolerate some broken CRLs.
16664 * Initialize all non-automatic variables each time one of the openssl
16665 sub-programs is started (this is necessary as they may be started
16666 multiple times from the "OpenSSL>" prompt).
16668 *Lennart Bang, Bodo Moeller*
16670 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16671 removing all other RSA functionality (this is what NO_RSA does). This
16672 is so (for example) those in the US can disable those operations covered
16673 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16678 * Non-copying interface to BIO pairs.
16679 (still largely untested)
16683 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16684 ASCII string. This was handled independently in various places before.
16688 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16689 UTF8 strings a character at a time.
16693 * Use client_version from client hello to select the protocol
16694 (s23_srvr.c) and for RSA client key exchange verification
16695 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16699 * Add various utility functions to handle SPKACs, these were previously
16700 handled by poking round in the structure internals. Added new function
16701 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16702 print, verify and generate SPKACs. Based on an original idea from
16703 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16707 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16711 * Allow the config file extension section to be overwritten on the
16712 command line. Based on an original idea from Massimiliano Pala
16713 <madwolf@comune.modena.it>. The new option is called -extensions
16714 and can be applied to ca, req and x509. Also -reqexts to override
16715 the request extensions in req and -crlexts to override the crl extensions
16720 * Add new feature to the SPKAC handling in ca. Now you can include
16721 the same field multiple times by preceding it by "XXXX." for example:
16724 this is the same syntax as used in the req config file.
16728 * Allow certificate extensions to be added to certificate requests. These
16729 are specified in a 'req_extensions' option of the req section of the
16730 config file. They can be printed out with the -text option to req but
16731 are otherwise ignored at present.
16735 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16736 data read consists of only the final block it would not decrypted because
16737 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16738 A misplaced 'break' also meant the decrypted final block might not be
16739 copied until the next read.
16743 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16744 a few extra parameters to the DH structure: these will be useful if
16745 for example we want the value of 'q' or implement X9.42 DH.
16749 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16750 provides hooks that allow the default DSA functions or functions on a
16751 "per key" basis to be replaced. This allows hardware acceleration and
16752 hardware key storage to be handled without major modification to the
16753 library. Also added low level modexp hooks and CRYPTO_EX structure and
16754 associated functions.
16758 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16759 as "read only": it can't be written to and the buffer it points to will
16760 not be freed. Reading from a read only BIO is much more efficient than
16761 a normal memory BIO. This was added because there are several times when
16762 an area of memory needs to be read from a BIO. The previous method was
16763 to create a memory BIO and write the data to it, this results in two
16764 copies of the data and an O(n^2) reading algorithm. There is a new
16765 function BIO_new_mem_buf() which creates a read only memory BIO from
16766 an area of memory. Also modified the PKCS#7 routines to use read only
16771 * Bugfix: ssl23_get_client_hello did not work properly when called in
16772 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16773 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16774 but a retry condition occurred while trying to read the rest.
16778 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16779 NID_pkcs7_encrypted by default: this was wrong since this should almost
16780 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16781 the encrypted data type: this is a more sensible place to put it and it
16782 allows the PKCS#12 code to be tidied up that duplicated this
16787 * Changed obj_dat.pl script so it takes its input and output files on
16788 the command line. This should avoid shell escape redirection problems
16793 * Initial support for certificate extension requests, these are included
16794 in things like Xenroll certificate requests. Included functions to allow
16795 extensions to be obtained and added.
16799 * -crlf option to s_client and s_server for sending newlines as
16800 CRLF (as required by many protocols).
16804 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16806 * Install libRSAglue.a when OpenSSL is built with RSAref.
16808 *Ralf S. Engelschall*
16810 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16812 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16814 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16819 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16820 DH parameters/keys (q is lost during that conversion, but the resulting
16821 DH parameters contain its length).
16823 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16824 much faster than DH_generate_parameters (which creates parameters
16825 where `p = 2*q + 1`), and also the smaller q makes DH computations
16826 much more efficient (160-bit exponentiation instead of 1024-bit
16827 exponentiation); so this provides a convenient way to support DHE
16828 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16829 utter importance to use
16830 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16832 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16833 when such DH parameters are used, because otherwise small subgroup
16834 attacks may become possible!
16838 * Avoid memory leak in i2d_DHparams.
16842 * Allow the -k option to be used more than once in the enc program:
16843 this allows the same encrypted message to be read by multiple recipients.
16847 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16848 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16849 it will always use the numerical form of the OID, even if it has a short
16854 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16855 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16856 otherwise bn_mod_exp was called. In the case of hardware keys for example
16857 no private key components need be present and it might store extra data
16858 in the RSA structure, which cannot be accessed from bn_mod_exp.
16859 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16860 private key operations.
16864 * Added support for SPARC Linux.
16868 * pem_password_cb function type incompatibly changed from
16869 typedef int pem_password_cb(char *buf, int size, int rwflag);
16871 ....(char *buf, int size, int rwflag, void *userdata);
16872 so that applications can pass data to their callbacks:
16873 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16874 additional void * argument, which is just handed through whenever
16875 the password callback is called.
16877 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16879 New function SSL_CTX_set_default_passwd_cb_userdata.
16881 Compatibility note: As many C implementations push function arguments
16882 onto the stack in reverse order, the new library version is likely to
16883 interoperate with programs that have been compiled with the old
16884 pem_password_cb definition (PEM_whatever takes some data that
16885 happens to be on the stack as its last argument, and the callback
16886 just ignores this garbage); but there is no guarantee whatsoever that
16889 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16890 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16891 problems not only on Windows, but also on some Unix platforms.
16892 To avoid problematic command lines, these definitions are now in an
16893 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16894 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16898 * MIPS III/IV assembler module is reimplemented.
16902 * More DES library cleanups: remove references to srand/rand and
16903 delete an unused file.
16907 * Add support for the free Netwide assembler (NASM) under Win32,
16908 since not many people have MASM (ml) and it can be hard to obtain.
