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 obsolete X9.31 RSA key generation related functions
27 BN_X931_generate_Xpq(), BN_X931_derive_prime_ex(), and
28 BN_X931_generate_prime_ex().
32 * Deprecated the type OCSP_REQ_CTX and the functions OCSP_REQ_CTX_new(),
33 OCSP_REQ_CTX_free(), OCSP_REQ_CTX_http(), OCSP_REQ_CTX_add1_header(),
34 OCSP_REQ_CTX_i2d(), OCSP_REQ_CTX_nbio(), OCSP_REQ_CTX_nbio_d2i(),
35 OCSP_REQ_CTX_get0_mem_bio() and OCSP_set_max_response_length(). These
36 were used to collect all necessary data to form a HTTP request, and to
37 perform the HTTP transfer with that request. With OpenSSL 3.0, the
38 type is OSSL_HTTP_REQ_CTX, and the deprecated functions are replaced
39 with OSSL_HTTP_REQ_CTX_new(), OSSL_HTTP_REQ_CTX_free(),
40 OSSL_HTTP_REQ_CTX_set_request_line(), OSSL_HTTP_REQ_CTX_add1_header(),
41 OSSL_HTTP_REQ_CTX_i2d(), OSSL_HTTP_REQ_CTX_nbio(),
42 OSSL_HTTP_REQ_CTX_sendreq_d2i(), OSSL_HTTP_REQ_CTX_get0_mem_bio() and
43 OSSL_HTTP_REQ_CTX_set_max_response_length().
45 *Rich Salz and Richard Levitte*
47 * Validation of SM2 keys has been separated from the validation of regular EC
48 keys, allowing to improve the SM2 validation process to reject loaded private
49 keys that are not conforming to the SM2 ISO standard.
50 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
55 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
56 switches: a validation failure triggers an early exit, returning a failure
57 exit status to the parent process.
61 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
62 to ignore unknown ciphers.
66 * The `-cipher-commands` and `-digest-commands` options
67 of the command line utility `list` have been deprecated.
68 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
72 * Deprecated all the libcrypto and libssl error string loading
73 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
74 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
75 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
76 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
77 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
78 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
79 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
80 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
81 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
82 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
84 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
85 now loads error strings automatically.
89 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
90 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
91 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
92 are to be used by servers requiring ephemeral DH keys. Instead applications
93 should consider using the built-in DH parameters that are available by
94 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
95 necessary then applications can use the alternative functions
96 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
97 replacement for the "callback" functions. The callback was originally useful
98 in order to have different parameters for export and non-export ciphersuites.
99 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
100 functions should be replaced by one of the other methods described above.
104 * The `-crypt` option to the `passwd` command line tool has been removed.
108 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
113 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
115 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
116 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
117 The inverse ciphers use AES decryption for wrapping, and
118 AES encryption for unwrapping.
122 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
123 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
124 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
125 OpenSSL 3.0 these are replaced by the more generic functions
126 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
127 The old versions have been converted to deprecated macros that just call the
132 * The security callback, which can be customised by application code, supports
133 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
134 in the "other" parameter. In most places this is what is passed. All these
135 places occur server side. However there was one client side call of this
136 security operation and it passed a DH object instead. This is incorrect
137 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
138 of the other locations. Therefore this client side call has been changed to
139 pass an EVP_PKEY instead.
143 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
144 interface. Their functionality remains unchanged.
148 * Added new option for 'openssl list', '-providers', which will display the
149 list of loaded providers, their names, version and status. It optionally
150 displays their gettable parameters.
154 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
155 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
156 type is internally recognised so the workaround is no longer needed.
158 Functionality is still retained as it is, but will only work with
159 EVP_PKEYs with a legacy internal key.
163 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
164 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
168 * Changed all "STACK" functions to be macros instead of inline functions. Macro
169 parameters are still checked for type safety at compile time via helper
174 * Remove the RAND_DRBG API
176 The RAND_DRBG API did not fit well into the new provider concept as
177 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
178 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
179 and some of its API calls are rather low-level. This holds in particular
180 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
182 Adding a compatibility layer to continue supporting the RAND_DRBG API as
183 a legacy API for a regular deprecation period turned out to come at the
184 price of complicating the new provider API unnecessarily. Since the
185 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
188 *Paul Dale and Matthias St. Pierre*
190 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
191 as well as actual hostnames.
195 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
196 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
197 conversely, silently ignore DTLS protocol version bounds when configuring
198 TLS-based contexts. The commands can be repeated to set bounds of both
199 types. The same applies with the corresponding "min_protocol" and
200 "max_protocol" command-line switches, in case some application uses both TLS
203 SSL_CTX instances that are created for a fixed protocol version (e.g.
204 `TLSv1_server_method()`) also silently ignore version bounds. Previously
205 attempts to apply bounds to these protocol versions would result in an
206 error. Now only the "version-flexible" SSL_CTX instances are subject to
207 limits in configuration files in command-line options.
211 * Deprecated the `ENGINE` API. Engines should be replaced with providers
216 * Reworked the recorded ERR codes to make better space for system errors.
217 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
218 given code is a system error (true) or an OpenSSL error (false).
222 * Reworked the test perl framework to better allow parallel testing.
224 *Nicola Tuveri and David von Oheimb*
226 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
227 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
231 * 'Configure' has been changed to figure out the configuration target if
232 none is given on the command line. Consequently, the 'config' script is
233 now only a mere wrapper. All documentation is changed to only mention
236 *Rich Salz and Richard Levitte*
238 * Added a library context `OSSL_LIB_CTX` that applications as well as
239 other libraries can use to form a separate context within which
240 libcrypto operations are performed.
242 There are two ways this can be used:
244 - Directly, by passing a library context to functions that take
245 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
247 - Indirectly, by creating a new library context and then assigning
248 it as the new default, with `OSSL_LIB_CTX_set0_default`.
250 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
251 apart from the functions directly related to `OSSL_LIB_CTX`, accept
252 NULL to indicate that the default library context should be used.
254 Library code that changes the default library context using
255 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
256 second call before returning to the caller.
258 _(Note: the library context was initially called `OPENSSL_CTX` and
259 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
263 * Handshake now fails if Extended Master Secret extension is dropped
268 * Dropped interactive mode from the `openssl` program. From now on,
269 running it without arguments is equivalent to `openssl help`.
273 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
274 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
275 While the old function names have been retained for backward compatibility
276 they should not be used in new developments
277 because their return values are confusing: Unlike other `_cmp()` functions
278 they do not return 0 in case their arguments are equal.
282 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
283 `EC_GROUP_get_field_type()`.
287 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
288 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
289 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
290 Applications should rely on the library automatically assigning a suitable
291 EC_METHOD internally upon EC_GROUP construction.
295 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
296 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
297 assigned internally without application intervention.
298 Users of EC_GROUP_new() should switch to a different suitable constructor.
302 * Add CAdES-BES signature verification support, mostly derived
303 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
305 *Filipe Raimundo da Silva*
307 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
311 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
312 functions are not widely used and now OpenSSL automatically perform this
313 conversion when needed.
317 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
318 EC_KEY_precompute_mult(). These functions are not widely used and
319 applications should instead switch to named curves which OpenSSL has
320 hardcoded lookup tables for.
324 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
325 should instead use the L<EC_POINT_mul(3)> function.
329 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
330 that are not applicable to the new provider model. Applications should
331 instead use EVP_default_properties_is_fips_enabled() and
332 EVP_default_properties_enable_fips().
336 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
337 is set, an unexpected EOF is ignored, it pretends a close notify was received
338 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
342 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
343 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
344 used and applications should instead use the
345 L<EC_POINT_set_affine_coordinates(3)> and
346 L<EC_POINT_get_affine_coordinates(3)> functions.
350 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
351 arrays to be more easily constructed via a series of utility functions.
352 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
353 the various push functions and finally convert to a passable OSSL_PARAM
354 array using OSSL_PARAM_BLD_to_param().
358 * The security strength of SHA1 and MD5 based signatures in TLS has been
359 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
360 working at the default security level of 1 and instead requires security
361 level 0. The security level can be changed either using the cipher string
362 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
366 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
367 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
368 internal keys, if they correspond to one of those built in types.
372 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
373 contain a provider side internal key.
377 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
378 They are old functions that we don't use, and that you could disable with
379 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
383 * Project text documents not yet having a proper file name extension
384 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
385 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
389 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
390 have been converted to Markdown with the goal to produce documents
391 which not only look pretty when viewed online in the browser, but
392 remain well readable inside a plain text editor.
394 To achieve this goal, a 'minimalistic' Markdown style has been applied
395 which avoids formatting elements that interfere too much with the
396 reading flow in the text file. For example, it
398 * avoids [ATX headings][] and uses [setext headings][] instead
399 (which works for `<h1>` and `<h2>` headings only).
400 * avoids [inline links][] and uses [reference links][] instead.
401 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
403 [ATX headings]: https://github.github.com/gfm/#atx-headings
404 [setext headings]: https://github.github.com/gfm/#setext-headings
405 [inline links]: https://github.github.com/gfm/#inline-link
406 [reference links]: https://github.github.com/gfm/#reference-link
407 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
408 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
410 *Matthias St. Pierre*
412 * The test suite is changed to preserve results of each test recipe.
413 A new directory test-runs/ with subdirectories named like the
414 test recipes are created in the build tree for this purpose.
418 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
419 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
420 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
422 *David von Oheimb, Martin Peylo*
424 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
425 The legacy OCSP-focused and only partly documented API is retained for
426 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
430 * Added `util/check-format.pl`, a tool for checking adherence to the
431 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
432 The checks performed are incomplete and yield some false positives.
433 Still the tool should be useful for detecting most typical glitches.
437 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
438 If domain name resolution yields multiple IP addresses all of them are tried
439 after `connect()` failures.
443 * All of the low level RSA functions have been deprecated including:
445 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
446 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
447 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
448 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
449 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
450 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
451 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
452 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
453 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
454 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
455 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
456 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
457 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
458 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
459 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
460 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
461 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
462 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
463 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
464 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
465 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
466 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
467 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
468 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
469 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
470 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
471 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
472 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
474 Use of these low level functions has been informally discouraged for a long
475 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
476 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
477 L<EVP_PKEY_decrypt(3)>.
481 * X509 certificates signed using SHA1 are no longer allowed at security
483 In TLS/SSL the default security level is 1. It can be set either
484 using the cipher string with `@SECLEVEL`, or calling
485 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
486 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
488 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
489 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
490 options of the commands.
494 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
495 modified to use PKEY APIs. These commands are now in maintenance mode
496 and no new features will be added to them.
500 * The command line utility rsautl has been deprecated.
501 Instead use the pkeyutl program.
505 * The command line utilities genrsa and rsa have been modified to use PKEY
506 APIs. They now write PKCS#8 keys by default. These commands are now in
507 maintenance mode and no new features will be added to them.
511 * All of the low level DH functions have been deprecated including:
513 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
514 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
515 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
516 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
517 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
518 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
519 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
520 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
521 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
522 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
523 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
524 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
525 DH_meth_get_generate_params and DH_meth_set_generate_params.
527 Use of these low level functions has been informally discouraged for a long
528 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
529 and L<EVP_PKEY_derive(3)>.
531 Additionally functions that read and write DH objects such as d2i_DHparams,
532 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
533 functions have also been deprecated. Applications should instead use the
534 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
536 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
537 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
538 `EVP_PKEY_set1_DH()` are also deprecated.
539 Applications should instead either read or write an
540 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
541 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
543 *Paul Dale and Matt Caswell*
545 * All of the low level DSA functions have been deprecated including:
547 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
548 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
549 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
550 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
551 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
552 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
553 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
554 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
555 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
556 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
557 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
558 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
559 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
560 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
561 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
562 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
563 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
565 Use of these low level functions has been informally discouraged for a long
566 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
567 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
571 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
572 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
573 This means that applications don't have to look at the curve NID and
574 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
575 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
576 a no-op when the EVP_PKEY is already of the given type.
578 Parameter and key generation is also reworked to make it possible
579 to generate EVP_PKEY_SM2 parameters and keys without having to go
580 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
581 However, code that does the latter will still work as before.
585 * Deprecated low level ECDH and ECDSA functions. These include:
587 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
588 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
591 Use of these low level functions has been informally discouraged for a long
592 time. Instead applications should use the EVP_PKEY_derive(3),
593 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
597 * Deprecated the EC_KEY_METHOD functions. These include:
599 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
600 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
601 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
602 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
603 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
604 EC_KEY_METHOD_get_verify.
606 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
610 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
611 and EVP_PKEY_decrypt() instead.
612 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
613 and EVP_PKEY_encrypt() instead.
617 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
618 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
619 a new formulation to include all the things it can be used for,
620 as well as words of caution.
624 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
625 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
629 * All of the low level HMAC functions have been deprecated including:
631 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
632 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
635 Use of these low level functions has been informally discouraged for a long
636 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
637 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
638 and L<EVP_MAC_final(3)>.
642 * Over two thousand fixes were made to the documentation, including:
643 - Common options (such as -rand/-writerand, TLS version control, etc)
644 were refactored and point to newly-enhanced descriptions in openssl.pod.
645 - Added style conformance for all options (with help from Richard Levitte),
646 documented all reported missing options, added a CI build to check
647 that all options are documented and that no unimplemented options
649 - Documented some internals, such as all use of environment variables.
650 - Addressed all internal broken L<> references.
654 * All of the low level CMAC functions have been deprecated including:
656 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
657 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
659 Use of these low level functions has been informally discouraged for a long
660 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
661 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
662 and L<EVP_MAC_final(3)>.
666 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
667 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
670 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
671 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
672 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
673 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
674 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
675 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
676 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
677 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
678 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
679 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
681 Use of these low level functions has been informally discouraged
682 for a long time. Applications should use the EVP_DigestInit_ex(3),
683 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
687 * Corrected the documentation of the return values from the `EVP_DigestSign*`
688 set of functions. The documentation mentioned negative values for some
689 errors, but this was never the case, so the mention of negative values
692 Code that followed the documentation and thereby check with something
693 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
697 * All of the low level cipher functions have been deprecated including:
699 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
700 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
701 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
702 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
703 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
704 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
705 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
706 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
707 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
708 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
709 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
710 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
711 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
712 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
713 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
714 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
715 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
716 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
717 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
718 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
719 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
720 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
721 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
722 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
723 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
724 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
725 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
726 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
727 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
729 Use of these low level functions has been informally discouraged for
730 a long time. Applications should use the high level EVP APIs, e.g.
731 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
732 equivalently named decrypt functions instead.
734 *Matt Caswell and Paul Dale*
736 * Removed include/openssl/opensslconf.h.in and replaced it with
737 include/openssl/configuration.h.in, which differs in not including
738 <openssl/macros.h>. A short header include/openssl/opensslconf.h
739 was added to include both.
741 This allows internal hacks where one might need to modify the set
742 of configured macros, for example this if deprecated symbols are
743 still supposed to be available internally:
745 #include <openssl/configuration.h>
747 #undef OPENSSL_NO_DEPRECATED
748 #define OPENSSL_SUPPRESS_DEPRECATED
750 #include <openssl/macros.h>
752 This should not be used by applications that use the exported
753 symbols, as that will lead to linking errors.
757 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
758 used in exponentiation with 512-bit moduli. No EC algorithms are
759 affected. Analysis suggests that attacks against 2-prime RSA1024,
760 3-prime RSA1536, and DSA1024 as a result of this defect would be very
761 difficult to perform and are not believed likely. Attacks against DH512
762 are considered just feasible. However, for an attack the target would
763 have to re-use the DH512 private key, which is not recommended anyway.
764 Also applications directly using the low level API BN_mod_exp may be
765 affected if they use BN_FLG_CONSTTIME.
770 * Most memory-debug features have been deprecated, and the functionality
771 replaced with no-ops.
775 * Added documentation for the STACK API.
779 * Introduced a new method type and API, OSSL_ENCODER, to
780 represent generic encoders. An implementation is expected to
781 be able to encode an object associated with a given name (such
782 as an algorithm name for an asymmetric key) into forms given by
783 implementation properties.
785 Encoders are primarily used from inside libcrypto, through
786 calls to functions like EVP_PKEY_print_private(),
787 PEM_write_bio_PrivateKey() and similar.
789 Encoders are specified in such a way that they can be made to
790 directly handle the provider side portion of an object, if this
791 provider side part comes from the same provider as the encoder
792 itself, but can also be made to handle objects in parametrized
793 form (as an OSSL_PARAM array of data). This allows a provider to
794 offer generic encoders as a service for any other provider.
798 * Added a .pragma directive to the syntax of configuration files, to
799 allow varying behavior in a supported and predictable manner.
800 Currently added pragma:
804 This allows dollar signs to be a keyword character unless it's
805 followed by a opening brace or parenthesis. This is useful for
806 platforms where dollar signs are commonly used in names, such as
807 volume names and system directory names on VMS.
811 * Added functionality to create an EVP_PKEY from user data. This
812 is effectively the same as creating a RSA, DH or DSA object and
813 then assigning them to an EVP_PKEY, but directly using algorithm
814 agnostic EVP functions. A benefit is that this should be future
815 proof for public key algorithms to come.
819 * Change the interpretation of the '--api' configuration option to
820 mean that this is a desired API compatibility level with no
821 further meaning. The previous interpretation, that this would
822 also mean to remove all deprecated symbols up to and including
823 the given version, no requires that 'no-deprecated' is also used
824 in the configuration.
826 When building applications, the desired API compatibility level
827 can be set with the OPENSSL_API_COMPAT macro like before. For
828 API compatibility version below 3.0, the old style numerical
829 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
830 For version 3.0 and on, the value is expected to be the decimal
831 value calculated from the major and minor version like this:
833 MAJOR * 10000 + MINOR * 100
837 -DOPENSSL_API_COMPAT=30000 For 3.0
838 -DOPENSSL_API_COMPAT=30200 For 3.2
840 To hide declarations that are deprecated up to and including the
841 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
842 given when building the application as well.
846 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
847 access to certificate and CRL stores via URIs and OSSL_STORE
850 This adds the following functions:
852 - X509_LOOKUP_store()
853 - X509_STORE_load_file()
854 - X509_STORE_load_path()
855 - X509_STORE_load_store()
856 - SSL_add_store_cert_subjects_to_stack()
857 - SSL_CTX_set_default_verify_store()
858 - SSL_CTX_load_verify_file()
859 - SSL_CTX_load_verify_dir()
860 - SSL_CTX_load_verify_store()
864 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
865 The presence of this system service is determined at run-time.
869 * Added functionality to create an EVP_PKEY context based on data
870 for methods from providers. This takes an algorithm name and a
871 property query string and simply stores them, with the intent
872 that any operation that uses this context will use those strings
873 to fetch the needed methods implicitly, thereby making the port
874 of application written for pre-3.0 OpenSSL easier.
878 * The undocumented function NCONF_WIN32() has been deprecated; for
879 conversion details see the HISTORY section of doc/man5/config.pod
883 * Introduced the new functions EVP_DigestSignInit_ex() and
884 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
885 EVP_DigestVerifyUpdate() have been converted to functions. See the man
886 pages for further details.
890 * Over two thousand fixes were made to the documentation, including:
891 adding missing command flags, better style conformance, documentation
894 *Rich Salz, Richard Levitte*
896 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
897 X25519, X448, Ed25519 and Ed448.
901 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
906 * Deprecated the public definition of `ERR_STATE` as well as the function
907 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
912 * Added ERR functionality to give callers access to the stored function
913 names that have replaced the older function code based functions.
915 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
916 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
917 ERR_peek_error_all() and ERR_peek_last_error_all().
919 These functions have become deprecated: ERR_get_error_line(),
920 ERR_get_error_line_data(), ERR_peek_error_line_data(),
921 ERR_peek_last_error_line_data() and ERR_func_error_string().
923 Users are recommended to use ERR_get_error_all(), or to pick information
924 with ERR_peek functions and finish off with getting the error code by using
929 * Extended testing to be verbose for failing tests only. The make variables
930 VERBOSE_FAILURE or VF can be used to enable this:
932 $ make VF=1 test # Unix
933 $ mms /macro=(VF=1) test ! OpenVMS
934 $ nmake VF=1 test # Windows
938 * Added the `-copy_extensions` option to the `x509` command for use with
939 `-req` and `-x509toreq`. When given with the `copy` or `copyall` argument,
940 all extensions in the request are copied to the certificate or vice versa.
942 *David von Oheimb*, *Kirill Stefanenkov <kirill_stefanenkov@rambler.ru>*
944 * Added the `-copy_extensions` option to the `req` command for use with
945 `-x509`. When given with the `copy` or `copyall` argument,
946 all extensions in the certification request are copied to the certificate.
950 * The `x509`, `req`, and `ca` commands now make sure that X.509v3 certificates
951 they generate are by default RFC 5280 compliant in the following sense:
952 There is a subjectKeyIdentifier extension with a hash value of the public key
953 and for not self-signed certs there is an authorityKeyIdentifier extension
954 with a keyIdentifier field or issuer information identifying the signing key.
955 This is done unless some configuration overrides the new default behavior,
956 such as `subjectKeyIdentifier = none` and `authorityKeyIdentifier = none`.
960 * Added several checks to `X509_verify_cert()` according to requirements in
961 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
962 (which may be done by using the CLI option `-x509_strict`):
963 * The basicConstraints of CA certificates must be marked critical.
964 * CA certificates must explicitly include the keyUsage extension.
965 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
966 * The issuer name of any certificate must not be empty.
967 * The subject name of CA certs, certs with keyUsage crlSign,
968 and certs without subjectAlternativeName must not be empty.
969 * If a subjectAlternativeName extension is given it must not be empty.
970 * The signatureAlgorithm field and the cert signature must be consistent.
971 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
972 must not be marked critical.
973 * The authorityKeyIdentifier must be given for X.509v3 certs
974 unless they are self-signed.
975 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
979 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
980 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
984 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
985 used even when parsing explicit parameters, when loading a encoded key
986 or calling `EC_GROUP_new_from_ecpkparameters()`/
987 `EC_GROUP_new_from_ecparameters()`.
988 This prevents bypass of security hardening and performance gains,
989 especially for curves with specialized EC_METHODs.
990 By default, if a key encoded with explicit parameters is loaded and later
991 encoded, the output is still encoded with explicit parameters, even if
992 internally a "named" EC_GROUP is used for computation.
996 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
997 this change, EC_GROUP_set_generator would accept order and/or cofactor as
998 NULL. After this change, only the cofactor parameter can be NULL. It also
999 does some minimal sanity checks on the passed order.
1004 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1005 An attack is simple, if the first CMS_recipientInfo is valid but the
1006 second CMS_recipientInfo is chosen ciphertext. If the second
1007 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1008 encryption key will be replaced by garbage, and the message cannot be
1009 decoded, but if the RSA decryption fails, the correct encryption key is
1010 used and the recipient will not notice the attack.
1011 As a work around for this potential attack the length of the decrypted
1012 key must be equal to the cipher default key length, in case the
1013 certifiate is not given and all recipientInfo are tried out.
1014 The old behaviour can be re-enabled in the CMS code by setting the
1015 CMS_DEBUG_DECRYPT flag.
1019 * Early start up entropy quality from the DEVRANDOM seed source has been
1020 improved for older Linux systems. The RAND subsystem will wait for
1021 /dev/random to be producing output before seeding from /dev/urandom.
1022 The seeded state is stored for future library initialisations using
1023 a system global shared memory segment. The shared memory identifier
1024 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1025 the desired value. The default identifier is 114.
1029 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
1030 when primes for RSA keys are computed.
1031 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1032 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1033 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1034 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1035 This avoids possible fingerprinting of newly generated RSA modules.
1039 * Correct the extended master secret constant on EBCDIC systems. Without this
1040 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1041 negotiate EMS will fail. Unfortunately this also means that TLS connections
1042 between EBCDIC systems with this fix, and EBCDIC systems without this
1043 fix will fail if they negotiate EMS.
1047 * Changed the library initialisation so that the config file is now loaded
1048 by default. This was already the case for libssl. It now occurs for both
1049 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1050 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1054 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1055 where the former acts as a replacement for `ERR_put_error()`, and the
1056 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1057 `ERR_raise_data()` adds more flexibility by taking a format string and
1058 an arbitrary number of arguments following it, to be processed with
1063 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1064 to check if a named provider is loaded and available. When called, it
1065 will also activate all fallback providers if such are still present.
1069 * Enforce a minimum DH modulus size of 512 bits.
1073 * Changed DH parameters to generate the order q subgroup instead of 2q.
1074 Previously generated DH parameters are still accepted by DH_check
1075 but DH_generate_key works around that by clearing bit 0 of the
1076 private key for those. This avoids leaking bit 0 of the private key.
1080 * Significantly reduce secure memory usage by the randomness pools.
1084 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1089 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1090 algorithms. An implementation of a key exchange algorithm can be obtained
1091 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1092 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1093 the older EVP_PKEY_derive_init() function. See the man pages for the new
1094 functions for further details.
1098 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1102 * Removed the function names from error messages and deprecated the
1105 * Removed NextStep support and the macro OPENSSL_UNISTD
1109 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1110 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1111 Also removed "export var as function" capability; we do not export
1112 variables, only functions.
1116 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1117 an error and 1 indicating success. In previous versions of OpenSSL this
1118 was a void type. If a key was set longer than the maximum possible this
1123 * Support SM2 signing and verification schemes with X509 certificate.
1127 * Use SHA256 as the default digest for TS query in the `ts` app.
1131 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1132 This checks that the salt length is at least 128 bits, the derived key
1133 length is at least 112 bits, and that the iteration count is at least 1000.
1134 For backwards compatibility these checks are disabled by default in the
1135 default provider, but are enabled by default in the fips provider.
1136 To enable or disable these checks use the control
1137 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1141 * Default cipher lists/suites are now available via a function, the
1142 #defines are deprecated.
1146 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1147 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1148 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1152 * Join the directories crypto/x509 and crypto/x509v3
1156 * Added command 'openssl kdf' that uses the EVP_KDF API.
1160 * Added command 'openssl mac' that uses the EVP_MAC API.
1164 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1165 as default directories. Also added the command 'openssl info'
1166 for scripting purposes.
1170 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1171 deprecated. These undocumented functions were never integrated into the EVP
1172 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1173 Bi-directional IGE mode. These modes were never formally standardised and
1174 usage of these functions is believed to be very small. In particular
1175 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1176 is ever used. The security implications are believed to be minimal, but
1177 this issue was never fixed for backwards compatibility reasons. New code
1178 should not use these modes.
1182 * Add prediction resistance to the DRBG reseeding process.
1186 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1187 mandated by IEEE Std 1619-2018.
1191 * Added newline escaping functionality to a filename when using openssl dgst.
1192 This output format is to replicate the output format found in the `*sum`
1193 checksum programs. This aims to preserve backward compatibility.
1195 *Matt Eaton, Richard Levitte, and Paul Dale*
1197 * Removed the heartbeat message in DTLS feature, as it has very
1198 little usage and doesn't seem to fulfill a valuable purpose.
1199 The configuration option is now deprecated.
1203 * Changed the output of 'openssl {digestname} < file' to display the
1204 digest name in its output.
1208 * Added a new generic trace API which provides support for enabling
1209 instrumentation through trace output. This feature is mainly intended
1210 as an aid for developers and is disabled by default. To utilize it,
1211 OpenSSL needs to be configured with the `enable-trace` option.
1213 If the tracing API is enabled, the application can activate trace output
1214 by registering BIOs as trace channels for a number of tracing and debugging
1217 The `openssl` program has been expanded to enable any of the types
1218 available via environment variables defined by the user, and serves as
1219 one possible example on how to use this functionality.
1221 *Richard Levitte & Matthias St. Pierre*
1223 * Added build tests for C++. These are generated files that only do one
1224 thing, to include one public OpenSSL head file each. This tests that
1225 the public header files can be usefully included in a C++ application.
1227 This test isn't enabled by default. It can be enabled with the option
1228 'enable-buildtest-c++'.
1232 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1236 * Add KMAC to EVP_MAC.
1240 * Added property based algorithm implementation selection framework to
1245 * Added SCA hardening for modular field inversion in EC_GROUP through
1246 a new dedicated field_inv() pointer in EC_METHOD.
1247 This also addresses a leakage affecting conversions from projective
1248 to affine coordinates.
1250 *Billy Bob Brumley, Nicola Tuveri*
1252 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1253 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1254 those algorithms that were already supported through the EVP_PKEY API
1255 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1256 and scrypt are now wrappers that call EVP_KDF.
1260 * Build devcrypto engine as a dynamic engine.
1262 *Eneas U de Queiroz*
1264 * Add keyed BLAKE2 to EVP_MAC.
1268 * Fix a bug in the computation of the endpoint-pair shared secret used
1269 by DTLS over SCTP. This breaks interoperability with older versions
1270 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1271 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1272 interoperability with such broken implementations. However, enabling
1273 this switch breaks interoperability with correct implementations.
1275 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1276 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1280 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1284 * Change the license to the Apache License v2.0.
1288 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1290 - Major releases (indicated by incrementing the MAJOR release number)
1291 may introduce incompatible API/ABI changes.
1292 - Minor releases (indicated by incrementing the MINOR release number)
1293 may introduce new features but retain API/ABI compatibility.
1294 - Patch releases (indicated by incrementing the PATCH number)
1295 are intended for bug fixes and other improvements of existing
1296 features only (like improving performance or adding documentation)
1297 and retain API/ABI compatibility.
1301 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1305 * Remove the 'dist' target and add a tarball building script. The
1306 'dist' target has fallen out of use, and it shouldn't be
1307 necessary to configure just to create a source distribution.
1311 * Recreate the OS390-Unix config target. It no longer relies on a
1312 special script like it did for OpenSSL pre-1.1.0.
1316 * Instead of having the source directories listed in Configure, add
1317 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1322 * Add GMAC to EVP_MAC.
1326 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1330 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1331 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1332 to facilitate the continued use of MACs through raw private keys in
1333 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1337 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1338 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1342 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1343 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1344 are retained for backwards compatibility.
1348 * AES-XTS mode now enforces that its two keys are different to mitigate
1349 the attacked described in "Efficient Instantiations of Tweakable
1350 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1351 Details of this attack can be obtained from:
1352 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1356 * Rename the object files, i.e. give them other names than in previous
1357 versions. Their names now include the name of the final product, as
1358 well as its type mnemonic (bin, lib, shlib).
1362 * Added new option for 'openssl list', '-objects', which will display the
1363 list of built in objects, i.e. OIDs with names.
1367 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1368 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1373 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1374 improves application performance by removing data copies and providing
1375 applications with zero-copy system calls such as sendfile and splice.
1379 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1380 option is set, openssl cleanses (zeroize) plaintext bytes from
1381 internal buffers after delivering them to the application. Note,
1382 the application is still responsible for cleansing other copies
1383 (e.g.: data received by SSL_read(3)).
1387 * `PKCS12_parse` now maintains the order of the parsed certificates
1388 when outputting them via `*ca` (rather than reversing it).
1392 * Deprecated pthread fork support methods. These were unused so no
1393 replacement is required.
1395 - OPENSSL_fork_prepare()
1396 - OPENSSL_fork_parent()
1397 - OPENSSL_fork_child()
1404 ### Changes between 1.1.1i and 1.1.1j [xx XXX xxxx]
1406 * Fixed SRP_Calc_client_key so that it uses constant time. The previous
1407 implementation called BN_mod_exp without setting BN_FLG_CONSTTIME. This
1408 could be exploited in a side channel attack to recover the password. Since
1409 the attack is local host only this is outside of the current OpenSSL
1410 threat model and therefore no CVE is assigned.
1412 Thanks to Mohammed Sabt and Daniel De Almeida Braga for reporting this
1417 ### Changes between 1.1.1h and 1.1.1i [8 Dec 2020]
1419 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1420 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1421 If an attacker can control both items being compared then this could lead
1422 to a possible denial of service attack. OpenSSL itself uses the
1423 GENERAL_NAME_cmp function for two purposes:
1424 1) Comparing CRL distribution point names between an available CRL and a
1425 CRL distribution point embedded in an X509 certificate
1426 2) When verifying that a timestamp response token signer matches the
1427 timestamp authority name (exposed via the API functions
1428 TS_RESP_verify_response and TS_RESP_verify_token)
1433 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1435 * Certificates with explicit curve parameters are now disallowed in
1436 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1440 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1441 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1442 conversely, silently ignore DTLS protocol version bounds when configuring
1443 TLS-based contexts. The commands can be repeated to set bounds of both
1444 types. The same applies with the corresponding "min_protocol" and
1445 "max_protocol" command-line switches, in case some application uses both TLS
1448 SSL_CTX instances that are created for a fixed protocol version (e.g.
1449 TLSv1_server_method()) also silently ignore version bounds. Previously
1450 attempts to apply bounds to these protocol versions would result in an
1451 error. Now only the "version-flexible" SSL_CTX instances are subject to
1452 limits in configuration files in command-line options.
1456 * Handshake now fails if Extended Master Secret extension is dropped
1461 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1463 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1465 * Fixed segmentation fault in SSL_check_chain()
1466 Server or client applications that call the SSL_check_chain() function
1467 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1468 dereference as a result of incorrect handling of the
1469 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1470 or unrecognised signature algorithm is received from the peer. This could
1471 be exploited by a malicious peer in a Denial of Service attack.
1476 * Added AES consttime code for no-asm configurations
1477 an optional constant time support for AES was added
1478 when building openssl for no-asm.
1479 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1480 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1481 At this time this feature is by default disabled.
1482 It will be enabled by default in 3.0.
1486 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1488 * Revert the change of EOF detection while reading in libssl to avoid
1489 regressions in applications depending on the current way of reporting
1490 the EOF. As the existing method is not fully accurate the change to
1491 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1492 branch and will be present in the 3.0 release.
1496 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1497 when primes for RSA keys are computed.
1498 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1499 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1500 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1501 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1502 This avoids possible fingerprinting of newly generated RSA modules.
1506 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1508 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1509 while reading in libssl then we would report an error back to the
1510 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1511 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1512 therefore give a hint as to what went wrong.
1516 * Check that ed25519 and ed448 are allowed by the security level. Previously
1517 signature algorithms not using an MD were not being checked that they were
1518 allowed by the security level.
1522 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1523 was not quite right. The behaviour was not consistent between resumption
1524 and normal handshakes, and also not quite consistent with historical
1525 behaviour. The behaviour in various scenarios has been clarified and
1526 it has been updated to make it match historical behaviour as closely as
1531 * *[VMS only]* The header files that the VMS compilers include automatically,
1532 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1533 that the C++ compiler doesn't understand. This is a shortcoming in the
1534 compiler, but can be worked around with `__cplusplus` guards.
1536 C++ applications that use OpenSSL libraries must be compiled using the
1537 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1538 functions. Otherwise, only functions with symbols of less than 31
1539 characters can be used, as the linker will not be able to successfully
1540 resolve symbols with longer names.
1544 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1545 The presence of this system service is determined at run-time.
1549 * Added newline escaping functionality to a filename when using openssl dgst.
1550 This output format is to replicate the output format found in the `*sum`
1551 checksum programs. This aims to preserve backward compatibility.
1553 *Matt Eaton, Richard Levitte, and Paul Dale*
1555 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1560 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1562 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1563 number generator (RNG). This was intended to include protection in the
1564 event of a fork() system call in order to ensure that the parent and child
1565 processes did not share the same RNG state. However this protection was not
1566 being used in the default case.
1568 A partial mitigation for this issue is that the output from a high
1569 precision timer is mixed into the RNG state so the likelihood of a parent
1570 and child process sharing state is significantly reduced.
1572 If an application already calls OPENSSL_init_crypto() explicitly using
1573 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1576 *Matthias St. Pierre*
1578 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1579 used even when parsing explicit parameters, when loading a encoded key
1580 or calling `EC_GROUP_new_from_ecpkparameters()`/
1581 `EC_GROUP_new_from_ecparameters()`.
1582 This prevents bypass of security hardening and performance gains,
1583 especially for curves with specialized EC_METHODs.
1584 By default, if a key encoded with explicit parameters is loaded and later
1585 encoded, the output is still encoded with explicit parameters, even if
1586 internally a "named" EC_GROUP is used for computation.
1590 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1591 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1592 NULL. After this change, only the cofactor parameter can be NULL. It also
1593 does some minimal sanity checks on the passed order.
1598 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1599 An attack is simple, if the first CMS_recipientInfo is valid but the
1600 second CMS_recipientInfo is chosen ciphertext. If the second
1601 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1602 encryption key will be replaced by garbage, and the message cannot be
1603 decoded, but if the RSA decryption fails, the correct encryption key is
1604 used and the recipient will not notice the attack.
1605 As a work around for this potential attack the length of the decrypted
1606 key must be equal to the cipher default key length, in case the
1607 certifiate is not given and all recipientInfo are tried out.
1608 The old behaviour can be re-enabled in the CMS code by setting the
1609 CMS_DEBUG_DECRYPT flag.
1614 * Early start up entropy quality from the DEVRANDOM seed source has been
1615 improved for older Linux systems. The RAND subsystem will wait for
1616 /dev/random to be producing output before seeding from /dev/urandom.
1617 The seeded state is stored for future library initialisations using
1618 a system global shared memory segment. The shared memory identifier
1619 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1620 the desired value. The default identifier is 114.
1624 * Correct the extended master secret constant on EBCDIC systems. Without this
1625 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1626 negotiate EMS will fail. Unfortunately this also means that TLS connections
1627 between EBCDIC systems with this fix, and EBCDIC systems without this
1628 fix will fail if they negotiate EMS.
1632 * Use Windows installation paths in the mingw builds
1634 Mingw isn't a POSIX environment per se, which means that Windows
1635 paths should be used for installation.
1640 * Changed DH_check to accept parameters with order q and 2q subgroups.
1641 With order 2q subgroups the bit 0 of the private key is not secret
1642 but DH_generate_key works around that by clearing bit 0 of the
1643 private key for those. This avoids leaking bit 0 of the private key.
1647 * Significantly reduce secure memory usage by the randomness pools.
1651 * Revert the DEVRANDOM_WAIT feature for Linux systems
1653 The DEVRANDOM_WAIT feature added a select() call to wait for the
1654 /dev/random device to become readable before reading from the
1655 /dev/urandom device.
1657 It turned out that this change had negative side effects on
1658 performance which were not acceptable. After some discussion it
1659 was decided to revert this feature and leave it up to the OS
1660 resp. the platform maintainer to ensure a proper initialization
1661 during early boot time.
1663 *Matthias St. Pierre*
1665 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1667 * Add build tests for C++. These are generated files that only do one
1668 thing, to include one public OpenSSL head file each. This tests that
1669 the public header files can be usefully included in a C++ application.
1671 This test isn't enabled by default. It can be enabled with the option
1672 'enable-buildtest-c++'.
1676 * Enable SHA3 pre-hashing for ECDSA and DSA.
1680 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1681 This changes the size when using the `genpkey` command when no size is given.
1682 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1683 generation commands to use 2048 bits by default.
1687 * Reorganize the manual pages to consistently have RETURN VALUES,
1688 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1689 util/fix-doc-nits accordingly.
1691 *Paul Yang, Joshua Lock*
1693 * Add the missing accessor EVP_PKEY_get0_engine()
1697 * Have commands like `s_client` and `s_server` output the signature scheme
1698 along with other cipher suite parameters when debugging.
1702 * Make OPENSSL_config() error agnostic again.