16909 This is currently experimental but it seems to work OK and pass all
16910 the tests. Check out INSTALL.W32 for info.
16914 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16915 without temporary keys kept an extra copy of the server key,
16916 and connections with temporary keys did not free everything in case
16921 * New function RSA_check_key and new openssl rsa option -check
16922 for verifying the consistency of RSA keys.
16924 *Ulf Moeller, Bodo Moeller*
16926 * Various changes to make Win32 compile work:
16927 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16928 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16929 comparison" warnings.
16930 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16934 * Add a debugging option to PKCS#5 v2 key generation function: when
16935 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16936 derived keys are printed to stderr.
16940 * Copy the flags in ASN1_STRING_dup().
16942 *Roman E. Pavlov <pre@mo.msk.ru>*
16944 * The x509 application mishandled signing requests containing DSA
16945 keys when the signing key was also DSA and the parameters didn't match.
16947 It was supposed to omit the parameters when they matched the signing key:
16948 the verifying software was then supposed to automatically use the CA's
16949 parameters if they were absent from the end user certificate.
16951 Omitting parameters is no longer recommended. The test was also
16952 the wrong way round! This was probably due to unusual behaviour in
16953 EVP_cmp_parameters() which returns 1 if the parameters match.
16954 This meant that parameters were omitted when they *didn't* match and
16955 the certificate was useless. Certificates signed with 'ca' didn't have
16958 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16960 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16961 The interface is as follows:
16962 Applications can use
16963 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16964 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16965 "off" is now the default.
16966 The library internally uses
16967 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16968 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16969 to disable memory-checking temporarily.
16971 Some inconsistent states that previously were possible (and were
16972 even the default) are now avoided.
16974 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16975 with each memory chunk allocated; this is occasionally more helpful
16976 than just having a counter.
16978 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16980 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16985 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16986 which largely parallels "options", but is for changing API behaviour,
16987 whereas "options" are about protocol behaviour.
16988 Initial "mode" flags are:
16990 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16991 a single record has been written.
16992 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16993 retries use the same buffer location.
16994 (But all of the contents must be
16999 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17002 * Fix problems with no-hmac etc.
17004 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17006 * New functions RSA_get_default_method(), RSA_set_method() and
17007 RSA_get_method(). These allows replacement of RSA_METHODs without having
17008 to mess around with the internals of an RSA structure.
17012 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17013 Also really enable memory leak checks in openssl.c and in some
17016 *Chad C. Mulligan, Bodo Moeller*
17018 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17019 up the length of negative integers. This has now been simplified to just
17020 store the length when it is first determined and use it later, rather
17021 than trying to keep track of where data is copied and updating it to
17023 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17025 * Add a new function PKCS7_signatureVerify. This allows the verification
17026 of a PKCS#7 signature but with the signing certificate passed to the
17027 function itself. This contrasts with PKCS7_dataVerify which assumes the
17028 certificate is present in the PKCS#7 structure. This isn't always the
17029 case: certificates can be omitted from a PKCS#7 structure and be
17030 distributed by "out of band" means (such as a certificate database).
17034 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17035 function prototypes in pem.h, also change util/mkdef.pl to add the
17036 necessary function names.
17040 * mk1mf.pl (used by Windows builds) did not properly read the
17041 options set by Configure in the top level Makefile, and Configure
17042 was not even able to write more than one option correctly.
17043 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17047 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17048 file to be loaded from a BIO or FILE pointer. The BIO version will
17049 for example allow memory BIOs to contain config info.
17053 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17054 Whoever hopes to achieve shared-library compatibility across versions
17055 must use this, not the compile-time macro.
17056 (Exercise 0.9.4: Which is the minimum library version required by
17058 Note: All this applies only to multi-threaded programs, others don't
17063 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17064 through a BIO pair triggered the default case, i.e.
17065 SSLerr(...,SSL_R_UNKNOWN_STATE).
17069 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17070 can use the SSL library even if none of the specific BIOs is
17075 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17076 for the encoded length.
17078 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17080 * Add initial documentation of the X509V3 functions.
17084 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17085 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17086 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17087 secure PKCS#8 private key format with a high iteration count.
17091 * Fix determination of Perl interpreter: A perl or perl5
17092 *directory* in $PATH was also accepted as the interpreter.
17094 *Ralf S. Engelschall*
17096 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17097 wrong with it but it was very old and did things like calling
17098 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17099 unusual formatting.
17103 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17104 to use the new extension code.
17108 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17109 with macros. This should make it easier to change their form, add extra
17110 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17115 * Add to configuration table a new entry that can specify an alternative
17116 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17117 according to Mark Crispin <MRC@Panda.COM>.
17121 * DES CBC did not update the IV. Weird.
17125 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17126 Changing the behaviour of the former might break existing programs --
17127 where IV updating is needed, des_ncbc_encrypt can be used.
17130 * When bntest is run from "make test" it drives bc to check its
17131 calculations, as well as internally checking them. If an internal check
17132 fails, it needs to cause bc to give a non-zero result or make test carries
17133 on without noticing the failure. Fixed.
17137 * DES library cleanups.
17141 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17142 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17143 ciphers. NOTE: although the key derivation function has been verified
17144 against some published test vectors it has not been extensively tested
17145 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17150 * Instead of "mkdir -p", which is not fully portable, use new
17151 Perl script "util/mkdir-p.pl".
17155 * Rewrite the way password based encryption (PBE) is handled. It used to
17156 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17157 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17158 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17159 the 'parameter' field of the AlgorithmIdentifier is passed to the
17160 underlying key generation function so it must do its own ASN1 parsing.
17161 This has also changed the EVP_PBE_CipherInit() function which now has a
17162 'parameter' argument instead of literal salt and iteration count values
17163 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17167 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17168 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17169 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17170 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17171 value was just used as a "magic string" and not used directly its
17172 value doesn't matter.
17176 * Introduce some semblance of const correctness to BN. Shame C doesn't
17181 * "linux-sparc64" configuration (ultrapenguin).
17183 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17184 "linux-sparc" configuration.
17186 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17188 * config now generates no-xxx options for missing ciphers.
17192 * Support the EBCDIC character set (work in progress).
17193 File ebcdic.c not yet included because it has a different license.
17195 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17197 * Support BS2000/OSD-POSIX.
17199 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17201 * Make callbacks for key generation use `void *` instead of `char *`.
17205 * Make S/MIME samples compile (not yet tested).
17209 * Additional typesafe stacks.
17213 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17217 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17219 * New configuration variant "sco5-gcc".
17221 * Updated some demos.
17223 *Sean O Riordain, Wade Scholine*
17225 * Add missing BIO_free at exit of pkcs12 application.
17229 * Fix memory leak in conf.c.
17233 * Updates for Win32 to assembler version of MD5.
17237 * Set #! path to perl in `apps/der_chop` to where we found it
17238 instead of using a fixed path.
17242 * SHA library changes for irix64-mips4-cc.
17246 * Improvements for VMS support.
17250 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17252 * Bignum library bug fix. IRIX 6 passes "make test" now!
17253 This also avoids the problems with SC4.2 and unpatched SC5.
17255 *Andy Polyakov <appro@fy.chalmers.se>*
17257 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17258 These are required because of the typesafe stack would otherwise break
17259 existing code. If old code used a structure member which used to be STACK
17260 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17261 sk_num or sk_value it would produce an error because the num, data members
17262 are not present in STACK_OF. Now it just produces a warning. sk_set
17263 replaces the old method of assigning a value to sk_value
17264 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17265 that does this will no longer work (and should use sk_set instead) but
17266 this could be regarded as a "questionable" behaviour anyway.
17270 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17271 correctly handle encrypted S/MIME data.
17275 * Change type of various DES function arguments from des_cblock
17276 (which means, in function argument declarations, pointer to char)
17277 to des_cblock * (meaning pointer to array with 8 char elements),
17278 which allows the compiler to do more typechecking; it was like
17279 that back in SSLeay, but with lots of ugly casts.
17281 Introduce new type const_des_cblock.
17285 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17286 problems: find RecipientInfo structure that matches recipient certificate
17287 and initialise the ASN1 structures properly based on passed cipher.
17291 * Belatedly make the BN tests actually check the results.
17295 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17296 to and from BNs: it was completely broken. New compilation option
17297 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17298 key elements as negative integers.
17302 * Reorganize and speed up MD5.
17304 *Andy Polyakov <appro@fy.chalmers.se>*
17308 *Richard Levitte <richard@levitte.org>*
17310 * New option -out to asn1parse to allow the parsed structure to be
17311 output to a file. This is most useful when combined with the -strparse
17312 option to examine the output of things like OCTET STRINGS.
17316 * Make SSL library a little more fool-proof by not requiring any longer
17317 that `SSL_set_{accept,connect}_state` be called before
17318 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17319 in many applications because usually everything *appeared* to work as
17320 intended anyway -- now it really works as intended).
17324 * Move openssl.cnf out of lib/.
17328 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17329 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17330 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17332 *Ralf S. Engelschall*
17334 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17335 handle PKCS#7 enveloped data properly.
17337 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17339 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17340 copying pointers. The cert_st handling is changed by this in
17341 various ways (and thus what used to be known as ctx->default_cert
17342 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17343 any longer when s->cert does not give us what we need).
17344 ssl_cert_instantiate becomes obsolete by this change.
17345 As soon as we've got the new code right (possibly it already is?),
17346 we have solved a couple of bugs of the earlier code where s->cert
17347 was used as if it could not have been shared with other SSL structures.
17349 Note that using the SSL API in certain dirty ways now will result
17350 in different behaviour than observed with earlier library versions:
17351 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17352 does not influence s as it used to.
17354 In order to clean up things more thoroughly, inside SSL_SESSION
17355 we don't use CERT any longer, but a new structure SESS_CERT
17356 that holds per-session data (if available); currently, this is
17357 the peer's certificate chain and, for clients, the server's certificate
17358 and temporary key. CERT holds only those values that can have
17359 meaningful defaults in an SSL_CTX.
17363 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17364 from the internal representation. Various PKCS#7 fixes: remove some
17365 evil casts and set the enc_dig_alg field properly based on the signing
17370 * Allow PKCS#12 password to be set from the command line or the
17371 environment. Let 'ca' get its config file name from the environment
17372 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17377 * Allow certificate policies extension to use an IA5STRING for the
17378 organization field. This is contrary to the PKIX definition but
17379 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17384 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17385 without disallowing inline assembler and the like for non-pedantic builds.
17389 * Support Borland C++ builder.
17391 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17397 * SHA-1 cleanups and performance enhancements.
17399 *Andy Polyakov <appro@fy.chalmers.se>*
17401 * Sparc v8plus assembler for the bignum library.
17403 *Andy Polyakov <appro@fy.chalmers.se>*
17405 * Accept any -xxx and +xxx compiler options in Configure.
17409 * Update HPUX configuration.
17413 * Add missing `sk_<type>_unshift()` function to safestack.h
17415 *Ralf S. Engelschall*
17417 * New function SSL_CTX_use_certificate_chain_file that sets the
17418 "extra_cert"s in addition to the certificate. (This makes sense
17419 only for "PEM" format files, as chains as a whole are not
17424 * Support verify_depth from the SSL API.
17425 x509_vfy.c had what can be considered an off-by-one-error:
17426 Its depth (which was not part of the external interface)
17427 was actually counting the number of certificates in a chain;
17428 now it really counts the depth.
17432 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17433 instead of X509err, which often resulted in confusing error
17434 messages since the error codes are not globally unique
17435 (e.g. an alleged error in ssl3_accept when a certificate
17436 didn't match the private key).
17438 * New function SSL_CTX_set_session_id_context that allows to set a default
17439 value (so that you don't need SSL_set_session_id_context for each
17440 connection using the SSL_CTX).