1706 * Do the error handling in RSA decryption constant time.
1710 * Prevent over long nonces in ChaCha20-Poly1305.
1712 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1713 for every encryption operation. RFC 7539 specifies that the nonce value
1714 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1715 and front pads the nonce with 0 bytes if it is less than 12
1716 bytes. However it also incorrectly allows a nonce to be set of up to 16
1717 bytes. In this case only the last 12 bytes are significant and any
1718 additional leading bytes are ignored.
1720 It is a requirement of using this cipher that nonce values are
1721 unique. Messages encrypted using a reused nonce value are susceptible to
1722 serious confidentiality and integrity attacks. If an application changes
1723 the default nonce length to be longer than 12 bytes and then makes a
1724 change to the leading bytes of the nonce expecting the new value to be a
1725 new unique nonce then such an application could inadvertently encrypt
1726 messages with a reused nonce.
1728 Additionally the ignored bytes in a long nonce are not covered by the
1729 integrity guarantee of this cipher. Any application that relies on the
1730 integrity of these ignored leading bytes of a long nonce may be further
1731 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1732 is safe because no such use sets such a long nonce value. However user
1733 applications that use this cipher directly and set a non-default nonce
1734 length to be longer than 12 bytes may be vulnerable.
1736 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1742 * Add DEVRANDOM_WAIT feature for Linux systems
1744 On older Linux systems where the getrandom() system call is not available,
1745 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1746 Contrary to getrandom(), the /dev/urandom device will not block during
1747 early boot when the kernel CSPRNG has not been seeded yet.
1749 To mitigate this known weakness, use select() to wait for /dev/random to
1750 become readable before reading from /dev/urandom.
1752 * Ensure that SM2 only uses SM3 as digest algorithm
1756 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1758 * Change the info callback signals for the start and end of a post-handshake
1759 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1760 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1761 confused by this and assume that a TLSv1.2 renegotiation has started. This
1762 can break KeyUpdate handling. Instead we no longer signal the start and end
1763 of a post handshake message exchange (although the messages themselves are
1764 still signalled). This could break some applications that were expecting
1765 the old signals. However without this KeyUpdate is not usable for many
1770 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1772 * Timing vulnerability in DSA signature generation
1774 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1775 timing side channel attack. An attacker could use variations in the signing
1776 algorithm to recover the private key.
1778 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1783 * Timing vulnerability in ECDSA signature generation
1785 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1786 timing side channel attack. An attacker could use variations in the signing
1787 algorithm to recover the private key.
1789 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1794 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1795 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1796 of two gigabytes and the error handling improved.
1798 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1799 categorized as a normal bug, not a security issue, because the DRBG reseeds
1800 automatically and is fully functional even without additional randomness
1801 provided by the application.
1803 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1805 * Add a new ClientHello callback. Provides a callback interface that gives
1806 the application the ability to adjust the nascent SSL object at the
1807 earliest stage of ClientHello processing, immediately after extensions have
1808 been collected but before they have been processed. In particular, this
1809 callback can adjust the supported TLS versions in response to the contents
1814 * Add SM2 base algorithm support.
1818 * s390x assembly pack: add (improved) hardware-support for the following
1819 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1820 aes-cfb/cfb8, aes-ecb.
1824 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1825 parameter is no longer accepted, as it leads to a corrupt table. NULL
1826 pem_str is reserved for alias entries only.
1830 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1831 step for prime curves. The new implementation is based on formulae from
1832 differential addition-and-doubling in homogeneous projective coordinates
1833 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1834 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1835 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1836 to work in projective coordinates.
1838 *Billy Bob Brumley, Nicola Tuveri*
1840 * Change generating and checking of primes so that the error rate of not
1841 being prime depends on the intended use based on the size of the input.
1842 For larger primes this will result in more rounds of Miller-Rabin.
1843 The maximal error rate for primes with more than 1080 bits is lowered
1846 *Kurt Roeckx, Annie Yousar*
1848 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1852 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1853 moving between systems, and to avoid confusion when a Windows build is
1854 done with mingw vs with MSVC. For POSIX installs, there's still a
1855 symlink or copy named 'tsget' to avoid that confusion as well.
1859 * Revert blinding in ECDSA sign and instead make problematic addition
1860 length-invariant. Switch even to fixed-length Montgomery multiplication.
1864 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1865 step for binary curves. The new implementation is based on formulae from
1866 differential addition-and-doubling in mixed Lopez-Dahab projective
1867 coordinates, modified to independently blind the operands.
1869 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1871 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1872 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1873 EC_METHODs to implement their own specialized "ladder step", to take
1874 advantage of more favorable coordinate systems or more efficient
1875 differential addition-and-doubling algorithms.
1877 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1879 * Modified the random device based seed sources to keep the relevant
1880 file descriptors open rather than reopening them on each access.
1881 This allows such sources to operate in a chroot() jail without
1882 the associated device nodes being available. This behaviour can be
1883 controlled using RAND_keep_random_devices_open().
1887 * Numerous side-channel attack mitigations have been applied. This may have
1888 performance impacts for some algorithms for the benefit of improved
1889 security. Specific changes are noted in this change log by their respective
1894 * AIX shared library support overhaul. Switch to AIX "natural" way of
1895 handling shared libraries, which means collecting shared objects of
1896 different versions and bitnesses in one common archive. This allows to
1897 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1898 doesn't affect the way 3rd party applications are linked, only how
1899 multi-version installation is managed.
1903 * Make ec_group_do_inverse_ord() more robust and available to other
1904 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1905 mitigations are applied to the fallback BN_mod_inverse().
1906 When using this function rather than BN_mod_inverse() directly, new
1907 EC cryptosystem implementations are then safer-by-default.
1911 * Add coordinate blinding for EC_POINT and implement projective
1912 coordinate blinding for generic prime curves as a countermeasure to
1913 chosen point SCA attacks.
1915 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1917 * Add blinding to ECDSA and DSA signatures to protect against side channel
1918 attacks discovered by Keegan Ryan (NCC Group).
1922 * Enforce checking in the `pkeyutl` command to ensure that the input
1923 length does not exceed the maximum supported digest length when performing
1924 a sign, verify or verifyrecover operation.
1928 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1929 I/O in combination with something like select() or poll() will hang. This
1930 can be turned off again using SSL_CTX_clear_mode().
1931 Many applications do not properly handle non-application data records, and
1932 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1933 around the problems in those applications, but can also break some.
1934 It's recommended to read the manpages about SSL_read(), SSL_write(),
1935 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1936 SSL_CTX_set_read_ahead() again.
1940 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1941 now allow empty (zero character) pass phrases.
1945 * Apply blinding to binary field modular inversion and remove patent
1946 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1950 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1951 binary and prime elliptic curves.
1955 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1956 constant time fixed point multiplication.
1960 * Revise elliptic curve scalar multiplication with timing attack
1961 defenses: ec_wNAF_mul redirects to a constant time implementation
1962 when computing fixed point and variable point multiplication (which
1963 in OpenSSL are mostly used with secret scalars in keygen, sign,
1964 ECDH derive operations).
1965 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1968 * Updated CONTRIBUTING
1972 * Updated DRBG / RAND to request nonce and additional low entropy
1973 randomness from the system.
1975 *Matthias St. Pierre*
1977 * Updated 'openssl rehash' to use OpenSSL consistent default.
1981 * Moved the load of the ssl_conf module to libcrypto, which helps
1982 loading engines that libssl uses before libssl is initialised.
1986 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1990 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1992 *Ingo Schwarze, Rich Salz*
1994 * Added output of accepting IP address and port for 'openssl s_server'
1998 * Added a new API for TLSv1.3 ciphersuites:
1999 SSL_CTX_set_ciphersuites()
2000 SSL_set_ciphersuites()
2004 * Memory allocation failures consistently add an error to the error
2009 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
2010 in libcrypto when run as setuid/setgid.
2014 * Load any config file by default when libssl is used.
2018 * Added new public header file <openssl/rand_drbg.h> and documentation
2019 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
2021 *Matthias St. Pierre*
2023 * QNX support removed (cannot find contributors to get their approval
2024 for the license change).
2028 * TLSv1.3 replay protection for early data has been implemented. See the
2029 SSL_read_early_data() man page for further details.
2033 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
2034 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
2035 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
2036 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
2037 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
2038 configuration has been separated out. See the ciphers man page or the
2039 SSL_CTX_set_ciphersuites() man page for more information.
2043 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
2044 in responder mode now supports the new "-multi" option, which
2045 spawns the specified number of child processes to handle OCSP
2046 requests. The "-timeout" option now also limits the OCSP
2047 responder's patience to wait to receive the full client request
2048 on a newly accepted connection. Child processes are respawned
2049 as needed, and the CA index file is automatically reloaded
2050 when changed. This makes it possible to run the "ocsp" responder
2051 as a long-running service, making the OpenSSL CA somewhat more
2052 feature-complete. In this mode, most diagnostic messages logged
2053 after entering the event loop are logged via syslog(3) rather than
2058 * Added support for X448 and Ed448. Heavily based on original work by
2063 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2064 objects loaded. This adds the functions OSSL_STORE_expect() and
2065 OSSL_STORE_find() as well as needed tools to construct searches and
2066 get the search data out of them.
2070 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2071 version of OpenSSL should review their configuration settings to ensure
2072 that they are still appropriate for TLSv1.3. For further information see:
2073 <https://wiki.openssl.org/index.php/TLS1.3>
2077 * Grand redesign of the OpenSSL random generator
2079 The default RAND method now utilizes an AES-CTR DRBG according to
2080 NIST standard SP 800-90Ar1. The new random generator is essentially
2081 a port of the default random generator from the OpenSSL FIPS 2.0
2082 object module. It is a hybrid deterministic random bit generator
2083 using an AES-CTR bit stream and which seeds and reseeds itself
2084 automatically using trusted system entropy sources.
2086 Some of its new features are:
2087 - Support for multiple DRBG instances with seed chaining.
2088 - The default RAND method makes use of a DRBG.
2089 - There is a public and private DRBG instance.
2090 - The DRBG instances are fork-safe.
2091 - Keep all global DRBG instances on the secure heap if it is enabled.
2092 - The public and private DRBG instance are per thread for lock free
2095 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2097 * Changed Configure so it only says what it does and doesn't dump
2098 so much data. Instead, ./configdata.pm should be used as a script
2099 to display all sorts of configuration data.
2103 * Added processing of "make variables" to Configure.
2107 * Added SHA512/224 and SHA512/256 algorithm support.
2111 * The last traces of Netware support, first removed in 1.1.0, have
2116 * Get rid of Makefile.shared, and in the process, make the processing
2117 of certain files (rc.obj, or the .def/.map/.opt files produced from
2118 the ordinal files) more visible and hopefully easier to trace and
2119 debug (or make silent).
2123 * Make it possible to have environment variable assignments as
2124 arguments to config / Configure.
2128 * Add multi-prime RSA (RFC 8017) support.
2132 * Add SM3 implemented according to GB/T 32905-2016
2133 *Jack Lloyd <jack.lloyd@ribose.com>,*
2134 *Ronald Tse <ronald.tse@ribose.com>,*
2135 *Erick Borsboom <erick.borsboom@ribose.com>*
2137 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2138 as documented in RFC6066.
2139 Based on a patch from Tomasz Moń
2141 *Filipe Raimundo da Silva*
2143 * Add SM4 implemented according to GB/T 32907-2016.
2144 *Jack Lloyd <jack.lloyd@ribose.com>,*
2145 *Ronald Tse <ronald.tse@ribose.com>,*
2146 *Erick Borsboom <erick.borsboom@ribose.com>*
2148 * Reimplement -newreq-nodes and ERR_error_string_n; the
2149 original author does not agree with the license change.
2153 * Add ARIA AEAD TLS support.
2157 * Some macro definitions to support VS6 have been removed. Visual
2158 Studio 6 has not worked since 1.1.0
2162 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2163 without clearing the errors.
2167 * Add "atfork" functions. If building on a system that without
2168 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2169 requirements. The RAND facility now uses/requires this.
2177 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2178 not possible to disable entirely. However, it's still possible to
2179 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2182 To disable, configure with 'no-ui-console'. 'no-ui' is still
2183 possible to use as an alias. Check at compile time with the
2184 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2185 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2189 * Add a STORE module, which implements a uniform and URI based reader of
2190 stores that can contain keys, certificates, CRLs and numerous other
2191 objects. The main API is loosely based on a few stdio functions,
2192 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2193 OSSL_STORE_error and OSSL_STORE_close.
2194 The implementation uses backends called "loaders" to implement arbitrary
2195 URI schemes. There is one built in "loader" for the 'file' scheme.
2199 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2200 then adjusted to work on FreeBSD 8.4 as well.
2201 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2202 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2206 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2207 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2208 error code calls like this:
2210 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2212 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2213 that can be encoded in C. For the foreseeable future, this will only
2216 *Richard Levitte and Tim Hudson*
2218 * Removed BSD cryptodev engine.
2222 * Add a build target 'build_all_generated', to build all generated files
2223 and only that. This can be used to prepare everything that requires
2224 things like perl for a system that lacks perl and then move everything
2225 to that system and do the rest of the build there.
2229 * In the UI interface, make it possible to duplicate the user data. This
2230 can be used by engines that need to retain the data for a longer time
2231 than just the call where this user data is passed.
2235 * Ignore the '-named_curve auto' value for compatibility of applications
2238 *Tomas Mraz <tmraz@fedoraproject.org>*
2240 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2241 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2242 alerts across multiple records (some of which could be empty). In practice
2243 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2244 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2245 support this at all. Supporting it adds significant complexity to the
2246 record layer, and its removal is unlikely to cause interoperability
2251 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2252 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2253 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2258 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2259 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2261 *Richard Levitte, Andy Polyakov*
2263 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2268 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2269 platform rather than 'mingw'.
2273 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2274 success if they are asked to add an object which already exists
2275 in the store. This change cascades to other functions which load
2276 certificates and CRLs.
2280 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2281 facilitate stack unwinding even from assembly subroutines.
2285 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2286 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2290 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2291 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2292 which is the minimum version we support.
2296 * Certificate time validation (X509_cmp_time) enforces stricter
2297 compliance with RFC 5280. Fractional seconds and timezone offsets
2298 are no longer allowed.
2302 * Add support for ARIA
2306 * s_client will now send the Server Name Indication (SNI) extension by
2307 default unless the new "-noservername" option is used. The server name is
2308 based on the host provided to the "-connect" option unless overridden by
2309 using "-servername".
2313 * Add support for SipHash
2317 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2318 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2319 prevent issues where no progress is being made and the peer continually
2320 sends unrecognised record types, using up resources processing them.
2324 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2325 using the algorithm defined in
2326 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2330 * Heartbeat support has been removed; the ABI is changed for now.
2332 *Richard Levitte, Rich Salz*
2334 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2338 * The RSA "null" method, which was partially supported to avoid patent
2339 issues, has been replaced to always returns NULL.
2346 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2348 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2349 used even when parsing explicit parameters, when loading a encoded key
2350 or calling `EC_GROUP_new_from_ecpkparameters()`/
2351 `EC_GROUP_new_from_ecparameters()`.
2352 This prevents bypass of security hardening and performance gains,
2353 especially for curves with specialized EC_METHODs.
2354 By default, if a key encoded with explicit parameters is loaded and later
2355 encoded, the output is still encoded with explicit parameters, even if
2356 internally a "named" EC_GROUP is used for computation.
2360 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2361 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2362 NULL. After this change, only the cofactor parameter can be NULL. It also
2363 does some minimal sanity checks on the passed order.
2368 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2369 An attack is simple, if the first CMS_recipientInfo is valid but the
2370 second CMS_recipientInfo is chosen ciphertext. If the second
2371 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2372 encryption key will be replaced by garbage, and the message cannot be
2373 decoded, but if the RSA decryption fails, the correct encryption key is
2374 used and the recipient will not notice the attack.
2375 As a work around for this potential attack the length of the decrypted
2376 key must be equal to the cipher default key length, in case the
2377 certifiate is not given and all recipientInfo are tried out.
2378 The old behaviour can be re-enabled in the CMS code by setting the
2379 CMS_DEBUG_DECRYPT flag.
2384 * Use Windows installation paths in the mingw builds
2386 Mingw isn't a POSIX environment per se, which means that Windows
2387 paths should be used for installation.
2392 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2394 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2395 This changes the size when using the `genpkey` command when no size is given.
2396 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2397 generation commands to use 2048 bits by default.
2401 * Prevent over long nonces in ChaCha20-Poly1305.
2403 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2404 for every encryption operation. RFC 7539 specifies that the nonce value
2405 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2406 and front pads the nonce with 0 bytes if it is less than 12
2407 bytes. However it also incorrectly allows a nonce to be set of up to 16
2408 bytes. In this case only the last 12 bytes are significant and any
2409 additional leading bytes are ignored.
2411 It is a requirement of using this cipher that nonce values are
2412 unique. Messages encrypted using a reused nonce value are susceptible to
2413 serious confidentiality and integrity attacks. If an application changes
2414 the default nonce length to be longer than 12 bytes and then makes a
2415 change to the leading bytes of the nonce expecting the new value to be a
2416 new unique nonce then such an application could inadvertently encrypt
2417 messages with a reused nonce.
2419 Additionally the ignored bytes in a long nonce are not covered by the
2420 integrity guarantee of this cipher. Any application that relies on the
2421 integrity of these ignored leading bytes of a long nonce may be further
2422 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2423 is safe because no such use sets such a long nonce value. However user
2424 applications that use this cipher directly and set a non-default nonce
2425 length to be longer than 12 bytes may be vulnerable.
2427 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2433 * Added SCA hardening for modular field inversion in EC_GROUP through
2434 a new dedicated field_inv() pointer in EC_METHOD.
2435 This also addresses a leakage affecting conversions from projective
2436 to affine coordinates.
2438 *Billy Bob Brumley, Nicola Tuveri*
2440 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2441 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2445 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2449 * Remove the 'dist' target and add a tarball building script. The
2450 'dist' target has fallen out of use, and it shouldn't be
2451 necessary to configure just to create a source distribution.
2455 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2457 * Timing vulnerability in DSA signature generation
2459 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2460 timing side channel attack. An attacker could use variations in the signing
2461 algorithm to recover the private key.
2463 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2468 * Timing vulnerability in ECDSA signature generation
2470 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2471 timing side channel attack. An attacker could use variations in the signing
2472 algorithm to recover the private key.
2474 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2479 * Add coordinate blinding for EC_POINT and implement projective
2480 coordinate blinding for generic prime curves as a countermeasure to
2481 chosen point SCA attacks.
2483 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2485 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2487 * Client DoS due to large DH parameter
2489 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2490 malicious server can send a very large prime value to the client. This will
2491 cause the client to spend an unreasonably long period of time generating a
2492 key for this prime resulting in a hang until the client has finished. This
2493 could be exploited in a Denial Of Service attack.
2495 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2500 * Cache timing vulnerability in RSA Key Generation
2502 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2503 a cache timing side channel attack. An attacker with sufficient access to
2504 mount cache timing attacks during the RSA key generation process could
2505 recover the private key.
2507 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2508 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2513 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2514 parameter is no longer accepted, as it leads to a corrupt table. NULL
2515 pem_str is reserved for alias entries only.
2519 * Revert blinding in ECDSA sign and instead make problematic addition
2520 length-invariant. Switch even to fixed-length Montgomery multiplication.
2524 * Change generating and checking of primes so that the error rate of not
2525 being prime depends on the intended use based on the size of the input.
2526 For larger primes this will result in more rounds of Miller-Rabin.
2527 The maximal error rate for primes with more than 1080 bits is lowered
2530 *Kurt Roeckx, Annie Yousar*
2532 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2536 * Add blinding to ECDSA and DSA signatures to protect against side channel
2537 attacks discovered by Keegan Ryan (NCC Group).
2541 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2542 now allow empty (zero character) pass phrases.
2546 * Certificate time validation (X509_cmp_time) enforces stricter
2547 compliance with RFC 5280. Fractional seconds and timezone offsets
2548 are no longer allowed.
2552 * Fixed a text canonicalisation bug in CMS
2554 Where a CMS detached signature is used with text content the text goes
2555 through a canonicalisation process first prior to signing or verifying a
2556 signature. This process strips trailing space at the end of lines, converts
2557 line terminators to CRLF and removes additional trailing line terminators
2558 at the end of a file. A bug in the canonicalisation process meant that
2559 some characters, such as form-feed, were incorrectly treated as whitespace
2560 and removed. This is contrary to the specification (RFC5485). This fix
2561 could mean that detached text data signed with an earlier version of
2562 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2563 signed with a fixed OpenSSL may fail to verify with an earlier version of
2564 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2565 and use the "-binary" flag (for the "cms" command line application) or set
2566 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2570 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2572 * Constructed ASN.1 types with a recursive definition could exceed the stack
2574 Constructed ASN.1 types with a recursive definition (such as can be found
2575 in PKCS7) could eventually exceed the stack given malicious input with
2576 excessive recursion. This could result in a Denial Of Service attack. There
2577 are no such structures used within SSL/TLS that come from untrusted sources
2578 so this is considered safe.
2580 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2586 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2588 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2589 effectively reduced to only comparing the least significant bit of each
2590 byte. This allows an attacker to forge messages that would be considered as
2591 authenticated in an amount of tries lower than that guaranteed by the
2592 security claims of the scheme. The module can only be compiled by the
2593 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2595 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2601 * Add a build target 'build_all_generated', to build all generated files
2602 and only that. This can be used to prepare everything that requires
2603 things like perl for a system that lacks perl and then move everything
2604 to that system and do the rest of the build there.
2608 * Backport SSL_OP_NO_RENGOTIATION
2610 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2611 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2612 changes this is no longer possible in 1.1.0. Therefore the new
2613 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2614 1.1.0 to provide equivalent functionality.
2616 Note that if an application built against 1.1.0h headers (or above) is run
2617 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2618 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2622 * Removed the OS390-Unix config target. It relied on a script that doesn't
2627 * rsaz_1024_mul_avx2 overflow bug on x86_64
2629 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2630 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2631 Analysis suggests that attacks against RSA and DSA as a result of this
2632 defect would be very difficult to perform and are not believed likely.
2633 Attacks against DH1024 are considered just feasible, because most of the
2634 work necessary to deduce information about a private key may be performed
2635 offline. The amount of resources required for such an attack would be
2636 significant. However, for an attack on TLS to be meaningful, the server
2637 would have to share the DH1024 private key among multiple clients, which is
2638 no longer an option since CVE-2016-0701.
2640 This only affects processors that support the AVX2 but not ADX extensions
2641 like Intel Haswell (4th generation).
2643 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2644 was originally found via the OSS-Fuzz project.
2649 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2651 * bn_sqrx8x_internal carry bug on x86_64
2653 There is a carry propagating bug in the x86_64 Montgomery squaring
2654 procedure. No EC algorithms are affected. Analysis suggests that attacks
2655 against RSA and DSA as a result of this defect would be very difficult to
2656 perform and are not believed likely. Attacks against DH are considered just
2657 feasible (although very difficult) because most of the work necessary to
2658 deduce information about a private key may be performed offline. The amount
2659 of resources required for such an attack would be very significant and
2660 likely only accessible to a limited number of attackers. An attacker would
2661 additionally need online access to an unpatched system using the target
2662 private key in a scenario with persistent DH parameters and a private
2663 key that is shared between multiple clients.
2665 This only affects processors that support the BMI1, BMI2 and ADX extensions
2666 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2668 This issue was reported to OpenSSL by the OSS-Fuzz project.
2673 * Malformed X.509 IPAddressFamily could cause OOB read
2675 If an X.509 certificate has a malformed IPAddressFamily extension,
2676 OpenSSL could do a one-byte buffer overread. The most likely result
2677 would be an erroneous display of the certificate in text format.
2679 This issue was reported to OpenSSL by the OSS-Fuzz project.
2684 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2686 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2687 platform rather than 'mingw'.
2691 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2692 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2693 which is the minimum version we support.
2697 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2699 * Encrypt-Then-Mac renegotiation crash
2701 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2702 negotiated where it was not in the original handshake (or vice-versa) then
2703 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2704 and servers are affected.
2706 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2711 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2713 * Truncated packet could crash via OOB read
2715 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2716 cipher is being used, then a truncated packet can cause that host to
2717 perform an out-of-bounds read, usually resulting in a crash.
2719 This issue was reported to OpenSSL by Robert Święcki of Google.
2724 * Bad (EC)DHE parameters cause a client crash
2726 If a malicious server supplies bad parameters for a DHE or ECDHE key
2727 exchange then this can result in the client attempting to dereference a
2728 NULL pointer leading to a client crash. This could be exploited in a Denial
2731 This issue was reported to OpenSSL by Guido Vranken.
2736 * BN_mod_exp may produce incorrect results on x86_64
2738 There is a carry propagating bug in the x86_64 Montgomery squaring
2739 procedure. No EC algorithms are affected. Analysis suggests that attacks
2740 against RSA and DSA as a result of this defect would be very difficult to
2741 perform and are not believed likely. Attacks against DH are considered just
2742 feasible (although very difficult) because most of the work necessary to
2743 deduce information about a private key may be performed offline. The amount
2744 of resources required for such an attack would be very significant and
2745 likely only accessible to a limited number of attackers. An attacker would
2746 additionally need online access to an unpatched system using the target
2747 private key in a scenario with persistent DH parameters and a private
2748 key that is shared between multiple clients. For example this can occur by
2749 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2750 similar to CVE-2015-3193 but must be treated as a separate problem.
2752 This issue was reported to OpenSSL by the OSS-Fuzz project.
2757 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2759 * ChaCha20/Poly1305 heap-buffer-overflow
2761 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2762 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2763 crash. This issue is not considered to be exploitable beyond a DoS.
2765 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2770 * CMS Null dereference
2772 Applications parsing invalid CMS structures can crash with a NULL pointer
2773 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2774 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2775 structure callback if an attempt is made to free certain invalid encodings.
2776 Only CHOICE structures using a callback which do not handle NULL value are
2779 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2784 * Montgomery multiplication may produce incorrect results
2786 There is a carry propagating bug in the Broadwell-specific Montgomery
2787 multiplication procedure that handles input lengths divisible by, but
2788 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2789 and DH private keys are impossible. This is because the subroutine in
2790 question is not used in operations with the private key itself and an input
2791 of the attacker's direct choice. Otherwise the bug can manifest itself as
2792 transient authentication and key negotiation failures or reproducible
2793 erroneous outcome of public-key operations with specially crafted input.
2794 Among EC algorithms only Brainpool P-512 curves are affected and one
2795 presumably can attack ECDH key negotiation. Impact was not analyzed in
2796 detail, because pre-requisites for attack are considered unlikely. Namely
2797 multiple clients have to choose the curve in question and the server has to
2798 share the private key among them, neither of which is default behaviour.
2799 Even then only clients that chose the curve will be affected.
2801 This issue was publicly reported as transient failures and was not
2802 initially recognized as a security issue. Thanks to Richard Morgan for
2803 providing reproducible case.
2808 * Removed automatic addition of RPATH in shared libraries and executables,
2809 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2813 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2815 * Fix Use After Free for large message sizes
2817 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2818 message larger than approx 16k is received then the underlying buffer to
2819 store the incoming message is reallocated and moved. Unfortunately a
2820 dangling pointer to the old location is left which results in an attempt to
2821 write to the previously freed location. This is likely to result in a
2822 crash, however it could potentially lead to execution of arbitrary code.
2824 This issue only affects OpenSSL 1.1.0a.
2826 This issue was reported to OpenSSL by Robert Święcki.
2831 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2833 * OCSP Status Request extension unbounded memory growth
2835 A malicious client can send an excessively large OCSP Status Request
2836 extension. If that client continually requests renegotiation, sending a
2837 large OCSP Status Request extension each time, then there will be unbounded
2838 memory growth on the server. This will eventually lead to a Denial Of
2839 Service attack through memory exhaustion. Servers with a default
2840 configuration are vulnerable even if they do not support OCSP. Builds using
2841 the "no-ocsp" build time option are not affected.
2843 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2848 * SSL_peek() hang on empty record
2850 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2851 sends an empty record. This could be exploited by a malicious peer in a
2852 Denial Of Service attack.
2854 This issue was reported to OpenSSL by Alex Gaynor.
2859 * Excessive allocation of memory in tls_get_message_header() and
2860 dtls1_preprocess_fragment()
2862 A (D)TLS message includes 3 bytes for its length in the header for the
2863 message. This would allow for messages up to 16Mb in length. Messages of
2864 this length are excessive and OpenSSL includes a check to ensure that a
2865 peer is sending reasonably sized messages in order to avoid too much memory
2866 being consumed to service a connection. A flaw in the logic of version
2867 1.1.0 means that memory for the message is allocated too early, prior to
2868 the excessive message length check. Due to way memory is allocated in
2869 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2870 to service a connection. This could lead to a Denial of Service through
2871 memory exhaustion. However, the excessive message length check still takes
2872 place, and this would cause the connection to immediately fail. Assuming
2873 that the application calls SSL_free() on the failed connection in a timely
2874 manner then the 21Mb of allocated memory will then be immediately freed
2875 again. Therefore the excessive memory allocation will be transitory in
2876 nature. This then means that there is only a security impact if:
2878 1) The application does not call SSL_free() in a timely manner in the event
2879 that the connection fails
2881 2) The application is working in a constrained environment where there is
2882 very little free memory
2884 3) The attacker initiates multiple connection attempts such that there are
2885 multiple connections in a state where memory has been allocated for the
2886 connection; SSL_free() has not yet been called; and there is insufficient
2887 memory to service the multiple requests.
2889 Except in the instance of (1) above any Denial Of Service is likely to be
2890 transitory because as soon as the connection fails the memory is
2891 subsequently freed again in the SSL_free() call. However there is an
2892 increased risk during this period of application crashes due to the lack of
2893 memory - which would then mean a more serious Denial of Service.
2895 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2896 (CVE-2016-6307 and CVE-2016-6308)
2900 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2901 had to be removed. Primary reason is that vendor assembler can't
2902 assemble our modules with -KPIC flag. As result it, assembly
2903 support, was not even available as option. But its lack means
2904 lack of side-channel resistant code, which is incompatible with
2905 security by todays standards. Fortunately gcc is readily available
2906 prepackaged option, which we firmly point at...
2910 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2912 * Windows command-line tool supports UTF-8 opt-in option for arguments
2913 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2914 (to any value) allows Windows user to access PKCS#12 file generated
2915 with Windows CryptoAPI and protected with non-ASCII password, as well
2916 as files generated under UTF-8 locale on Linux also protected with
2921 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2922 have been disabled by default and removed from DEFAULT, just like RC4.
2923 See the RC4 item below to re-enable both.
2927 * The method for finding the storage location for the Windows RAND seed file
2928 has changed. First we check %RANDFILE%. If that is not set then we check
2929 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2930 all else fails we fall back to C:\.
2934 * The EVP_EncryptUpdate() function has had its return type changed from void
2935 to int. A return of 0 indicates and error while a return of 1 indicates
2940 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2941 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2942 off the constant time implementation for RSA, DSA and DH have been made
2943 no-ops and deprecated.
2947 * Windows RAND implementation was simplified to only get entropy by
2948 calling CryptGenRandom(). Various other RAND-related tickets
2951 *Joseph Wylie Yandle, Rich Salz*
2953 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2954 and `OPENSSL_LH_`, respectively. The old names are available
2955 with API compatibility. They new names are now completely documented.
2959 * Unify TYPE_up_ref(obj) methods signature.
2960 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2961 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2962 int (instead of void) like all others TYPE_up_ref() methods.
2963 So now these methods also check the return value of CRYPTO_atomic_add(),
2964 and the validity of object reference counter.
2966 *fdasilvayy@gmail.com*
2968 * With Windows Visual Studio builds, the .pdb files are installed
2969 alongside the installed libraries and executables. For a static
2970 library installation, ossl_static.pdb is the associate compiler
2971 generated .pdb file to be used when linking programs.
2975 * Remove openssl.spec. Packaging files belong with the packagers.
2979 * Automatic Darwin/OSX configuration has had a refresh, it will now
2980 recognise x86_64 architectures automatically. You can still decide
2981 to build for a different bitness with the environment variable
2982 KERNEL_BITS (can be 32 or 64), for example:
2984 KERNEL_BITS=32 ./config
2988 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2989 256 bit AES and HMAC with SHA256.
2993 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2997 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
3001 * To enable users to have their own config files and build file templates,
3002 Configure looks in the directory indicated by the environment variable
3003 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
3004 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
3005 name and is used as is.
3009 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
3010 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
3011 X509_CERT_FILE_CTX was removed.
3015 * "shared" builds are now the default. To create only static libraries use
3016 the "no-shared" Configure option.
3020 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
3021 All of these option have not worked for some while and are fundamental
3026 * Make various cleanup routines no-ops and mark them as deprecated. Most
3027 global cleanup functions are no longer required because they are handled
3028 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
3029 Explicitly de-initing can cause problems (e.g. where a library that uses
3030 OpenSSL de-inits, but an application is still using it). The affected
3031 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
3032 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
3033 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
3034 COMP_zlib_cleanup().
3038 * --strict-warnings no longer enables runtime debugging options
3039 such as REF_DEBUG. Instead, debug options are automatically
3040 enabled with '--debug' builds.
3042 *Andy Polyakov, Emilia Käsper*
3044 * Made DH and DH_METHOD opaque. The structures for managing DH objects
3045 have been moved out of the public header files. New functions for managing
3046 these have been added.
3050 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3051 objects have been moved out of the public header files. New
3052 functions for managing these have been added.
3056 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3057 have been moved out of the public header files. New functions for managing
3058 these have been added.
3062 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3063 moved out of the public header files. New functions for managing these
3068 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3072 * Removed the mk1mf build scripts.
3076 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3077 it is always safe to #include a header now.
3081 * Removed the aged BC-32 config and all its supporting scripts
3085 * Removed support for Ultrix, Netware, and OS/2.
3089 * Add support for HKDF.
3091 *Alessandro Ghedini*
3093 * Add support for blake2b and blake2s
3097 * Added support for "pipelining". Ciphers that have the
3098 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3099 encryptions/decryptions simultaneously. There are currently no built-in
3100 ciphers with this property but the expectation is that engines will be able
3101 to offer it to significantly improve throughput. Support has been extended
3102 into libssl so that multiple records for a single connection can be
3103 processed in one go (for >=TLS 1.1).
3107 * Added the AFALG engine. This is an async capable engine which is able to
3108 offload work to the Linux kernel. In this initial version it only supports
3109 AES128-CBC. The kernel must be version 4.1.0 or greater.
3113 * OpenSSL now uses a new threading API. It is no longer necessary to
3114 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3115 are two supported threading models: pthreads and windows threads. It is
3116 also possible to configure OpenSSL at compile time for "no-threads". The
3117 old threading API should no longer be used. The functions have been
3118 replaced with "no-op" compatibility macros.
3120 *Alessandro Ghedini, Matt Caswell*
3122 * Modify behavior of ALPN to invoke callback after SNI/servername
3123 callback, such that updates to the SSL_CTX affect ALPN.
3127 * Add SSL_CIPHER queries for authentication and key-exchange.
3131 * Changes to the DEFAULT cipherlist:
3132 - Prefer (EC)DHE handshakes over plain RSA.
3133 - Prefer AEAD ciphers over legacy ciphers.
3134 - Prefer ECDSA over RSA when both certificates are available.
3135 - Prefer TLSv1.2 ciphers/PRF.
3136 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3141 * Change the ECC default curve list to be this, in order: x25519,
3142 secp256r1, secp521r1, secp384r1.
3146 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3147 disabled by default. They can be re-enabled using the
3148 enable-weak-ssl-ciphers option to Configure.
3152 * If the server has ALPN configured, but supports no protocols that the
3153 client advertises, send a fatal "no_application_protocol" alert.
3154 This behaviour is SHALL in RFC 7301, though it isn't universally
3155 implemented by other servers.
3159 * Add X25519 support.
3160 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3161 for public and private key encoding using the format documented in
3162 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3163 key generation and key derivation.
3165 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3170 * Deprecate SRP_VBASE_get_by_user.
3171 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3172 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3173 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3174 seed, even if the seed is configured.
3176 Users should use SRP_VBASE_get1_by_user instead. Note that in
3177 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3178 also that even though configuring the SRP seed attempts to hide
3179 invalid usernames by continuing the handshake with fake
3180 credentials, this behaviour is not constant time and no strong
3181 guarantees are made that the handshake is indistinguishable from
3182 that of a valid user.
3186 * Configuration change; it's now possible to build dynamic engines
3187 without having to build shared libraries and vice versa. This
3188 only applies to the engines in `engines/`, those in `crypto/engine/`
3189 will always be built into libcrypto (i.e. "static").
3191 Building dynamic engines is enabled by default; to disable, use
3192 the configuration option "disable-dynamic-engine".
3194 The only requirements for building dynamic engines are the
3195 presence of the DSO module and building with position independent
3196 code, so they will also automatically be disabled if configuring
3197 with "disable-dso" or "disable-pic".
3199 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3200 are also taken away from openssl/opensslconf.h, as they are
3205 * Configuration change; if there is a known flag to compile
3206 position independent code, it will always be applied on the
3207 libcrypto and libssl object files, and never on the application
3208 object files. This means other libraries that use routines from
3209 libcrypto / libssl can be made into shared libraries regardless
3210 of how OpenSSL was configured.
3212 If this isn't desirable, the configuration options "disable-pic"
3213 or "no-pic" can be used to disable the use of PIC. This will
3214 also disable building shared libraries and dynamic engines.
3218 * Removed JPAKE code. It was experimental and has no wide use.
3222 * The INSTALL_PREFIX Makefile variable has been renamed to
3223 DESTDIR. That makes for less confusion on what this variable
3224 is for. Also, the configuration option --install_prefix is
3229 * Heartbeat for TLS has been removed and is disabled by default
3230 for DTLS; configure with enable-heartbeats. Code that uses the
3231 old #define's might need to be updated.
3233 *Emilia Käsper, Rich Salz*
3235 * Rename REF_CHECK to REF_DEBUG.
3239 * New "unified" build system
3241 The "unified" build system is aimed to be a common system for all
3242 platforms we support. With it comes new support for VMS.
3244 This system builds supports building in a different directory tree
3245 than the source tree. It produces one Makefile (for unix family
3246 or lookalikes), or one descrip.mms (for VMS).
3248 The source of information to make the Makefile / descrip.mms is
3249 small files called 'build.info', holding the necessary
3250 information for each directory with source to compile, and a
3251 template in Configurations, like unix-Makefile.tmpl or
3254 With this change, the library names were also renamed on Windows
3255 and on VMS. They now have names that are closer to the standard
3256 on Unix, and include the major version number, and in certain
3257 cases, the architecture they are built for. See "Notes on shared
3258 libraries" in INSTALL.
3260 We rely heavily on the perl module Text::Template.
3264 * Added support for auto-initialisation and de-initialisation of the library.
3265 OpenSSL no longer requires explicit init or deinit routines to be called,
3266 except in certain circumstances. See the OPENSSL_init_crypto() and
3267 OPENSSL_init_ssl() man pages for further information.
3271 * The arguments to the DTLSv1_listen function have changed. Specifically the
3272 "peer" argument is now expected to be a BIO_ADDR object.
3274 * Rewrite of BIO networking library. The BIO library lacked consistent
3275 support of IPv6, and adding it required some more extensive
3276 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3277 which hold all types of addresses and chains of address information.
3278 It also introduces a new API, with functions like BIO_socket,
3279 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3280 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3281 have been adapted accordingly.
3285 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3290 * CRIME protection: disable compression by default, even if OpenSSL is
3291 compiled with zlib enabled. Applications can still enable compression
3292 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3293 using the SSL_CONF library to configure compression.
3297 * The signature of the session callback configured with
3298 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3299 was explicitly marked as `const unsigned char*` instead of
3304 * Always DPURIFY. Remove the use of uninitialized memory in the
3305 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3309 * Removed many obsolete configuration items, including
3310 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3311 MD2_CHAR, MD2_INT, MD2_LONG
3313 IDEA_SHORT, IDEA_LONG
3314 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3316 *Rich Salz, with advice from Andy Polyakov*
3318 * Many BN internals have been moved to an internal header file.
3320 *Rich Salz with help from Andy Polyakov*
3322 * Configuration and writing out the results from it has changed.
3323 Files such as Makefile include/openssl/opensslconf.h and are now
3324 produced through general templates, such as Makefile.in and
3325 crypto/opensslconf.h.in and some help from the perl module
3328 Also, the center of configuration information is no longer
3329 Makefile. Instead, Configure produces a perl module in
3330 configdata.pm which holds most of the config data (in the hash
3331 table %config), the target data that comes from the target
3332 configuration in one of the `Configurations/*.conf` files (in
3337 * To clarify their intended purposes, the Configure options
3338 --prefix and --openssldir change their semantics, and become more
3339 straightforward and less interdependent.
3341 --prefix shall be used exclusively to give the location INSTALLTOP
3342 where programs, scripts, libraries, include files and manuals are
3343 going to be installed. The default is now /usr/local.
3345 --openssldir shall be used exclusively to give the default
3346 location OPENSSLDIR where certificates, private keys, CRLs are
3347 managed. This is also where the default openssl.cnf gets
3349 If the directory given with this option is a relative path, the
3350 values of both the --prefix value and the --openssldir value will
3351 be combined to become OPENSSLDIR.
3352 The default for --openssldir is INSTALLTOP/ssl.
3354 Anyone who uses --openssldir to specify where OpenSSL is to be
3355 installed MUST change to use --prefix instead.
3359 * The GOST engine was out of date and therefore it has been removed. An up
3360 to date GOST engine is now being maintained in an external repository.
3361 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3362 support for GOST ciphersuites (these are only activated if a GOST engine
3367 * EGD is no longer supported by default; use enable-egd when
3370 *Ben Kaduk and Rich Salz*
3372 * The distribution now has Makefile.in files, which are used to
3373 create Makefile's when Configure is run. *Configure must be run
3374 before trying to build now.*
3378 * The return value for SSL_CIPHER_description() for error conditions
3383 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3385 Obtaining and performing DNSSEC validation of TLSA records is
3386 the application's responsibility. The application provides
3387 the TLSA records of its choice to OpenSSL, and these are then
3388 used to authenticate the peer.
3390 The TLSA records need not even come from DNS. They can, for
3391 example, be used to implement local end-entity certificate or
3392 trust-anchor "pinning", where the "pin" data takes the form
3393 of TLSA records, which can augment or replace verification
3394 based on the usual WebPKI public certification authorities.
3398 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3399 continues to support deprecated interfaces in default builds.
3400 However, applications are strongly advised to compile their
3401 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3402 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3403 or the 1.1.0 releases.
3405 In environments in which all applications have been ported to
3406 not use any deprecated interfaces OpenSSL's Configure script
3407 should be used with the --api=1.1.0 option to entirely remove
3408 support for the deprecated features from the library and
3409 unconditionally disable them in the installed headers.
3410 Essentially the same effect can be achieved with the "no-deprecated"
3411 argument to Configure, except that this will always restrict
3412 the build to just the latest API, rather than a fixed API
3415 As applications are ported to future revisions of the API,
3416 they should update their compile-time OPENSSL_API_COMPAT define
3417 accordingly, but in most cases should be able to continue to
3418 compile with later releases.
3420 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3421 0x10000000L and 0x00908000L, respectively. However those
3422 versions did not support the OPENSSL_API_COMPAT feature, and
3423 so applications are not typically tested for explicit support
3424 of just the undeprecated features of either release.
3428 * Add support for setting the minimum and maximum supported protocol.
3429 It can bet set via the SSL_set_min_proto_version() and
3430 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3431 MaxProtocol. It's recommended to use the new APIs to disable
3432 protocols instead of disabling individual protocols using
3433 SSL_set_options() or SSL_CONF's Protocol. This change also
3434 removes support for disabling TLS 1.2 in the OpenSSL TLS
3435 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3439 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3443 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3444 and integrates ECDSA and ECDH functionality into EC. Implementations can
3445 now redirect key generation and no longer need to convert to or from
3448 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3449 include the ec.h header file instead.
3453 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3454 ciphers who are no longer supported and drops support the ephemeral RSA key
3455 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3459 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3460 opaque. For HMAC_CTX, the following constructors and destructors
3463 HMAC_CTX *HMAC_CTX_new(void);
3464 void HMAC_CTX_free(HMAC_CTX *ctx);
3466 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3467 destroy such methods has been added. See EVP_MD_meth_new(3) and
3468 EVP_CIPHER_meth_new(3) for documentation.
3471 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3472 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3473 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3474 an already created structure.
3475 2) For consistency with the majority of our object creators and
3476 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3477 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3478 for deprecated builds.
3482 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3483 cryptographic operations to be performed asynchronously as long as an
3484 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3485 further details. Libssl has also had this capability integrated with the
3486 introduction of the new mode SSL_MODE_ASYNC and associated error
3487 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3488 pages. This work was developed in partnership with Intel Corp.
3492 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3493 always enabled now. If you want to disable the support you should
3494 exclude it using the list of supported ciphers. This also means that the
3495 "-no_ecdhe" option has been removed from s_server.
3499 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3500 SSL_{CTX_}set1_curves() which can set a list.
3504 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3505 curve you want to support using SSL_{CTX_}set1_curves().
3509 * State machine rewrite. The state machine code has been significantly
3510 refactored in order to remove much duplication of code and solve issues
3511 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3512 further details). This change does have some associated API changes.
3513 Notably the SSL_state() function has been removed and replaced by
3514 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3515 SSL_set_state() has been removed altogether. The previous handshake states
3516 defined in ssl.h and ssl3.h have also been removed.
3520 * All instances of the string "ssleay" in the public API were replaced
3521 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3522 Some error codes related to internal RSA_eay API's were renamed.
3526 * The demo files in crypto/threads were moved to demo/threads.
3530 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3533 *Matt Caswell, Rich Salz*
3535 * New ASN.1 embed macro.
3537 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3538 structure is not allocated: it is part of the parent. That is instead of
3546 This reduces memory fragmentation and make it impossible to accidentally
3547 set a mandatory field to NULL.
3549 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3550 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3551 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3556 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3560 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3561 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3562 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3563 DES and RC4 ciphersuites.
3567 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3568 This changes the decoding behaviour for some invalid messages,
3569 though the change is mostly in the more lenient direction, and
3570 legacy behaviour is preserved as much as possible.
3574 * Fix no-stdio build.
3575 *David Woodhouse <David.Woodhouse@intel.com> and also*
3576 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3578 * New testing framework
3579 The testing framework has been largely rewritten and is now using
3580 perl and the perl modules Test::Harness and an extended variant of
3581 Test::More called OpenSSL::Test to do its work. All test scripts in
3582 test/ have been rewritten into test recipes, and all direct calls to
3583 executables in test/Makefile have become individual recipes using the
3584 simplified testing OpenSSL::Test::Simple.
3586 For documentation on our testing modules, do:
3588 perldoc test/testlib/OpenSSL/Test/Simple.pm
3589 perldoc test/testlib/OpenSSL/Test.pm
3593 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3594 are used; the latter aborts on memory leaks (usually checked on exit).
3595 Some undocumented "set malloc, etc., hooks" functions were removed
3596 and others were changed. All are now documented.
3600 * In DSA_generate_parameters_ex, if the provided seed is too short,
3603 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3605 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3606 from RFC4279, RFC4785, RFC5487, RFC5489.
3608 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3609 original RSA_PSK patch.
3613 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3614 era flag was never set throughout the codebase (only read). Also removed
3615 SSL3_FLAGS_POP_BUFFER which was only used if
3616 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3620 * Changed the default name options in the "ca", "crl", "req" and "x509"
3621 to be "oneline" instead of "compat".
3625 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3626 not aware of clients that still exhibit this bug, and the workaround
3627 hasn't been working properly for a while.
3631 * The return type of BIO_number_read() and BIO_number_written() as well as
3632 the corresponding num_read and num_write members in the BIO structure has
3633 changed from unsigned long to uint64_t. On platforms where an unsigned
3634 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3639 * Given the pervasive nature of TLS extensions it is inadvisable to run
3640 OpenSSL without support for them. It also means that maintaining
3641 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3642 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3646 * Removed support for the two export grade static DH ciphersuites
3647 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3648 were newly added (along with a number of other static DH ciphersuites) to
3649 1.0.2. However the two export ones have *never* worked since they were
3650 introduced. It seems strange in any case to be adding new export
3651 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3655 * Version negotiation has been rewritten. In particular SSLv23_method(),
3656 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3657 and turned into macros which simply call the new preferred function names
3658 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3659 should use the new names instead. Also as part of this change the ssl23.h
3660 header file has been removed.
3664 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3665 code and the associated standard is no longer considered fit-for-purpose.
3669 * RT2547 was closed. When generating a private key, try to make the
3670 output file readable only by the owner. This behavior change might
3671 be noticeable when interacting with other software.
3673 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3678 * Added HTTP GET support to the ocsp command.
3682 * Changed default digest for the dgst and enc commands from MD5 to
3687 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3691 * Added support for TLS extended master secret from
3692 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3693 initial patch which was a great help during development.
3697 * All libssl internal structures have been removed from the public header
3698 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3699 now redundant). Users should not attempt to access internal structures
3700 directly. Instead they should use the provided API functions.
3704 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3705 Access to deprecated functions can be re-enabled by running config with
3706 "enable-deprecated". In addition applications wishing to use deprecated
3707 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3708 will, by default, disable some transitive includes that previously existed
3709 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3713 * Added support for OCB mode. OpenSSL has been granted a patent license
3714 compatible with the OpenSSL license for use of OCB. Details are available
3715 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3716 for OCB can be removed by calling config with no-ocb.
3720 * SSLv2 support has been removed. It still supports receiving a SSLv2
3721 compatible client hello.
3725 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3726 done while fixing the error code for the key-too-small case.
3728 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3730 * CA.sh has been removed; use CA.pl instead.
3734 * Removed old DES API.
3738 * Remove various unsupported platforms:
3744 Sinix/ReliantUNIX RM400
3749 16-bit platforms such as WIN16
3753 * Clean up OPENSSL_NO_xxx #define's
3754 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3755 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3756 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3757 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3758 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3759 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3760 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3761 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3762 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3763 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3767 * Cleaned up dead code
3768 Remove all but one '#ifdef undef' which is to be looked at.
3772 * Clean up calling of xxx_free routines.
3773 Just like free(), fix most of the xxx_free routines to accept
3774 NULL. Remove the non-null checks from callers. Save much code.
3778 * Add secure heap for storage of private keys (when possible).
3779 Add BIO_s_secmem(), CBIGNUM, etc.
3780 Contributed by Akamai Technologies under our Corporate CLA.
3784 * Experimental support for a new, fast, unbiased prime candidate generator,
3785 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3787 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3789 * New output format NSS in the sess_id command line tool. This allows
3790 exporting the session id and the master key in NSS keylog format.
3792 *Martin Kaiser <martin@kaiser.cx>*
3794 * Harmonize version and its documentation. -f flag is used to display
3797 *mancha <mancha1@zoho.com>*
3799 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3800 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3802 *mancha <mancha1@zoho.com>*
3804 * Fix some double frees. These are not thought to be exploitable.
3806 *mancha <mancha1@zoho.com>*
3808 * A missing bounds check in the handling of the TLS heartbeat extension
3809 can be used to reveal up to 64k of memory to a connected client or
3812 Thanks for Neel Mehta of Google Security for discovering this bug and to
3813 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3814 preparing the fix ([CVE-2014-0160])
3816 *Adam Langley, Bodo Moeller*
3818 * Fix for the attack described in the paper "Recovering OpenSSL
3819 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3820 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3821 <http://eprint.iacr.org/2014/140>
3823 Thanks to Yuval Yarom and Naomi Benger for discovering this
3824 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3826 *Yuval Yarom and Naomi Benger*
3828 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3829 this fixes a limitation in previous versions of OpenSSL.
3833 * Experimental encrypt-then-mac support.
3835 Experimental support for encrypt then mac from
3836 draft-gutmann-tls-encrypt-then-mac-02.txt
3838 To enable it set the appropriate extension number (0x42 for the test
3839 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3841 For non-compliant peers (i.e. just about everything) this should have no
3844 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3848 * Add EVP support for key wrapping algorithms, to avoid problems with
3849 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3850 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3851 algorithms and include tests cases.
3855 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3860 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3861 MGF1 digest and OAEP label.
3865 * Make openssl verify return errors.
3867 *Chris Palmer <palmer@google.com> and Ben Laurie*
3869 * New function ASN1_TIME_diff to calculate the difference between two
3870 ASN1_TIME structures or one structure and the current time.
3874 * Update fips_test_suite to support multiple command line options. New
3875 test to induce all self test errors in sequence and check expected
3880 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3881 sign or verify all in one operation.
3885 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3886 test programs and fips_test_suite. Includes functionality to parse
3887 the minimal script output of fipsalgest.pl directly.
3891 * Add authorisation parameter to FIPS_module_mode_set().
3895 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3899 * Use separate DRBG fields for internal and external flags. New function
3900 FIPS_drbg_health_check() to perform on demand health checking. Add
3901 generation tests to fips_test_suite with reduced health check interval to
3902 demonstrate periodic health checking. Add "nodh" option to
3903 fips_test_suite to skip very slow DH test.
3907 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3912 * More extensive health check for DRBG checking many more failure modes.
3913 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3914 combination: call this in fips_test_suite.
3918 * Add support for canonical generation of DSA parameter 'g'. See
3921 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3922 POST to handle HMAC cases.
3926 * Add functions FIPS_module_version() and FIPS_module_version_text()
3927 to return numerical and string versions of the FIPS module number.
3931 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3932 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3933 outside the validated module in the FIPS capable OpenSSL.
3937 * Minor change to DRBG entropy callback semantics. In some cases
3938 there is no multiple of the block length between min_len and
3939 max_len. Allow the callback to return more than max_len bytes
3940 of entropy but discard any extra: it is the callback's responsibility
3941 to ensure that the extra data discarded does not impact the
3942 requested amount of entropy.
3946 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3947 information in FIPS186-3, SP800-57 and SP800-131A.
3951 * CCM support via EVP. Interface is very similar to GCM case except we
3952 must supply all data in one chunk (i.e. no update, final) and the
3953 message length must be supplied if AAD is used. Add algorithm test
3958 * Initial version of POST overhaul. Add POST callback to allow the status
3959 of POST to be monitored and/or failures induced. Modify fips_test_suite
3960 to use callback. Always run all selftests even if one fails.
3964 * XTS support including algorithm test driver in the fips_gcmtest program.
3965 Note: this does increase the maximum key length from 32 to 64 bytes but
3966 there should be no binary compatibility issues as existing applications
3967 will never use XTS mode.
3971 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3972 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3973 performs algorithm blocking for unapproved PRNG types. Also do not
3974 set PRNG type in FIPS_mode_set(): leave this to the application.
3975 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3976 the standard OpenSSL PRNG: set additional data to a date time vector.
3980 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3981 This shouldn't present any incompatibility problems because applications
3982 shouldn't be using these directly and any that are will need to rethink
3983 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3987 * Extensive self tests and health checking required by SP800-90 DRBG.
3988 Remove strength parameter from FIPS_drbg_instantiate and always
3989 instantiate at maximum supported strength.
3993 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3997 * New algorithm test program fips_dhvs to handle DH primitives only testing.
4001 * New function DH_compute_key_padded() to compute a DH key and pad with
4002 leading zeroes if needed: this complies with SP800-56A et al.
4006 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
4007 anything, incomplete, subject to change and largely untested at present.
4011 * Modify fipscanisteronly build option to only build the necessary object
4012 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
4016 * Add experimental option FIPSSYMS to give all symbols in
4017 fipscanister.o and FIPS or fips prefix. This will avoid
4018 conflicts with future versions of OpenSSL. Add perl script
4019 util/fipsas.pl to preprocess assembly language source files
4020 and rename any affected symbols.
4024 * Add selftest checks and algorithm block of non-fips algorithms in
4025 FIPS mode. Remove DES2 from selftests.
4029 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
4030 return internal method without any ENGINE dependencies. Add new
4031 tiny fips sign and verify functions.
4035 * New build option no-ec2m to disable characteristic 2 code.
4039 * New build option "fipscanisteronly". This only builds fipscanister.o
4040 and (currently) associated fips utilities. Uses the file Makefile.fips
4041 instead of Makefile.org as the prototype.
4045 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4046 Update fips_gcmtest to use IV generator.
4050 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4051 setting output buffer to NULL. The `*Final` function must be
4052 called although it will not retrieve any additional data. The tag
4053 can be set or retrieved with a ctrl. The IV length is by default 12
4054 bytes (96 bits) but can be set to an alternative value. If the IV
4055 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4060 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4061 underlying do_cipher function handles all cipher semantics itself
4062 including padding and finalisation. This is useful if (for example)
4063 an ENGINE cipher handles block padding itself. The behaviour of
4064 do_cipher is subtly changed if this flag is set: the return value
4065 is the number of characters written to the output buffer (zero is
4066 no longer an error code) or a negative error code. Also if the
4067 input buffer is NULL and length 0 finalisation should be performed.
4071 * If a candidate issuer certificate is already part of the constructed
4072 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4076 * Improve forward-security support: add functions
4078 void SSL_CTX_set_not_resumable_session_callback(
4079 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4080 void SSL_set_not_resumable_session_callback(
4081 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4083 for use by SSL/TLS servers; the callback function will be called whenever a
4084 new session is created, and gets to decide whether the session may be
4085 cached to make it resumable (return 0) or not (return 1). (As by the
4086 SSL/TLS protocol specifications, the session_id sent by the server will be
4087 empty to indicate that the session is not resumable; also, the server will
4088 not generate RFC 4507 (RFC 5077) session tickets.)
4090 A simple reasonable callback implementation is to return is_forward_secure.
4091 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4092 by the SSL/TLS server library, indicating whether it can provide forward
4095 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4097 * New -verify_name option in command line utilities to set verification
4102 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4103 Add CMAC pkey methods.
4107 * Experimental renegotiation in s_server -www mode. If the client
4108 browses /reneg connection is renegotiated. If /renegcert it is
4109 renegotiated requesting a certificate.
4113 * Add an "external" session cache for debugging purposes to s_server. This
4114 should help trace issues which normally are only apparent in deployed
4115 multi-process servers.
4119 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4120 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4121 BIO_set_cipher() and some obscure PEM functions were changed so they
4122 can now return an error. The RAND changes required a change to the
4123 RAND_METHOD structure.
4127 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4128 a gcc attribute to warn if the result of a function is ignored. This
4129 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4130 whose return value is often ignored.
4134 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4135 These allow SCTs (signed certificate timestamps) to be requested and
4136 validated when establishing a connection.
4138 *Rob Percival <robpercival@google.com>*
4143 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4145 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4146 used even when parsing explicit parameters, when loading a encoded key
4147 or calling `EC_GROUP_new_from_ecpkparameters()`/
4148 `EC_GROUP_new_from_ecparameters()`.
4149 This prevents bypass of security hardening and performance gains,
4150 especially for curves with specialized EC_METHODs.
4151 By default, if a key encoded with explicit parameters is loaded and later
4152 encoded, the output is still encoded with explicit parameters, even if
4153 internally a "named" EC_GROUP is used for computation.
4157 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4158 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4159 NULL. After this change, only the cofactor parameter can be NULL. It also
4160 does some minimal sanity checks on the passed order.
4165 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4166 An attack is simple, if the first CMS_recipientInfo is valid but the
4167 second CMS_recipientInfo is chosen ciphertext. If the second
4168 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4169 encryption key will be replaced by garbage, and the message cannot be
4170 decoded, but if the RSA decryption fails, the correct encryption key is
4171 used and the recipient will not notice the attack.
4172 As a work around for this potential attack the length of the decrypted
4173 key must be equal to the cipher default key length, in case the
4174 certifiate is not given and all recipientInfo are tried out.
4175 The old behaviour can be re-enabled in the CMS code by setting the
4176 CMS_DEBUG_DECRYPT flag.
4181 * Document issue with installation paths in diverse Windows builds
4183 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4184 binaries and run-time config file.
4189 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4191 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4192 This changes the size when using the `genpkey` command when no size is given.
4193 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4194 generation commands to use 2048 bits by default.
4198 * Add FIPS support for Android Arm 64-bit
4200 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4201 Module in Version 2.0.10. For some reason, the corresponding target
4202 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4203 built with FIPS support on Android Arm 64-bit. This omission has been
4206 *Matthias St. Pierre*
4208 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4210 * 0-byte record padding oracle
4212 If an application encounters a fatal protocol error and then calls
4213 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4214 then OpenSSL can respond differently to the calling application if a 0 byte
4215 record is received with invalid padding compared to if a 0 byte record is
4216 received with an invalid MAC. If the application then behaves differently
4217 based on that in a way that is detectable to the remote peer, then this
4218 amounts to a padding oracle that could be used to decrypt data.
4220 In order for this to be exploitable "non-stitched" ciphersuites must be in
4221 use. Stitched ciphersuites are optimised implementations of certain
4222 commonly used ciphersuites. Also the application must call SSL_shutdown()
4223 twice even if a protocol error has occurred (applications should not do
4224 this but some do anyway).
4226 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4227 Aviram, with additional investigation by Steven Collison and Andrew
4228 Hourselt. It was reported to OpenSSL on 10th December 2018.
4233 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4237 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4239 * Microarchitecture timing vulnerability in ECC scalar multiplication
4241 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4242 shown to be vulnerable to a microarchitecture timing side channel attack.
4243 An attacker with sufficient access to mount local timing attacks during
4244 ECDSA signature generation could recover the private key.
4246 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4247 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4253 * Timing vulnerability in DSA signature generation
4255 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4256 timing side channel attack. An attacker could use variations in the signing
4257 algorithm to recover the private key.
4259 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4264 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4265 Module, accidentally introduced while backporting security fixes from the
4266 development branch and hindering the use of ECC in FIPS mode.
4270 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4272 * Client DoS due to large DH parameter
4274 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4275 malicious server can send a very large prime value to the client. This will
4276 cause the client to spend an unreasonably long period of time generating a
4277 key for this prime resulting in a hang until the client has finished. This
4278 could be exploited in a Denial Of Service attack.
4280 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4285 * Cache timing vulnerability in RSA Key Generation
4287 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4288 a cache timing side channel attack. An attacker with sufficient access to
4289 mount cache timing attacks during the RSA key generation process could
4290 recover the private key.
4292 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4293 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4298 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4299 parameter is no longer accepted, as it leads to a corrupt table. NULL
4300 pem_str is reserved for alias entries only.
4304 * Revert blinding in ECDSA sign and instead make problematic addition
4305 length-invariant. Switch even to fixed-length Montgomery multiplication.
4309 * Change generating and checking of primes so that the error rate of not
4310 being prime depends on the intended use based on the size of the input.
4311 For larger primes this will result in more rounds of Miller-Rabin.
4312 The maximal error rate for primes with more than 1080 bits is lowered
4315 *Kurt Roeckx, Annie Yousar*
4317 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4321 * Add blinding to ECDSA and DSA signatures to protect against side channel
4322 attacks discovered by Keegan Ryan (NCC Group).
4326 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4327 now allow empty (zero character) pass phrases.
4331 * Certificate time validation (X509_cmp_time) enforces stricter
4332 compliance with RFC 5280. Fractional seconds and timezone offsets
4333 are no longer allowed.
4337 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4339 * Constructed ASN.1 types with a recursive definition could exceed the stack
4341 Constructed ASN.1 types with a recursive definition (such as can be found
4342 in PKCS7) could eventually exceed the stack given malicious input with
4343 excessive recursion. This could result in a Denial Of Service attack. There
4344 are no such structures used within SSL/TLS that come from untrusted sources
4345 so this is considered safe.
4347 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4353 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4355 * Read/write after SSL object in error state
4357 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4358 mechanism. The intent was that if a fatal error occurred during a handshake
4359 then OpenSSL would move into the error state and would immediately fail if
4360 you attempted to continue the handshake. This works as designed for the
4361 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4362 SSL_connect()), however due to a bug it does not work correctly if
4363 SSL_read() or SSL_write() is called directly. In that scenario, if the
4364 handshake fails then a fatal error will be returned in the initial function
4365 call. If SSL_read()/SSL_write() is subsequently called by the application
4366 for the same SSL object then it will succeed and the data is passed without
4367 being decrypted/encrypted directly from the SSL/TLS record layer.
4369 In order to exploit this issue an application bug would have to be present
4370 that resulted in a call to SSL_read()/SSL_write() being issued after having
4371 already received a fatal error.
4373 This issue was reported to OpenSSL by David Benjamin (Google).
4378 * rsaz_1024_mul_avx2 overflow bug on x86_64
4380 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4381 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4382 Analysis suggests that attacks against RSA and DSA as a result of this
4383 defect would be very difficult to perform and are not believed likely.
4384 Attacks against DH1024 are considered just feasible, because most of the
4385 work necessary to deduce information about a private key may be performed
4386 offline. The amount of resources required for such an attack would be
4387 significant. However, for an attack on TLS to be meaningful, the server
4388 would have to share the DH1024 private key among multiple clients, which is
4389 no longer an option since CVE-2016-0701.
4391 This only affects processors that support the AVX2 but not ADX extensions
4392 like Intel Haswell (4th generation).
4394 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4395 was originally found via the OSS-Fuzz project.
4400 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4402 * bn_sqrx8x_internal carry bug on x86_64
4404 There is a carry propagating bug in the x86_64 Montgomery squaring
4405 procedure. No EC algorithms are affected. Analysis suggests that attacks
4406 against RSA and DSA as a result of this defect would be very difficult to
4407 perform and are not believed likely. Attacks against DH are considered just
4408 feasible (although very difficult) because most of the work necessary to
4409 deduce information about a private key may be performed offline. The amount
4410 of resources required for such an attack would be very significant and
4411 likely only accessible to a limited number of attackers. An attacker would
4412 additionally need online access to an unpatched system using the target
4413 private key in a scenario with persistent DH parameters and a private
4414 key that is shared between multiple clients.
4416 This only affects processors that support the BMI1, BMI2 and ADX extensions
4417 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4419 This issue was reported to OpenSSL by the OSS-Fuzz project.
4424 * Malformed X.509 IPAddressFamily could cause OOB read
4426 If an X.509 certificate has a malformed IPAddressFamily extension,
4427 OpenSSL could do a one-byte buffer overread. The most likely result
4428 would be an erroneous display of the certificate in text format.
4430 This issue was reported to OpenSSL by the OSS-Fuzz project.
4435 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4437 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4438 platform rather than 'mingw'.
4442 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4444 * Truncated packet could crash via OOB read
4446 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4447 cipher is being used, then a truncated packet can cause that host to
4448 perform an out-of-bounds read, usually resulting in a crash.
4450 This issue was reported to OpenSSL by Robert Święcki of Google.
4455 * BN_mod_exp may produce incorrect results on x86_64
4457 There is a carry propagating bug in the x86_64 Montgomery squaring
4458 procedure. No EC algorithms are affected. Analysis suggests that attacks
4459 against RSA and DSA as a result of this defect would be very difficult to
4460 perform and are not believed likely. Attacks against DH are considered just
4461 feasible (although very difficult) because most of the work necessary to
4462 deduce information about a private key may be performed offline. The amount
4463 of resources required for such an attack would be very significant and
4464 likely only accessible to a limited number of attackers. An attacker would
4465 additionally need online access to an unpatched system using the target
4466 private key in a scenario with persistent DH parameters and a private
4467 key that is shared between multiple clients. For example this can occur by
4468 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4469 similar to CVE-2015-3193 but must be treated as a separate problem.
4471 This issue was reported to OpenSSL by the OSS-Fuzz project.
4476 * Montgomery multiplication may produce incorrect results
4478 There is a carry propagating bug in the Broadwell-specific Montgomery
4479 multiplication procedure that handles input lengths divisible by, but
4480 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4481 and DH private keys are impossible. This is because the subroutine in
4482 question is not used in operations with the private key itself and an input
4483 of the attacker's direct choice. Otherwise the bug can manifest itself as
4484 transient authentication and key negotiation failures or reproducible
4485 erroneous outcome of public-key operations with specially crafted input.
4486 Among EC algorithms only Brainpool P-512 curves are affected and one
4487 presumably can attack ECDH key negotiation. Impact was not analyzed in
4488 detail, because pre-requisites for attack are considered unlikely. Namely
4489 multiple clients have to choose the curve in question and the server has to
4490 share the private key among them, neither of which is default behaviour.
4491 Even then only clients that chose the curve will be affected.
4493 This issue was publicly reported as transient failures and was not
4494 initially recognized as a security issue. Thanks to Richard Morgan for
4495 providing reproducible case.
4500 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4501 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4502 prevent issues where no progress is being made and the peer continually
4503 sends unrecognised record types, using up resources processing them.
4507 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4509 * Missing CRL sanity check
4511 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4512 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4513 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4515 This issue only affects the OpenSSL 1.0.2i
4520 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4522 * OCSP Status Request extension unbounded memory growth
4524 A malicious client can send an excessively large OCSP Status Request
4525 extension. If that client continually requests renegotiation, sending a
4526 large OCSP Status Request extension each time, then there will be unbounded
4527 memory growth on the server. This will eventually lead to a Denial Of
4528 Service attack through memory exhaustion. Servers with a default
4529 configuration are vulnerable even if they do not support OCSP. Builds using
4530 the "no-ocsp" build time option are not affected.
4532 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4537 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4540 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4546 * OOB write in MDC2_Update()
4548 An overflow can occur in MDC2_Update() either if called directly or
4549 through the EVP_DigestUpdate() function using MDC2. If an attacker
4550 is able to supply very large amounts of input data after a previous
4551 call to EVP_EncryptUpdate() with a partial block then a length check
4552 can overflow resulting in a heap corruption.
4554 The amount of data needed is comparable to SIZE_MAX which is impractical
4557 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4562 * Malformed SHA512 ticket DoS
4564 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4565 DoS attack where a malformed ticket will result in an OOB read which will
4568 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4569 a custom server callback and ticket lookup mechanism.
4571 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4576 * OOB write in BN_bn2dec()
4578 The function BN_bn2dec() does not check the return value of BN_div_word().
4579 This can cause an OOB write if an application uses this function with an
4580 overly large BIGNUM. This could be a problem if an overly large certificate
4581 or CRL is printed out from an untrusted source. TLS is not affected because
4582 record limits will reject an oversized certificate before it is parsed.
4584 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4589 * OOB read in TS_OBJ_print_bio()
4591 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4592 the total length the OID text representation would use and not the amount
4593 of data written. This will result in OOB reads when large OIDs are
4596 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4601 * Pointer arithmetic undefined behaviour
4603 Avoid some undefined pointer arithmetic
4605 A common idiom in the codebase is to check limits in the following manner:
4608 Where "p" points to some malloc'd data of SIZE bytes and
4611 "len" here could be from some externally supplied data (e.g. from a TLS
4614 The rules of C pointer arithmetic are such that "p + len" is only well
4615 defined where len <= SIZE. Therefore the above idiom is actually
4616 undefined behaviour.
4618 For example this could cause problems if some malloc implementation
4619 provides an address for "p" such that "p + len" actually overflows for
4620 values of len that are too big and therefore p + len < limit.
4622 This issue was reported to OpenSSL by Guido Vranken
4627 * Constant time flag not preserved in DSA signing
4629 Operations in the DSA signing algorithm should run in constant time in
4630 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4631 implementation means that a non-constant time codepath is followed for
4632 certain operations. This has been demonstrated through a cache-timing
4633 attack to be sufficient for an attacker to recover the private DSA key.
4635 This issue was reported by César Pereida (Aalto University), Billy Brumley
4636 (Tampere University of Technology), and Yuval Yarom (The University of
4637 Adelaide and NICTA).
4642 * DTLS buffered message DoS
4644 In a DTLS connection where handshake messages are delivered out-of-order
4645 those messages that OpenSSL is not yet ready to process will be buffered
4646 for later use. Under certain circumstances, a flaw in the logic means that
4647 those messages do not get removed from the buffer even though the handshake
4648 has been completed. An attacker could force up to approx. 15 messages to
4649 remain in the buffer when they are no longer required. These messages will
4650 be cleared when the DTLS connection is closed. The default maximum size for
4651 a message is 100k. Therefore the attacker could force an additional 1500k
4652 to be consumed per connection. By opening many simulataneous connections an
4653 attacker could cause a DoS attack through memory exhaustion.
4655 This issue was reported to OpenSSL by Quan Luo.
4660 * DTLS replay protection DoS
4662 A flaw in the DTLS replay attack protection mechanism means that records
4663 that arrive for future epochs update the replay protection "window" before
4664 the MAC for the record has been validated. This could be exploited by an
4665 attacker by sending a record for the next epoch (which does not have to
4666 decrypt or have a valid MAC), with a very large sequence number. This means
4667 that all subsequent legitimate packets are dropped causing a denial of
4668 service for a specific DTLS connection.
4670 This issue was reported to OpenSSL by the OCAP audit team.
4675 * Certificate message OOB reads
4677 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4678 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4679 theoretical DoS risk but this has not been observed in practice on common
4682 The messages affected are client certificate, client certificate request
4683 and server certificate. As a result the attack can only be performed
4684 against a client or a server which enables client authentication.
4686 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4691 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4693 * Prevent padding oracle in AES-NI CBC MAC check
4695 A MITM attacker can use a padding oracle attack to decrypt traffic
4696 when the connection uses an AES CBC cipher and the server support
4699 This issue was introduced as part of the fix for Lucky 13 padding
4700 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4701 constant time by making sure that always the same bytes are read and
4702 compared against either the MAC or padding bytes. But it no longer
4703 checked that there was enough data to have both the MAC and padding
4706 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4711 * Fix EVP_EncodeUpdate overflow
4713 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4714 Base64 encoding of binary data. If an attacker is able to supply very large
4715 amounts of input data then a length check can overflow resulting in a heap
4718 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4719 the `PEM_write_bio*` family of functions. These are mainly used within the
4720 OpenSSL command line applications, so any application which processes data
4721 from an untrusted source and outputs it as a PEM file should be considered
4722 vulnerable to this issue. User applications that call these APIs directly
4723 with large amounts of untrusted data may also be vulnerable.
4725 This issue was reported by Guido Vranken.
4730 * Fix EVP_EncryptUpdate overflow
4732 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4733 is able to supply very large amounts of input data after a previous call to
4734 EVP_EncryptUpdate() with a partial block then a length check can overflow
4735 resulting in a heap corruption. Following an analysis of all OpenSSL
4736 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4737 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4738 the first called function after an EVP_EncryptInit(), and therefore that
4739 specific call must be safe. The second form is where the length passed to
4740 EVP_EncryptUpdate() can be seen from the code to be some small value and
4741 therefore there is no possibility of an overflow. Since all instances are
4742 one of these two forms, it is believed that there can be no overflows in
4743 internal code due to this problem. It should be noted that
4744 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4745 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4746 of these calls have also been analysed too and it is believed there are no
4747 instances in internal usage where an overflow could occur.
4749 This issue was reported by Guido Vranken.
4754 * Prevent ASN.1 BIO excessive memory allocation
4756 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4757 a short invalid encoding can cause allocation of large amounts of memory
4758 potentially consuming excessive resources or exhausting memory.
4760 Any application parsing untrusted data through d2i BIO functions is
4761 affected. The memory based functions such as d2i_X509() are *not* affected.
4762 Since the memory based functions are used by the TLS library, TLS
4763 applications are not affected.
4765 This issue was reported by Brian Carpenter.
4772 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4773 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4774 in arbitrary stack data being returned in the buffer.
4776 This issue was reported by Guido Vranken.
4781 * Modify behavior of ALPN to invoke callback after SNI/servername
4782 callback, such that updates to the SSL_CTX affect ALPN.
4786 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4791 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4792 methods are enabled and ssl2 is disabled the methods return NULL.
4796 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4798 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4799 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4800 provide any "EXPORT" or "LOW" strength ciphers.
4804 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4805 is by default disabled at build-time. Builds that are not configured with
4806 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4807 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4808 will need to explicitly call either of:
4810 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4812 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4814 as appropriate. Even if either of those is used, or the application
4815 explicitly uses the version-specific SSLv2_method() or its client and
4816 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4817 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4818 ciphers, and SSLv2 56-bit DES are no longer available.
4823 * Fix a double-free in DSA code
4825 A double free bug was discovered when OpenSSL parses malformed DSA private
4826 keys and could lead to a DoS attack or memory corruption for applications
4827 that receive DSA private keys from untrusted sources. This scenario is
4830 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4836 * Disable SRP fake user seed to address a server memory leak.
4838 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4840 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4841 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4842 was changed to ignore the "fake user" SRP seed, even if the seed
4845 Users should use SRP_VBASE_get1_by_user instead. Note that in
4846 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4847 also that even though configuring the SRP seed attempts to hide
4848 invalid usernames by continuing the handshake with fake
4849 credentials, this behaviour is not constant time and no strong
4850 guarantees are made that the handshake is indistinguishable from
4851 that of a valid user.
4856 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4858 In the BN_hex2bn function the number of hex digits is calculated using an
4859 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4860 large values of `i` this can result in `bn_expand` not allocating any
4861 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4862 field as NULL leading to a subsequent NULL ptr deref. For very large values
4863 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4864 In this case memory is allocated to the internal BIGNUM data field, but it
4865 is insufficiently sized leading to heap corruption. A similar issue exists
4866 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4867 is ever called by user applications with very large untrusted hex/dec data.
4868 This is anticipated to be a rare occurrence.
4870 All OpenSSL internal usage of these functions use data that is not expected
4871 to be untrusted, e.g. config file data or application command line
4872 arguments. If user developed applications generate config file data based
4873 on untrusted data then it is possible that this could also lead to security
4874 consequences. This is also anticipated to be rare.
4876 This issue was reported to OpenSSL by Guido Vranken.
4881 * Fix memory issues in `BIO_*printf` functions
4883 The internal `fmtstr` function used in processing a "%s" format string in
4884 the `BIO_*printf` functions could overflow while calculating the length of a
4885 string and cause an OOB read when printing very long strings.
4887 Additionally the internal `doapr_outch` function can attempt to write to an
4888 OOB memory location (at an offset from the NULL pointer) in the event of a
4889 memory allocation failure. In 1.0.2 and below this could be caused where
4890 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4891 could be in processing a very long "%s" format string. Memory leaks can
4894 The first issue may mask the second issue dependent on compiler behaviour.
4895 These problems could enable attacks where large amounts of untrusted data
4896 is passed to the `BIO_*printf` functions. If applications use these functions
4897 in this way then they could be vulnerable. OpenSSL itself uses these
4898 functions when printing out human-readable dumps of ASN.1 data. Therefore
4899 applications that print this data could be vulnerable if the data is from
4900 untrusted sources. OpenSSL command line applications could also be
4901 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4902 as command line arguments.