17444 * OAEP decoding bug fix.
17448 * Support INSTALL_PREFIX for package builders, as proposed by
17453 * New Configure options "threads" and "no-threads". For systems
17454 where the proper compiler options are known (currently Solaris
17455 and Linux), "threads" is the default.
17459 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17463 * Install various scripts to $(OPENSSLDIR)/misc, not to
17464 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17465 such as /usr/local/bin.
17469 * "make linux-shared" to build shared libraries.
17471 *Niels Poppe <niels@netbox.org>*
17473 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17477 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17478 extension adding in x509 utility.
17482 * Remove NOPROTO sections and error code comments.
17486 * Partial rewrite of the DEF file generator to now parse the ANSI
17491 * New Configure options --prefix=DIR and --openssldir=DIR.
17495 * Complete rewrite of the error code script(s). It is all now handled
17496 by one script at the top level which handles error code gathering,
17497 header rewriting and C source file generation. It should be much better
17498 than the old method: it now uses a modified version of Ulf's parser to
17499 read the ANSI prototypes in all header files (thus the old K&R definitions
17500 aren't needed for error creation any more) and do a better job of
17501 translating function codes into names. The old 'ASN1 error code embedded
17502 in a comment' is no longer necessary and it doesn't use .err files which
17503 have now been deleted. Also the error code call doesn't have to appear all
17504 on one line (which resulted in some large lines...).
17508 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17512 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17513 0 (which usually indicates a closed connection), but continue reading.
17517 * Fix some race conditions.
17521 * Add support for CRL distribution points extension. Add Certificate
17522 Policies and CRL distribution points documentation.
17526 * Move the autogenerated header file parts to crypto/opensslconf.h.
17530 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17531 8 of keying material. Merlin has also confirmed interop with this fix
17532 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17534 *Merlin Hughes <merlin@baltimore.ie>*
17536 * Fix lots of warnings.
17538 *Richard Levitte <levitte@stacken.kth.se>*
17540 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17541 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17543 *Richard Levitte <levitte@stacken.kth.se>*
17545 * Fix problems with sizeof(long) == 8.
17547 *Andy Polyakov <appro@fy.chalmers.se>*
17549 * Change functions to ANSI C.
17553 * Fix typos in error codes.
17555 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17557 * Remove defunct assembler files from Configure.
17561 * SPARC v8 assembler BIGNUM implementation.
17563 *Andy Polyakov <appro@fy.chalmers.se>*
17565 * Support for Certificate Policies extension: both print and set.
17566 Various additions to support the r2i method this uses.
17570 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17571 return a const string when you are expecting an allocated buffer.
17575 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17576 types DirectoryString and DisplayText.
17580 * Add code to allow r2i extensions to access the configuration database,
17581 add an LHASH database driver and add several ctx helper functions.
17585 * Fix an evil bug in bn_expand2() which caused various BN functions to
17586 fail when they extended the size of a BIGNUM.
17590 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17591 support typesafe stack.
17595 * Fix typo in SSL_[gs]et_options().
17597 *Nils Frostberg <nils@medcom.se>*
17599 * Delete various functions and files that belonged to the (now obsolete)
17600 old X509V3 handling code.
17604 * New Configure option "rsaref".
17608 * Don't auto-generate pem.h.
17612 * Introduce type-safe ASN.1 SETs.
17616 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17618 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17620 * Introduce type-safe STACKs. This will almost certainly break lots of code
17621 that links with OpenSSL (well at least cause lots of warnings), but fear
17622 not: the conversion is trivial, and it eliminates loads of evil casts. A
17623 few STACKed things have been converted already. Feel free to convert more.
17624 In the fullness of time, I'll do away with the STACK type altogether.
17628 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17629 specified in `<certfile>` by updating the entry in the index.txt file.
17630 This way one no longer has to edit the index.txt file manually for
17631 revoking a certificate. The -revoke option does the gory details now.
17633 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17635 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17636 `-text` option at all and this way the `-noout -text` combination was
17637 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17639 *Ralf S. Engelschall*
17641 * Make sure a corresponding plain text error message exists for the
17642 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17643 verify callback function determined that a certificate was revoked.
17645 *Ralf S. Engelschall*
17647 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17648 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17649 all available ciphers including rc5, which was forgotten until now.
17650 In order to let the testing shell script know which algorithms
17651 are available, a new (up to now undocumented) command
17652 `openssl list-cipher-commands` is used.
17656 * Bugfix: s_client occasionally would sleep in select() when
17657 it should have checked SSL_pending() first.
17661 * New functions DSA_do_sign and DSA_do_verify to provide access to
17662 the raw DSA values prior to ASN.1 encoding.
17666 * Tweaks to Configure
17668 *Niels Poppe <niels@netbox.org>*
17670 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17675 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17679 * New config option to avoid instructions that are illegal on the 80386.
17680 The default code is faster, but requires at least a 486.
17684 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17685 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17686 same as SSL2_VERSION anyway.
17690 * New "-showcerts" option for s_client.
17694 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17695 application. Various cleanups and fixes.
17699 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17700 modify error routines to work internally. Add error codes and PBE init
17701 to library startup routines.
17705 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17706 packing functions to asn1 and evp. Changed function names and error
17707 codes along the way.
17711 * PKCS12 integration: and so it begins... First of several patches to
17712 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17713 objects to objects.h
17717 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17718 and display support for Thawte strong extranet extension.
17722 * Add LinuxPPC support.
17724 *Jeff Dubrule <igor@pobox.org>*
17726 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17727 bn_div_words in alpha.s.
17729 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17731 * Make sure the RSA OAEP test is skipped under -DRSAref because
17732 OAEP isn't supported when OpenSSL is built with RSAref.
17734 *Ulf Moeller <ulf@fitug.de>*
17736 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17737 so they no longer are missing under -DNOPROTO.
17739 *Soren S. Jorvang <soren@t.dk>*
17741 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17743 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17744 doesn't work when the session is reused. Coming soon!