4904 Libssl is not considered directly vulnerable. Additionally certificates etc
4905 received via remote connections via libssl are also unlikely to be able to
4906 trigger these issues because of message size limits enforced within libssl.
4908 This issue was reported to OpenSSL Guido Vranken.
4913 * Side channel attack on modular exponentiation
4915 A side-channel attack was found which makes use of cache-bank conflicts on
4916 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4917 of RSA keys. The ability to exploit this issue is limited as it relies on
4918 an attacker who has control of code in a thread running on the same
4919 hyper-threaded core as the victim thread which is performing decryptions.
4921 This issue was reported to OpenSSL by Yuval Yarom, The University of
4922 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4923 Nadia Heninger, University of Pennsylvania with more information at
4924 <http://cachebleed.info>.
4929 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4930 if no keysize is specified with default_bits. This fixes an
4931 omission in an earlier change that changed all RSA/DSA key generation
4932 commands to use 2048 bits by default.
4936 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4938 * DH small subgroups
4940 Historically OpenSSL only ever generated DH parameters based on "safe"
4941 primes. More recently (in version 1.0.2) support was provided for
4942 generating X9.42 style parameter files such as those required for RFC 5114
4943 support. The primes used in such files may not be "safe". Where an
4944 application is using DH configured with parameters based on primes that are
4945 not "safe" then an attacker could use this fact to find a peer's private
4946 DH exponent. This attack requires that the attacker complete multiple
4947 handshakes in which the peer uses the same private DH exponent. For example
4948 this could be used to discover a TLS server's private DH exponent if it's
4949 reusing the private DH exponent or it's using a static DH ciphersuite.
4951 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4952 TLS. It is not on by default. If the option is not set then the server
4953 reuses the same private DH exponent for the life of the server process and
4954 would be vulnerable to this attack. It is believed that many popular
4955 applications do set this option and would therefore not be at risk.
4957 The fix for this issue adds an additional check where a "q" parameter is
4958 available (as is the case in X9.42 based parameters). This detects the
4959 only known attack, and is the only possible defense for static DH
4960 ciphersuites. This could have some performance impact.
4962 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4963 default and cannot be disabled. This could have some performance impact.
4965 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4970 * SSLv2 doesn't block disabled ciphers
4972 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4973 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4974 been disabled, provided that the SSLv2 protocol was not also disabled via
4977 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4978 and Sebastian Schinzel.
4983 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4985 * BN_mod_exp may produce incorrect results on x86_64
4987 There is a carry propagating bug in the x86_64 Montgomery squaring
4988 procedure. No EC algorithms are affected. Analysis suggests that attacks
4989 against RSA and DSA as a result of this defect would be very difficult to
4990 perform and are not believed likely. Attacks against DH are considered just
4991 feasible (although very difficult) because most of the work necessary to
4992 deduce information about a private key may be performed offline. The amount
4993 of resources required for such an attack would be very significant and
4994 likely only accessible to a limited number of attackers. An attacker would
4995 additionally need online access to an unpatched system using the target
4996 private key in a scenario with persistent DH parameters and a private
4997 key that is shared between multiple clients. For example this can occur by
4998 default in OpenSSL DHE based SSL/TLS ciphersuites.
5000 This issue was reported to OpenSSL by Hanno Böck.
5005 * Certificate verify crash with missing PSS parameter
5007 The signature verification routines will crash with a NULL pointer
5008 dereference if presented with an ASN.1 signature using the RSA PSS
5009 algorithm and absent mask generation function parameter. Since these
5010 routines are used to verify certificate signature algorithms this can be
5011 used to crash any certificate verification operation and exploited in a
5012 DoS attack. Any application which performs certificate verification is
5013 vulnerable including OpenSSL clients and servers which enable client
5016 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
5021 * X509_ATTRIBUTE memory leak
5023 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
5024 memory. This structure is used by the PKCS#7 and CMS routines so any
5025 application which reads PKCS#7 or CMS data from untrusted sources is
5026 affected. SSL/TLS is not affected.
5028 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
5034 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
5035 This changes the decoding behaviour for some invalid messages,
5036 though the change is mostly in the more lenient direction, and
5037 legacy behaviour is preserved as much as possible.
5041 * In DSA_generate_parameters_ex, if the provided seed is too short,
5044 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5046 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5048 * Alternate chains certificate forgery
5050 During certificate verification, OpenSSL will attempt to find an
5051 alternative certificate chain if the first attempt to build such a chain
5052 fails. An error in the implementation of this logic can mean that an
5053 attacker could cause certain checks on untrusted certificates to be
5054 bypassed, such as the CA flag, enabling them to use a valid leaf
5055 certificate to act as a CA and "issue" an invalid certificate.
5057 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5062 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5064 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5065 incompatibility in the handling of HMAC. The previous ABI has now been
5070 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5072 * Malformed ECParameters causes infinite loop
5074 When processing an ECParameters structure OpenSSL enters an infinite loop
5075 if the curve specified is over a specially malformed binary polynomial
5078 This can be used to perform denial of service against any
5079 system which processes public keys, certificate requests or
5080 certificates. This includes TLS clients and TLS servers with
5081 client authentication enabled.
5083 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5088 * Exploitable out-of-bounds read in X509_cmp_time
5090 X509_cmp_time does not properly check the length of the ASN1_TIME
5091 string and can read a few bytes out of bounds. In addition,
5092 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5095 An attacker can use this to craft malformed certificates and CRLs of
5096 various sizes and potentially cause a segmentation fault, resulting in
5097 a DoS on applications that verify certificates or CRLs. TLS clients
5098 that verify CRLs are affected. TLS clients and servers with client
5099 authentication enabled may be affected if they use custom verification
5102 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5103 independently by Hanno Böck.
5108 * PKCS7 crash with missing EnvelopedContent
5110 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5111 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5112 with missing content and trigger a NULL pointer dereference on parsing.
5114 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5115 structures from untrusted sources are affected. OpenSSL clients and
5116 servers are not affected.
5118 This issue was reported to OpenSSL by Michal Zalewski (Google).
5123 * CMS verify infinite loop with unknown hash function
5125 When verifying a signedData message the CMS code can enter an infinite loop
5126 if presented with an unknown hash function OID. This can be used to perform
5127 denial of service against any system which verifies signedData messages using
5129 This issue was reported to OpenSSL by Johannes Bauer.
5134 * Race condition handling NewSessionTicket
5136 If a NewSessionTicket is received by a multi-threaded client when attempting to
5137 reuse a previous ticket then a race condition can occur potentially leading to
5138 a double free of the ticket data.
5143 * Only support 256-bit or stronger elliptic curves with the
5144 'ecdh_auto' setting (server) or by default (client). Of supported
5145 curves, prefer P-256 (both).
5149 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5151 * ClientHello sigalgs DoS fix
5153 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5154 invalid signature algorithms extension a NULL pointer dereference will
5155 occur. This can be exploited in a DoS attack against the server.
5157 This issue was was reported to OpenSSL by David Ramos of Stanford
5161 *Stephen Henson and Matt Caswell*
5163 * Multiblock corrupted pointer fix
5165 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5166 feature only applies on 64 bit x86 architecture platforms that support AES
5167 NI instructions. A defect in the implementation of "multiblock" can cause
5168 OpenSSL's internal write buffer to become incorrectly set to NULL when
5169 using non-blocking IO. Typically, when the user application is using a
5170 socket BIO for writing, this will only result in a failed connection.
5171 However if some other BIO is used then it is likely that a segmentation
5172 fault will be triggered, thus enabling a potential DoS attack.
5174 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5179 * Segmentation fault in DTLSv1_listen fix
5181 The DTLSv1_listen function is intended to be stateless and processes the
5182 initial ClientHello from many peers. It is common for user code to loop
5183 over the call to DTLSv1_listen until a valid ClientHello is received with
5184 an associated cookie. A defect in the implementation of DTLSv1_listen means
5185 that state is preserved in the SSL object from one invocation to the next
5186 that can lead to a segmentation fault. Errors processing the initial
5187 ClientHello can trigger this scenario. An example of such an error could be
5188 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5191 This issue was reported to OpenSSL by Per Allansson.
5196 * Segmentation fault in ASN1_TYPE_cmp fix
5198 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5199 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5200 certificate signature algorithm consistency this can be used to crash any
5201 certificate verification operation and exploited in a DoS attack. Any
5202 application which performs certificate verification is vulnerable including
5203 OpenSSL clients and servers which enable client authentication.
5208 * Segmentation fault for invalid PSS parameters fix
5210 The signature verification routines will crash with a NULL pointer
5211 dereference if presented with an ASN.1 signature using the RSA PSS
5212 algorithm and invalid parameters. Since these routines are used to verify
5213 certificate signature algorithms this can be used to crash any
5214 certificate verification operation and exploited in a DoS attack. Any
5215 application which performs certificate verification is vulnerable including
5216 OpenSSL clients and servers which enable client authentication.
5218 This issue was was reported to OpenSSL by Brian Carpenter.
5223 * ASN.1 structure reuse memory corruption fix
5225 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5226 memory corruption via an invalid write. Such reuse is and has been
5227 strongly discouraged and is believed to be rare.
5229 Applications that parse structures containing CHOICE or ANY DEFINED BY
5230 components may be affected. Certificate parsing (d2i_X509 and related
5231 functions) are however not affected. OpenSSL clients and servers are
5237 * PKCS7 NULL pointer dereferences fix
5239 The PKCS#7 parsing code does not handle missing outer ContentInfo
5240 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5241 missing content and trigger a NULL pointer dereference on parsing.
5243 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5244 otherwise parse PKCS#7 structures from untrusted sources are
5245 affected. OpenSSL clients and servers are not affected.
5247 This issue was reported to OpenSSL by Michal Zalewski (Google).
5252 * DoS via reachable assert in SSLv2 servers fix
5254 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5255 servers that both support SSLv2 and enable export cipher suites by sending
5256 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5258 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5259 (OpenSSL development team).
5264 * Empty CKE with client auth and DHE fix
5266 If client auth is used then a server can seg fault in the event of a DHE
5267 ciphersuite being selected and a zero length ClientKeyExchange message
5268 being sent by the client. This could be exploited in a DoS attack.
5273 * Handshake with unseeded PRNG fix
5275 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5276 with an unseeded PRNG. The conditions are:
5277 - The client is on a platform where the PRNG has not been seeded
5278 automatically, and the user has not seeded manually
5279 - A protocol specific client method version has been used (i.e. not
5280 SSL_client_methodv23)
5281 - A ciphersuite is used that does not require additional random data from
5282 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5284 If the handshake succeeds then the client random that has been used will
5285 have been generated from a PRNG with insufficient entropy and therefore the
5286 output may be predictable.
5288 For example using the following command with an unseeded openssl will
5289 succeed on an unpatched platform:
5291 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5296 * Use After Free following d2i_ECPrivatekey error fix
5298 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5299 could cause a use after free condition. This, in turn, could cause a double
5300 free in several private key parsing functions (such as d2i_PrivateKey
5301 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5302 for applications that receive EC private keys from untrusted
5303 sources. This scenario is considered rare.
5305 This issue was discovered by the BoringSSL project and fixed in their
5311 * X509_to_X509_REQ NULL pointer deref fix
5313 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5314 the certificate key is invalid. This function is rarely used in practice.
5316 This issue was discovered by Brian Carpenter.
5321 * Removed the export ciphers from the DEFAULT ciphers
5325 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5327 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5328 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5329 So far those who have to target multiple platforms would compromise
5330 and argue that binary targeting say ARMv5 would still execute on
5331 ARMv8. "Universal" build resolves this compromise by providing
5332 near-optimal performance even on newer platforms.
5336 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5337 (other platforms pending).
5339 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5341 * Add support for the SignedCertificateTimestampList certificate and
5342 OCSP response extensions from RFC6962.
5346 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5347 for corner cases. (Certain input points at infinity could lead to
5348 bogus results, with non-infinity inputs mapped to infinity too.)
5352 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5353 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5354 common cases are optimized and there still is room for further
5355 improvements. Vector Permutation AES for Altivec is also added.
5359 * Add support for little-endian ppc64 Linux target.
5361 *Marcelo Cerri (IBM)*
5363 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5364 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5365 are optimized and there still is room for further improvements.
5366 Both 32- and 64-bit modes are supported.
5368 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5370 * Improved ARMv7 NEON support.
5374 * Support for SPARC Architecture 2011 crypto extensions, first
5375 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5376 SHA256/512, MD5, GHASH and modular exponentiation.
5378 *Andy Polyakov, David Miller*
5380 * Accelerated modular exponentiation for Intel processors, a.k.a.
5383 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5385 * Support for new and upcoming Intel processors, including AVX2,
5386 BMI and SHA ISA extensions. This includes additional "stitched"
5387 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5390 This work was sponsored by Intel Corp.
5394 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5395 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5396 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5400 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5401 this fixes a limitation in previous versions of OpenSSL.
5405 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5406 MGF1 digest and OAEP label.
5410 * Add EVP support for key wrapping algorithms, to avoid problems with
5411 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5412 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5413 algorithms and include tests cases.
5417 * Add functions to allocate and set the fields of an ECDSA_METHOD
5420 *Douglas E. Engert, Steve Henson*
5422 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5423 difference in days and seconds between two tm or ASN1_TIME structures.
5427 * Add -rev test option to s_server to just reverse order of characters
5428 received by client and send back to server. Also prints an abbreviated
5429 summary of the connection parameters.
5433 * New option -brief for s_client and s_server to print out a brief summary
5434 of connection parameters.
5438 * Add callbacks for arbitrary TLS extensions.
5440 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5442 * New option -crl_download in several openssl utilities to download CRLs
5443 from CRLDP extension in certificates.
5447 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5451 * New function X509_CRL_diff to generate a delta CRL from the difference
5452 of two full CRLs. Add support to "crl" utility.
5456 * New functions to set lookup_crls function and to retrieve
5457 X509_STORE from X509_STORE_CTX.
5461 * Print out deprecated issuer and subject unique ID fields in
5466 * Extend OCSP I/O functions so they can be used for simple general purpose
5467 HTTP as well as OCSP. New wrapper function which can be used to download
5468 CRLs using the OCSP API.
5472 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5476 * `SSL_CONF*` functions. These provide a common framework for application
5477 configuration using configuration files or command lines.
5481 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5482 message callback and prints the results. Needs compile time option
5483 "enable-ssl-trace". New options to s_client and s_server to enable
5488 * New ctrl and macro to retrieve supported points extensions.
5489 Print out extension in s_server and s_client.
5493 * New functions to retrieve certificate signature and signature
5498 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5503 * New Suite B modes for TLS code. These use and enforce the requirements
5504 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5505 only use Suite B curves. The Suite B modes can be set by using the
5506 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5510 * New chain verification flags for Suite B levels of security. Check
5511 algorithms are acceptable when flags are set in X509_verify_cert.
5515 * Make tls1_check_chain return a set of flags indicating checks passed
5516 by a certificate chain. Add additional tests to handle client
5517 certificates: checks for matching certificate type and issuer name
5522 * If an attempt is made to use a signature algorithm not in the peer
5523 preference list abort the handshake. If client has no suitable
5524 signature algorithms in response to a certificate request do not
5525 use the certificate.
5529 * If server EC tmp key is not in client preference list abort handshake.
5533 * Add support for certificate stores in CERT structure. This makes it
5534 possible to have different stores per SSL structure or one store in
5535 the parent SSL_CTX. Include distinct stores for certificate chain
5536 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5537 to build and store a certificate chain in CERT structure: returning
5538 an error if the chain cannot be built: this will allow applications
5539 to test if a chain is correctly configured.
5541 Note: if the CERT based stores are not set then the parent SSL_CTX
5542 store is used to retain compatibility with existing behaviour.
5546 * New function ssl_set_client_disabled to set a ciphersuite disabled
5547 mask based on the current session, check mask when sending client
5548 hello and checking the requested ciphersuite.
5552 * New ctrls to retrieve and set certificate types in a certificate
5553 request message. Print out received values in s_client. If certificate
5554 types is not set with custom values set sensible values based on
5555 supported signature algorithms.
5559 * Support for distinct client and server supported signature algorithms.
5563 * Add certificate callback. If set this is called whenever a certificate
5564 is required by client or server. An application can decide which
5565 certificate chain to present based on arbitrary criteria: for example
5566 supported signature algorithms. Add very simple example to s_server.
5567 This fixes many of the problems and restrictions of the existing client
5568 certificate callback: for example you can now clear an existing
5569 certificate and specify the whole chain.
5573 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5574 the certificate can be used for (if anything). Set valid_flags field
5575 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5576 to have similar checks in it.
5578 Add new "cert_flags" field to CERT structure and include a "strict mode".
5579 This enforces some TLS certificate requirements (such as only permitting
5580 certificate signature algorithms contained in the supported algorithms
5581 extension) which some implementations ignore: this option should be used
5582 with caution as it could cause interoperability issues.
5586 * Update and tidy signature algorithm extension processing. Work out
5587 shared signature algorithms based on preferences and peer algorithms
5588 and print them out in s_client and s_server. Abort handshake if no
5589 shared signature algorithms.
5593 * Add new functions to allow customised supported signature algorithms
5594 for SSL and SSL_CTX structures. Add options to s_client and s_server
5599 * New function SSL_certs_clear() to delete all references to certificates
5600 from an SSL structure. Before this once a certificate had been added
5601 it couldn't be removed.
5605 * Integrate hostname, email address and IP address checking with certificate
5606 verification. New verify options supporting checking in openssl utility.
5610 * Fixes and wildcard matching support to hostname and email checking
5611 functions. Add manual page.
5613 *Florian Weimer (Red Hat Product Security Team)*
5615 * New functions to check a hostname email or IP address against a
5616 certificate. Add options x509 utility to print results of checks against
5621 * Fix OCSP checking.
5623 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5625 * Initial experimental support for explicitly trusted non-root CAs.
5626 OpenSSL still tries to build a complete chain to a root but if an
5627 intermediate CA has a trust setting included that is used. The first
5628 setting is used: whether to trust (e.g., -addtrust option to the x509
5633 * Add -trusted_first option which attempts to find certificates in the
5634 trusted store even if an untrusted chain is also supplied.
5638 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5639 platform support for Linux and Android.
5643 * Support for linux-x32, ILP32 environment in x86_64 framework.
5647 * Experimental multi-implementation support for FIPS capable OpenSSL.
5648 When in FIPS mode the approved implementations are used as normal,
5649 when not in FIPS mode the internal unapproved versions are used instead.
5650 This means that the FIPS capable OpenSSL isn't forced to use the
5651 (often lower performance) FIPS implementations outside FIPS mode.
5655 * Transparently support X9.42 DH parameters when calling
5656 PEM_read_bio_DHparameters. This means existing applications can handle
5657 the new parameter format automatically.
5661 * Initial experimental support for X9.42 DH parameter format: mainly
5662 to support use of 'q' parameter for RFC5114 parameters.
5666 * Add DH parameters from RFC5114 including test data to dhtest.
5670 * Support for automatic EC temporary key parameter selection. If enabled
5671 the most preferred EC parameters are automatically used instead of
5672 hardcoded fixed parameters. Now a server just has to call:
5673 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5674 support ECDH and use the most appropriate parameters.
5678 * Enhance and tidy EC curve and point format TLS extension code. Use
5679 static structures instead of allocation if default values are used.
5680 New ctrls to set curves we wish to support and to retrieve shared curves.
5681 Print out shared curves in s_server. New options to s_server and s_client
5682 to set list of supported curves.
5686 * New ctrls to retrieve supported signature algorithms and
5687 supported curve values as an array of NIDs. Extend openssl utility
5688 to print out received values.
5692 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5693 between NIDs and the more common NIST names such as "P-256". Enhance
5694 ecparam utility and ECC method to recognise the NIST names for curves.
5698 * Enhance SSL/TLS certificate chain handling to support different
5699 chains for each certificate instead of one chain in the parent SSL_CTX.
5703 * Support for fixed DH ciphersuite client authentication: where both
5704 server and client use DH certificates with common parameters.
5708 * Support for fixed DH ciphersuites: those requiring DH server
5713 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5715 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5716 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5717 X509_CINF_get_signature were reverted post internal team review.
5722 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5724 * OCSP Status Request extension unbounded memory growth
5726 A malicious client can send an excessively large OCSP Status Request
5727 extension. If that client continually requests renegotiation, sending a
5728 large OCSP Status Request extension each time, then there will be unbounded
5729 memory growth on the server. This will eventually lead to a Denial Of
5730 Service attack through memory exhaustion. Servers with a default
5731 configuration are vulnerable even if they do not support OCSP. Builds using
5732 the "no-ocsp" build time option are not affected.
5734 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5739 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5742 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5748 * OOB write in MDC2_Update()
5750 An overflow can occur in MDC2_Update() either if called directly or
5751 through the EVP_DigestUpdate() function using MDC2. If an attacker
5752 is able to supply very large amounts of input data after a previous
5753 call to EVP_EncryptUpdate() with a partial block then a length check
5754 can overflow resulting in a heap corruption.
5756 The amount of data needed is comparable to SIZE_MAX which is impractical
5759 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5764 * Malformed SHA512 ticket DoS
5766 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5767 DoS attack where a malformed ticket will result in an OOB read which will
5770 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5771 a custom server callback and ticket lookup mechanism.
5773 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5778 * OOB write in BN_bn2dec()
5780 The function BN_bn2dec() does not check the return value of BN_div_word().
5781 This can cause an OOB write if an application uses this function with an
5782 overly large BIGNUM. This could be a problem if an overly large certificate
5783 or CRL is printed out from an untrusted source. TLS is not affected because
5784 record limits will reject an oversized certificate before it is parsed.
5786 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5791 * OOB read in TS_OBJ_print_bio()
5793 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5794 the total length the OID text representation would use and not the amount
5795 of data written. This will result in OOB reads when large OIDs are
5798 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5803 * Pointer arithmetic undefined behaviour
5805 Avoid some undefined pointer arithmetic
5807 A common idiom in the codebase is to check limits in the following manner:
5810 Where "p" points to some malloc'd data of SIZE bytes and
5813 "len" here could be from some externally supplied data (e.g. from a TLS
5816 The rules of C pointer arithmetic are such that "p + len" is only well
5817 defined where len <= SIZE. Therefore the above idiom is actually
5818 undefined behaviour.
5820 For example this could cause problems if some malloc implementation
5821 provides an address for "p" such that "p + len" actually overflows for
5822 values of len that are too big and therefore p + len < limit.
5824 This issue was reported to OpenSSL by Guido Vranken
5829 * Constant time flag not preserved in DSA signing
5831 Operations in the DSA signing algorithm should run in constant time in
5832 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5833 implementation means that a non-constant time codepath is followed for
5834 certain operations. This has been demonstrated through a cache-timing
5835 attack to be sufficient for an attacker to recover the private DSA key.
5837 This issue was reported by César Pereida (Aalto University), Billy Brumley
5838 (Tampere University of Technology), and Yuval Yarom (The University of
5839 Adelaide and NICTA).
5844 * DTLS buffered message DoS
5846 In a DTLS connection where handshake messages are delivered out-of-order
5847 those messages that OpenSSL is not yet ready to process will be buffered
5848 for later use. Under certain circumstances, a flaw in the logic means that
5849 those messages do not get removed from the buffer even though the handshake
5850 has been completed. An attacker could force up to approx. 15 messages to
5851 remain in the buffer when they are no longer required. These messages will
5852 be cleared when the DTLS connection is closed. The default maximum size for
5853 a message is 100k. Therefore the attacker could force an additional 1500k
5854 to be consumed per connection. By opening many simulataneous connections an
5855 attacker could cause a DoS attack through memory exhaustion.
5857 This issue was reported to OpenSSL by Quan Luo.
5862 * DTLS replay protection DoS
5864 A flaw in the DTLS replay attack protection mechanism means that records
5865 that arrive for future epochs update the replay protection "window" before
5866 the MAC for the record has been validated. This could be exploited by an
5867 attacker by sending a record for the next epoch (which does not have to
5868 decrypt or have a valid MAC), with a very large sequence number. This means
5869 that all subsequent legitimate packets are dropped causing a denial of
5870 service for a specific DTLS connection.
5872 This issue was reported to OpenSSL by the OCAP audit team.
5877 * Certificate message OOB reads
5879 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5880 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5881 theoretical DoS risk but this has not been observed in practice on common
5884 The messages affected are client certificate, client certificate request
5885 and server certificate. As a result the attack can only be performed
5886 against a client or a server which enables client authentication.
5888 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5893 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5895 * Prevent padding oracle in AES-NI CBC MAC check
5897 A MITM attacker can use a padding oracle attack to decrypt traffic
5898 when the connection uses an AES CBC cipher and the server support
5901 This issue was introduced as part of the fix for Lucky 13 padding
5902 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5903 constant time by making sure that always the same bytes are read and
5904 compared against either the MAC or padding bytes. But it no longer
5905 checked that there was enough data to have both the MAC and padding
5908 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5913 * Fix EVP_EncodeUpdate overflow
5915 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5916 Base64 encoding of binary data. If an attacker is able to supply very large
5917 amounts of input data then a length check can overflow resulting in a heap
5920 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5921 the `PEM_write_bio*` family of functions. These are mainly used within the
5922 OpenSSL command line applications, so any application which processes data
5923 from an untrusted source and outputs it as a PEM file should be considered
5924 vulnerable to this issue. User applications that call these APIs directly
5925 with large amounts of untrusted data may also be vulnerable.
5927 This issue was reported by Guido Vranken.
5932 * Fix EVP_EncryptUpdate overflow
5934 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5935 is able to supply very large amounts of input data after a previous call to
5936 EVP_EncryptUpdate() with a partial block then a length check can overflow
5937 resulting in a heap corruption. Following an analysis of all OpenSSL
5938 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5939 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5940 the first called function after an EVP_EncryptInit(), and therefore that
5941 specific call must be safe. The second form is where the length passed to
5942 EVP_EncryptUpdate() can be seen from the code to be some small value and
5943 therefore there is no possibility of an overflow. Since all instances are
5944 one of these two forms, it is believed that there can be no overflows in
5945 internal code due to this problem. It should be noted that
5946 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5947 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5948 of these calls have also been analysed too and it is believed there are no
5949 instances in internal usage where an overflow could occur.
5951 This issue was reported by Guido Vranken.
5956 * Prevent ASN.1 BIO excessive memory allocation
5958 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5959 a short invalid encoding can casuse allocation of large amounts of memory
5960 potentially consuming excessive resources or exhausting memory.
5962 Any application parsing untrusted data through d2i BIO functions is
5963 affected. The memory based functions such as d2i_X509() are *not* affected.
5964 Since the memory based functions are used by the TLS library, TLS
5965 applications are not affected.
5967 This issue was reported by Brian Carpenter.
5974 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5975 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5976 in arbitrary stack data being returned in the buffer.
5978 This issue was reported by Guido Vranken.
5983 * Modify behavior of ALPN to invoke callback after SNI/servername
5984 callback, such that updates to the SSL_CTX affect ALPN.
5988 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5993 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5994 methods are enabled and ssl2 is disabled the methods return NULL.
5998 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
6000 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
6001 Builds that are not configured with "enable-weak-ssl-ciphers" will not
6002 provide any "EXPORT" or "LOW" strength ciphers.
6006 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
6007 is by default disabled at build-time. Builds that are not configured with
6008 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
6009 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
6010 will need to explicitly call either of:
6012 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
6014 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
6016 as appropriate. Even if either of those is used, or the application
6017 explicitly uses the version-specific SSLv2_method() or its client and
6018 server variants, SSLv2 ciphers vulnerable to exhaustive search key
6019 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
6020 ciphers, and SSLv2 56-bit DES are no longer available.
6025 * Fix a double-free in DSA code
6027 A double free bug was discovered when OpenSSL parses malformed DSA private
6028 keys and could lead to a DoS attack or memory corruption for applications
6029 that receive DSA private keys from untrusted sources. This scenario is
6032 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
6038 * Disable SRP fake user seed to address a server memory leak.
6040 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
6042 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
6043 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
6044 was changed to ignore the "fake user" SRP seed, even if the seed
6047 Users should use SRP_VBASE_get1_by_user instead. Note that in
6048 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6049 also that even though configuring the SRP seed attempts to hide
6050 invalid usernames by continuing the handshake with fake
6051 credentials, this behaviour is not constant time and no strong
6052 guarantees are made that the handshake is indistinguishable from
6053 that of a valid user.
6058 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6060 In the BN_hex2bn function the number of hex digits is calculated using an
6061 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6062 large values of `i` this can result in `bn_expand` not allocating any
6063 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6064 field as NULL leading to a subsequent NULL ptr deref. For very large values
6065 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6066 In this case memory is allocated to the internal BIGNUM data field, but it
6067 is insufficiently sized leading to heap corruption. A similar issue exists
6068 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6069 is ever called by user applications with very large untrusted hex/dec data.
6070 This is anticipated to be a rare occurrence.
6072 All OpenSSL internal usage of these functions use data that is not expected
6073 to be untrusted, e.g. config file data or application command line
6074 arguments. If user developed applications generate config file data based
6075 on untrusted data then it is possible that this could also lead to security
6076 consequences. This is also anticipated to be rare.
6078 This issue was reported to OpenSSL by Guido Vranken.
6083 * Fix memory issues in `BIO_*printf` functions
6085 The internal `fmtstr` function used in processing a "%s" format string in
6086 the `BIO_*printf` functions could overflow while calculating the length of a
6087 string and cause an OOB read when printing very long strings.
6089 Additionally the internal `doapr_outch` function can attempt to write to an
6090 OOB memory location (at an offset from the NULL pointer) in the event of a
6091 memory allocation failure. In 1.0.2 and below this could be caused where
6092 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6093 could be in processing a very long "%s" format string. Memory leaks can
6096 The first issue may mask the second issue dependent on compiler behaviour.
6097 These problems could enable attacks where large amounts of untrusted data
6098 is passed to the `BIO_*printf` functions. If applications use these functions
6099 in this way then they could be vulnerable. OpenSSL itself uses these
6100 functions when printing out human-readable dumps of ASN.1 data. Therefore
6101 applications that print this data could be vulnerable if the data is from
6102 untrusted sources. OpenSSL command line applications could also be
6103 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6104 as command line arguments.
6106 Libssl is not considered directly vulnerable. Additionally certificates etc
6107 received via remote connections via libssl are also unlikely to be able to
6108 trigger these issues because of message size limits enforced within libssl.
6110 This issue was reported to OpenSSL Guido Vranken.
6115 * Side channel attack on modular exponentiation
6117 A side-channel attack was found which makes use of cache-bank conflicts on
6118 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6119 of RSA keys. The ability to exploit this issue is limited as it relies on
6120 an attacker who has control of code in a thread running on the same
6121 hyper-threaded core as the victim thread which is performing decryptions.
6123 This issue was reported to OpenSSL by Yuval Yarom, The University of
6124 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6125 Nadia Heninger, University of Pennsylvania with more information at
6126 <http://cachebleed.info>.
6131 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6132 if no keysize is specified with default_bits. This fixes an
6133 omission in an earlier change that changed all RSA/DSA key generation
6134 commands to use 2048 bits by default.
6138 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6140 * Protection for DH small subgroup attacks
6142 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6143 switched on by default and cannot be disabled. This could have some
6148 * SSLv2 doesn't block disabled ciphers
6150 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6151 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6152 been disabled, provided that the SSLv2 protocol was not also disabled via
6155 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6156 and Sebastian Schinzel.
6161 * Reject DH handshakes with parameters shorter than 1024 bits.
6165 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6167 * Certificate verify crash with missing PSS parameter
6169 The signature verification routines will crash with a NULL pointer
6170 dereference if presented with an ASN.1 signature using the RSA PSS
6171 algorithm and absent mask generation function parameter. Since these
6172 routines are used to verify certificate signature algorithms this can be
6173 used to crash any certificate verification operation and exploited in a
6174 DoS attack. Any application which performs certificate verification is
6175 vulnerable including OpenSSL clients and servers which enable client
6178 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6183 * X509_ATTRIBUTE memory leak
6185 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6186 memory. This structure is used by the PKCS#7 and CMS routines so any
6187 application which reads PKCS#7 or CMS data from untrusted sources is
6188 affected. SSL/TLS is not affected.
6190 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6196 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6197 This changes the decoding behaviour for some invalid messages,
6198 though the change is mostly in the more lenient direction, and
6199 legacy behaviour is preserved as much as possible.
6203 * In DSA_generate_parameters_ex, if the provided seed is too short,
6204 use a random seed, as already documented.
6206 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6208 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6210 * Alternate chains certificate forgery
6212 During certificate verfification, OpenSSL will attempt to find an
6213 alternative certificate chain if the first attempt to build such a chain
6214 fails. An error in the implementation of this logic can mean that an
6215 attacker could cause certain checks on untrusted certificates to be
6216 bypassed, such as the CA flag, enabling them to use a valid leaf
6217 certificate to act as a CA and "issue" an invalid certificate.
6219 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6225 * Race condition handling PSK identify hint
6227 If PSK identity hints are received by a multi-threaded client then
6228 the values are wrongly updated in the parent SSL_CTX structure. This can
6229 result in a race condition potentially leading to a double free of the
6235 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6237 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6238 incompatibility in the handling of HMAC. The previous ABI has now been
6241 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6243 * Malformed ECParameters causes infinite loop
6245 When processing an ECParameters structure OpenSSL enters an infinite loop
6246 if the curve specified is over a specially malformed binary polynomial
6249 This can be used to perform denial of service against any
6250 system which processes public keys, certificate requests or
6251 certificates. This includes TLS clients and TLS servers with
6252 client authentication enabled.
6254 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6259 * Exploitable out-of-bounds read in X509_cmp_time
6261 X509_cmp_time does not properly check the length of the ASN1_TIME
6262 string and can read a few bytes out of bounds. In addition,
6263 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6266 An attacker can use this to craft malformed certificates and CRLs of
6267 various sizes and potentially cause a segmentation fault, resulting in
6268 a DoS on applications that verify certificates or CRLs. TLS clients
6269 that verify CRLs are affected. TLS clients and servers with client
6270 authentication enabled may be affected if they use custom verification
6273 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6274 independently by Hanno Böck.
6279 * PKCS7 crash with missing EnvelopedContent
6281 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6282 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6283 with missing content and trigger a NULL pointer dereference on parsing.
6285 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6286 structures from untrusted sources are affected. OpenSSL clients and
6287 servers are not affected.
6289 This issue was reported to OpenSSL by Michal Zalewski (Google).
6294 * CMS verify infinite loop with unknown hash function
6296 When verifying a signedData message the CMS code can enter an infinite loop
6297 if presented with an unknown hash function OID. This can be used to perform
6298 denial of service against any system which verifies signedData messages using
6300 This issue was reported to OpenSSL by Johannes Bauer.
6305 * Race condition handling NewSessionTicket
6307 If a NewSessionTicket is received by a multi-threaded client when attempting to
6308 reuse a previous ticket then a race condition can occur potentially leading to
6309 a double free of the ticket data.
6314 * Reject DH handshakes with parameters shorter than 768 bits.
6316 *Kurt Roeckx and Emilia Kasper*
6318 * dhparam: generate 2048-bit parameters by default.
6320 *Kurt Roeckx and Emilia Kasper*
6322 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6324 * Segmentation fault in ASN1_TYPE_cmp fix
6326 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6327 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6328 certificate signature algorithm consistency this can be used to crash any
6329 certificate verification operation and exploited in a DoS attack. Any
6330 application which performs certificate verification is vulnerable including
6331 OpenSSL clients and servers which enable client authentication.
6336 * ASN.1 structure reuse memory corruption fix
6338 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6339 memory corruption via an invalid write. Such reuse is and has been
6340 strongly discouraged and is believed to be rare.
6342 Applications that parse structures containing CHOICE or ANY DEFINED BY
6343 components may be affected. Certificate parsing (d2i_X509 and related
6344 functions) are however not affected. OpenSSL clients and servers are
6350 * PKCS7 NULL pointer dereferences fix
6352 The PKCS#7 parsing code does not handle missing outer ContentInfo
6353 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6354 missing content and trigger a NULL pointer dereference on parsing.
6356 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6357 otherwise parse PKCS#7 structures from untrusted sources are
6358 affected. OpenSSL clients and servers are not affected.
6360 This issue was reported to OpenSSL by Michal Zalewski (Google).
6365 * DoS via reachable assert in SSLv2 servers fix
6367 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6368 servers that both support SSLv2 and enable export cipher suites by sending
6369 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6371 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6372 (OpenSSL development team).
6377 * Use After Free following d2i_ECPrivatekey error fix
6379 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6380 could cause a use after free condition. This, in turn, could cause a double
6381 free in several private key parsing functions (such as d2i_PrivateKey
6382 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6383 for applications that receive EC private keys from untrusted
6384 sources. This scenario is considered rare.
6386 This issue was discovered by the BoringSSL project and fixed in their
6392 * X509_to_X509_REQ NULL pointer deref fix
6394 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6395 the certificate key is invalid. This function is rarely used in practice.
6397 This issue was discovered by Brian Carpenter.
6402 * Removed the export ciphers from the DEFAULT ciphers
6406 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6408 * Build fixes for the Windows and OpenVMS platforms
6410 *Matt Caswell and Richard Levitte*
6412 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6414 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6415 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6416 dereference. This could lead to a Denial Of Service attack. Thanks to
6417 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6422 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6423 dtls1_buffer_record function under certain conditions. In particular this
6424 could occur if an attacker sent repeated DTLS records with the same
6425 sequence number but for the next epoch. The memory leak could be exploited
6426 by an attacker in a Denial of Service attack through memory exhaustion.
6427 Thanks to Chris Mueller for reporting this issue.
6432 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6433 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6434 method would be set to NULL which could later result in a NULL pointer
6435 dereference. Thanks to Frank Schmirler for reporting this issue.
6440 * Abort handshake if server key exchange message is omitted for ephemeral
6443 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6444 reporting this issue.
6449 * Remove non-export ephemeral RSA code on client and server. This code
6450 violated the TLS standard by allowing the use of temporary RSA keys in
6451 non-export ciphersuites and could be used by a server to effectively
6452 downgrade the RSA key length used to a value smaller than the server
6453 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6454 INRIA or reporting this issue.
6459 * Fixed issue where DH client certificates are accepted without verification.
6460 An OpenSSL server will accept a DH certificate for client authentication
6461 without the certificate verify message. This effectively allows a client to
6462 authenticate without the use of a private key. This only affects servers
6463 which trust a client certificate authority which issues certificates
6464 containing DH keys: these are extremely rare and hardly ever encountered.
6465 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6471 * Ensure that the session ID context of an SSL is updated when its
6472 SSL_CTX is updated via SSL_set_SSL_CTX.
6474 The session ID context is typically set from the parent SSL_CTX,
6475 and can vary with the CTX.
6479 * Fix various certificate fingerprint issues.
6481 By using non-DER or invalid encodings outside the signed portion of a
6482 certificate the fingerprint can be changed without breaking the signature.
6483 Although no details of the signed portion of the certificate can be changed
6484 this can cause problems with some applications: e.g. those using the
6485 certificate fingerprint for blacklists.
6487 1. Reject signatures with non zero unused bits.
6489 If the BIT STRING containing the signature has non zero unused bits reject
6490 the signature. All current signature algorithms require zero unused bits.
6492 2. Check certificate algorithm consistency.
6494 Check the AlgorithmIdentifier inside TBS matches the one in the
6495 certificate signature. NB: this will result in signature failure
6496 errors for some broken certificates.
6498 Thanks to Konrad Kraszewski from Google for reporting this issue.
6500 3. Check DSA/ECDSA signatures use DER.
6502 Re-encode DSA/ECDSA signatures and compare with the original received
6503 signature. Return an error if there is a mismatch.
6505 This will reject various cases including garbage after signature
6506 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6507 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6508 (negative or with leading zeroes).
6510 Further analysis was conducted and fixes were developed by Stephen Henson
6511 of the OpenSSL core team.
6517 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6518 results on some platforms, including x86_64. This bug occurs at random
6519 with a very low probability, and is not known to be exploitable in any
6520 way, though its exact impact is difficult to determine. Thanks to Pieter
6521 Wuille (Blockstream) who reported this issue and also suggested an initial
6522 fix. Further analysis was conducted by the OpenSSL development team and
6523 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6524 the OpenSSL core team.
6529 * Do not resume sessions on the server if the negotiated protocol
6530 version does not match the session's version. Resuming with a different
6531 version, while not strictly forbidden by the RFC, is of questionable
6532 sanity and breaks all known clients.
6534 *David Benjamin, Emilia Käsper*
6536 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6537 early CCS messages during renegotiation. (Note that because
6538 renegotiation is encrypted, this early CCS was not exploitable.)
6542 * Tighten client-side session ticket handling during renegotiation:
6543 ensure that the client only accepts a session ticket if the server sends
6544 the extension anew in the ServerHello. Previously, a TLS client would
6545 reuse the old extension state and thus accept a session ticket if one was
6546 announced in the initial ServerHello.
6548 Similarly, ensure that the client requires a session ticket if one
6549 was advertised in the ServerHello. Previously, a TLS client would
6550 ignore a missing NewSessionTicket message.
6554 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6558 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6559 sends a carefully crafted handshake message, to cause OpenSSL to fail
6560 to free up to 64k of memory causing a memory leak. This could be
6561 exploited in a Denial Of Service attack. This issue affects OpenSSL
6562 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6563 whether SRTP is used or configured. Implementations of OpenSSL that
6564 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6566 The fix was developed by the OpenSSL team.
6571 * Session Ticket Memory Leak.
6573 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6574 integrity of that ticket is first verified. In the event of a session
6575 ticket integrity check failing, OpenSSL will fail to free memory
6576 causing a memory leak. By sending a large number of invalid session
6577 tickets an attacker could exploit this issue in a Denial Of Service
6583 * Build option no-ssl3 is incomplete.
6585 When OpenSSL is configured with "no-ssl3" as a build option, servers
6586 could accept and complete a SSL 3.0 handshake, and clients could be
6587 configured to send them.
6590 *Akamai and the OpenSSL team*
6592 * Add support for TLS_FALLBACK_SCSV.
6593 Client applications doing fallback retries should call
6594 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6597 *Adam Langley, Bodo Moeller*
6599 * Add additional DigestInfo checks.
6601 Re-encode DigestInto in DER and check against the original when
6602 verifying RSA signature: this will reject any improperly encoded
6603 DigestInfo structures.
6605 Note: this is a precautionary measure and no attacks are currently known.
6609 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6611 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6612 SRP code can be overrun an internal buffer. Add sanity check that
6613 g, A, B < N to SRP code.
6615 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6616 Group for discovering this issue.
6621 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6622 TLS 1.0 instead of higher protocol versions when the ClientHello message
6623 is badly fragmented. This allows a man-in-the-middle attacker to force a
6624 downgrade to TLS 1.0 even if both the server and the client support a
6625 higher protocol version, by modifying the client's TLS records.
6627 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6628 researching this issue.
6633 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6634 to a denial of service attack. A malicious server can crash the client
6635 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6636 ciphersuite and sending carefully crafted handshake messages.
6638 Thanks to Felix Gröbert (Google) for discovering and researching this
6644 * By sending carefully crafted DTLS packets an attacker could cause openssl
6645 to leak memory. This can be exploited through a Denial of Service attack.
6646 Thanks to Adam Langley for discovering and researching this issue.
6651 * An attacker can force openssl to consume large amounts of memory whilst
6652 processing DTLS handshake messages. This can be exploited through a
6653 Denial of Service attack.
6654 Thanks to Adam Langley for discovering and researching this issue.
6659 * An attacker can force an error condition which causes openssl to crash
6660 whilst processing DTLS packets due to memory being freed twice. This
6661 can be exploited through a Denial of Service attack.
6662 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6668 * If a multithreaded client connects to a malicious server using a resumed
6669 session and the server sends an ec point format extension it could write
6670 up to 255 bytes to freed memory.
6672 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6678 * A malicious server can crash an OpenSSL client with a null pointer
6679 dereference (read) by specifying an SRP ciphersuite even though it was not
6680 properly negotiated with the client. This can be exploited through a
6681 Denial of Service attack.
6683 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6684 discovering and researching this issue.
6689 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6690 X509_name_oneline, X509_name_print_ex et al. to leak some information
6691 from the stack. Applications may be affected if they echo pretty printing
6692 output to the attacker.
6694 Thanks to Ivan Fratric (Google) for discovering this issue.
6697 *Emilia Käsper, and Steve Henson*
6699 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6700 for corner cases. (Certain input points at infinity could lead to
6701 bogus results, with non-infinity inputs mapped to infinity too.)
6705 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6707 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6708 handshake can force the use of weak keying material in OpenSSL
6709 SSL/TLS clients and servers.
6711 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6712 researching this issue. ([CVE-2014-0224])
6714 *KIKUCHI Masashi, Steve Henson*
6716 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6717 OpenSSL DTLS client the code can be made to recurse eventually crashing
6720 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6723 *Imre Rad, Steve Henson*
6725 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6726 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6727 client or server. This is potentially exploitable to run arbitrary
6728 code on a vulnerable client or server.
6730 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6732 *Jüri Aedla, Steve Henson*
6734 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6735 are subject to a denial of service attack.
6737 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6738 this issue. ([CVE-2014-3470])
6740 *Felix Gröbert, Ivan Fratric, Steve Henson*
6742 * Harmonize version and its documentation. -f flag is used to display
6745 *mancha <mancha1@zoho.com>*
6747 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6748 in i2d_ECPrivateKey.
6750 *mancha <mancha1@zoho.com>*
6752 * Fix some double frees. These are not thought to be exploitable.
6754 *mancha <mancha1@zoho.com>*
6756 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6758 * A missing bounds check in the handling of the TLS heartbeat extension
6759 can be used to reveal up to 64k of memory to a connected client or
6762 Thanks for Neel Mehta of Google Security for discovering this bug and to
6763 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6764 preparing the fix ([CVE-2014-0160])
6766 *Adam Langley, Bodo Moeller*
6768 * Fix for the attack described in the paper "Recovering OpenSSL
6769 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6770 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6771 <http://eprint.iacr.org/2014/140>
6773 Thanks to Yuval Yarom and Naomi Benger for discovering this
6774 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6776 *Yuval Yarom and Naomi Benger*
6778 * TLS pad extension: draft-agl-tls-padding-03
6780 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6781 TLS client Hello record length value would otherwise be > 255 and
6782 less that 512 pad with a dummy extension containing zeroes so it
6783 is at least 512 bytes long.
6785 *Adam Langley, Steve Henson*
6787 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6789 * Fix for TLS record tampering bug. A carefully crafted invalid
6790 handshake could crash OpenSSL with a NULL pointer exception.
6791 Thanks to Anton Johansson for reporting this issues.
6794 * Keep original DTLS digest and encryption contexts in retransmission
6795 structures so we can use the previous session parameters if they need
6796 to be resent. ([CVE-2013-6450])
6800 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6801 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6802 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6803 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6804 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6805 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6807 *Rob Stradling, Adam Langley*
6809 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6811 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6812 supporting platforms or when small records were transferred.
6814 *Andy Polyakov, Steve Henson*
6816 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6818 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6820 This addresses the flaw in CBC record processing discovered by
6821 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6822 at: <http://www.isg.rhul.ac.uk/tls/>
6824 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6825 Security Group at Royal Holloway, University of London
6826 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6827 Emilia Käsper for the initial patch.
6830 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6832 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6833 ciphersuites which can be exploited in a denial of service attack.
6834 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6835 and detecting this bug and to Wolfgang Ettlinger
6836 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6841 * Return an error when checking OCSP signatures when key is NULL.
6842 This fixes a DoS attack. ([CVE-2013-0166])
6846 * Make openssl verify return errors.
6848 *Chris Palmer <palmer@google.com> and Ben Laurie*
6850 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6851 the right response is stapled. Also change SSL_get_certificate()
6852 so it returns the certificate actually sent.
6853 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6855 *Rob Stradling <rob.stradling@comodo.com>*
6857 * Fix possible deadlock when decoding public keys.
6861 * Don't use TLS 1.0 record version number in initial client hello
6866 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6868 * Sanity check record length before skipping explicit IV in TLS
6869 1.2, 1.1 and DTLS to fix DoS attack.
6871 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6872 fuzzing as a service testing platform.
6877 * Initialise tkeylen properly when encrypting CMS messages.
6878 Thanks to Solar Designer of Openwall for reporting this issue.
6882 * In FIPS mode don't try to use composite ciphers as they are not
6887 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6889 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6890 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6891 mean any application compiled against OpenSSL 1.0.0 headers setting
6892 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6893 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6894 0x10000000L Any application which was previously compiled against
6895 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6896 will need to be recompiled as a result. Letting be results in
6897 inability to disable specifically TLS 1.1 and in client context,
6898 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6902 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6903 disable just protocol X, but all protocols above X *if* there are
6904 protocols *below* X still enabled. In more practical terms it means
6905 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6906 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6907 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6912 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6914 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6915 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6916 in CRYPTO_realloc_clean.
6918 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6919 issue and to Adam Langley <agl@chromium.org> for fixing it.
6922 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6924 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6928 * Workarounds for some broken servers that "hang" if a client hello
6929 record length exceeds 255 bytes.
6931 1. Do not use record version number > TLS 1.0 in initial client
6932 hello: some (but not all) hanging servers will now work.
6933 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6934 the number of ciphers sent in the client hello. This should be
6935 set to an even number, such as 50, for example by passing:
6936 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6937 Most broken servers should now work.
6938 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6939 TLS 1.2 client support entirely.
6943 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6947 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6949 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6950 STRING form instead of a DigestInfo.
6954 * The format used for MDC2 RSA signatures is inconsistent between EVP
6955 and the RSA_sign/RSA_verify functions. This was made more apparent when
6956 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6957 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6958 the correct format in RSA_verify so both forms transparently work.
6962 * Some servers which support TLS 1.0 can choke if we initially indicate
6963 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6964 encrypted premaster secret. As a workaround use the maximum permitted
6965 client version in client hello, this should keep such servers happy
6966 and still work with previous versions of OpenSSL.
6970 * Add support for TLS/DTLS heartbeats.
6972 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6974 * Add support for SCTP.
6976 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6978 * Improved PRNG seeding for VOS.
6980 *Paul Green <Paul.Green@stratus.com>*
6982 * Extensive assembler packs updates, most notably:
6984 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6985 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6986 - x86_64: bit-sliced AES implementation;
6987 - ARM: NEON support, contemporary platforms optimizations;
6988 - s390x: z196 support;
6989 - `*`: GHASH and GF(2^m) multiplication implementations;
6993 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6994 (removal of unnecessary code)
6996 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6998 * Add TLS key material exporter from RFC 5705.
7002 * Add DTLS-SRTP negotiation from RFC 5764.
7006 * Add Next Protocol Negotiation,
7007 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
7008 disabled with a no-npn flag to config or Configure. Code donated
7011 *Adam Langley <agl@google.com> and Ben Laurie*
7013 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
7014 NIST-P256, NIST-P521, with constant-time single point multiplication on
7015 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
7016 required to use this (present in gcc 4.4 and later, for 64-bit builds).
7017 Code made available under Apache License version 2.0.
7019 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
7020 line to include this in your build of OpenSSL, and run "make depend" (or
7021 "make update"). This enables the following EC_METHODs:
7023 EC_GFp_nistp224_method()
7024 EC_GFp_nistp256_method()
7025 EC_GFp_nistp521_method()
7027 EC_GROUP_new_by_curve_name() will automatically use these (while
7028 EC_GROUP_new_curve_GFp() currently prefers the more flexible
7031 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
7033 * Use type ossl_ssize_t instead of ssize_t which isn't available on
7034 all platforms. Move ssize_t definition from e_os.h to the public
7035 header file e_os2.h as it now appears in public header file cms.h
7039 * New -sigopt option to the ca, req and x509 utilities. Additional
7040 signature parameters can be passed using this option and in
7045 * Add RSA PSS signing function. This will generate and set the
7046 appropriate AlgorithmIdentifiers for PSS based on those in the
7047 corresponding EVP_MD_CTX structure. No application support yet.
7051 * Support for companion algorithm specific ASN1 signing routines.
7052 New function ASN1_item_sign_ctx() signs a pre-initialised
7053 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7054 the appropriate parameters.
7058 * Add new algorithm specific ASN1 verification initialisation function
7059 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7060 handling will be the same no matter what EVP_PKEY_METHOD is used.
7061 Add a PSS handler to support verification of PSS signatures: checked
7062 against a number of sample certificates.
7066 * Add signature printing for PSS. Add PSS OIDs.
7068 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7070 * Add algorithm specific signature printing. An individual ASN1 method
7071 can now print out signatures instead of the standard hex dump.
7073 More complex signatures (e.g. PSS) can print out more meaningful
7074 information. Include DSA version that prints out the signature
7079 * Password based recipient info support for CMS library: implementing
7084 * Split password based encryption into PBES2 and PBKDF2 functions. This
7085 neatly separates the code into cipher and PBE sections and is required
7086 for some algorithms that split PBES2 into separate pieces (such as
7087 password based CMS).
7091 * Session-handling fixes:
7092 - Fix handling of connections that are resuming with a session ID,
7093 but also support Session Tickets.
7094 - Fix a bug that suppressed issuing of a new ticket if the client
7095 presented a ticket with an expired session.
7096 - Try to set the ticket lifetime hint to something reasonable.
7097 - Make tickets shorter by excluding irrelevant information.
7098 - On the client side, don't ignore renewed tickets.
7100 *Adam Langley, Bodo Moeller (Google)*
7102 * Fix PSK session representation.
7106 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7108 This work was sponsored by Intel.
7112 * Add GCM support to TLS library. Some custom code is needed to split
7113 the IV between the fixed (from PRF) and explicit (from TLS record)
7114 portions. This adds all GCM ciphersuites supported by RFC5288 and
7115 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7116 add a special AESGCM string for GCM only.
7120 * Expand range of ctrls for AES GCM. Permit setting invocation
7121 field on decrypt and retrieval of invocation field only on encrypt.
7125 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7126 As required by RFC5289 these ciphersuites cannot be used if for
7127 versions of TLS earlier than 1.2.
7131 * For FIPS capable OpenSSL interpret a NULL default public key method
7132 as unset and return the appropriate default but do *not* set the default.
7133 This means we can return the appropriate method in applications that
7134 switch between FIPS and non-FIPS modes.
7138 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7139 ENGINE is used then we cannot handle that in the FIPS module so we
7140 keep original code iff non-FIPS operations are allowed.
7144 * Add -attime option to openssl utilities.
7146 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7148 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7152 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7153 FIPS EC methods unconditionally for now.
7157 * New build option no-ec2m to disable characteristic 2 code.
7161 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7162 all cases can be covered as some introduce binary incompatibilities.
7166 * Redirect RSA operations to FIPS module including keygen,
7167 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7171 * Add similar low level API blocking to ciphers.
7175 * Low level digest APIs are not approved in FIPS mode: any attempt
7176 to use these will cause a fatal error. Applications that *really* want
7177 to use them can use the `private_*` version instead.
7181 * Redirect cipher operations to FIPS module for FIPS builds.
7185 * Redirect digest operations to FIPS module for FIPS builds.
7189 * Update build system to add "fips" flag which will link in fipscanister.o
7190 for static and shared library builds embedding a signature if needed.
7194 * Output TLS supported curves in preference order instead of numerical
7195 order. This is currently hardcoded for the highest order curves first.
7196 This should be configurable so applications can judge speed vs strength.
7200 * Add TLS v1.2 server support for client authentication.
7204 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7209 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7210 FIPS modules versions.
7214 * Add TLS v1.2 client side support for client authentication. Keep cache
7215 of handshake records longer as we don't know the hash algorithm to use
7216 until after the certificate request message is received.
7220 * Initial TLS v1.2 client support. Add a default signature algorithms
7221 extension including all the algorithms we support. Parse new signature
7222 format in client key exchange. Relax some ECC signing restrictions for
7223 TLS v1.2 as indicated in RFC5246.
7227 * Add server support for TLS v1.2 signature algorithms extension. Switch
7228 to new signature format when needed using client digest preference.
7229 All server ciphersuites should now work correctly in TLS v1.2. No client
7230 support yet and no support for client certificates.
7234 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7235 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7236 ciphersuites. At present only RSA key exchange ciphersuites work with
7237 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7238 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7239 and version checking.
7243 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7244 with this defined it will not be affected by any changes to ssl internal
7245 structures. Add several utility functions to allow openssl application
7246 to work with OPENSSL_NO_SSL_INTERN defined.
7250 * A long standing patch to add support for SRP from EdelWeb (Peter
7251 Sylvester and Christophe Renou) was integrated.
7252 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7253 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7256 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7260 * Permit abbreviated handshakes when renegotiating using the function
7261 SSL_renegotiate_abbreviated().
7263 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7265 * Add call to ENGINE_register_all_complete() to
7266 ENGINE_load_builtin_engines(), so some implementations get used
7267 automatically instead of needing explicit application support.
7271 * Add support for TLS key exporter as described in RFC5705.
7273 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7275 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7276 a few changes are required:
7278 Add SSL_OP_NO_TLSv1_1 flag.
7279 Add TLSv1_1 methods.
7280 Update version checking logic to handle version 1.1.
7281 Add explicit IV handling (ported from DTLS code).
7282 Add command line options to s_client/s_server.
7289 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7291 * X509_ATTRIBUTE memory leak
7293 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7294 memory. This structure is used by the PKCS#7 and CMS routines so any
7295 application which reads PKCS#7 or CMS data from untrusted sources is
7296 affected. SSL/TLS is not affected.
7298 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7304 * Race condition handling PSK identify hint
7306 If PSK identity hints are received by a multi-threaded client then
7307 the values are wrongly updated in the parent SSL_CTX structure. This can
7308 result in a race condition potentially leading to a double free of the
7314 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7316 * Malformed ECParameters causes infinite loop
7318 When processing an ECParameters structure OpenSSL enters an infinite loop
7319 if the curve specified is over a specially malformed binary polynomial
7322 This can be used to perform denial of service against any
7323 system which processes public keys, certificate requests or
7324 certificates. This includes TLS clients and TLS servers with
7325 client authentication enabled.
7327 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7332 * Exploitable out-of-bounds read in X509_cmp_time
7334 X509_cmp_time does not properly check the length of the ASN1_TIME
7335 string and can read a few bytes out of bounds. In addition,
7336 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7339 An attacker can use this to craft malformed certificates and CRLs of
7340 various sizes and potentially cause a segmentation fault, resulting in
7341 a DoS on applications that verify certificates or CRLs. TLS clients
7342 that verify CRLs are affected. TLS clients and servers with client
7343 authentication enabled may be affected if they use custom verification
7346 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7347 independently by Hanno Böck.
7352 * PKCS7 crash with missing EnvelopedContent
7354 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7355 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7356 with missing content and trigger a NULL pointer dereference on parsing.
7358 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7359 structures from untrusted sources are affected. OpenSSL clients and
7360 servers are not affected.
7362 This issue was reported to OpenSSL by Michal Zalewski (Google).
7367 * CMS verify infinite loop with unknown hash function
7369 When verifying a signedData message the CMS code can enter an infinite loop
7370 if presented with an unknown hash function OID. This can be used to perform
7371 denial of service against any system which verifies signedData messages using
7373 This issue was reported to OpenSSL by Johannes Bauer.
7378 * Race condition handling NewSessionTicket
7380 If a NewSessionTicket is received by a multi-threaded client when attempting to
7381 reuse a previous ticket then a race condition can occur potentially leading to
7382 a double free of the ticket data.
7387 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7389 * Segmentation fault in ASN1_TYPE_cmp fix
7391 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7392 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7393 certificate signature algorithm consistency this can be used to crash any
7394 certificate verification operation and exploited in a DoS attack. Any
7395 application which performs certificate verification is vulnerable including
7396 OpenSSL clients and servers which enable client authentication.
7401 * ASN.1 structure reuse memory corruption fix
7403 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7404 memory corruption via an invalid write. Such reuse is and has been
7405 strongly discouraged and is believed to be rare.
7407 Applications that parse structures containing CHOICE or ANY DEFINED BY
7408 components may be affected. Certificate parsing (d2i_X509 and related
7409 functions) are however not affected. OpenSSL clients and servers are
7415 * PKCS7 NULL pointer dereferences fix
7417 The PKCS#7 parsing code does not handle missing outer ContentInfo
7418 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7419 missing content and trigger a NULL pointer dereference on parsing.
7421 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7422 otherwise parse PKCS#7 structures from untrusted sources are
7423 affected. OpenSSL clients and servers are not affected.
7425 This issue was reported to OpenSSL by Michal Zalewski (Google).
7430 * DoS via reachable assert in SSLv2 servers fix
7432 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7433 servers that both support SSLv2 and enable export cipher suites by sending
7434 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7436 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7437 (OpenSSL development team).
7442 * Use After Free following d2i_ECPrivatekey error fix
7444 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7445 could cause a use after free condition. This, in turn, could cause a double
7446 free in several private key parsing functions (such as d2i_PrivateKey
7447 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7448 for applications that receive EC private keys from untrusted
7449 sources. This scenario is considered rare.
7451 This issue was discovered by the BoringSSL project and fixed in their
7457 * X509_to_X509_REQ NULL pointer deref fix
7459 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7460 the certificate key is invalid. This function is rarely used in practice.
7462 This issue was discovered by Brian Carpenter.
7467 * Removed the export ciphers from the DEFAULT ciphers
7471 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7473 * Build fixes for the Windows and OpenVMS platforms
7475 *Matt Caswell and Richard Levitte*
7477 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7479 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7480 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7481 dereference. This could lead to a Denial Of Service attack. Thanks to
7482 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7487 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7488 dtls1_buffer_record function under certain conditions. In particular this
7489 could occur if an attacker sent repeated DTLS records with the same
7490 sequence number but for the next epoch. The memory leak could be exploited
7491 by an attacker in a Denial of Service attack through memory exhaustion.
7492 Thanks to Chris Mueller for reporting this issue.
7497 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7498 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7499 method would be set to NULL which could later result in a NULL pointer
7500 dereference. Thanks to Frank Schmirler for reporting this issue.
7505 * Abort handshake if server key exchange message is omitted for ephemeral
7508 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7509 reporting this issue.
7514 * Remove non-export ephemeral RSA code on client and server. This code
7515 violated the TLS standard by allowing the use of temporary RSA keys in
7516 non-export ciphersuites and could be used by a server to effectively
7517 downgrade the RSA key length used to a value smaller than the server
7518 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7519 INRIA or reporting this issue.
7524 * Fixed issue where DH client certificates are accepted without verification.
7525 An OpenSSL server will accept a DH certificate for client authentication
7526 without the certificate verify message. This effectively allows a client to
7527 authenticate without the use of a private key. This only affects servers
7528 which trust a client certificate authority which issues certificates
7529 containing DH keys: these are extremely rare and hardly ever encountered.
7530 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7536 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7537 results on some platforms, including x86_64. This bug occurs at random
7538 with a very low probability, and is not known to be exploitable in any
7539 way, though its exact impact is difficult to determine. Thanks to Pieter
7540 Wuille (Blockstream) who reported this issue and also suggested an initial
7541 fix. Further analysis was conducted by the OpenSSL development team and
7542 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7543 the OpenSSL core team.
7548 * Fix various certificate fingerprint issues.
7550 By using non-DER or invalid encodings outside the signed portion of a
7551 certificate the fingerprint can be changed without breaking the signature.
7552 Although no details of the signed portion of the certificate can be changed
7553 this can cause problems with some applications: e.g. those using the
7554 certificate fingerprint for blacklists.
7556 1. Reject signatures with non zero unused bits.
7558 If the BIT STRING containing the signature has non zero unused bits reject
7559 the signature. All current signature algorithms require zero unused bits.
7561 2. Check certificate algorithm consistency.
7563 Check the AlgorithmIdentifier inside TBS matches the one in the
7564 certificate signature. NB: this will result in signature failure
7565 errors for some broken certificates.
7567 Thanks to Konrad Kraszewski from Google for reporting this issue.
7569 3. Check DSA/ECDSA signatures use DER.
7571 Reencode DSA/ECDSA signatures and compare with the original received
7572 signature. Return an error if there is a mismatch.
7574 This will reject various cases including garbage after signature
7575 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7576 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7577 (negative or with leading zeroes).
7579 Further analysis was conducted and fixes were developed by Stephen Henson
7580 of the OpenSSL core team.
7586 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7588 * Session Ticket Memory Leak.
7590 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7591 integrity of that ticket is first verified. In the event of a session
7592 ticket integrity check failing, OpenSSL will fail to free memory
7593 causing a memory leak. By sending a large number of invalid session
7594 tickets an attacker could exploit this issue in a Denial Of Service
7600 * Build option no-ssl3 is incomplete.
7602 When OpenSSL is configured with "no-ssl3" as a build option, servers
7603 could accept and complete a SSL 3.0 handshake, and clients could be
7604 configured to send them.
7607 *Akamai and the OpenSSL team*
7609 * Add support for TLS_FALLBACK_SCSV.
7610 Client applications doing fallback retries should call
7611 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7614 *Adam Langley, Bodo Moeller*
7616 * Add additional DigestInfo checks.
7618 Reencode DigestInto in DER and check against the original when
7619 verifying RSA signature: this will reject any improperly encoded
7620 DigestInfo structures.
7622 Note: this is a precautionary measure and no attacks are currently known.
7626 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7628 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7629 to a denial of service attack. A malicious server can crash the client
7630 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7631 ciphersuite and sending carefully crafted handshake messages.
7633 Thanks to Felix Gröbert (Google) for discovering and researching this
7639 * By sending carefully crafted DTLS packets an attacker could cause openssl
7640 to leak memory. This can be exploited through a Denial of Service attack.
7641 Thanks to Adam Langley for discovering and researching this issue.
7646 * An attacker can force openssl to consume large amounts of memory whilst
7647 processing DTLS handshake messages. This can be exploited through a
7648 Denial of Service attack.
7649 Thanks to Adam Langley for discovering and researching this issue.
7654 * An attacker can force an error condition which causes openssl to crash
7655 whilst processing DTLS packets due to memory being freed twice. This
7656 can be exploited through a Denial of Service attack.
7657 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7663 * If a multithreaded client connects to a malicious server using a resumed
7664 session and the server sends an ec point format extension it could write
7665 up to 255 bytes to freed memory.
7667 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7673 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7674 X509_name_oneline, X509_name_print_ex et al. to leak some information
7675 from the stack. Applications may be affected if they echo pretty printing
7676 output to the attacker.
7678 Thanks to Ivan Fratric (Google) for discovering this issue.
7681 *Emilia Käsper, and Steve Henson*
7683 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7684 for corner cases. (Certain input points at infinity could lead to
7685 bogus results, with non-infinity inputs mapped to infinity too.)
7689 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7691 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7692 handshake can force the use of weak keying material in OpenSSL
7693 SSL/TLS clients and servers.
7695 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7696 researching this issue. ([CVE-2014-0224])
7698 *KIKUCHI Masashi, Steve Henson*
7700 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7701 OpenSSL DTLS client the code can be made to recurse eventually crashing
7704 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7707 *Imre Rad, Steve Henson*
7709 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7710 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7711 client or server. This is potentially exploitable to run arbitrary
7712 code on a vulnerable client or server.
7714 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7716 *Jüri Aedla, Steve Henson*
7718 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7719 are subject to a denial of service attack.
7721 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7722 this issue. ([CVE-2014-3470])
7724 *Felix Gröbert, Ivan Fratric, Steve Henson*
7726 * Harmonize version and its documentation. -f flag is used to display
7729 *mancha <mancha1@zoho.com>*
7731 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7732 in i2d_ECPrivateKey.
7734 *mancha <mancha1@zoho.com>*
7736 * Fix some double frees. These are not thought to be exploitable.
7738 *mancha <mancha1@zoho.com>*
7740 * Fix for the attack described in the paper "Recovering OpenSSL
7741 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7742 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7743 <http://eprint.iacr.org/2014/140>
7745 Thanks to Yuval Yarom and Naomi Benger for discovering this
7746 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7748 *Yuval Yarom and Naomi Benger*
7750 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7752 * Keep original DTLS digest and encryption contexts in retransmission
7753 structures so we can use the previous session parameters if they need
7754 to be resent. ([CVE-2013-6450])
7758 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7759 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7760 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7761 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7762 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7763 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7765 *Rob Stradling, Adam Langley*
7767 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7769 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7771 This addresses the flaw in CBC record processing discovered by
7772 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7773 at: <http://www.isg.rhul.ac.uk/tls/>
7775 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7776 Security Group at Royal Holloway, University of London
7777 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7778 Emilia Käsper for the initial patch.
7781 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7783 * Return an error when checking OCSP signatures when key is NULL.
7784 This fixes a DoS attack. ([CVE-2013-0166])
7788 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7789 the right response is stapled. Also change SSL_get_certificate()
7790 so it returns the certificate actually sent.
7791 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7792 (This is a backport)
7794 *Rob Stradling <rob.stradling@comodo.com>*
7796 * Fix possible deadlock when decoding public keys.
7800 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7802 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7805 * Sanity check record length before skipping explicit IV in DTLS
7808 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7809 fuzzing as a service testing platform.
7814 * Initialise tkeylen properly when encrypting CMS messages.
7815 Thanks to Solar Designer of Openwall for reporting this issue.
7819 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7821 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7822 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7823 in CRYPTO_realloc_clean.
7825 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7826 issue and to Adam Langley <agl@chromium.org> for fixing it.
7829 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7831 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7833 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7834 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7835 content decryption and always return the same error. Note: this attack
7836 needs on average 2^20 messages so it only affects automated senders. The
7837 old behaviour can be re-enabled in the CMS code by setting the
7838 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7839 an MMA defence is not necessary.
7840 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7841 this issue. ([CVE-2012-0884])
7845 * Fix CVE-2011-4619: make sure we really are receiving a
7846 client hello before rejecting multiple SGC restarts. Thanks to
7847 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7851 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7853 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7854 Thanks to Antonio Martin, Enterprise Secure Access Research and
7855 Development, Cisco Systems, Inc. for discovering this bug and
7856 preparing a fix. ([CVE-2012-0050])
7860 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7862 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7863 of the Vaudenay padding oracle attack on CBC mode encryption
7864 which enables an efficient plaintext recovery attack against
7865 the OpenSSL implementation of DTLS. Their attack exploits timing
7866 differences arising during decryption processing. A research
7867 paper describing this attack can be found at:
7868 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7869 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7870 Security Group at Royal Holloway, University of London
7871 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7872 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7873 for preparing the fix. ([CVE-2011-4108])
7875 *Robin Seggelmann, Michael Tuexen*
7877 * Clear bytes used for block padding of SSL 3.0 records.
7880 *Adam Langley (Google)*
7882 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7883 Kadianakis <desnacked@gmail.com> for discovering this issue and
7884 Adam Langley for preparing the fix. ([CVE-2011-4619])
7886 *Adam Langley (Google)*
7888 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7890 *Andrey Kulikov <amdeich@gmail.com>*
7892 * Prevent malformed RFC3779 data triggering an assertion failure.
7893 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7894 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7896 *Rob Austein <sra@hactrn.net>*
7898 * Improved PRNG seeding for VOS.
7900 *Paul Green <Paul.Green@stratus.com>*
7902 * Fix ssl_ciph.c set-up race.
7904 *Adam Langley (Google)*
7906 * Fix spurious failures in ecdsatest.c.
7908 *Emilia Käsper (Google)*
7910 * Fix the BIO_f_buffer() implementation (which was mixing different
7911 interpretations of the `..._len` fields).
7913 *Adam Langley (Google)*
7915 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7916 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7917 threads won't reuse the same blinding coefficients.
7919 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7920 lock to call BN_BLINDING_invert_ex, and avoids one use of
7921 BN_BLINDING_update for each BN_BLINDING structure (previously,
7922 the last update always remained unused).
7924 *Emilia Käsper (Google)*
7926 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7928 *Bob Buckholz (Google)*
7930 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7932 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7933 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7935 *Kaspar Brand <ossl@velox.ch>*
7937 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7938 for multi-threaded use of ECDH. ([CVE-2011-3210])
7940 *Adam Langley (Google)*
7942 * Fix x509_name_ex_d2i memory leak on bad inputs.
7946 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7947 signature public key algorithm by using OID xref utilities instead.
7948 Before this you could only use some ECC ciphersuites with SHA1 only.
7952 * Add protection against ECDSA timing attacks as mentioned in the paper
7953 by Billy Bob Brumley and Nicola Tuveri, see:
7954 <http://eprint.iacr.org/2011/232.pdf>
7956 *Billy Bob Brumley and Nicola Tuveri*
7958 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7960 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7962 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7964 * Fix bug in string printing code: if *any* escaping is enabled we must
7965 escape the escape character (backslash) or the resulting string is
7970 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7972 * Disable code workaround for ancient and obsolete Netscape browsers
7973 and servers: an attacker can use it in a ciphersuite downgrade attack.
7974 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7978 * Fixed J-PAKE implementation error, originally discovered by
7979 Sebastien Martini, further info and confirmation from Stefan
7980 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7984 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7986 * Fix extension code to avoid race conditions which can result in a buffer
7987 overrun vulnerability: resumed sessions must not be modified as they can
7988 be shared by multiple threads. CVE-2010-3864
7992 * Fix WIN32 build system to correctly link an ENGINE directory into
7997 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7999 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
8002 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
8004 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
8006 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
8007 context. The operation can be customised via the ctrl mechanism in
8008 case ENGINEs want to include additional functionality.
8012 * Tolerate yet another broken PKCS#8 key format: private key value negative.
8016 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
8017 output hashes compatible with older versions of OpenSSL.
8019 *Willy Weisz <weisz@vcpc.univie.ac.at>*
8021 * Fix compression algorithm handling: if resuming a session use the
8022 compression algorithm of the resumed session instead of determining
8023 it from client hello again. Don't allow server to change algorithm.
8027 * Add load_crls() function to commands tidying load_certs() too. Add option
8028 to verify utility to allow additional CRLs to be included.
8032 * Update OCSP request code to permit adding custom headers to the request:
8033 some responders need this.
8037 * The function EVP_PKEY_sign() returns <=0 on error: check return code
8040 *Julia Lawall <julia@diku.dk>*
8042 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
8043 needlessly dereferenced structures, used obsolete functions and
8044 didn't handle all updated verify codes correctly.
8048 * Disable MD2 in the default configuration.
8052 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8053 indicate the initial BIO being pushed or popped. This makes it possible
8054 to determine whether the BIO is the one explicitly called or as a result
8055 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8056 it handles reference counts correctly and doesn't zero out the I/O bio
8057 when it is not being explicitly popped. WARNING: applications which
8058 included workarounds for the old buggy behaviour will need to be modified
8059 or they could free up already freed BIOs.
8063 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8064 renaming to all platforms (within the 0.9.8 branch, this was
8065 done conditionally on Netware platforms to avoid a name clash).
8067 *Guenter <lists@gknw.net>*
8069 * Add ECDHE and PSK support to DTLS.
8071 *Michael Tuexen <tuexen@fh-muenster.de>*
8073 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8078 * Add "missing" function EVP_MD_flags() (without this the only way to
8079 retrieve a digest flags is by accessing the structure directly. Update
8080 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8081 or cipher is registered as in the "from" argument. Print out all
8082 registered digests in the dgst usage message instead of manually
8083 attempting to work them out.
8087 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8088 this allows the use of compression and extensions. Change default cipher
8089 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8090 by default unless an application cipher string requests it.
8094 * Alter match criteria in PKCS12_parse(). It used to try to use local
8095 key ids to find matching certificates and keys but some PKCS#12 files
8096 don't follow the (somewhat unwritten) rules and this strategy fails.
8097 Now just gather all certificates together and the first private key
8098 then look for the first certificate that matches the key.
8102 * Support use of registered digest and cipher names for dgst and cipher
8103 commands instead of having to add each one as a special case. So now
8110 openssl dgst -sha256 foo
8112 and this works for ENGINE based algorithms too.
8116 * Update Gost ENGINE to support parameter files.
8118 *Victor B. Wagner <vitus@cryptocom.ru>*
8120 * Support GeneralizedTime in ca utility.
8122 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8124 * Enhance the hash format used for certificate directory links. The new
8125 form uses the canonical encoding (meaning equivalent names will work
8126 even if they aren't identical) and uses SHA1 instead of MD5. This form
8127 is incompatible with the older format and as a result c_rehash should
8128 be used to rebuild symbolic links.
8132 * Make PKCS#8 the default write format for private keys, replacing the
8133 traditional format. This form is standardised, more secure and doesn't
8134 include an implicit MD5 dependency.
8138 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8139 committed to OpenSSL should pass this lot as a minimum.
8143 * Add session ticket override functionality for use by EAP-FAST.
8145 *Jouni Malinen <j@w1.fi>*
8147 * Modify HMAC functions to return a value. Since these can be implemented
8148 in an ENGINE errors can occur.
8152 * Type-checked OBJ_bsearch_ex.
8156 * Type-checked OBJ_bsearch. Also some constification necessitated
8157 by type-checking. Still to come: TXT_DB, bsearch(?),
8158 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8163 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8164 seconds to a tm structure directly, instead of going through OS
8165 specific date routines. This avoids any issues with OS routines such
8166 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8167 and X509_time_adj_ex() to cover the extended range. The existing
8168 X509_time_adj() is still usable and will no longer have any date issues.
8172 * Delta CRL support. New use deltas option which will attempt to locate
8173 and search any appropriate delta CRLs available.
8175 This work was sponsored by Google.
8179 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8180 code and add additional score elements. Validate alternate CRL paths
8181 as part of the CRL checking and indicate a new error "CRL path validation
8182 error" in this case. Applications wanting additional details can use
8183 the verify callback and check the new "parent" field. If this is not
8184 NULL CRL path validation is taking place. Existing applications won't
8185 see this because it requires extended CRL support which is off by
8188 This work was sponsored by Google.
8192 * Support for freshest CRL extension.
8194 This work was sponsored by Google.
8198 * Initial indirect CRL support. Currently only supported in the CRLs
8199 passed directly and not via lookup. Process certificate issuer
8200 CRL entry extension and lookup CRL entries by bother issuer name
8201 and serial number. Check and process CRL issuer entry in IDP extension.