17748 * Fix a security hole, that allows sessions to be reused in the wrong
17749 context thus bypassing client cert protection! All software that uses
17750 client certs and session caches in multiple contexts NEEDS PATCHING to
17751 allow session reuse! A fuller solution is in the works.
17753 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17755 * Some more source tree cleanups (removed obsolete files
17756 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17757 permission on "config" script to be executable) and a fix for the INSTALL
17760 *Ulf Moeller <ulf@fitug.de>*
17762 * Remove some legacy and erroneous uses of malloc, free instead of
17765 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17767 * Make rsa_oaep_test return non-zero on error.
17769 *Ulf Moeller <ulf@fitug.de>*
17771 * Add support for native Solaris shared libraries. Configure
17772 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17773 if someone would make that last step automatic.
17775 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17777 * ctx_size was not built with the right compiler during "make links". Fixed.
17781 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17782 except NULL ciphers". This means the default cipher list will no longer
17783 enable NULL ciphers. They need to be specifically enabled e.g. with
17784 the string "DEFAULT:eNULL".
17788 * Fix to RSA private encryption routines: if p < q then it would
17789 occasionally produce an invalid result. This will only happen with
17790 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17794 * Be less restrictive and allow also `perl util/perlpath.pl
17795 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17796 because this way one can also use an interpreter named `perl5` (which is
17797 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17798 installed as `perl`).
17800 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17802 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17804 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17806 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17807 advapi32.lib to Win32 build and change the pem test comparison
17808 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17809 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17810 and crypto/des/ede_cbcm_enc.c.
17814 * DES quad checksum was broken on big-endian architectures. Fixed.
17818 * Comment out two functions in bio.h that aren't implemented. Fix up the
17819 Win32 test batch file so it (might) work again. The Win32 test batch file
17820 is horrible: I feel ill....
17824 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17825 in e_os.h. Audit of header files to check ANSI and non ANSI
17826 sections: 10 functions were absent from non ANSI section and not exported
17827 from Windows DLLs. Fixed up libeay.num for new functions.
17831 * Make `openssl version` output lines consistent.
17833 *Ralf S. Engelschall*
17835 * Fix Win32 symbol export lists for BIO functions: Added
17836 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17837 to ms/libeay{16,32}.def.
17839 *Ralf S. Engelschall*
17841 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17842 fine under Unix and passes some trivial tests I've now added. But the
17843 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17844 added to make sure no one expects that this stuff really works in the
17845 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17846 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17849 *Ralf S. Engelschall*
17851 * Fix the generation of two part addresses in perl.
17853 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17855 * Add config entry for Linux on MIPS.
17857 *John Tobey <jtobey@channel1.com>*
17859 * Make links whenever Configure is run, unless we are on Windoze.
17863 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17864 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17869 * Add a useful kludge to allow package maintainers to specify compiler and
17870 other platforms details on the command line without having to patch the
17871 Configure script every time: One now can use
17872 `perl Configure <id>:<details>`,
17873 i.e. platform ids are allowed to have details appended
17874 to them (separated by colons). This is treated as there would be a static
17875 pre-configured entry in Configure's %table under key `<id>` with value
17876 `<details>` and `perl Configure <id>` is called. So, when you want to
17877 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17878 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17879 now, which overrides the FreeBSD-elf entry on-the-fly.
17881 *Ralf S. Engelschall*
17883 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17887 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17888 on the `perl Configure ...` command line. This way one can compile
17889 OpenSSL libraries with Position Independent Code (PIC) which is needed
17890 for linking it into DSOs.
17892 *Ralf S. Engelschall*
17894 * Remarkably, export ciphers were totally broken and no-one had noticed!
17899 * Cleaned up the LICENSE document: The official contact for any license
17900 questions now is the OpenSSL core team under openssl-core@openssl.org.
17901 And add a paragraph about the dual-license situation to make sure people
17902 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17903 to the OpenSSL toolkit.
17905 *Ralf S. Engelschall*
17907 * General source tree makefile cleanups: Made `making xxx in yyy...`
17908 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17909 Additionally cleaned up the `make links` target: Remove unnecessary
17910 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17911 to speed processing and no longer clutter the display with confusing
17912 stuff. Instead only the actually done links are displayed.
17914 *Ralf S. Engelschall*
17916 * Permit null encryption ciphersuites, used for authentication only. It used
17917 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17918 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17923 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17924 signed attributes when verifying signatures (this would break them),
17925 the detached data encoding was wrong and public keys obtained using
17926 X509_get_pubkey() weren't freed.
17930 * Add text documentation for the BUFFER functions. Also added a work around
17931 to a Win95 console bug. This was triggered by the password read stuff: the
17932 last character typed gets carried over to the next fread(). If you were
17933 generating a new cert request using 'req' for example then the last
17934 character of the passphrase would be CR which would then enter the first
17939 * Added the new 'Includes OpenSSL Cryptography Software' button as
17940 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17941 button and can be used by applications based on OpenSSL to show the
17942 relationship to the OpenSSL project.
17944 *Ralf S. Engelschall*
17946 * Remove confusing variables in function signatures in files
17947 ssl/ssl_lib.c and ssl/ssl.h.
17949 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17951 * Don't install bss_file.c under PREFIX/include/
17953 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17955 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17956 functions that return function pointers and has support for NT specific
17957 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17958 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17959 unsigned to signed types: this was killing the Win32 compile.
17963 * Add new certificate file to stack functions,
17964 SSL_add_dir_cert_subjects_to_stack() and
17965 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17966 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17967 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17968 This means that Apache-SSL and similar packages don't have to mess around
17969 to add as many CAs as they want to the preferred list.
17973 * Experiment with doxygen documentation. Currently only partially applied to
17975 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17976 openssl.doxy as the configuration file.
17980 * Get rid of remaining C++-style comments which strict C compilers hate.
17982 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17984 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17985 compiled in by default: it has problems with large keys.