8203 This work was sponsored by Google.
8207 * Add support for distinct certificate and CRL paths. The CRL issuer
8208 certificate is validated separately in this case. Only enabled if
8209 an extended CRL support flag is set: this flag will enable additional
8210 CRL functionality in future.
8212 This work was sponsored by Google.
8216 * Add support for policy mappings extension.
8218 This work was sponsored by Google.
8222 * Fixes to pathlength constraint, self issued certificate handling,
8223 policy processing to align with RFC3280 and PKITS tests.
8225 This work was sponsored by Google.
8229 * Support for name constraints certificate extension. DN, email, DNS
8230 and URI types are currently supported.
8232 This work was sponsored by Google.
8236 * To cater for systems that provide a pointer-based thread ID rather
8237 than numeric, deprecate the current numeric thread ID mechanism and
8238 replace it with a structure and associated callback type. This
8239 mechanism allows a numeric "hash" to be extracted from a thread ID in
8240 either case, and on platforms where pointers are larger than 'long',
8241 mixing is done to help ensure the numeric 'hash' is usable even if it
8242 can't be guaranteed unique. The default mechanism is to use "&errno"
8243 as a pointer-based thread ID to distinguish between threads.
8245 Applications that want to provide their own thread IDs should now use
8246 CRYPTO_THREADID_set_callback() to register a callback that will call
8247 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8249 Note that ERR_remove_state() is now deprecated, because it is tied
8250 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8251 to free the current thread's error state should be replaced by
8252 ERR_remove_thread_state(NULL).
8254 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8255 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8256 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8257 application was previously providing a numeric thread callback that
8258 was inappropriate for distinguishing threads, then uniqueness might
8259 have been obtained with &errno that happened immediately in the
8260 intermediate development versions of OpenSSL; this is no longer the
8261 case, the numeric thread callback will now override the automatic use
8264 *Geoff Thorpe, with help from Bodo Moeller*
8266 * Initial support for different CRL issuing certificates. This covers a
8267 simple case where the self issued certificates in the chain exist and
8268 the real CRL issuer is higher in the existing chain.
8270 This work was sponsored by Google.
8274 * Removed effectively defunct crypto/store from the build.
8278 * Revamp of STACK to provide stronger type-checking. Still to come:
8279 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8280 ASN1_STRING, CONF_VALUE.
8284 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8285 RAM on SSL connections. This option can save about 34k per idle SSL.
8289 * Revamp of LHASH to provide stronger type-checking. Still to come:
8290 STACK, TXT_DB, bsearch, qsort.
8294 * Initial support for Cryptographic Message Syntax (aka CMS) based
8295 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8296 support for data, signedData, compressedData, digestedData and
8297 encryptedData, envelopedData types included. Scripts to check against
8298 RFC4134 examples draft and interop and consistency checks of many
8299 content types and variants.
8303 * Add options to enc utility to support use of zlib compression BIO.
8307 * Extend mk1mf to support importing of options and assembly language
8308 files from Configure script, currently only included in VC-WIN32.
8309 The assembly language rules can now optionally generate the source
8310 files from the associated perl scripts.
8314 * Implement remaining functionality needed to support GOST ciphersuites.
8315 Interop testing has been performed using CryptoPro implementations.
8317 *Victor B. Wagner <vitus@cryptocom.ru>*
8319 * s390x assembler pack.
8323 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8328 * Implement Opaque PRF Input TLS extension as specified in
8329 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8330 official specification yet and no extension type assignment by
8331 IANA exists, this extension (for now) will have to be explicitly
8332 enabled when building OpenSSL by providing the extension number
8333 to use. For example, specify an option
8335 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8337 to the "config" or "Configure" script to enable the extension,
8338 assuming extension number 0x9527 (which is a completely arbitrary
8339 and unofficial assignment based on the MD5 hash of the Internet
8340 Draft). Note that by doing so, you potentially lose
8341 interoperability with other TLS implementations since these might
8342 be using the same extension number for other purposes.
8344 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8345 opaque PRF input value to use in the handshake. This will create
8346 an internal copy of the length-'len' string at 'src', and will
8347 return non-zero for success.
8349 To get more control and flexibility, provide a callback function
8352 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8353 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8357 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8360 Callback function 'cb' will be called in handshakes, and is
8361 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8362 Argument 'arg' is for application purposes (the value as given to
8363 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8364 be provided to the callback function). The callback function
8365 has to return non-zero to report success: usually 1 to use opaque
8366 PRF input just if possible, or 2 to enforce use of the opaque PRF
8367 input. In the latter case, the library will abort the handshake
8368 if opaque PRF input is not successfully negotiated.
8370 Arguments 'peerinput' and 'len' given to the callback function
8371 will always be NULL and 0 in the case of a client. A server will
8372 see the client's opaque PRF input through these variables if
8373 available (NULL and 0 otherwise). Note that if the server
8374 provides an opaque PRF input, the length must be the same as the
8375 length of the client's opaque PRF input.
8377 Note that the callback function will only be called when creating
8378 a new session (session resumption can resume whatever was
8379 previously negotiated), and will not be called in SSL 2.0
8380 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8381 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8382 for applications that need to enforce opaque PRF input.
8386 * Update ssl code to support digests other than SHA1+MD5 for handshake
8389 *Victor B. Wagner <vitus@cryptocom.ru>*
8391 * Add RFC4507 support to OpenSSL. This includes the corrections in
8392 RFC4507bis. The encrypted ticket format is an encrypted encoded
8393 SSL_SESSION structure, that way new session features are automatically
8396 If a client application caches session in an SSL_SESSION structure
8397 support is transparent because tickets are now stored in the encoded
8400 The SSL_CTX structure automatically generates keys for ticket
8401 protection in servers so again support should be possible
8402 with no application modification.
8404 If a client or server wishes to disable RFC4507 support then the option
8405 SSL_OP_NO_TICKET can be set.
8407 Add a TLS extension debugging callback to allow the contents of any client
8408 or server extensions to be examined.
8410 This work was sponsored by Google.
8414 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8415 OpenSSL should now compile cleanly on gcc 4.2
8417 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8419 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8420 support including streaming MAC support: this is required for GOST
8421 ciphersuite support.
8423 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8425 * Add option -stream to use PKCS#7 streaming in smime utility. New
8426 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8427 to output in BER and PEM format.
8431 * Experimental support for use of HMAC via EVP_PKEY interface. This
8432 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8433 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8434 ENGINE support for HMAC keys which are unextractable. New -mac and
8435 -macopt options to dgst utility.
8439 * New option -sigopt to dgst utility. Update dgst to use
8440 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8441 alternative signing parameters such as X9.31 or PSS in the dgst
8446 * Change ssl_cipher_apply_rule(), the internal function that does
8447 the work each time a ciphersuite string requests enabling
8448 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8449 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8450 the order of disabled ciphersuites such that those ciphersuites
8451 that most recently went from enabled to disabled not only stay
8452 in order with respect to each other, but also have higher priority
8453 than other disabled ciphersuites the next time ciphersuites are
8456 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8457 the same ciphersuites as with "HIGH" alone, but in a specific
8458 order where the PSK ciphersuites come first (since they are the
8459 most recently disabled ciphersuites when "HIGH" is parsed).
8461 Also, change ssl_create_cipher_list() (using this new
8462 functionality) such that between otherwise identical
8463 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8468 * Change ssl_create_cipher_list() so that it automatically
8469 arranges the ciphersuites in reasonable order before starting
8470 to process the rule string. Thus, the definition for "DEFAULT"
8471 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8472 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8473 This makes it much easier to arrive at a reasonable default order
8474 in applications for which anonymous ciphers are OK (meaning
8475 that you can't actually use DEFAULT).
8477 *Bodo Moeller; suggested by Victor Duchovni*
8479 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8480 processing) into multiple integers instead of setting
8481 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8482 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8483 (These masks as well as the individual bit definitions are hidden
8484 away into the non-exported interface ssl/ssl_locl.h, so this
8485 change to the definition of the SSL_CIPHER structure shouldn't
8486 affect applications.) This give us more bits for each of these
8487 categories, so there is no longer a need to coagulate AES128 and
8488 AES256 into a single algorithm bit, and to coagulate Camellia128
8489 and Camellia256 into a single algorithm bit, which has led to all
8492 Thus, among other things, the kludge introduced in 0.9.7m and
8493 0.9.8e for masking out AES256 independently of AES128 or masking
8494 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8496 With the change, we also introduce new ciphersuite aliases that
8497 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8502 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8503 Use the leftmost N bytes of the signature input if the input is
8504 larger than the prime q (with N being the size in bytes of q).
8508 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8509 it yet and it is largely untested.
8513 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8517 * Initial incomplete changes to avoid need for function casts in OpenSSL
8518 some compilers (gcc 4.2 and later) reject their use. Safestack is
8519 reimplemented. Update ASN1 to avoid use of legacy functions.
8523 * Win32/64 targets are linked with Winsock2.
8527 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8528 to external functions. This can be used to increase CRL handling
8529 efficiency especially when CRLs are very large by (for example) storing
8530 the CRL revoked certificates in a database.
8534 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8535 new CRLs added to a directory can be used. New command line option
8536 -verify_return_error to s_client and s_server. This causes real errors
8537 to be returned by the verify callback instead of carrying on no matter
8538 what. This reflects the way a "real world" verify callback would behave.
8542 * GOST engine, supporting several GOST algorithms and public key formats.
8543 Kindly donated by Cryptocom.
8547 * Partial support for Issuing Distribution Point CRL extension. CRLs
8548 partitioned by DP are handled but no indirect CRL or reason partitioning
8549 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8550 selected via a scoring technique which handles IDP and AKID in CRLs.
8554 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8555 will ultimately be used for all verify operations: this will remove the
8556 X509_STORE dependency on certificate verification and allow alternative
8557 lookup methods. X509_STORE based implementations of these two callbacks.
8561 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8562 Modify get_crl() to find a valid (unexpired) CRL if possible.
8566 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8567 this would be called X509_CRL_cmp() but that name is already used by
8568 a function that just compares CRL issuer names. Cache several CRL
8569 extensions in X509_CRL structure and cache CRLDP in X509.
8573 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8574 this maps equivalent X509_NAME structures into a consistent structure.
8575 Name comparison can then be performed rapidly using memcmp().
8579 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8584 * Allow digests to supply their own micalg string for S/MIME type using
8585 the ctrl EVP_MD_CTRL_MICALG.
8589 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8590 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8591 ctrl. It can then customise the structure before and/or after signing
8596 * New function OBJ_add_sigid() to allow application defined signature OIDs
8597 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8598 to free up any added signature OIDs.
8602 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8603 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8604 digest and cipher tables. New options added to openssl utility:
8605 list-message-digest-algorithms and list-cipher-algorithms.
8609 * Change the array representation of binary polynomials: the list
8610 of degrees of non-zero coefficients is now terminated with -1.
8611 Previously it was terminated with 0, which was also part of the
8612 value; thus, the array representation was not applicable to
8613 polynomials where t^0 has coefficient zero. This change makes
8614 the array representation useful in a more general context.
8618 * Various modifications and fixes to SSL/TLS cipher string
8619 handling. For ECC, the code now distinguishes between fixed ECDH
8620 with RSA certificates on the one hand and with ECDSA certificates
8621 on the other hand, since these are separate ciphersuites. The
8622 unused code for Fortezza ciphersuites has been removed.
8624 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8625 (not "ECDHE"). For consistency with the code for DH
8626 certificates, use of ECDH certificates is now considered ECDH
8627 authentication, not RSA or ECDSA authentication (the latter is
8628 merely the CA's signing algorithm and not actively used in the
8631 The temporary ciphersuite alias "ECCdraft" is no longer
8632 available, and ECC ciphersuites are no longer excluded from "ALL"
8633 and "DEFAULT". The following aliases now exist for RFC 4492
8634 ciphersuites, most of these by analogy with the DH case:
8636 kECDHr - ECDH cert, signed with RSA
8637 kECDHe - ECDH cert, signed with ECDSA
8638 kECDH - ECDH cert (signed with either RSA or ECDSA)
8639 kEECDH - ephemeral ECDH
8640 ECDH - ECDH cert or ephemeral ECDH
8646 AECDH - anonymous ECDH
8647 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8651 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8652 Use correct micalg parameters depending on digest(s) in signed message.
8656 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8657 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8661 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8662 an engine to register a method. Add ENGINE lookups for methods and
8663 functional reference processing.
8667 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8668 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8673 * New -resign option to smime utility. This adds one or more signers
8674 to an existing PKCS#7 signedData structure. Also -md option to use an
8675 alternative message digest algorithm for signing.
8679 * Tidy up PKCS#7 routines and add new functions to make it easier to
8680 create PKCS7 structures containing multiple signers. Update smime
8681 application to support multiple signers.
8685 * New -macalg option to pkcs12 utility to allow setting of an alternative
8690 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8691 Reorganize PBE internals to lookup from a static table using NIDs,
8692 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8693 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8694 PRF which will be automatically used with PBES2.
8698 * Replace the algorithm specific calls to generate keys in "req" with the
8703 * Update PKCS#7 enveloped data routines to use new API. This is now
8704 supported by any public key method supporting the encrypt operation. A
8705 ctrl is added to allow the public key algorithm to examine or modify
8706 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8711 * Add a ctrl to asn1 method to allow a public key algorithm to express
8712 a default digest type to use. In most cases this will be SHA1 but some
8713 algorithms (such as GOST) need to specify an alternative digest. The
8714 return value indicates how strong the preference is 1 means optional and
8715 2 is mandatory (that is it is the only supported type). Modify
8716 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8717 use the default md. Update openssl utilities to use the default digest
8718 type for signing if it is not explicitly indicated.
8722 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8723 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8724 signing method from the key type. This effectively removes the link
8725 between digests and public key types.
8729 * Add an OID cross reference table and utility functions. Its purpose is to
8730 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8731 rsaEncryption. This will allow some of the algorithm specific hackery
8732 needed to use the correct OID to be removed.
8736 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8737 structures for PKCS7_sign(). They are now set up by the relevant public
8742 * Add provisional EC pkey method with support for ECDSA and ECDH.
8746 * Add support for key derivation (agreement) in the API, DH method and
8751 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8752 public and private key formats. As a side effect these add additional
8753 command line functionality not previously available: DSA signatures can be
8754 generated and verified using pkeyutl and DH key support and generation in
8761 *Oliver Tappe <zooey@hirschkaefer.de>*
8763 * New make target "install_html_docs" installs HTML renditions of the
8766 *Oliver Tappe <zooey@hirschkaefer.de>*
8768 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8769 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8770 support key and parameter generation and add initial key generation
8771 functionality for RSA.
8775 * Add functions for main EVP_PKEY_method operations. The undocumented
8776 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8777 `EVP_PKEY_{encrypt,decrypt}_old`.
8781 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8782 key API, doesn't do much yet.
8786 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8787 public key algorithms. New option to openssl utility:
8788 "list-public-key-algorithms" to print out info.
8792 * Implement the Supported Elliptic Curves Extension for
8793 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8797 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8798 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8802 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8803 utilities such as rsa, dsa, dsaparam etc except they process any key
8808 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8809 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8810 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8815 * Initial support for pluggable public key ASN1.
8816 De-spaghettify the public key ASN1 handling. Move public and private
8817 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8818 algorithm specific handling to a single module within the relevant
8819 algorithm directory. Add functions to allow (near) opaque processing
8820 of public and private key structures.
8824 * Implement the Supported Point Formats Extension for
8825 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8829 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8830 for the psk identity [hint] and the psk callback functions to the
8831 SSL_SESSION, SSL and SSL_CTX structure.
8834 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8838 SSL_CTX_use_psk_identity_hint
8839 SSL_get_psk_identity_hint
8840 SSL_get_psk_identity
8841 SSL_use_psk_identity_hint
8843 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8845 * Add RFC 3161 compliant time stamp request creation, response generation
8846 and response verification functionality.
8848 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8850 * Add initial support for TLS extensions, specifically for the server_name
8851 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8852 have new members for a host name. The SSL data structure has an
8853 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8854 stored in that context to allow for session resumption, even after the
8855 SSL has been switched to a new SSL_CTX in reaction to a client's
8856 server_name extension.
8858 New functions (subject to change):
8860 SSL_get_servername()
8861 SSL_get_servername_type()
8864 New CTRL codes and macros (subject to change):
8866 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8867 - SSL_CTX_set_tlsext_servername_callback()
8868 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8869 - SSL_CTX_set_tlsext_servername_arg()
8870 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8872 openssl s_client has a new '-servername ...' option.
8874 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8875 '-key2 ...', '-servername_fatal' (subject to change). This allows
8876 testing the HostName extension for a specific single host name ('-cert'
8877 and '-key' remain fallbacks for handshakes without HostName
8878 negotiation). If the unrecognized_name alert has to be sent, this by
8879 default is a warning; it becomes fatal with the '-servername_fatal'
8882 *Peter Sylvester, Remy Allais, Christophe Renou*
8884 * Whirlpool hash implementation is added.
8888 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8889 bn(64,32). Because of instruction set limitations it doesn't have
8890 any negative impact on performance. This was done mostly in order
8891 to make it possible to share assembler modules, such as bn_mul_mont
8892 implementations, between 32- and 64-bit builds without hassle.
8896 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8897 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8902 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8903 dedicated Montgomery multiplication procedure, is introduced.
8904 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8905 "64-bit" performance on certain 32-bit targets.
8909 * New option SSL_OP_NO_COMP to disable use of compression selectively
8910 in SSL structures. New SSL ctrl to set maximum send fragment size.
8911 Save memory by setting the I/O buffer sizes dynamically instead of
8912 using the maximum available value.
8916 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8917 in addition to the text details.
8921 * Very, very preliminary EXPERIMENTAL support for printing of general
8922 ASN1 structures. This currently produces rather ugly output and doesn't
8923 handle several customised structures at all.
8927 * Integrated support for PVK file format and some related formats such
8928 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8929 these in the 'rsa' and 'dsa' utilities.
8933 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8937 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8938 place for the (very old) "NETSCAPE" format certificates which are now
8939 handled using new ASN1 code equivalents.
8943 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8944 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8945 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8949 * Modify CRL distribution points extension code to print out previously
8950 unsupported fields. Enhance extension setting code to allow setting of
8955 * Add print and set support for Issuing Distribution Point CRL extension.
8959 * Change 'Configure' script to enable Camellia by default.
8966 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8968 * When rejecting SSL/TLS records due to an incorrect version number, never
8969 update s->server with a new major version number. As of
8970 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8971 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8972 the previous behavior could result in a read attempt at NULL when
8973 receiving specific incorrect SSL/TLS records once record payload
8974 protection is active. ([CVE-2010-0740])
8976 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8978 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8979 could be crashed if the relevant tables were not present (e.g. chrooted).
8981 *Tomas Hoger <thoger@redhat.com>*
8983 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8985 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8987 *Martin Olsson, Neel Mehta*
8989 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8990 accommodate for stack sorting, always a write lock!).
8994 * On some versions of WIN32 Heap32Next is very slow. This can cause
8995 excessive delays in the RAND_poll(): over a minute. As a workaround
8996 include a time check in the inner Heap32Next loop too.
9000 * The code that handled flushing of data in SSL/TLS originally used the
9001 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
9002 the problem outlined in PR#1949. The fix suggested there however can
9003 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
9004 of Apache). So instead simplify the code to flush unconditionally.
9005 This should be fine since flushing with no data to flush is a no op.
9009 * Handle TLS versions 2.0 and later properly and correctly use the
9010 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
9011 off ancient servers have a habit of sticking around for a while...
9015 * Modify compression code so it frees up structures without using the
9016 ex_data callbacks. This works around a problem where some applications
9017 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
9018 restarting) then use compression (e.g. SSL with compression) later.
9019 This results in significant per-connection memory leaks and
9020 has caused some security issues including CVE-2008-1678 and
9025 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
9026 change when encrypting or decrypting.
9030 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
9031 connect and renegotiate with servers which do not support RI.
9032 Until RI is more widely deployed this option is enabled by default.
9036 * Add "missing" ssl ctrls to clear options and mode.
9040 * If client attempts to renegotiate and doesn't support RI respond with
9041 a no_renegotiation alert as required by RFC5746. Some renegotiating
9042 TLS clients will continue a connection gracefully when they receive
9043 the alert. Unfortunately OpenSSL mishandled this alert and would hang
9044 waiting for a server hello which it will never receive. Now we treat a
9045 received no_renegotiation alert as a fatal error. This is because
9046 applications requesting a renegotiation might well expect it to succeed
9047 and would have no code in place to handle the server denying it so the
9048 only safe thing to do is to terminate the connection.
9052 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9053 peer supports secure renegotiation and 0 otherwise. Print out peer
9054 renegotiation support in s_client/s_server.
9058 * Replace the highly broken and deprecated SPKAC certification method with
9059 the updated NID creation version. This should correctly handle UTF8.
9063 * Implement RFC5746. Re-enable renegotiation but require the extension
9064 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9065 turns out to be a bad idea. It has been replaced by
9066 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9067 SSL_CTX_set_options(). This is really not recommended unless you
9068 know what you are doing.
9070 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9072 * Fixes to stateless session resumption handling. Use initial_ctx when
9073 issuing and attempting to decrypt tickets in case it has changed during
9074 servername handling. Use a non-zero length session ID when attempting
9075 stateless session resumption: this makes it possible to determine if
9076 a resumption has occurred immediately after receiving server hello
9077 (several places in OpenSSL subtly assume this) instead of later in
9082 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9083 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9084 fixes for a few places where the return code is not checked
9087 *Julia Lawall <julia@diku.dk>*
9089 * Add --strict-warnings option to Configure script to include devteam
9090 warnings in other configurations.
9094 * Add support for --libdir option and LIBDIR variable in makefiles. This
9095 makes it possible to install openssl libraries in locations which
9096 have names other than "lib", for example "/usr/lib64" which some
9099 *Steve Henson, based on patch from Jeremy Utley*
9101 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9102 X690 8.9.12 and can produce some misleading textual output of OIDs.
9104 *Steve Henson, reported by Dan Kaminsky*
9106 * Delete MD2 from algorithm tables. This follows the recommendation in
9107 several standards that it is not used in new applications due to
9108 several cryptographic weaknesses. For binary compatibility reasons
9109 the MD2 API is still compiled in by default.
9113 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9118 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9119 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9122 *Guenter <lists@gknw.net>*
9124 * Fix the server certificate chain building code to use X509_verify_cert(),
9125 it used to have an ad-hoc builder which was unable to cope with anything
9126 other than a simple chain.
9128 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9130 * Don't check self signed certificate signatures in X509_verify_cert()
9131 by default (a flag can override this): it just wastes time without
9132 adding any security. As a useful side effect self signed root CAs
9133 with non-FIPS digests are now usable in FIPS mode.
9137 * In dtls1_process_out_of_seq_message() the check if the current message
9138 is already buffered was missing. For every new message was memory
9139 allocated, allowing an attacker to perform an denial of service attack
9140 with sending out of seq handshake messages until there is no memory
9141 left. Additionally every future message was buffered, even if the
9142 sequence number made no sense and would be part of another handshake.
9143 So only messages with sequence numbers less than 10 in advance will be
9144 buffered. ([CVE-2009-1378])
9146 *Robin Seggelmann, discovered by Daniel Mentz*
9148 * Records are buffered if they arrive with a future epoch to be
9149 processed after finishing the corresponding handshake. There is
9150 currently no limitation to this buffer allowing an attacker to perform
9151 a DOS attack with sending records with future epochs until there is no
9152 memory left. This patch adds the pqueue_size() function to determine
9153 the size of a buffer and limits the record buffer to 100 entries.
9156 *Robin Seggelmann, discovered by Daniel Mentz*
9158 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9159 parent structure is freed. ([CVE-2009-1379])
9163 * Handle non-blocking I/O properly in SSL_shutdown() call.
9165 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9167 * Add `2.5.4.*` OIDs
9169 *Ilya O. <vrghost@gmail.com>*
9171 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9173 * Disable renegotiation completely - this fixes a severe security
9174 problem ([CVE-2009-3555]) at the cost of breaking all
9175 renegotiation. Renegotiation can be re-enabled by setting
9176 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9177 run-time. This is really not recommended unless you know what
9182 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9184 * Don't set val to NULL when freeing up structures, it is freed up by
9185 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9186 zeroing past the valid field. ([CVE-2009-0789])
9188 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9190 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9191 checked correctly. This would allow some invalid signed attributes to
9192 appear to verify correctly. ([CVE-2009-0591])
9194 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9196 * Reject UniversalString and BMPString types with invalid lengths. This
9197 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9198 a legal length. ([CVE-2009-0590])
9202 * Set S/MIME signing as the default purpose rather than setting it
9203 unconditionally. This allows applications to override it at the store
9208 * Permit restricted recursion of ASN1 strings. This is needed in practice
9209 to handle some structures.
9213 * Improve efficiency of mem_gets: don't search whole buffer each time
9216 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9218 * New -hex option for openssl rand.
9222 * Print out UTF8String and NumericString when parsing ASN1.
9226 * Support NumericString type for name components.
9230 * Allow CC in the environment to override the automatically chosen
9231 compiler. Note that nothing is done to ensure flags work with the
9236 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9238 * Properly check EVP_VerifyFinal() and similar return values
9241 *Ben Laurie, Bodo Moeller, Google Security Team*
9243 * Enable TLS extensions by default.
9247 * Allow the CHIL engine to be loaded, whether the application is
9248 multithreaded or not. (This does not release the developer from the
9249 obligation to set up the dynamic locking callbacks.)
9251 *Sander Temme <sander@temme.net>*
9253 * Use correct exit code if there is an error in dgst command.
9255 *Steve Henson; problem pointed out by Roland Dirlewanger*
9257 * Tweak Configure so that you need to say "experimental-jpake" to enable
9258 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9262 * Add experimental JPAKE support, including demo authentication in
9263 s_client and s_server.
9267 * Set the comparison function in v3_addr_canonize().
9269 *Rob Austein <sra@hactrn.net>*
9271 * Add support for XMPP STARTTLS in s_client.
9273 *Philip Paeps <philip@freebsd.org>*
9275 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9276 to ensure that even with this option, only ciphersuites in the
9277 server's preference list will be accepted. (Note that the option
9278 applies only when resuming a session, so the earlier behavior was
9279 just about the algorithm choice for symmetric cryptography.)
9283 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9285 * Fix NULL pointer dereference if a DTLS server received
9286 ChangeCipherSpec as first record ([CVE-2009-1386]).
9290 * Fix a state transition in s3_srvr.c and d1_srvr.c
9291 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9295 * The fix in 0.9.8c that supposedly got rid of unsafe
9296 double-checked locking was incomplete for RSA blinding,
9297 addressing just one layer of what turns out to have been
9298 doubly unsafe triple-checked locking.
9300 So now fix this for real by retiring the MONT_HELPER macro
9301 in crypto/rsa/rsa_eay.c.
9303 *Bodo Moeller; problem pointed out by Marius Schilder*
9305 * Various precautionary measures:
9307 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9309 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9310 (NB: This would require knowledge of the secret session ticket key
9311 to exploit, in which case you'd be SOL either way.)
9313 - Change bn_nist.c so that it will properly handle input BIGNUMs
9314 outside the expected range.
9316 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9319 *Neel Mehta, Bodo Moeller*
9321 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9322 the load fails. Useful for distros.
9324 *Ben Laurie and the FreeBSD team*
9326 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9330 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9334 * Expand ENGINE to support engine supplied SSL client certificate functions.
9336 This work was sponsored by Logica.
9340 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9341 keystores. Support for SSL/TLS client authentication too.
9342 Not compiled unless enable-capieng specified to Configure.
9344 This work was sponsored by Logica.
9348 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9349 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9350 attribute creation routines such as certificate requests and PKCS#12
9355 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9357 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9358 handshake which could lead to a client crash as found using the
9359 Codenomicon TLS test suite ([CVE-2008-1672])
9361 *Steve Henson, Mark Cox*
9363 * Fix double free in TLS server name extensions which could lead to
9364 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9368 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9370 Clear the error queue to ensure that error entries left from
9371 older function calls do not interfere with the correct operation.
9373 *Lutz Jaenicke, Erik de Castro Lopo*
9375 * Remove root CA certificates of commercial CAs:
9377 The OpenSSL project does not recommend any specific CA and does not
9378 have any policy with respect to including or excluding any CA.
9379 Therefore it does not make any sense to ship an arbitrary selection
9380 of root CA certificates with the OpenSSL software.
9384 * RSA OAEP patches to fix two separate invalid memory reads.
9385 The first one involves inputs when 'lzero' is greater than
9386 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9387 before the beginning of from). The second one involves inputs where
9388 the 'db' section contains nothing but zeroes (there is a one-byte
9389 invalid read after the end of 'db').
9391 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9393 * Partial backport from 0.9.9-dev:
9395 Introduce bn_mul_mont (dedicated Montgomery multiplication
9396 procedure) as a candidate for BIGNUM assembler implementation.
9397 While 0.9.9-dev uses assembler for various architectures, only
9398 x86_64 is available by default here in the 0.9.8 branch, and
9399 32-bit x86 is available through a compile-time setting.
9401 To try the 32-bit x86 assembler implementation, use Configure
9402 option "enable-montasm" (which exists only for this backport).
9404 As "enable-montasm" for 32-bit x86 disclaims code stability
9405 anyway, in this constellation we activate additional code
9406 backported from 0.9.9-dev for further performance improvements,
9407 namely BN_from_montgomery_word. (To enable this otherwise,
9408 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9410 *Andy Polyakov (backport partially by Bodo Moeller)*
9412 * Add TLS session ticket callback. This allows an application to set
9413 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9414 values. This is useful for key rollover for example where several key
9415 sets may exist with different names.
9419 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9420 This was broken until now in 0.9.8 releases, such that the only way
9421 a registered ENGINE could be used (assuming it initialises
9422 successfully on the host) was to explicitly set it as the default
9423 for the relevant algorithms. This is in contradiction with 0.9.7
9424 behaviour and the documentation. With this fix, when an ENGINE is
9425 registered into a given algorithm's table of implementations, the
9426 'uptodate' flag is reset so that auto-discovery will be used next
9427 time a new context for that algorithm attempts to select an
9430 *Ian Lister (tweaked by Geoff Thorpe)*
9432 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9433 implementation in the following ways:
9435 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9438 Lack of BER streaming support means one pass streaming processing is
9439 only supported if data is detached: setting the streaming flag is
9440 ignored for embedded content.
9442 CMS support is disabled by default and must be explicitly enabled
9443 with the enable-cms configuration option.
9447 * Update the GMP engine glue to do direct copies between BIGNUM and
9448 mpz_t when openssl and GMP use the same limb size. Otherwise the
9449 existing "conversion via a text string export" trick is still used.
9451 *Paul Sheer <paulsheer@gmail.com>*
9453 * Zlib compression BIO. This is a filter BIO which compressed and
9454 uncompresses any data passed through it.
9458 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9459 RFC3394 compatible AES key wrapping.
9463 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9464 sets string data without copying. X509_ALGOR_set0() and
9465 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9466 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9467 from an X509_ATTRIBUTE structure optionally checking it occurs only
9468 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9473 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9474 to get the expected BN_FLG_CONSTTIME behavior.
9476 *Bodo Moeller (Google)*
9480 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9481 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9482 - added some more tests to do_tests.pl
9483 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9484 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9485 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9486 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9487 - various changes to netware.pl to enable gcc-cross builds on Win32
9489 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9490 - various changes to fix missing prototype warnings
9491 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9492 - added AES, WHIRLPOOL and CPUID assembler code to build files
9493 - added missing AES assembler make rules to mk1mf.pl
9494 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9496 *Guenter Knauf <eflash@gmx.net>*
9498 * Implement certificate status request TLS extension defined in RFC3546.
9499 A client can set the appropriate parameters and receive the encoded
9500 OCSP response via a callback. A server can query the supplied parameters
9501 and set the encoded OCSP response in the callback. Add simplified examples
9502 to s_client and s_server.
9506 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9509 + Binary incompatibility of ssl_ctx_st structure
9510 + DTLS interoperation with non-compliant servers
9511 + Don't call get_session_cb() without proposed session
9512 + Fix ia64 assembler code
9514 *Andy Polyakov, Steve Henson*
9516 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9518 * DTLS Handshake overhaul. There were longstanding issues with
9519 OpenSSL DTLS implementation, which were making it impossible for
9520 RFC 4347 compliant client to communicate with OpenSSL server.
9521 Unfortunately just fixing these incompatibilities would "cut off"
9522 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9523 server keeps tolerating non RFC compliant syntax. The opposite is
9524 not true, 0.9.8f client can not communicate with earlier server.
9525 This update even addresses CVE-2007-4995.
9529 * Changes to avoid need for function casts in OpenSSL: some compilers
9530 (gcc 4.2 and later) reject their use.
9531 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9534 * Add RFC4507 support to OpenSSL. This includes the corrections in
9535 RFC4507bis. The encrypted ticket format is an encrypted encoded
9536 SSL_SESSION structure, that way new session features are automatically
9539 If a client application caches session in an SSL_SESSION structure
9540 support is transparent because tickets are now stored in the encoded
9543 The SSL_CTX structure automatically generates keys for ticket
9544 protection in servers so again support should be possible
9545 with no application modification.
9547 If a client or server wishes to disable RFC4507 support then the option
9548 SSL_OP_NO_TICKET can be set.
9550 Add a TLS extension debugging callback to allow the contents of any client
9551 or server extensions to be examined.
9553 This work was sponsored by Google.
9557 * Add initial support for TLS extensions, specifically for the server_name
9558 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9559 have new members for a host name. The SSL data structure has an
9560 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9561 stored in that context to allow for session resumption, even after the
9562 SSL has been switched to a new SSL_CTX in reaction to a client's
9563 server_name extension.
9565 New functions (subject to change):
9567 SSL_get_servername()
9568 SSL_get_servername_type()
9571 New CTRL codes and macros (subject to change):
9573 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9574 - SSL_CTX_set_tlsext_servername_callback()
9575 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9576 - SSL_CTX_set_tlsext_servername_arg()
9577 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9579 openssl s_client has a new '-servername ...' option.
9581 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9582 '-key2 ...', '-servername_fatal' (subject to change). This allows
9583 testing the HostName extension for a specific single host name ('-cert'
9584 and '-key' remain fallbacks for handshakes without HostName
9585 negotiation). If the unrecognized_name alert has to be sent, this by
9586 default is a warning; it becomes fatal with the '-servername_fatal'
9589 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9591 * Add AES and SSE2 assembly language support to VC++ build.
9595 * Mitigate attack on final subtraction in Montgomery reduction.
9599 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9600 (which previously caused an internal error).
9604 * Squeeze another 10% out of IGE mode when in != out.
9608 * AES IGE mode speedup.
9610 *Dean Gaudet (Google)*
9612 * Add the Korean symmetric 128-bit cipher SEED (see
9613 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9614 add SEED ciphersuites from RFC 4162:
9616 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9617 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9618 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9619 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9621 To minimize changes between patchlevels in the OpenSSL 0.9.8
9622 series, SEED remains excluded from compilation unless OpenSSL
9623 is configured with 'enable-seed'.
9625 *KISA, Bodo Moeller*
9627 * Mitigate branch prediction attacks, which can be practical if a
9628 single processor is shared, allowing a spy process to extract
9629 information. For detailed background information, see
9630 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9631 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9632 and Necessary Software Countermeasures"). The core of the change
9633 are new versions BN_div_no_branch() and
9634 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9635 respectively, which are slower, but avoid the security-relevant
9636 conditional branches. These are automatically called by BN_div()
9637 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9638 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9639 remove a conditional branch.
9641 BN_FLG_CONSTTIME is the new name for the previous
9642 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9643 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9644 in the exponent causes BN_mod_exp_mont() to use the alternative
9645 implementation in BN_mod_exp_mont_consttime().) The old name
9646 remains as a deprecated alias.
9648 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9649 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9650 constant-time implementations for more than just exponentiation.
9651 Here too the old name is kept as a deprecated alias.
9653 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9654 the BN_BLINDING structure gets an independent copy of the
9655 modulus. This means that the previous `BIGNUM *m` argument to
9656 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9657 essentially becomes `const BIGNUM *m`, although we can't actually
9658 change this in the header file before 0.9.9. It allows
9659 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9660 enable BN_FLG_CONSTTIME.
9662 *Matthew D Wood (Intel Corp)*
9664 * In the SSL/TLS server implementation, be strict about session ID
9665 context matching (which matters if an application uses a single
9666 external cache for different purposes). Previously,
9667 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9668 set. This did ensure strict client verification, but meant that,
9669 with applications using a single external cache for quite
9670 different requirements, clients could circumvent ciphersuite
9671 restrictions for a given session ID context by starting a session
9672 in a different context.
9676 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9677 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9678 authentication-only ciphersuites.
9682 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9683 not complete and could lead to a possible single byte overflow
9684 ([CVE-2007-5135]) [Ben Laurie]
9686 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9688 * Since AES128 and AES256 (and similarly Camellia128 and
9689 Camellia256) share a single mask bit in the logic of
9690 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9691 kludge to work properly if AES128 is available and AES256 isn't
9692 (or if Camellia128 is available and Camellia256 isn't).
9696 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9697 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9698 When a point or a seed is encoded in a BIT STRING, we need to
9699 prevent the removal of trailing zero bits to get the proper DER
9700 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9701 of a NamedBitList, for which trailing 0 bits need to be removed.)
9705 * Have SSL/TLS server implementation tolerate "mismatched" record
9706 protocol version while receiving ClientHello even if the
9707 ClientHello is fragmented. (The server can't insist on the
9708 particular protocol version it has chosen before the ServerHello
9709 message has informed the client about his choice.)
9713 * Add RFC 3779 support.
9715 *Rob Austein for ARIN, Ben Laurie*
9717 * Load error codes if they are not already present instead of using a
9718 static variable. This allows them to be cleanly unloaded and reloaded.
9719 Improve header file function name parsing.
9723 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9724 or CAPABILITY handshake as required by RFCs.
9728 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9730 * Introduce limits to prevent malicious keys being able to
9731 cause a denial of service. ([CVE-2006-2940])
9733 *Steve Henson, Bodo Moeller*
9735 * Fix ASN.1 parsing of certain invalid structures that can result
9736 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9738 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9739 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9741 * Fix SSL client code which could crash if connecting to a
9742 malicious SSLv2 server. ([CVE-2006-4343])
9744 *Tavis Ormandy and Will Drewry, Google Security Team*
9746 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9747 match only those. Before that, "AES256-SHA" would be interpreted
9748 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9749 the same strength classification in 0.9.7h) as we currently only
9750 have a single AES bit in the ciphersuite description bitmap.
9751 That change, however, also applied to ciphersuite strings such as
9752 "RC4-MD5" that intentionally matched multiple ciphersuites --
9753 namely, SSL 2.0 ciphersuites in addition to the more common ones
9754 from SSL 3.0/TLS 1.0.
9756 So we change the selection algorithm again: Naming an explicit
9757 ciphersuite selects this one ciphersuite, and any other similar
9758 ciphersuite (same bitmap) from *other* protocol versions.
9759 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9760 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9762 Since SSL 2.0 does not have any ciphersuites for which the
9763 128/256 bit distinction would be relevant, this works for now.
9764 The proper fix will be to use different bits for AES128 and
9765 AES256, which would have avoided the problems from the beginning;
9766 however, bits are scarce, so we can only do this in a new release
9767 (not just a patchlevel) when we can change the SSL_CIPHER
9768 definition to split the single 'unsigned long mask' bitmap into
9769 multiple values to extend the available space.
9773 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9775 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9776 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9778 * Add AES IGE and biIGE modes.
9782 * Change the Unix randomness entropy gathering to use poll() when
9783 possible instead of select(), since the latter has some
9784 undesirable limitations.
9786 *Darryl Miles via Richard Levitte and Bodo Moeller*
9788 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9789 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9790 cannot be implicitly activated as part of, e.g., the "AES" alias.
9791 However, please upgrade to OpenSSL 0.9.9[-dev] for
9792 non-experimental use of the ECC ciphersuites to get TLS extension
9793 support, which is required for curve and point format negotiation
9794 to avoid potential handshake problems.
9798 * Disable rogue ciphersuites:
9800 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9801 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9802 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9804 The latter two were purportedly from
9805 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9808 Also deactivate the remaining ciphersuites from
9809 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9810 unofficial, and the ID has long expired.
9814 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9815 dual-core machines) and other potential thread-safety issues.
9819 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9820 versions), which is now available for royalty-free use
9821 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9822 Also, add Camellia TLS ciphersuites from RFC 4132.
9824 To minimize changes between patchlevels in the OpenSSL 0.9.8
9825 series, Camellia remains excluded from compilation unless OpenSSL
9826 is configured with 'enable-camellia'.
9830 * Disable the padding bug check when compression is in use. The padding
9831 bug check assumes the first packet is of even length, this is not
9832 necessarily true if compression is enabled and can result in false
9833 positives causing handshake failure. The actual bug test is ancient
9834 code so it is hoped that implementations will either have fixed it by
9835 now or any which still have the bug do not support compression.
9839 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9841 * When applying a cipher rule check to see if string match is an explicit
9842 cipher suite and only match that one cipher suite if it is.
9846 * Link in manifests for VC++ if needed.
9848 *Austin Ziegler <halostatue@gmail.com>*
9850 * Update support for ECC-based TLS ciphersuites according to
9851 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9852 TLS extensions, which are supported starting with the 0.9.9
9853 branch, not in the OpenSSL 0.9.8 branch).
9857 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9858 opaque EVP_CIPHER_CTX handling.
9862 * Fixes and enhancements to zlib compression code. We now only use
9863 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9864 to conform with the standards mentioned here:
9865 <http://www.zlib.net/DLL_FAQ.txt>
9866 Static zlib linking now works on Windows and the new --with-zlib-include
9867 --with-zlib-lib options to Configure can be used to supply the location
9868 of the headers and library. Gracefully handle case where zlib library
9873 * Several fixes and enhancements to the OID generation code. The old code
9874 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9875 handle numbers larger than ULONG_MAX, truncated printing and had a
9876 non standard OBJ_obj2txt() behaviour.
9880 * Add support for building of engines under engine/ as shared libraries
9881 under VC++ build system.
9885 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9886 Hopefully, we will not see any false combination of paths any more.
9890 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9892 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9893 (part of SSL_OP_ALL). This option used to disable the
9894 countermeasure against man-in-the-middle protocol-version
9895 rollback in the SSL 2.0 server implementation, which is a bad
9896 idea. ([CVE-2005-2969])
9898 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9899 for Information Security, National Institute of Advanced Industrial
9900 Science and Technology [AIST], Japan)*
9902 * Add two function to clear and return the verify parameter flags.
9906 * Keep cipherlists sorted in the source instead of sorting them at
9907 runtime, thus removing the need for a lock.
9911 * Avoid some small subgroup attacks in Diffie-Hellman.
9913 *Nick Mathewson and Ben Laurie*
9915 * Add functions for well-known primes.
9919 * Extended Windows CE support.
9921 *Satoshi Nakamura and Andy Polyakov*
9923 * Initialize SSL_METHOD structures at compile time instead of during
9924 runtime, thus removing the need for a lock.
9928 * Make PKCS7_decrypt() work even if no certificate is supplied by
9929 attempting to decrypt each encrypted key in turn. Add support to
9934 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9936 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9939 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9943 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9944 key into the same file any more.
9948 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9952 * Add -utf8 command line and config file option to 'ca'.
9954 *Stefan <stf@udoma.org*
9956 * Removed the macro des_crypt(), as it seems to conflict with some
9957 libraries. Use DES_crypt().
9961 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9962 involves renaming the source and generated shared-libs for
9963 both. The engines will accept the corrected or legacy ids
9964 ('ncipher' and '4758_cca' respectively) when binding. NB,
9965 this only applies when building 'shared'.
9967 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9969 * Add attribute functions to EVP_PKEY structure. Modify
9970 PKCS12_create() to recognize a CSP name attribute and
9971 use it. Make -CSP option work again in pkcs12 utility.
9975 * Add new functionality to the bn blinding code:
9976 - automatic re-creation of the BN_BLINDING parameters after
9977 a fixed number of uses (currently 32)
9978 - add new function for parameter creation
9979 - introduce flags to control the update behaviour of the
9980 BN_BLINDING parameters
9981 - hide BN_BLINDING structure
9982 Add a second BN_BLINDING slot to the RSA structure to improve
9983 performance when a single RSA object is shared among several
9988 * Add support for DTLS.
9990 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9992 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9993 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9997 * Remove buggy and incomplete DH cert support from
9998 ssl/ssl_rsa.c and ssl/s3_both.c
10002 * Use SHA-1 instead of MD5 as the default digest algorithm for
10003 the `apps/openssl` commands.
10007 * Compile clean with "-Wall -Wmissing-prototypes
10008 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
10009 DEBUG_SAFESTACK must also be set.
10013 * Change ./Configure so that certain algorithms can be disabled by default.
10014 The new counterpiece to "no-xxx" is "enable-xxx".
10016 The patented RC5 and MDC2 algorithms will now be disabled unless
10017 "enable-rc5" and "enable-mdc2", respectively, are specified.
10019 (IDEA remains enabled despite being patented. This is because IDEA
10020 is frequently required for interoperability, and there is no license
10021 fee for non-commercial use. As before, "no-idea" can be used to
10022 avoid this algorithm.)
10026 * Add processing of proxy certificates (see RFC 3820). This work was
10027 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
10028 EGEE (Enabling Grids for E-science in Europe).
10032 * RC4 performance overhaul on modern architectures/implementations, such
10033 as Intel P4, IA-64 and AMD64.
10037 * New utility extract-section.pl. This can be used specify an alternative
10038 section number in a pod file instead of having to treat each file as
10039 a separate case in Makefile. This can be done by adding two lines to the
10042 =for comment openssl_section:XXX
10044 The blank line is mandatory.
10048 * New arguments -certform, -keyform and -pass for s_client and s_server
10049 to allow alternative format key and certificate files and passphrase
10054 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10055 update associated structures and add various utility functions.
10057 Add new policy related verify parameters, include policy checking in
10058 standard verify code. Enhance 'smime' application with extra parameters
10059 to support policy checking and print out.
10063 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10064 Nehemiah processors. These extensions support AES encryption in hardware
10065 as well as RNG (though RNG support is currently disabled).
10067 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10069 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10073 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10075 *Andy Polyakov and a number of other people*
10077 * Improved PowerPC platform support. Most notably BIGNUM assembler
10078 implementation contributed by IBM.
10080 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10082 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10083 exponent rather than 'unsigned long'. There is a corresponding change to
10084 the new 'rsa_keygen' element of the RSA_METHOD structure.
10086 *Jelte Jansen, Geoff Thorpe*
10088 * Functionality for creating the initial serial number file is now
10089 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10091 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10092 number file to 1, which is bound to cause problems. To avoid
10093 the problems while respecting compatibility between different 0.9.7
10094 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10095 CA.pl for serial number initialization. With the new release 0.9.8,
10096 we can fix the problem directly in the 'ca' utility.)
10100 * Reduced header interdependencies by declaring more opaque objects in
10101 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10102 give fewer recursive includes, which could break lazy source code - so
10103 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10104 developers should define this symbol when building and using openssl to
10105 ensure they track the recommended behaviour, interfaces, [etc], but
10106 backwards-compatible behaviour prevails when this isn't defined.
10110 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10114 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10115 This will generate a random key of the appropriate length based on the
10116 cipher context. The EVP_CIPHER can provide its own random key generation
10117 routine to support keys of a specific form. This is used in the des and
10118 3des routines to generate a key of the correct parity. Update S/MIME
10119 code to use new functions and hence generate correct parity DES keys.
10120 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10121 valid (weak or incorrect parity).
10125 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10126 as looking them up. This is useful when the verified structure may contain
10127 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10128 present unless the new PKCS7_NO_CRL flag is asserted.
10132 * Extend ASN1 oid configuration module. It now additionally accepts the
10135 shortName = some long name, 1.2.3.4
10139 * Reimplemented the BN_CTX implementation. There is now no more static
10140 limitation on the number of variables it can handle nor the depth of the
10141 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10142 information can now expand as required, and rather than having a single
10143 static array of bignums, BN_CTX now uses a linked-list of such arrays
10144 allowing it to expand on demand whilst maintaining the usefulness of
10145 BN_CTX's "bundling".
10149 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10150 to allow all RSA operations to function using a single BN_CTX.
10154 * Preliminary support for certificate policy evaluation and checking. This
10155 is initially intended to pass the tests outlined in "Conformance Testing
10156 of Relying Party Client Certificate Path Processing Logic" v1.07.
10160 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10161 remained unused and not that useful. A variety of other little bignum
10162 tweaks and fixes have also been made continuing on from the audit (see
10167 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10168 associated ASN1, EVP and SSL functions and old ASN1 macros.
10172 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10173 and this should never fail. So the return value from the use of
10174 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10175 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10179 * BN_CTX_get() should return zero-valued bignums, providing the same
10180 initialised value as BN_new().
10182 *Geoff Thorpe, suggested by Ulf Möller*
10184 * Support for inhibitAnyPolicy certificate extension.
10188 * An audit of the BIGNUM code is underway, for which debugging code is
10189 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10190 is considered valid when processing BIGNUMs, and causes execution to
10191 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10192 further steps are taken to deliberately pollute unused data in BIGNUM
10193 structures to try and expose faulty code further on. For now, openssl will
10194 (in its default mode of operation) continue to tolerate the inconsistent
10195 forms that it has tolerated in the past, but authors and packagers should
10196 consider trying openssl and their own applications when compiled with
10197 these debugging symbols defined. It will help highlight potential bugs in
10198 their own code, and will improve the test coverage for OpenSSL itself. At
10199 some point, these tighter rules will become openssl's default to improve
10200 maintainability, though the assert()s and other overheads will remain only
10201 in debugging configurations. See bn.h for more details.
10203 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10205 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10206 that can only be obtained through BN_CTX_new() (which implicitly
10207 initialises it). The presence of this function only made it possible
10208 to overwrite an existing structure (and cause memory leaks).
10212 * Because of the callback-based approach for implementing LHASH as a
10213 template type, lh_insert() adds opaque objects to hash-tables and
10214 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10215 to clean up those corresponding objects before destroying the hash table
10216 (and losing the object pointers). So some over-zealous constifications in
10217 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10218 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10219 prototyped to have "const" restrictions on the object pointers they are
10220 given (and so aren't required to cast them away any more).
10224 * The tmdiff.h API was so ugly and minimal that our own timing utility
10225 (speed) prefers to use its own implementation. The two implementations
10226 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10227 its object type properly exposed (MS_TM) instead of casting to/from
10228 `char *`. This may still change yet if someone realises MS_TM and
10230 aren't necessarily the greatest nomenclatures - but this is what was used
10231 internally to the implementation so I've used that for now.
10235 * Ensure that deprecated functions do not get compiled when
10236 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10237 the self-tests were still using deprecated key-generation functions so
10238 these have been updated also.
10242 * Reorganise PKCS#7 code to separate the digest location functionality
10243 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10244 New function PKCS7_set_digest() to set the digest type for PKCS#7
10245 digestedData type. Add additional code to correctly generate the
10246 digestedData type and add support for this type in PKCS7 initialization
10251 * New function PKCS7_set0_type_other() this initializes a PKCS7
10252 structure of type "other".
10256 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10257 sure the loop does correctly stop and breaking ("division by zero")
10258 modulus operations are not performed. The (pre-generated) prime
10259 table crypto/bn/bn_prime.h was already correct, but it could not be
10260 re-generated on some platforms because of the "division by zero"
10261 situation in the script.
10263 *Ralf S. Engelschall*
10265 * Update support for ECC-based TLS ciphersuites according to
10266 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10267 SHA-1 now is only used for "small" curves (where the
10268 representation of a field element takes up to 24 bytes); for
10269 larger curves, the field element resulting from ECDH is directly
10270 used as premaster secret.
10272 *Douglas Stebila (Sun Microsystems Laboratories)*
10274 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10275 curve secp160r1 to the tests.
10277 *Douglas Stebila (Sun Microsystems Laboratories)*
10279 * Add the possibility to load symbols globally with DSO.
10281 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10283 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10284 control of the error stack.
10288 * Add support for STORE in ENGINE.
10292 * Add the STORE type. The intention is to provide a common interface
10293 to certificate and key stores, be they simple file-based stores, or
10294 HSM-type store, or LDAP stores, or...
10295 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10299 * Add a generic structure called OPENSSL_ITEM. This can be used to
10300 pass a list of arguments to any function as well as provide a way
10301 for a function to pass data back to the caller.
10305 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10306 works like BUF_strdup() but can be used to duplicate a portion of
10307 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10312 * Add the function sk_find_ex() which works like sk_find(), but will
10313 return an index to an element even if an exact match couldn't be
10314 found. The index is guaranteed to point at the element where the
10315 searched-for key would be inserted to preserve sorting order.
10319 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10320 takes an extra flags argument for optional functionality. Currently,
10321 the following flags are defined:
10323 OBJ_BSEARCH_VALUE_ON_NOMATCH
10324 This one gets OBJ_bsearch_ex() to return a pointer to the first
10325 element where the comparing function returns a negative or zero
10328 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10329 This one gets OBJ_bsearch_ex() to return a pointer to the first
10330 element where the comparing function returns zero. This is useful
10331 if there are more than one element where the comparing function
10336 * Make it possible to create self-signed certificates with 'openssl ca'
10337 in such a way that the self-signed certificate becomes part of the
10338 CA database and uses the same mechanisms for serial number generation
10339 as all other certificate signing. The new flag '-selfsign' enables
10340 this functionality. Adapt CA.sh and CA.pl.in.
10344 * Add functionality to check the public key of a certificate request
10345 against a given private. This is useful to check that a certificate
10346 request can be signed by that key (self-signing).
10350 * Make it possible to have multiple active certificates with the same
10351 subject in the CA index file. This is done only if the keyword
10352 'unique_subject' is set to 'no' in the main CA section (default
10353 if 'CA_default') of the configuration file. The value is saved
10354 with the database itself in a separate index attribute file,
10355 named like the index file with '.attr' appended to the name.
10359 * Generate multi-valued AVAs using '+' notation in config files for
10364 * Support for nameConstraints certificate extension.
10368 * Support for policyConstraints certificate extension.
10372 * Support for policyMappings certificate extension.
10376 * Make sure the default DSA_METHOD implementation only uses its
10377 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10378 and change its own handlers to be NULL so as to remove unnecessary
10379 indirection. This lets alternative implementations fallback to the
10380 default implementation more easily.
10384 * Support for directoryName in GeneralName related extensions
10389 * Make it possible to link applications using Makefile.shared.
10390 Make that possible even when linking against static libraries!
10394 * Support for single pass processing for S/MIME signing. This now
10395 means that S/MIME signing can be done from a pipe, in addition
10396 cleartext signing (multipart/signed type) is effectively streaming
10397 and the signed data does not need to be all held in memory.
10399 This is done with a new flag PKCS7_STREAM. When this flag is set
10400 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10401 is done after the data is output (and digests calculated) in
10402 SMIME_write_PKCS7().
10406 * Add full support for -rpath/-R, both in shared libraries and
10407 applications, at least on the platforms where it's known how
10412 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10413 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10414 will now compute a table of multiples of the generator that
10415 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10416 faster (notably in the case of a single point multiplication,
10417 scalar * generator).
10419 *Nils Larsch, Bodo Moeller*
10421 * IPv6 support for certificate extensions. The various extensions
10422 which use the IP:a.b.c.d can now take IPv6 addresses using the
10423 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10428 * Added an ENGINE that implements RSA by performing private key
10429 exponentiations with the GMP library. The conversions to and from
10430 GMP's mpz_t format aren't optimised nor are any montgomery forms
10431 cached, and on x86 it appears OpenSSL's own performance has caught up.
10432 However there are likely to be other architectures where GMP could
10433 provide a boost. This ENGINE is not built in by default, but it can be
10434 specified at Configure time and should be accompanied by the necessary
10435 linker additions, eg;
10436 ./config -DOPENSSL_USE_GMP -lgmp
10440 * "openssl engine" will not display ENGINE/DSO load failure errors when
10441 testing availability of engines with "-t" - the old behaviour is
10442 produced by increasing the feature's verbosity with "-tt".
10446 * ECDSA routines: under certain error conditions uninitialized BN objects
10447 could be freed. Solution: make sure initialization is performed early
10448 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10453 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10454 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10455 software implementations. For DSA and DH, parameter generation can
10456 also be overridden by providing the appropriate method callbacks.
10460 * Change the "progress" mechanism used in key-generation and
10461 primality testing to functions that take a new BN_GENCB pointer in
10462 place of callback/argument pairs. The new API functions have `_ex`
10463 postfixes and the older functions are reimplemented as wrappers for
10464 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10465 declarations of the old functions to help (graceful) attempts to
10466 migrate to the new functions. Also, the new key-generation API
10467 functions operate on a caller-supplied key-structure and return
10468 success/failure rather than returning a key or NULL - this is to
10469 help make "keygen" another member function of RSA_METHOD etc.
10471 Example for using the new callback interface:
10473 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10474 void *my_arg = ...;
10477 BN_GENCB_set(&my_cb, my_callback, my_arg);
10479 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10480 /* For the meaning of a, b in calls to my_callback(), see the
10481 * documentation of the function that calls the callback.
10482 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10483 * my_callback should return 1 if it wants BN_is_prime_ex()
10484 * to continue, or 0 to stop.
10489 * Change the ZLIB compression method to be stateful, and make it
10490 available to TLS with the number defined in
10491 draft-ietf-tls-compression-04.txt.
10495 * Add the ASN.1 structures and functions for CertificatePair, which
10496 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10498 CertificatePair ::= SEQUENCE {
10499 forward [0] Certificate OPTIONAL,
10500 reverse [1] Certificate OPTIONAL,
10501 -- at least one of the pair shall be present -- }
10503 Also implement the PEM functions to read and write certificate
10504 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10506 This needed to be defined, mostly for the sake of the LDAP
10507 attribute crossCertificatePair, but may prove useful elsewhere as
10512 * Make it possible to inhibit symlinking of shared libraries in
10513 Makefile.shared, for Cygwin's sake.
10517 * Extend the BIGNUM API by creating a function
10518 void BN_set_negative(BIGNUM *a, int neg);
10519 and a macro that behave like
10520 int BN_is_negative(const BIGNUM *a);
10522 to avoid the need to access 'a->neg' directly in applications.
10526 * Implement fast modular reduction for pseudo-Mersenne primes
10527 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10528 EC_GROUP_new_curve_GFp() will now automatically use this
10531 *Nils Larsch <nla@trustcenter.de>*
10533 * Add new lock type (CRYPTO_LOCK_BN).
10537 * Change the ENGINE framework to automatically load engines
10538 dynamically from specific directories unless they could be
10539 found to already be built in or loaded. Move all the
10540 current engines except for the cryptodev one to a new
10541 directory engines/.
10542 The engines in engines/ are built as shared libraries if
10543 the "shared" options was given to ./Configure or ./config.
10544 Otherwise, they are inserted in libcrypto.a.
10545 /usr/local/ssl/engines is the default directory for dynamic
10546 engines, but that can be overridden at configure time through
10547 the usual use of --prefix and/or --openssldir, and at run
10548 time with the environment variable OPENSSL_ENGINES.
10550 *Geoff Thorpe and Richard Levitte*
10552 * Add Makefile.shared, a helper makefile to build shared
10553 libraries. Adapt Makefile.org.
10557 * Add version info to Win32 DLLs.
10559 *Peter 'Luna' Runestig" <peter@runestig.com>*
10561 * Add new 'medium level' PKCS#12 API. Certificates and keys
10562 can be added using this API to created arbitrary PKCS#12
10563 files while avoiding the low level API.
10565 New options to PKCS12_create(), key or cert can be NULL and
10566 will then be omitted from the output file. The encryption
10567 algorithm NIDs can be set to -1 for no encryption, the mac
10568 iteration count can be set to 0 to omit the mac.
10570 Enhance pkcs12 utility by making the -nokeys and -nocerts
10571 options work when creating a PKCS#12 file. New option -nomac
10572 to omit the mac, NONE can be set for an encryption algorithm.
10573 New code is modified to use the enhanced PKCS12_create()
10574 instead of the low level API.
10578 * Extend ASN1 encoder to support indefinite length constructed
10579 encoding. This can output sequences tags and octet strings in
10580 this form. Modify pk7_asn1.c to support indefinite length
10581 encoding. This is experimental and needs additional code to
10582 be useful, such as an ASN1 bio and some enhanced streaming
10585 Extend template encode functionality so that tagging is passed
10586 down to the template encoder.
10590 * Let 'openssl req' fail if an argument to '-newkey' is not
10591 recognized instead of using RSA as a default.
10595 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10596 As these are not official, they are not included in "ALL";
10597 the "ECCdraft" ciphersuite group alias can be used to select them.
10599 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10601 * Add ECDH engine support.
10603 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10605 * Add ECDH in new directory crypto/ecdh/.
10607 *Douglas Stebila (Sun Microsystems Laboratories)*
10609 * Let BN_rand_range() abort with an error after 100 iterations
10610 without success (which indicates a broken PRNG).
10614 * Change BN_mod_sqrt() so that it verifies that the input value
10615 is really the square of the return value. (Previously,
10616 BN_mod_sqrt would show GIGO behaviour.)
10620 * Add named elliptic curves over binary fields from X9.62, SECG,
10621 and WAP/WTLS; add OIDs that were still missing.
10623 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10625 * Extend the EC library for elliptic curves over binary fields
10626 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10629 EC_GF2m_simple_method
10633 EC_GROUP_new_curve_GF2m
10634 EC_GROUP_set_curve_GF2m
10635 EC_GROUP_get_curve_GF2m
10636 EC_POINT_set_affine_coordinates_GF2m
10637 EC_POINT_get_affine_coordinates_GF2m
10638 EC_POINT_set_compressed_coordinates_GF2m
10640 Point compression for binary fields is disabled by default for
10641 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10644 As binary polynomials are represented as BIGNUMs, various members
10645 of the EC_GROUP and EC_POINT data structures can be shared
10646 between the implementations for prime fields and binary fields;
10647 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10648 are essentially identical to their `..._GFp` counterparts.
10649 (For simplicity, the `..._GFp` prefix has been dropped from
10650 various internal method names.)
10652 An internal 'field_div' method (similar to 'field_mul' and
10653 'field_sqr') has been added; this is used only for binary fields.
10655 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10657 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10658 through methods ('mul', 'precompute_mult').
10660 The generic implementations (now internally called 'ec_wNAF_mul'
10661 and 'ec_wNAF_precomputed_mult') remain the default if these
10662 methods are undefined.
10664 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10666 * New function EC_GROUP_get_degree, which is defined through
10667 EC_METHOD. For curves over prime fields, this returns the bit
10668 length of the modulus.
10670 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10672 * New functions EC_GROUP_dup, EC_POINT_dup.
10673 (These simply call ..._new and ..._copy).
10675 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10677 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10678 Polynomials are represented as BIGNUMs (where the sign bit is not
10679 used) in the following functions [macros]:
10682 BN_GF2m_sub [= BN_GF2m_add]
10683 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10684 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10685 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10687 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10688 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10689 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10690 BN_GF2m_cmp [= BN_ucmp]
10692 (Note that only the 'mod' functions are actually for fields GF(2^m).
10693 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10695 For some functions, an the irreducible polynomial defining a
10696 field can be given as an 'unsigned int[]' with strictly
10697 decreasing elements giving the indices of those bits that are set;
10698 i.e., p[] represents the polynomial
10699 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10701 p[0] > p[1] > ... > p[k] = 0.
10702 This applies to the following functions:
10705 BN_GF2m_mod_mul_arr
10706 BN_GF2m_mod_sqr_arr
10707 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10708 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10709 BN_GF2m_mod_exp_arr
10710 BN_GF2m_mod_sqrt_arr
10711 BN_GF2m_mod_solve_quad_arr
10715 Conversion can be performed by the following functions:
10720 bntest.c has additional tests for binary polynomial arithmetic.
10722 Two implementations for BN_GF2m_mod_div() are available.
10723 The default algorithm simply uses BN_GF2m_mod_inv() and
10724 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10725 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10726 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10728 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10730 * Add new error code 'ERR_R_DISABLED' that can be used when some
10731 functionality is disabled at compile-time.
10733 *Douglas Stebila <douglas.stebila@sun.com>*
10735 * Change default behaviour of 'openssl asn1parse' so that more
10736 information is visible when viewing, e.g., a certificate:
10738 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10739 mode the content of non-printable OCTET STRINGs is output in a
10740 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10741 avoid the appearance of a printable string.
10743 *Nils Larsch <nla@trustcenter.de>*
10745 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10747 EC_GROUP_set_asn1_flag()
10748 EC_GROUP_get_asn1_flag()
10749 EC_GROUP_set_point_conversion_form()
10750 EC_GROUP_get_point_conversion_form()
10751 These control ASN1 encoding details:
10752 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10753 has been set to OPENSSL_EC_NAMED_CURVE.
10754 - Points are encoded in uncompressed form by default; options for
10755 asn1_for are as for point2oct, namely
10756 POINT_CONVERSION_COMPRESSED
10757 POINT_CONVERSION_UNCOMPRESSED
10758 POINT_CONVERSION_HYBRID
10760 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10762 EC_GROUP_set_seed()
10763 EC_GROUP_get0_seed()
10764 EC_GROUP_get_seed_len()
10765 This is used only for ASN1 purposes (so far).
10767 *Nils Larsch <nla@trustcenter.de>*
10769 * Add 'field_type' member to EC_METHOD, which holds the NID
10770 of the appropriate field type OID. The new function
10771 EC_METHOD_get_field_type() returns this value.
10773 *Nils Larsch <nla@trustcenter.de>*
10776 EC_POINT_point2bn()
10777 EC_POINT_bn2point()
10778 EC_POINT_point2hex()
10779 EC_POINT_hex2point()
10780 providing useful interfaces to EC_POINT_point2oct() and
10781 EC_POINT_oct2point().
10783 *Nils Larsch <nla@trustcenter.de>*
10785 * Change internals of the EC library so that the functions
10786 EC_GROUP_set_generator()
10787 EC_GROUP_get_generator()
10788 EC_GROUP_get_order()
10789 EC_GROUP_get_cofactor()
10790 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10791 to methods, which would lead to unnecessary code duplication when
10792 adding different types of curves.
10794 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10796 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10797 arithmetic, and such that modified wNAFs are generated
10798 (which avoid length expansion in many cases).
10802 * Add a function EC_GROUP_check_discriminant() (defined via
10803 EC_METHOD) that verifies that the curve discriminant is non-zero.
10805 Add a function EC_GROUP_check() that makes some sanity tests
10806 on a EC_GROUP, its generator and order. This includes
10807 EC_GROUP_check_discriminant().
10809 *Nils Larsch <nla@trustcenter.de>*
10811 * Add ECDSA in new directory crypto/ecdsa/.
10813 Add applications 'openssl ecparam' and 'openssl ecdsa'
10814 (these are based on 'openssl dsaparam' and 'openssl dsa').
10816 ECDSA support is also included in various other files across the
10817 library. Most notably,
10818 - 'openssl req' now has a '-newkey ecdsa:file' option;
10819 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10820 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10821 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10822 them suitable for ECDSA where domain parameters must be
10823 extracted before the specific public key;
10824 - ECDSA engine support has been added.
10826 *Nils Larsch <nla@trustcenter.de>*
10828 * Include some named elliptic curves, and add OIDs from X9.62,
10829 SECG, and WAP/WTLS. Each curve can be obtained from the new
10831 EC_GROUP_new_by_curve_name(),
10832 and the list of available named curves can be obtained with
10833 EC_get_builtin_curves().
10834 Also add a 'curve_name' member to EC_GROUP objects, which can be
10836 EC_GROUP_set_curve_name()
10837 EC_GROUP_get_curve_name()
10839 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10841 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10842 was actually never needed) and in BN_mul(). The removal in BN_mul()
10843 required a small change in bn_mul_part_recursive() and the addition
10844 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10845 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10846 bn_sub_words() and bn_add_words() except they take arrays with
10851 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10853 * Cleanse PEM buffers before freeing them since they may contain
10856 *Benjamin Bennett <ben@psc.edu>*
10858 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10859 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10860 authentication-only ciphersuites.
10864 * Since AES128 and AES256 share a single mask bit in the logic of
10865 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10866 kludge to work properly if AES128 is available and AES256 isn't.
10870 * Expand security boundary to match 1.1.1 module.
10874 * Remove redundant features: hash file source, editing of test vectors
10875 modify fipsld to use external fips_premain.c signature.
10879 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10880 run algorithm test programs.
10884 * Make algorithm test programs more tolerant of whitespace.
10888 * Have SSL/TLS server implementation tolerate "mismatched" record
10889 protocol version while receiving ClientHello even if the
10890 ClientHello is fragmented. (The server can't insist on the
10891 particular protocol version it has chosen before the ServerHello
10892 message has informed the client about his choice.)
10896 * Load error codes if they are not already present instead of using a
10897 static variable. This allows them to be cleanly unloaded and reloaded.
10901 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10903 * Introduce limits to prevent malicious keys being able to
10904 cause a denial of service. ([CVE-2006-2940])
10906 *Steve Henson, Bodo Moeller*
10908 * Fix ASN.1 parsing of certain invalid structures that can result
10909 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10911 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10912 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10914 * Fix SSL client code which could crash if connecting to a
10915 malicious SSLv2 server. ([CVE-2006-4343])
10917 *Tavis Ormandy and Will Drewry, Google Security Team*
10919 * Change ciphersuite string processing so that an explicit
10920 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10921 will no longer include "AES128-SHA"), and any other similar
10922 ciphersuite (same bitmap) from *other* protocol versions (so that
10923 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10924 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10925 changes from 0.9.8b and 0.9.8d.
10929 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10931 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10932 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10934 * Change the Unix randomness entropy gathering to use poll() when
10935 possible instead of select(), since the latter has some
10936 undesirable limitations.
10938 *Darryl Miles via Richard Levitte and Bodo Moeller*
10940 * Disable rogue ciphersuites:
10942 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10943 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10944 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10946 The latter two were purportedly from
10947 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10950 Also deactivate the remaining ciphersuites from
10951 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10952 unofficial, and the ID has long expired.
10956 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10957 dual-core machines) and other potential thread-safety issues.
10961 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10963 * Adapt fipsld and the build system to link against the validated FIPS
10964 module in FIPS mode.
10968 * Fixes for VC++ 2005 build under Windows.
10972 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10973 from a Windows bash shell such as MSYS. It is autodetected from the
10974 "config" script when run from a VC++ environment. Modify standard VC++
10975 build to use fipscanister.o from the GNU make build.
10979 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10981 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10982 The value now differs depending on if you build for FIPS or not.
10983 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10984 safely run with a non-FIPSed libcrypto, as it may crash because of
10985 the difference induced by this change.
10989 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10991 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10992 (part of SSL_OP_ALL). This option used to disable the
10993 countermeasure against man-in-the-middle protocol-version
10994 rollback in the SSL 2.0 server implementation, which is a bad
10995 idea. ([CVE-2005-2969])
10997 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10998 for Information Security, National Institute of Advanced Industrial
10999 Science and Technology [AIST, Japan)]*
11001 * Minimal support for X9.31 signatures and PSS padding modes. This is
11002 mainly for FIPS compliance and not fully integrated at this stage.
11006 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
11007 the exponentiation using a fixed-length exponent. (Otherwise,
11008 the information leaked through timing could expose the secret key
11009 after many signatures; cf. Bleichenbacher's attack on DSA with
11014 * Make a new fixed-window mod_exp implementation the default for
11015 RSA, DSA, and DH private-key operations so that the sequence of
11016 squares and multiplies and the memory access pattern are
11017 independent of the particular secret key. This will mitigate
11018 cache-timing and potential related attacks.
11020 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
11021 and this is automatically used by BN_mod_exp_mont() if the new flag
11022 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
11023 will use this BN flag for private exponents unless the flag
11024 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
11025 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
11027 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
11029 * Change the client implementation for SSLv23_method() and
11030 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
11031 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
11032 (Previously, the SSL 2.0 backwards compatible Client Hello
11033 message format would be used even with SSL_OP_NO_SSLv2.)
11037 * Add support for smime-type MIME parameter in S/MIME messages which some
11042 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
11043 a threadsafe manner. Modify rsa code to use new function and add calls
11044 to dsa and dh code (which had race conditions before).
11048 * Include the fixed error library code in the C error file definitions
11049 instead of fixing them up at runtime. This keeps the error code
11050 structures constant.
11054 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11056 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11059 * Fixes for newer kerberos headers. NB: the casts are needed because
11060 the 'length' field is signed on one version and unsigned on another
11061 with no (?) obvious way to tell the difference, without these VC++
11062 complains. Also the "definition" of FAR (blank) is no longer included
11063 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11064 some needed definitions.
11068 * Undo Cygwin change.
11072 * Added support for proxy certificates according to RFC 3820.
11073 Because they may be a security thread to unaware applications,
11074 they must be explicitly allowed in run-time. See
11075 docs/HOWTO/proxy_certificates.txt for further information.
11079 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11081 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11082 server and client random values. Previously
11083 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11084 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11086 This change has negligible security impact because:
11088 1. Server and client random values still have 24 bytes of pseudo random
11091 2. Server and client random values are sent in the clear in the initial
11094 3. The master secret is derived using the premaster secret (48 bytes in
11095 size for static RSA ciphersuites) as well as client server and random
11098 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11101 *Stephen Henson, reported by UK NISCC*
11103 * Use Windows randomness collection on Cygwin.
11107 * Fix hang in EGD/PRNGD query when communication socket is closed
11108 prematurely by EGD/PRNGD.
11110 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11112 * Prompt for pass phrases when appropriate for PKCS12 input format.
11116 * Back-port of selected performance improvements from development
11117 branch, as well as improved support for PowerPC platforms.
11121 * Add lots of checks for memory allocation failure, error codes to indicate
11122 failure and freeing up memory if a failure occurs.
11124 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11126 * Add new -passin argument to dgst.
11130 * Perform some character comparisons of different types in X509_NAME_cmp:
11131 this is needed for some certificates that re-encode DNs into UTF8Strings
11132 (in violation of RFC3280) and can't or won't issue name rollover
11137 * Make an explicit check during certificate validation to see that
11138 the CA setting in each certificate on the chain is correct. As a
11139 side effect always do the following basic checks on extensions,
11140 not just when there's an associated purpose to the check:
11142 - if there is an unhandled critical extension (unless the user
11143 has chosen to ignore this fault)
11144 - if the path length has been exceeded (if one is set at all)
11145 - that certain extensions fit the associated purpose (if one has
11150 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11152 * Avoid a race condition when CRLs are checked in a multi threaded
11153 environment. This would happen due to the reordering of the revoked
11154 entries during signature checking and serial number lookup. Now the
11155 encoding is cached and the serial number sort performed under a lock.
11156 Add new STACK function sk_is_sorted().
11160 * Add Delta CRL to the extension code.
11164 * Various fixes to s3_pkt.c so alerts are sent properly.
11166 *David Holmes <d.holmes@f5.com>*
11168 * Reduce the chances of duplicate issuer name and serial numbers (in
11169 violation of RFC3280) using the OpenSSL certificate creation utilities.
11170 This is done by creating a random 64 bit value for the initial serial
11171 number when a serial number file is created or when a self signed
11172 certificate is created using 'openssl req -x509'. The initial serial
11173 number file is created using 'openssl x509 -next_serial' in CA.pl
11174 rather than being initialized to 1.
11178 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11180 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11181 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11183 *Joe Orton, Steve Henson*
11185 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11188 *Joe Orton, Steve Henson*
11190 * Make it possible to have multiple active certificates with the same
11191 subject in the CA index file. This is done only if the keyword
11192 'unique_subject' is set to 'no' in the main CA section (default
11193 if 'CA_default') of the configuration file. The value is saved
11194 with the database itself in a separate index attribute file,
11195 named like the index file with '.attr' appended to the name.
11199 * X509 verify fixes. Disable broken certificate workarounds when
11200 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11201 keyUsage extension present. Don't accept CRLs with unhandled critical
11202 extensions: since verify currently doesn't process CRL extensions this
11203 rejects a CRL with *any* critical extensions. Add new verify error codes
11208 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11209 A clarification of RFC2560 will require the use of OCTET STRINGs and
11210 some implementations cannot handle the current raw format. Since OpenSSL
11211 copies and compares OCSP nonces as opaque blobs without any attempt at
11212 parsing them this should not create any compatibility issues.
11216 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11217 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11218 this HMAC (and other) operations are several times slower than OpenSSL
11223 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11225 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11227 * Use the correct content when signing type "other".
11231 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11233 * Fix various bugs revealed by running the NISCC test suite:
11235 Stop out of bounds reads in the ASN1 code when presented with
11236 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11238 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11240 If verify callback ignores invalid public key errors don't try to check
11241 certificate signature with the NULL public key.
11245 * New -ignore_err option in ocsp application to stop the server
11246 exiting on the first error in a request.
11250 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11251 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11256 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11257 extra data after the compression methods not only for TLS 1.0
11258 but also for SSL 3.0 (as required by the specification).
11260 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11262 * Change X509_certificate_type() to mark the key as exported/exportable
11263 when it's 512 *bits* long, not 512 bytes.
11267 * Change AES_cbc_encrypt() so it outputs exact multiple of
11268 blocks during encryption.
11272 * Various fixes to base64 BIO and non blocking I/O. On write
11273 flushes were not handled properly if the BIO retried. On read
11274 data was not being buffered properly and had various logic bugs.
11275 This also affects blocking I/O when the data being decoded is a
11280 * Various S/MIME bugfixes and compatibility changes:
11281 output correct application/pkcs7 MIME type if
11282 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11283 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11284 of files as .eml work). Correctly handle very long lines in MIME
11289 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11291 * Countermeasure against the Klima-Pokorny-Rosa extension of
11292 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11293 a protocol version number mismatch like a decryption error
11294 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11298 * Turn on RSA blinding by default in the default implementation
11299 to avoid a timing attack. Applications that don't want it can call
11300 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11301 They would be ill-advised to do so in most cases.
11303 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11305 * Change RSA blinding code so that it works when the PRNG is not
11306 seeded (in this case, the secret RSA exponent is abused as
11307 an unpredictable seed -- if it is not unpredictable, there
11308 is no point in blinding anyway). Make RSA blinding thread-safe
11309 by remembering the creator's thread ID in rsa->blinding and
11310 having all other threads use local one-time blinding factors
11311 (this requires more computation than sharing rsa->blinding, but
11312 avoids excessive locking; and if an RSA object is not shared
11313 between threads, blinding will still be very fast).