17989 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17990 DH private keys and/or callback functions which directly correspond to
17991 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17992 is needed for applications which have to configure certificates on a
17993 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17995 For the RSA certificate situation is makes no difference, but
17996 for the DSA certificate situation this fixes the "no shared cipher"
17997 problem where the OpenSSL cipher selection procedure failed because the
17998 temporary keys were not overtaken from the context and the API provided
17999 no way to reconfigure them.
18000 The new functions now let applications reconfigure the stuff and they
18001 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18002 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18003 non-public-API function ssl_cert_instantiate() is used as a helper
18004 function and also to reduce code redundancy inside ssl_rsa.c.
18006 *Ralf S. Engelschall*
18008 * Move s_server -dcert and -dkey options out of the undocumented feature
18009 area because they are useful for the DSA situation and should be
18010 recognized by the users.
18012 *Ralf S. Engelschall*
18014 * Fix the cipher decision scheme for export ciphers: the export bits are
18015 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18016 SSL_EXP_MASK. So, the original variable has to be used instead of the
18017 already masked variable.
18019 *Richard Levitte <levitte@stacken.kth.se>*
18021 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18023 *Richard Levitte <levitte@stacken.kth.se>*
18025 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18026 from `int` to `unsigned int` because it is a length and initialized by
18027 EVP_DigestFinal() which expects an `unsigned int *`.
18029 *Richard Levitte <levitte@stacken.kth.se>*
18031 * Don't hard-code path to Perl interpreter on shebang line of Configure
18032 script. Instead use the usual Shell->Perl transition trick.
18034 *Ralf S. Engelschall*
18036 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18037 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18038 -noout -modulus` as it's already the case for `openssl rsa -noout
18039 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18040 currently the public key is printed (a decision which was already done by
18041 `openssl dsa -modulus` in the past) which serves a similar purpose.
18042 Additionally the NO_RSA no longer completely removes the whole -modulus
18043 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18046 *Ralf S. Engelschall*
18048 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18049 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18051 *Arne Ansper <arne@ats.cyber.ee>*
18053 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18054 to be added. Now both 'req' and 'ca' can use new objects defined in the
18059 * Add cool BIO that does syslog (or event log on NT).
18061 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18063 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18064 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18065 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18066 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18070 * Add preliminary config info for new extension code.
18074 * Make RSA_NO_PADDING really use no padding.
18076 *Ulf Moeller <ulf@fitug.de>*
18078 * Generate errors when private/public key check is done.
18082 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18083 for some CRL extensions and new objects added.
18087 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18088 key usage extension and fuller support for authority key id.
18092 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18093 padding method for RSA, which is recommended for new applications in PKCS
18094 #1 v2.0 (RFC 2437, October 1998).
18095 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18096 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18097 against Bleichbacher's attack on RSA.
18098 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18101 * Updates to the new SSL compression code
18103 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18105 * Fix so that the version number in the master secret, when passed
18106 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18107 (because the server will not accept higher), that the version number
18108 is 0x03,0x01, not 0x03,0x00
18110 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18112 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18113 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18114 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18118 * Support for RAW extensions where an arbitrary extension can be
18119 created by including its DER encoding. See `apps/openssl.cnf` for
18124 * Make sure latest Perl versions don't interpret some generated C array
18125 code as Perl array code in the crypto/err/err_genc.pl script.
18127 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18129 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18130 not many people have the assembler. Various Win32 compilation fixes and
18131 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18132 build instructions.
18136 * Modify configure script 'Configure' to automatically create crypto/date.h
18137 file under Win32 and also build pem.h from pem.org. New script
18138 util/mkfiles.pl to create the MINFO file on environments that can't do a
18139 'make files': perl util/mkfiles.pl >MINFO should work.
18143 * Major rework of DES function declarations, in the pursuit of correctness
18144 and purity. As a result, many evil casts evaporated, and some weirdness,
18145 too. You may find this causes warnings in your code. Zapping your evil
18146 casts will probably fix them. Mostly.
18150 * Fix for a typo in asn1.h. Bug fix to object creation script
18151 obj_dat.pl. It considered a zero in an object definition to mean
18152 "end of object": none of the objects in objects.h have any zeros
18153 so it wasn't spotted.
18155 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18157 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18158 Masking (CBCM). In the absence of test vectors, the best I have been able
18159 to do is check that the decrypt undoes the encrypt, so far. Send me test
18160 vectors if you have them.
18164 * Correct calculation of key length for export ciphers (too much space was
18165 allocated for null ciphers). This has not been tested!
18169 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18170 message is now correct (it understands "crypto" and "ssl" on its
18171 command line). There is also now an "update" option. This will update
18172 the util/ssleay.num and util/libeay.num files with any new functions.
18174 perl util/mkdef.pl crypto ssl update
18175 it will update them.
18179 * Overhauled the Perl interface:
18180 - ported BN stuff to OpenSSL's different BN library
18181 - made the perl/ source tree CVS-aware
18182 - renamed the package from SSLeay to OpenSSL (the files still contain
18183 their history because I've copied them in the repository)
18184 - removed obsolete files (the test scripts will be replaced
18185 by better Test::Harness variants in the future)
18187 *Ralf S. Engelschall*
18189 * First cut for a very conservative source tree cleanup:
18190 1. merge various obsolete readme texts into doc/ssleay.txt
18191 where we collect the old documents and readme texts.
18192 2. remove the first part of files where I'm already sure that we no
18193 longer need them because of three reasons: either they are just temporary
18194 files which were left by Eric or they are preserved original files where
18195 I've verified that the diff is also available in the CVS via "cvs diff
18196 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18197 the crypto/md/ stuff).
18199 *Ralf S. Engelschall*
18201 * More extension code. Incomplete support for subject and issuer alt
18202 name, issuer and authority key id. Change the i2v function parameters
18203 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18204 what that's for :-) Fix to ASN1 macro which messed up
18205 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18209 * Preliminary support for ENUMERATED type. This is largely copied from the
18214 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18216 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18218 * Make sure `make rehash` target really finds the `openssl` program.
18220 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18222 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18223 like to hear about it if this slows down other processors.
18227 * Add CygWin32 platform information to Configure script.
18229 *Alan Batie <batie@aahz.jf.intel.com>*
18231 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18233 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18235 * New program nseq to manipulate netscape certificate sequences
18239 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18244 * Fixes to BN code. Previously the default was to define BN_RECURSION
18245 but the BN code had some problems that would cause failures when
18246 doing certificate verification and some other functions.
18248 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18250 * Add ASN1 and PEM code to support netscape certificate sequences.
18254 * Add ASN1 and PEM code to support netscape certificate sequences.
18258 * Add several PKIX and private extended key usage OIDs.
18262 * Modify the 'ca' program to handle the new extension code. Modify
18263 openssl.cnf for new extension format, add comments.
18267 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18268 and add a sample to openssl.cnf so req -x509 now adds appropriate
18273 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18274 error code, add initial support to X509_print() and x509 application.
18278 * Takes a deep breath and start adding X509 V3 extension support code. Add
18279 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18280 stuff is currently isolated and isn't even compiled yet.
18284 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18285 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18286 Removed the versions check from X509 routines when loading extensions:
18287 this allows certain broken certificates that don't set the version
18288 properly to be processed.
18292 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18293 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18294 can still be regenerated with "make depend".
18298 * Spelling mistake in C version of CAST-128.
18300 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18302 * Changes to the error generation code. The perl script err-code.pl
18303 now reads in the old error codes and retains the old numbers, only
18304 adding new ones if necessary. It also only changes the .err files if new
18305 codes are added. The makefiles have been modified to only insert errors
18306 when needed (to avoid needlessly modifying header files). This is done
18307 by only inserting errors if the .err file is newer than the auto generated
18308 C file. To rebuild all the error codes from scratch (the old behaviour)
18309 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18310 or delete all the .err files.
18314 * CAST-128 was incorrectly implemented for short keys. The C version has
18315 been fixed, but is untested. The assembler versions are also fixed, but
18316 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18317 to regenerate it if needed.
18318 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18319 Hagino <itojun@kame.net>*
18321 * File was opened incorrectly in randfile.c.
18323 *Ulf Möller <ulf@fitug.de>*
18325 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18326 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18327 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18328 al: it's just almost always a UTCTime. Note this patch adds new error
18329 codes so do a "make errors" if there are problems.
18333 * Correct Linux 1 recognition in config.
18335 *Ulf Möller <ulf@fitug.de>*
18337 * Remove pointless MD5 hash when using DSA keys in ca.
18339 *Anonymous <nobody@replay.com>*
18341 * Generate an error if given an empty string as a cert directory. Also
18342 generate an error if handed NULL (previously returned 0 to indicate an
18343 error, but didn't set one).
18345 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18347 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18351 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18352 parameters. This was causing a warning which killed off the Win32 compile.
18356 * Remove C++ style comments from crypto/bn/bn_local.h.
18358 *Neil Costigan <neil.costigan@celocom.com>*
18360 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18361 based on a text string, looking up short and long names and finally
18362 "dot" format. The "dot" format stuff didn't work. Added new function
18363 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18364 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18365 OID is not part of the table.
18369 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18370 X509_LOOKUP_by_alias().
18374 * Sort openssl functions by name.
18378 * Get the `gendsa` command working and add it to the `list` command. Remove
18379 encryption from sample DSA keys (in case anyone is interested the password
18384 * Make *all* `*_free` functions accept a NULL pointer.
18386 *Frans Heymans <fheymans@isaserver.be>*
18388 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18391 *Anonymous <nobody@replay.com>*
18393 * s_server should send the CAfile as acceptable CAs, not its own cert.
18395 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18397 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18399 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18401 * Temp key "for export" tests were wrong in s3_srvr.c.
18403 *Anonymous <nobody@replay.com>*
18405 * Add prototype for temp key callback functions
18406 SSL_CTX_set_tmp_{rsa,dh}_callback().
18410 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18411 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18415 * X509_name_add_entry() freed the wrong thing after an error.
18417 *Arne Ansper <arne@ats.cyber.ee>*
18419 * rsa_eay.c would attempt to free a NULL context.
18421 *Arne Ansper <arne@ats.cyber.ee>*
18423 * BIO_s_socket() had a broken should_retry() on Windoze.
18425 *Arne Ansper <arne@ats.cyber.ee>*
18427 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18429 *Arne Ansper <arne@ats.cyber.ee>*
18431 * Make sure the already existing X509_STORE->depth variable is initialized
18432 in X509_STORE_new(), but document the fact that this variable is still
18433 unused in the certificate verification process.
18435 *Ralf S. Engelschall*
18437 * Fix the various library and `apps/` files to free up pkeys obtained from
18438 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18442 * Fix reference counting in X509_PUBKEY_get(). This makes
18443 demos/maurice/example2.c work, amongst others, probably.
18445 *Steve Henson and Ben Laurie*
18447 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18448 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18449 are no longer created. This way we have a single and consistent command
18450 line interface `openssl <command>`, similar to `cvs <command>`.
18452 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18454 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18455 BIT STRING wrapper always have zero unused bits.
18459 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18463 * Make the top-level INSTALL documentation easier to understand.
18467 * Makefiles updated to exit if an error occurs in a sub-directory
18468 make (including if user presses ^C) [Paul Sutton]
18470 * Make Montgomery context stuff explicit in RSA data structure.
18474 * Fix build order of pem and err to allow for generated pem.h.
18478 * Fix renumbering bug in X509_NAME_delete_entry().
18482 * Enhanced the err-ins.pl script so it makes the error library number
18483 global and can add a library name. This is needed for external ASN1 and
18484 other error libraries.
18488 * Fixed sk_insert which never worked properly.
18492 * Fix ASN1 macros so they can handle indefinite length constructed
18493 EXPLICIT tags. Some non standard certificates use these: they can now
18498 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18499 into a single doc/ssleay.txt bundle. This way the information is still
18500 preserved but no longer messes up this directory. Now it's new room for
18501 the new set of documentation files.
18503 *Ralf S. Engelschall*
18505 * SETs were incorrectly DER encoded. This was a major pain, because they
18506 shared code with SEQUENCEs, which aren't coded the same. This means that
18507 almost everything to do with SETs or SEQUENCEs has either changed name or
18508 number of arguments.
18510 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18512 * Fix test data to work with the above.
18516 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18517 was already fixed by Eric for 0.9.1 it seems.
18519 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18521 * Autodetect FreeBSD3.
18525 * Fix various bugs in Configure. This affects the following platforms:
18529 unixware-2.0-pentium
18534 * Eliminate generated files from CVS. Reorder tests to regenerate files
18535 before they are needed.
18539 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18543 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18545 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18546 changed SSLeay to OpenSSL in version strings.
18548 *Ralf S. Engelschall*
18550 * Some fixups to the top-level documents.
18554 * Fixed the nasty bug where rsaref.h was not found under compile-time
18555 because the symlink to include/ was missing.
18557 *Ralf S. Engelschall*
18559 * Incorporated the popular no-RSA/DSA-only patches
18560 which allow to compile a RSA-free SSLeay.
18562 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18564 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18565 when "ssleay" is still not found.
18567 *Ralf S. Engelschall*
18569 * Added more platforms to Configure: Cray T3E, HPUX 11,
18571 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18573 * Updated the README file.
18575 *Ralf S. Engelschall*
18577 * Added various .cvsignore files in the CVS repository subdirs
18578 to make a "cvs update" really silent.
18580 *Ralf S. Engelschall*
18582 * Recompiled the error-definition header files and added
18583 missing symbols to the Win32 linker tables.
18585 *Ralf S. Engelschall*
18587 * Cleaned up the top-level documents;
18588 o new files: CHANGES and LICENSE
18589 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18590 o merged COPYRIGHT into LICENSE
18591 o removed obsolete TODO file
18592 o renamed MICROSOFT to INSTALL.W32
18594 *Ralf S. Engelschall*
18596 * Removed dummy files from the 0.9.1b source tree:
18598 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18599 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18600 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18601 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18602 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18605 *Ralf S. Engelschall*
18607 * Added various platform portability fixes.
18611 * The Genesis of the OpenSSL rpject:
18612 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18613 Young and Tim J. Hudson created while they were working for C2Net until
18616 *The OpenSSL Project*
18618 ### Changes between 0.9.0b and 0.9.1b [not released]
18620 * Updated a few CA certificates under certs/
18624 * Changed some BIGNUM api stuff.
18628 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18629 DGUX x86, Linux Alpha, etc.
18633 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18634 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18639 * Add -strparse option to asn1pars program which parses nested
18642 *Dr Stephen Henson <shenson@bigfoot.com>*
18644 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18648 * DSA fix for "ca" program.
18652 * Added "-genkey" option to "dsaparam" program.
18656 * Added RIPE MD160 (rmd160) message digest.
18660 * Added -a (all) option to "ssleay version" command.
18664 * Added PLATFORM define which is the id given to Configure.
18668 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18672 * Extended the ASN.1 parser routines.
18676 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18680 * Added a BN_CTX to the BN library.
18684 * Fixed the weak key values in DES library
18688 * Changed API in EVP library for cipher aliases.
18692 * Added support for RC2/64bit cipher.
18696 * Converted the lhash library to the crypto/mem.c functions.
18700 * Added more recognized ASN.1 object ids.
18704 * Added more RSA padding checks for SSL/TLS.
18708 * Added BIO proxy/filter functionality.
18712 * Added extra_certs to SSL_CTX which can be used
18713 send extra CA certificates to the client in the CA cert chain sending
18714 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18718 * Now Fortezza is denied in the authentication phase because
18719 this is key exchange mechanism is not supported by SSLeay at all.
18723 * Additional PKCS1 checks.
18727 * Support the string "TLSv1" for all TLS v1 ciphers.
18731 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18732 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18736 * Fixed a few memory leaks.
18740 * Fixed various code and comment typos.
18744 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18745 bytes sent in the client random.
18747 *Edward Bishop <ebishop@spyglass.com>*
18751 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18752 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18753 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18754 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18755 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18756 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18757 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18758 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18759 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18760 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18761 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18762 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18763 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18764 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18765 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18766 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18767 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18768 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18769 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18770 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18771 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18772 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18773 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18774 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18775 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18776 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18777 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18778 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18779 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18780 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18781 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18782 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18783 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18784 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18785 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18786 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18787 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18788 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18789 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18790 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18791 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18792 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18793 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18794 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18795 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18796 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18797 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18798 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18799 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18800 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18801 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18802 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18803 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18804 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18805 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18806 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18807 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18808 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18809 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18810 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18811 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18812 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18813 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18814 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18815 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18816 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18817 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18818 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18819 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18820 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18821 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18822 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18823 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18824 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18825 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18826 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18827 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18828 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18829 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18830 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18831 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18832 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18833 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18834 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18835 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18836 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18837 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18838 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18839 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18840 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18841 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18842 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18843 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18844 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18845 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18846 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18847 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18848 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18849 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18850 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18851 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18852 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18853 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18854 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18855 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18856 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18857 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18858 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18859 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18860 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18861 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18862 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18863 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18864 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18865 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18866 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18867 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18868 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18869 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18870 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18871 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18872 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18873 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18874 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18875 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18876 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18877 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18878 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18879 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18880 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18881 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18882 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18883 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18884 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18885 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18886 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18887 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18888 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18889 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18890 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18891 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18892 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18893 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18894 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18895 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18896 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18897 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18898 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18899 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18900 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18901 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18902 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18903 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18904 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18905 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18906 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18907 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18908 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18909 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18910 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18911 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18912 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655