11317 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11318 ENGINE as defaults for all supported algorithms irrespective of
11319 the 'flags' parameter. 'flags' is now honoured, so applications
11320 should make sure they are passing it correctly.
11324 * Target "mingw" now allows native Windows code to be generated in
11325 the Cygwin environment as well as with the MinGW compiler.
11329 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11331 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11332 via timing by performing a MAC computation even if incorrect
11333 block cipher padding has been found. This is a countermeasure
11334 against active attacks where the attacker has to distinguish
11335 between bad padding and a MAC verification error. ([CVE-2003-0078])
11337 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11338 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11339 Martin Vuagnoux (EPFL, Ilion)*
11341 * Make the no-err option work as intended. The intention with no-err
11342 is not to have the whole error stack handling routines removed from
11343 libcrypto, it's only intended to remove all the function name and
11344 reason texts, thereby removing some of the footprint that may not
11345 be interesting if those errors aren't displayed anyway.
11347 NOTE: it's still possible for any application or module to have its
11348 own set of error texts inserted. The routines are there, just not
11349 used by default when no-err is given.
11353 * Add support for FreeBSD on IA64.
11355 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11357 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11358 Kerberos function mit_des_cbc_cksum(). Before this change,
11359 the value returned by DES_cbc_cksum() was like the one from
11360 mit_des_cbc_cksum(), except the bytes were swapped.
11362 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11364 * Allow an application to disable the automatic SSL chain building.
11365 Before this a rather primitive chain build was always performed in
11366 ssl3_output_cert_chain(): an application had no way to send the
11367 correct chain if the automatic operation produced an incorrect result.
11369 Now the chain builder is disabled if either:
11371 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11373 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11375 The reasoning behind this is that an application would not want the
11376 auto chain building to take place if extra chain certificates are
11377 present and it might also want a means of sending no additional
11378 certificates (for example the chain has two certificates and the
11383 * Add the possibility to build without the ENGINE framework.
11385 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11387 * Under Win32 gmtime() can return NULL: check return value in
11388 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11392 * DSA routines: under certain error conditions uninitialized BN objects
11393 could be freed. Solution: make sure initialization is performed early
11394 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11395 Nils Larsch <nla@trustcenter.de> via PR#459)
11399 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11400 checked on reconnect on the client side, therefore session resumption
11401 could still fail with a "ssl session id is different" error. This
11402 behaviour is masked when SSL_OP_ALL is used due to
11403 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11404 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11405 followup to PR #377.
11409 * IA-32 assembler support enhancements: unified ELF targets, support
11410 for SCO/Caldera platforms, fix for Cygwin shared build.
11414 * Add support for FreeBSD on sparc64. As a consequence, support for
11415 FreeBSD on non-x86 processors is separate from x86 processors on
11416 the config script, much like the NetBSD support.
11418 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11420 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11422 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11425 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11426 code (06) was taken as the first octet of the session ID and the last
11427 octet was ignored consequently. As a result SSLv2 client side session
11428 caching could not have worked due to the session ID mismatch between
11430 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11435 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11436 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11441 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11442 seems that in spite of existing for more than a year, many application
11443 author have done nothing to provide the necessary callbacks, which
11444 means that this particular engine will not work properly anywhere.
11445 This is a very unfortunate situation which forces us, in the name
11446 of usability, to give the hw_ncipher.c a static lock, which is part
11448 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11449 appear in 0.9.8 or later. We EXPECT application authors to have
11450 dealt properly with this when 0.9.8 is released (unless we actually
11451 make such changes in the libcrypto locking code that changes will
11452 have to be made anyway).
11456 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11457 octets have been read, EOF or an error occurs. Without this change
11458 some truncated ASN1 structures will not produce an error.
11462 * Disable Heimdal support, since it hasn't been fully implemented.
11463 Still give the possibility to force the use of Heimdal, but with
11464 warnings and a request that patches get sent to openssl-dev.
11468 * Add the VC-CE target, introduce the WINCE sysname, and add
11469 INSTALL.WCE and appropriate conditionals to make it build.
11471 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11473 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11474 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11475 edit numbers of the version.
11477 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11479 * Introduce safe string copy and catenation functions
11480 (BUF_strlcpy() and BUF_strlcat()).
11482 *Ben Laurie (CHATS) and Richard Levitte*
11484 * Avoid using fixed-size buffers for one-line DNs.
11486 *Ben Laurie (CHATS)*
11488 * Add BUF_MEM_grow_clean() to avoid information leakage when
11489 resizing buffers containing secrets, and use where appropriate.
11491 *Ben Laurie (CHATS)*
11493 * Avoid using fixed size buffers for configuration file location.
11495 *Ben Laurie (CHATS)*
11497 * Avoid filename truncation for various CA files.
11499 *Ben Laurie (CHATS)*
11501 * Use sizeof in preference to magic numbers.
11503 *Ben Laurie (CHATS)*
11505 * Avoid filename truncation in cert requests.
11507 *Ben Laurie (CHATS)*
11509 * Add assertions to check for (supposedly impossible) buffer
11512 *Ben Laurie (CHATS)*
11514 * Don't cache truncated DNS entries in the local cache (this could
11515 potentially lead to a spoofing attack).
11517 *Ben Laurie (CHATS)*
11519 * Fix various buffers to be large enough for hex/decimal
11520 representations in a platform independent manner.
11522 *Ben Laurie (CHATS)*
11524 * Add CRYPTO_realloc_clean() to avoid information leakage when
11525 resizing buffers containing secrets, and use where appropriate.
11527 *Ben Laurie (CHATS)*
11529 * Add BIO_indent() to avoid much slightly worrying code to do
11532 *Ben Laurie (CHATS)*
11534 * Convert sprintf()/BIO_puts() to BIO_printf().
11536 *Ben Laurie (CHATS)*
11538 * buffer_gets() could terminate with the buffer only half
11541 *Ben Laurie (CHATS)*
11543 * Add assertions to prevent user-supplied crypto functions from
11544 overflowing internal buffers by having large block sizes, etc.
11546 *Ben Laurie (CHATS)*
11548 * New OPENSSL_assert() macro (similar to assert(), but enabled
11551 *Ben Laurie (CHATS)*
11553 * Eliminate unused copy of key in RC4.
11555 *Ben Laurie (CHATS)*
11557 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11559 *Ben Laurie (CHATS)*
11561 * Fix off-by-one error in EGD path.
11563 *Ben Laurie (CHATS)*
11565 * If RANDFILE path is too long, ignore instead of truncating.
11567 *Ben Laurie (CHATS)*
11569 * Eliminate unused and incorrectly sized X.509 structure
11572 *Ben Laurie (CHATS)*
11574 * Eliminate unused and dangerous function knumber().
11576 *Ben Laurie (CHATS)*
11578 * Eliminate unused and dangerous structure, KSSL_ERR.
11580 *Ben Laurie (CHATS)*
11582 * Protect against overlong session ID context length in an encoded
11583 session object. Since these are local, this does not appear to be
11586 *Ben Laurie (CHATS)*
11588 * Change from security patch (see 0.9.6e below) that did not affect
11589 the 0.9.6 release series:
11591 Remote buffer overflow in SSL3 protocol - an attacker could
11592 supply an oversized master key in Kerberos-enabled versions.
11595 *Ben Laurie (CHATS)*
11597 * Change the SSL kerb5 codes to match RFC 2712.
11601 * Make -nameopt work fully for req and add -reqopt switch.
11603 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11605 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11607 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11609 * Make sure tests can be performed even if the corresponding algorithms
11610 have been removed entirely. This was also the last step to make
11611 OpenSSL compilable with DJGPP under all reasonable conditions.
11613 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11615 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11616 to allow version independent disabling of normally unselected ciphers,
11617 which may be activated as a side-effect of selecting a single cipher.
11619 (E.g., cipher list string "RSA" enables ciphersuites that are left
11620 out of "ALL" because they do not provide symmetric encryption.
11621 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11623 *Lutz Jaenicke, Bodo Moeller*
11625 * Add appropriate support for separate platform-dependent build
11626 directories. The recommended way to make a platform-dependent
11627 build directory is the following (tested on Linux), maybe with
11630 # Place yourself outside of the OpenSSL source tree. In
11631 # this example, the environment variable OPENSSL_SOURCE
11632 # is assumed to contain the absolute OpenSSL source directory.
11633 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11634 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11635 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11636 mkdir -p `dirname $F`
11637 ln -s $OPENSSL_SOURCE/$F $F
11640 To be absolutely sure not to disturb the source tree, a "make clean"
11641 is a good thing. If it isn't successful, don't worry about it,
11642 it probably means the source directory is very clean.
11646 * Make sure any ENGINE control commands make local copies of string
11647 pointers passed to them whenever necessary. Otherwise it is possible
11648 the caller may have overwritten (or deallocated) the original string
11649 data when a later ENGINE operation tries to use the stored values.
11651 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11653 * Improve diagnostics in file reading and command-line digests.
11655 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11657 * Add AES modes CFB and OFB to the object database. Correct an
11658 error in AES-CFB decryption.
11662 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11663 allows existing EVP_CIPHER_CTX structures to be reused after
11664 calling `EVP_*Final()`. This behaviour is used by encryption
11665 BIOs and some applications. This has the side effect that
11666 applications must explicitly clean up cipher contexts with
11667 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11671 * Check the values of dna and dnb in bn_mul_recursive before calling
11672 bn_mul_comba (a non zero value means the a or b arrays do not contain
11673 n2 elements) and fallback to bn_mul_normal if either is not zero.
11677 * Fix escaping of non-ASCII characters when using the -subj option
11678 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11682 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11683 form for "surname", serialNumber has no short form.
11684 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11685 therefore remove "mail" short name for "internet 7".
11686 The OID for unique identifiers in X509 certificates is
11687 x500UniqueIdentifier, not uniqueIdentifier.
11688 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11692 * Add an "init" command to the ENGINE config module and auto initialize
11693 ENGINEs. Without any "init" command the ENGINE will be initialized
11694 after all ctrl commands have been executed on it. If init=1 the
11695 ENGINE is initialized at that point (ctrls before that point are run
11696 on the uninitialized ENGINE and after on the initialized one). If
11697 init=0 then the ENGINE will not be initialized at all.
11701 * Fix the 'app_verify_callback' interface so that the user-defined
11702 argument is actually passed to the callback: In the
11703 SSL_CTX_set_cert_verify_callback() prototype, the callback
11704 declaration has been changed from
11707 int (*cb)(X509_STORE_CTX *,void *);
11708 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11709 i=s->ctx->app_verify_callback(&ctx)
11710 has been changed into
11711 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11713 To update applications using SSL_CTX_set_cert_verify_callback(),
11714 a dummy argument can be added to their callback functions.
11716 *D. K. Smetters <smetters@parc.xerox.com>*
11718 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11720 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11722 * Add and OPENSSL_LOAD_CONF define which will cause
11723 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11724 This allows older applications to transparently support certain
11725 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11726 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11727 load the config file and OPENSSL_add_all_algorithms_conf() which will
11728 always load it have also been added.
11732 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11733 Adjust NIDs and EVP layer.
11735 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11737 * Config modules support in openssl utility.
11739 Most commands now load modules from the config file,
11740 though in a few (such as version) this isn't done
11741 because it couldn't be used for anything.
11743 In the case of ca and req the config file used is
11744 the same as the utility itself: that is the -config
11745 command line option can be used to specify an
11750 * Move default behaviour from OPENSSL_config(). If appname is NULL
11751 use "openssl_conf" if filename is NULL use default openssl config file.
11755 * Add an argument to OPENSSL_config() to allow the use of an alternative
11756 config section name. Add a new flag to tolerate a missing config file
11757 and move code to CONF_modules_load_file().
11761 * Support for crypto accelerator cards from Accelerated Encryption
11762 Processing, www.aep.ie. (Use engine 'aep')
11763 The support was copied from 0.9.6c [engine] and adapted/corrected
11764 to work with the new engine framework.
11766 *AEP Inc. and Richard Levitte*
11768 * Support for SureWare crypto accelerator cards from Baltimore
11769 Technologies. (Use engine 'sureware')
11770 The support was copied from 0.9.6c [engine] and adapted
11771 to work with the new engine framework.
11775 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11776 make the newer ENGINE framework commands for the CHIL engine work.
11778 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11780 * Make it possible to produce shared libraries on ReliantUNIX.
11782 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11784 * Add the configuration target debug-linux-ppro.
11785 Make 'openssl rsa' use the general key loading routines
11786 implemented in `apps.c`, and make those routines able to
11787 handle the key format FORMAT_NETSCAPE and the variant
11790 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11792 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11794 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11796 * Add -keyform to rsautl, and document -engine.
11798 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11800 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11801 BIO_R_NO_SUCH_FILE error code rather than the generic
11802 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11806 * Add new functions
11807 ERR_peek_last_error
11808 ERR_peek_last_error_line
11809 ERR_peek_last_error_line_data.
11810 These are similar to
11812 ERR_peek_error_line
11813 ERR_peek_error_line_data,
11814 but report on the latest error recorded rather than the first one
11815 still in the error queue.
11817 *Ben Laurie, Bodo Moeller*
11819 * default_algorithms option in ENGINE config module. This allows things
11821 default_algorithms = ALL
11822 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11826 * Preliminary ENGINE config module.
11830 * New experimental application configuration code.
11834 * Change the AES code to follow the same name structure as all other
11835 symmetric ciphers, and behave the same way. Move everything to
11836 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11838 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11840 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11842 *Ben Laurie and Theo de Raadt*
11844 * Add option to output public keys in req command.
11846 *Massimiliano Pala madwolf@openca.org*
11848 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11849 (up to about 10% better than before for P-192 and P-224).
11853 * New functions/macros
11855 SSL_CTX_set_msg_callback(ctx, cb)
11856 SSL_CTX_set_msg_callback_arg(ctx, arg)
11857 SSL_set_msg_callback(ssl, cb)
11858 SSL_set_msg_callback_arg(ssl, arg)
11860 to request calling a callback function
11862 void cb(int write_p, int version, int content_type,
11863 const void *buf, size_t len, SSL *ssl, void *arg)
11865 whenever a protocol message has been completely received
11866 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11867 protocol version according to which the SSL library interprets
11868 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11869 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11870 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11871 specification (change_cipher_spec(20), alert(21), handshake(22)).
11872 'buf' and 'len' point to the actual message, 'ssl' to the
11873 SSL object, and 'arg' is the application-defined value set by
11874 SSL[_CTX]_set_msg_callback_arg().
11876 'openssl s_client' and 'openssl s_server' have new '-msg' options
11877 to enable a callback that displays all protocol messages.
11881 * Change the shared library support so shared libraries are built as
11882 soon as the corresponding static library is finished, and thereby get
11883 openssl and the test programs linked against the shared library.
11884 This still only happens when the keyword "shard" has been given to
11885 the configuration scripts.
11887 NOTE: shared library support is still an experimental thing, and
11888 backward binary compatibility is still not guaranteed.
11890 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11892 * Add support for Subject Information Access extension.
11894 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11896 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11897 additional bytes when new memory had to be allocated, not just
11898 when reusing an existing buffer.
11902 * New command line and configuration option 'utf8' for the req command.
11903 This allows field values to be specified as UTF8 strings.
11907 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11908 runs for the former and machine-readable output for the latter.
11912 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11913 of the e-mail address in the DN (i.e., it will go into a certificate
11914 extension only). The new configuration file option 'email_in_dn = no'
11915 has the same effect.
11917 *Massimiliano Pala madwolf@openca.org*
11919 * Change all functions with names starting with `des_` to be starting
11920 with `DES_` instead. Add wrappers that are compatible with libdes,
11921 but are named `_ossl_old_des_*`. Finally, add macros that map the
11922 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11923 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11924 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11927 Since we provide two compatibility mappings, the user needs to
11928 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11929 compatibility is desired. The default (i.e., when that macro
11930 isn't defined) is OpenSSL 0.9.6c compatibility.
11932 There are also macros that enable and disable the support of old
11933 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11934 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11935 are defined, the default will apply: to support the old des routines.
11937 In either case, one must include openssl/des.h to get the correct
11938 definitions. Do not try to just include openssl/des_old.h, that
11941 NOTE: This is a major break of an old API into a new one. Software
11942 authors are encouraged to switch to the `DES_` style functions. Some
11943 time in the future, des_old.h and the libdes compatibility functions
11944 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11945 default), and then completely removed.
11949 * Test for certificates which contain unsupported critical extensions.
11950 If such a certificate is found during a verify operation it is
11951 rejected by default: this behaviour can be overridden by either
11952 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11953 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11954 X509_supported_extension() has also been added which returns 1 if a
11955 particular extension is supported.
11959 * Modify the behaviour of EVP cipher functions in similar way to digests
11960 to retain compatibility with existing code.
11964 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11965 compatibility with existing code. In particular the 'ctx' parameter does
11966 not have to be to be initialized before the call to EVP_DigestInit() and
11967 it is tidied up after a call to EVP_DigestFinal(). New function
11968 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11969 EVP_MD_CTX_copy() changed to not require the destination to be
11970 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11971 requires the destination to be valid.
11973 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11974 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11978 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11979 so that complete 'Handshake' protocol structures are kept in memory
11980 instead of overwriting 'msg_type' and 'length' with 'body' data.
11984 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11986 *Massimo Santin via Richard Levitte*
11988 * Major restructuring to the underlying ENGINE code. This includes
11989 reduction of linker bloat, separation of pure "ENGINE" manipulation
11990 (initialisation, etc) from functionality dealing with implementations
11991 of specific crypto interfaces. This change also introduces integrated
11992 support for symmetric ciphers and digest implementations - so ENGINEs
11993 can now accelerate these by providing EVP_CIPHER and EVP_MD
11994 implementations of their own. This is detailed in
11995 [crypto/engine/README.md](crypto/engine/README.md)
11996 as it couldn't be adequately described here. However, there are a few
11997 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11998 were changed in the original introduction of ENGINE code have now
11999 reverted back - the hooking from this code to ENGINE is now a good
12000 deal more passive and at run-time, operations deal directly with
12001 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
12002 dereferencing through an ENGINE pointer any more. Also, the ENGINE
12003 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
12004 they were not being used by the framework as there is no concept of a
12005 BIGNUM_METHOD and they could not be generalised to the new
12006 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
12007 ENGINE_cpy() has been removed as it cannot be consistently defined in
12012 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
12016 * Change mkdef.pl to sort symbols that get the same entry number,
12017 and make sure the automatically generated functions `ERR_load_*`
12018 become part of libeay.num as well.
12022 * New function SSL_renegotiate_pending(). This returns true once
12023 renegotiation has been requested (either SSL_renegotiate() call
12024 or HelloRequest/ClientHello received from the peer) and becomes
12025 false once a handshake has been completed.
12026 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
12027 sends a HelloRequest, but does not ensure that a handshake takes
12028 place. SSL_renegotiate_pending() is useful for checking if the
12029 client has followed the request.)
12033 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
12034 By default, clients may request session resumption even during
12035 renegotiation (if session ID contexts permit); with this option,
12036 session resumption is possible only in the first handshake.
12038 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
12039 more bits available for options that should not be part of
12040 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
12044 * Add some demos for certificate and certificate request creation.
12048 * Make maximum certificate chain size accepted from the peer application
12049 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12050 "Douglas E. Engert" <deengert@anl.gov>.
12054 * Add support for shared libraries for Unixware-7
12055 (Boyd Lynn Gerber <gerberb@zenez.com>).
12059 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12060 be done prior to destruction. Use this to unload error strings from
12061 ENGINEs that load their own error strings. NB: This adds two new API
12062 functions to "get" and "set" this destroy handler in an ENGINE.
12066 * Alter all existing ENGINE implementations (except "openssl" and
12067 "openbsd") to dynamically instantiate their own error strings. This
12068 makes them more flexible to be built both as statically-linked ENGINEs
12069 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12070 Also, add stub code to each that makes building them as self-contained
12071 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12075 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12076 implementations into applications that are completely implemented in
12077 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12078 commands that can be used to configure what shared-library to load and
12079 to control aspects of the way it is handled. Also, made an update to
12080 the [README-Engine.md](README-Engine.md) file
12081 that brings its information up-to-date and
12082 provides some information and instructions on the "dynamic" ENGINE
12083 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12087 * Make it possible to unload ranges of ERR strings with a new
12088 "ERR_unload_strings" function.
12092 * Add a copy() function to EVP_MD.
12096 * Make EVP_MD routines take a context pointer instead of just the
12097 md_data void pointer.
12101 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12102 that the digest can only process a single chunk of data
12103 (typically because it is provided by a piece of
12104 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12105 is only going to provide a single chunk of data, and hence the
12106 framework needn't accumulate the data for oneshot drivers.
12110 * As with "ERR", make it possible to replace the underlying "ex_data"
12111 functions. This change also alters the storage and management of global
12112 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12113 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12114 index counters. The API functions that use this state have been changed
12115 to take a "class_index" rather than pointers to the class's local STACK
12116 and counter, and there is now an API function to dynamically create new
12117 classes. This centralisation allows us to (a) plug a lot of the
12118 thread-safety problems that existed, and (b) makes it possible to clean
12119 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12120 such data would previously have always leaked in application code and
12121 workarounds were in place to make the memory debugging turn a blind eye
12122 to it. Application code that doesn't use this new function will still
12123 leak as before, but their memory debugging output will announce it now
12124 rather than letting it slide.
12126 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12127 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12128 has a return value to indicate success or failure.
12132 * Make it possible to replace the underlying "ERR" functions such that the
12133 global state (2 LHASH tables and 2 locks) is only used by the "default"
12134 implementation. This change also adds two functions to "get" and "set"
12135 the implementation prior to it being automatically set the first time
12136 any other ERR function takes place. Ie. an application can call "get",
12137 pass the return value to a module it has just loaded, and that module
12138 can call its own "set" function using that value. This means the
12139 module's "ERR" operations will use (and modify) the error state in the
12140 application and not in its own statically linked copy of OpenSSL code.
12144 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12145 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12146 the operation, and provides a more encapsulated way for external code
12147 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12148 to use these functions rather than manually incrementing the counts.
12150 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12154 * Add EVP test program.
12158 * Add symmetric cipher support to ENGINE. Expect the API to change!
12162 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12163 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12164 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12165 These allow a CRL to be built without having to access X509_CRL fields
12166 directly. Modify 'ca' application to use new functions.
12170 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12171 bug workarounds. Rollback attack detection is a security feature.
12172 The problem will only arise on OpenSSL servers when TLSv1 is not
12173 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12174 Software authors not wanting to support TLSv1 will have special reasons
12175 for their choice and can explicitly enable this option.
12177 *Bodo Moeller, Lutz Jaenicke*
12179 * Rationalise EVP so it can be extended: don't include a union of
12180 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12181 (similar to those existing for EVP_CIPHER_CTX).
12186 EVP_MD_CTX_init(&md); /* new function call */
12187 EVP_DigestInit(&md, EVP_sha1());
12188 EVP_DigestUpdate(&md, in, len);
12189 EVP_DigestFinal(&md, out, NULL);
12190 EVP_MD_CTX_cleanup(&md); /* new function call */
12194 * Make DES key schedule conform to the usual scheme, as well as
12195 correcting its structure. This means that calls to DES functions
12196 now have to pass a pointer to a des_key_schedule instead of a
12197 plain des_key_schedule (which was actually always a pointer
12200 des_key_schedule ks;
12202 des_set_key_checked(..., &ks);
12203 des_ncbc_encrypt(..., &ks, ...);
12205 (Note that a later change renames 'des_...' into 'DES_...'.)
12209 * Initial reduction of linker bloat: the use of some functions, such as
12210 PEM causes large amounts of unused functions to be linked in due to
12211 poor organisation. For example pem_all.c contains every PEM function
12212 which has a knock on effect of linking in large amounts of (unused)
12213 ASN1 code. Grouping together similar functions and splitting unrelated
12214 functions prevents this.
12218 * Cleanup of EVP macros.
12222 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12223 correct `_ecb suffix`.
12227 * Add initial OCSP responder support to ocsp application. The
12228 revocation information is handled using the text based index
12229 use by the ca application. The responder can either handle
12230 requests generated internally, supplied in files (for example
12231 via a CGI script) or using an internal minimal server.
12235 * Add configuration choices to get zlib compression for TLS.
12239 * Changes to Kerberos SSL for RFC 2712 compliance:
12240 1. Implemented real KerberosWrapper, instead of just using
12241 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12242 2. Implemented optional authenticator field of KerberosWrapper.
12244 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12245 and authenticator structs; see crypto/krb5/.
12247 Generalized Kerberos calls to support multiple Kerberos libraries.
12248 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12249 via Richard Levitte*
12251 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12252 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12253 values for each of the key sizes rather than having just
12254 parameters (and 'speed' generating keys each time).
12258 * Speed up EVP routines.
12261 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12262 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12263 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12264 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12266 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12267 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12268 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12271 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12273 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12277 * Added the OS2-EMX target.
12279 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12281 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12282 New functions to support `NCONF `routines in extension code.
12283 New function `CONF_set_nconf()`
12284 to allow functions which take an `NCONF` to also handle the old `LHASH`
12285 structure: this means that the old `CONF` compatible routines can be
12286 retained (in particular w.rt. extensions) without having to duplicate the
12287 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12291 * Enhance the general user interface with mechanisms for inner control
12292 and with possibilities to have yes/no kind of prompts.
12296 * Change all calls to low level digest routines in the library and
12297 applications to use EVP. Add missing calls to HMAC_cleanup() and
12298 don't assume HMAC_CTX can be copied using memcpy().
12300 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12302 * Add the possibility to control engines through control names but with
12303 arbitrary arguments instead of just a string.
12304 Change the key loaders to take a UI_METHOD instead of a callback
12305 function pointer. NOTE: this breaks binary compatibility with earlier
12306 versions of OpenSSL [engine].
12307 Adapt the nCipher code for these new conditions and add a card insertion
12312 * Enhance the general user interface with mechanisms to better support
12313 dialog box interfaces, application-defined prompts, the possibility
12314 to use defaults (for example default passwords from somewhere else)
12315 and interrupts/cancellations.
12319 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12320 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12324 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12325 tidy up some unnecessarily weird code in 'sk_new()').
12327 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12329 * Change the key loading routines for ENGINEs to use the same kind
12330 callback (pem_password_cb) as all other routines that need this
12335 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12336 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12337 than this minimum value is recommended.
12341 * New random seeder for OpenVMS, using the system process statistics
12342 that are easily reachable.
12346 * Windows apparently can't transparently handle global
12347 variables defined in DLLs. Initialisations such as:
12349 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12351 won't compile. This is used by the any applications that need to
12352 declare their own ASN1 modules. This was fixed by adding the option
12353 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12354 needed for static libraries under Win32.
12358 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12359 setting of purpose and trust fields. New X509_STORE trust and
12360 purpose functions and tidy up setting in other SSL functions.
12364 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12365 structure. These are inherited by X509_STORE_CTX when it is
12366 initialised. This allows various defaults to be set in the
12367 X509_STORE structure (such as flags for CRL checking and custom
12368 purpose or trust settings) for functions which only use X509_STORE_CTX
12369 internally such as S/MIME.
12371 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12372 trust settings if they are not set in X509_STORE. This allows X509_STORE
12373 purposes and trust (in S/MIME for example) to override any set by default.
12375 Add command line options for CRL checking to smime, s_client and s_server
12380 * Initial CRL based revocation checking. If the CRL checking flag(s)
12381 are set then the CRL is looked up in the X509_STORE structure and
12382 its validity and signature checked, then if the certificate is found
12383 in the CRL the verify fails with a revoked error.
12385 Various new CRL related callbacks added to X509_STORE_CTX structure.
12387 Command line options added to 'verify' application to support this.
12389 This needs some additional work, such as being able to handle multiple
12390 CRLs with different times, extension based lookup (rather than just
12391 by subject name) and ultimately more complete V2 CRL extension
12396 * Add a general user interface API (crypto/ui/). This is designed
12397 to replace things like des_read_password and friends (backward
12398 compatibility functions using this new API are provided).
12399 The purpose is to remove prompting functions from the DES code
12400 section as well as provide for prompting through dialog boxes in
12401 a window system and the like.
12405 * Add "ex_data" support to ENGINE so implementations can add state at a
12406 per-structure level rather than having to store it globally.
12410 * Make it possible for ENGINE structures to be copied when retrieved by
12411 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12412 This causes the "original" ENGINE structure to act like a template,
12413 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12414 operational state can be localised to each ENGINE structure, despite the
12415 fact they all share the same "methods". New ENGINE structures returned in
12416 this case have no functional references and the return value is the single
12417 structural reference. This matches the single structural reference returned
12418 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12423 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12424 needs to match any other type at all we need to manually clear the
12429 * Changes to the "openssl engine" utility to include;
12430 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12431 about an ENGINE's available control commands.
12432 - executing control commands from command line arguments using the
12433 '-pre' and '-post' switches. '-post' is only used if '-t' is
12434 specified and the ENGINE is successfully initialised. The syntax for
12435 the individual commands are colon-separated, for example;
12436 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12440 * New dynamic control command support for ENGINEs. ENGINEs can now
12441 declare their own commands (numbers), names (strings), descriptions,
12442 and input types for run-time discovery by calling applications. A
12443 subset of these commands are implicitly classed as "executable"
12444 depending on their input type, and only these can be invoked through
12445 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12446 can be based on user input, config files, etc). The distinction is
12447 that "executable" commands cannot return anything other than a boolean
12448 result and can only support numeric or string input, whereas some
12449 discoverable commands may only be for direct use through
12450 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12451 pointers, or other custom uses. The "executable" commands are to
12452 support parameterisations of ENGINE behaviour that can be
12453 unambiguously defined by ENGINEs and used consistently across any
12454 OpenSSL-based application. Commands have been added to all the
12455 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12456 control over shared-library paths without source code alterations.
12460 * Changed all ENGINE implementations to dynamically allocate their
12461 ENGINEs rather than declaring them statically. Apart from this being
12462 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12463 this also allows the implementations to compile without using the
12464 internal engine_int.h header.
12468 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12469 'const' value. Any code that should be able to modify a RAND_METHOD
12470 should already have non-const pointers to it (ie. they should only
12471 modify their own ones).
12475 * Made a variety of little tweaks to the ENGINE code.
12476 - "atalla" and "ubsec" string definitions were moved from header files
12477 to C code. "nuron" string definitions were placed in variables
12478 rather than hard-coded - allowing parameterisation of these values
12479 later on via ctrl() commands.
12480 - Removed unused "#if 0"'d code.
12481 - Fixed engine list iteration code so it uses ENGINE_free() to release
12482 structural references.
12483 - Constified the RAND_METHOD element of ENGINE structures.
12484 - Constified various get/set functions as appropriate and added
12485 missing functions (including a catch-all ENGINE_cpy that duplicates
12486 all ENGINE values onto a new ENGINE except reference counts/state).
12487 - Removed NULL parameter checks in get/set functions. Setting a method
12488 or function to NULL is a way of cancelling out a previously set
12489 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12490 and doesn't justify the extra error symbols and code.
12491 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12492 flags from engine_int.h to engine.h.
12493 - Changed prototypes for ENGINE handler functions (init(), finish(),
12494 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12498 * Implement binary inversion algorithm for BN_mod_inverse in addition
12499 to the algorithm using long division. The binary algorithm can be
12500 used only if the modulus is odd. On 32-bit systems, it is faster
12501 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12502 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12503 up to 450 bits. In 64-bit environments, the binary algorithm
12504 appears to be advantageous for much longer moduli; here we use it
12505 for moduli up to 2048 bits.
12509 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12510 could not support the combine flag in choice fields.
12514 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12515 extensions from a certificate request to the certificate.
12519 * Allow multiple 'certopt' and 'nameopt' options to be separated
12520 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12521 file: this allows the display of the certificate about to be
12522 signed to be customised, to allow certain fields to be included
12523 or excluded and extension details. The old system didn't display
12524 multicharacter strings properly, omitted fields not in the policy
12525 and couldn't display additional details such as extensions.
12529 * Function EC_POINTs_mul for multiple scalar multiplication
12530 of an arbitrary number of elliptic curve points
12531 \sum scalars[i]*points[i],
12532 optionally including the generator defined for the EC_GROUP:
12533 scalar*generator + \sum scalars[i]*points[i].
12535 EC_POINT_mul is a simple wrapper function for the typical case
12536 that the point list has just one item (besides the optional
12541 * First EC_METHODs for curves over GF(p):
12543 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12544 operations and provides various method functions that can also
12545 operate with faster implementations of modular arithmetic.
12547 EC_GFp_mont_method() reuses most functions that are part of
12548 EC_GFp_simple_method, but uses Montgomery arithmetic.
12550 *Bodo Moeller; point addition and point doubling
12551 implementation directly derived from source code provided by
12552 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12554 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12555 crypto/ec/ec_lib.c):
12557 Curves are EC_GROUP objects (with an optional group generator)
12558 based on EC_METHODs that are built into the library.
12560 Points are EC_POINT objects based on EC_GROUP objects.
12562 Most of the framework would be able to handle curves over arbitrary
12563 finite fields, but as there are no obvious types for fields other
12564 than GF(p), some functions are limited to that for now.
12568 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12569 that the file contains a complete HTTP response.
12573 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12574 change the def and num file printf format specifier from "%-40sXXX"
12575 to "%-39s XXX". The latter will always guarantee a space after the
12576 field while the former will cause them to run together if the field
12577 is 40 of more characters long.
12581 * Constify the cipher and digest 'method' functions and structures
12582 and modify related functions to take constant EVP_MD and EVP_CIPHER
12587 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12588 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12592 * Modify `EVP_Digest*()` routines so they now return values. Although the
12593 internal software routines can never fail additional hardware versions
12598 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12600 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12601 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12604 ERR_R_NESTED_ASN1_ERROR
12606 ERR_R_MISSING_ASN1_EOS
12607 were 4 .. 9, conflicting with
12608 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12610 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12611 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12613 Add new error code 'ERR_R_INTERNAL_ERROR'.
12617 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12622 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12623 sets the subject name for a new request or supersedes the
12624 subject name in a given request. Formats that can be parsed are
12625 'CN=Some Name, OU=myOU, C=IT'
12627 'CN=Some Name/OU=myOU/C=IT'.
12629 Add options '-batch' and '-verbose' to 'openssl req'.
12631 *Massimiliano Pala <madwolf@hackmasters.net>*
12633 * Introduce the possibility to access global variables through
12634 functions on platform were that's the best way to handle exporting
12635 global variables in shared libraries. To enable this functionality,
12636 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12637 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12638 is normally done by Configure or something similar).
12640 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12641 in the source file (foo.c) like this:
12643 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12644 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12646 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12647 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12649 OPENSSL_DECLARE_GLOBAL(int,foo);
12650 #define foo OPENSSL_GLOBAL_REF(foo)
12651 OPENSSL_DECLARE_GLOBAL(double,bar);
12652 #define bar OPENSSL_GLOBAL_REF(bar)
12654 The #defines are very important, and therefore so is including the
12655 header file everywhere where the defined globals are used.
12657 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12658 of ASN.1 items, but that structure is a bit different.
12660 The largest change is in util/mkdef.pl which has been enhanced with
12661 better and easier to understand logic to choose which symbols should
12662 go into the Windows .def files as well as a number of fixes and code
12663 cleanup (among others, algorithm keywords are now sorted
12664 lexicographically to avoid constant rewrites).
12668 * In BN_div() keep a copy of the sign of 'num' before writing the
12669 result to 'rm' because if rm==num the value will be overwritten
12670 and produce the wrong result if 'num' is negative: this caused
12671 problems with BN_mod() and BN_nnmod().
12675 * Function OCSP_request_verify(). This checks the signature on an
12676 OCSP request and verifies the signer certificate. The signer
12677 certificate is just checked for a generic purpose and OCSP request
12682 * Add OCSP_check_validity() function to check the validity of OCSP
12683 responses. OCSP responses are prepared in real time and may only
12684 be a few seconds old. Simply checking that the current time lies
12685 between thisUpdate and nextUpdate max reject otherwise valid responses
12686 caused by either OCSP responder or client clock inaccuracy. Instead
12687 we allow thisUpdate and nextUpdate to fall within a certain period of
12688 the current time. The age of the response can also optionally be
12689 checked. Two new options -validity_period and -status_age added to
12694 * If signature or public key algorithm is unrecognized print out its
12695 OID rather that just UNKNOWN.
12699 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12700 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12701 ID to be generated from the issuer certificate alone which can then be
12702 passed to OCSP_id_issuer_cmp().
12706 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12707 ASN1 modules to export functions returning ASN1_ITEM pointers
12708 instead of the ASN1_ITEM structures themselves. This adds several
12709 new macros which allow the underlying ASN1 function/structure to
12710 be accessed transparently. As a result code should not use ASN1_ITEM
12711 references directly (such as &X509_it) but instead use the relevant
12712 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12713 use of the new ASN1 code on platforms where exporting structures
12714 is problematical (for example in shared libraries) but exporting
12715 functions returning pointers to structures is not.
12719 * Add support for overriding the generation of SSL/TLS session IDs.
12720 These callbacks can be registered either in an SSL_CTX or per SSL.
12721 The purpose of this is to allow applications to control, if they wish,
12722 the arbitrary values chosen for use as session IDs, particularly as it
12723 can be useful for session caching in multiple-server environments. A
12724 command-line switch for testing this (and any client code that wishes
12725 to use such a feature) has been added to "s_server".
12727 *Geoff Thorpe, Lutz Jaenicke*
12729 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12730 of the form `#if defined(...) || defined(...) || ...` and
12731 `#if !defined(...) && !defined(...) && ...`. This also avoids
12732 the growing number of special cases it was previously handling.
12736 * Make all configuration macros available for application by making
12737 sure they are available in opensslconf.h, by giving them names starting
12738 with `OPENSSL_` to avoid conflicts with other packages and by making
12739 sure e_os2.h will cover all platform-specific cases together with
12741 Additionally, it is now possible to define configuration/platform-
12742 specific names (called "system identities"). In the C code, these
12743 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12744 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12745 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12750 * New option -set_serial to 'req' and 'x509' this allows the serial
12751 number to use to be specified on the command line. Previously self
12752 signed certificates were hard coded with serial number 0 and the
12753 CA options of 'x509' had to use a serial number in a file which was
12758 * New options to 'ca' utility to support V2 CRL entry extensions.
12759 Currently CRL reason, invalidity date and hold instruction are
12760 supported. Add new CRL extensions to V3 code and some new objects.
12764 * New function EVP_CIPHER_CTX_set_padding() this is used to
12765 disable standard block padding (aka PKCS#5 padding) in the EVP
12766 API, which was previously mandatory. This means that the data is
12767 not padded in any way and so the total length much be a multiple
12768 of the block size, otherwise an error occurs.
12772 * Initial (incomplete) OCSP SSL support.
12776 * New function OCSP_parse_url(). This splits up a URL into its host,
12777 port and path components: primarily to parse OCSP URLs. New -url
12778 option to ocsp utility.
12782 * New nonce behavior. The return value of OCSP_check_nonce() now
12783 reflects the various checks performed. Applications can decide
12784 whether to tolerate certain situations such as an absent nonce
12785 in a response when one was present in a request: the ocsp application
12786 just prints out a warning. New function OCSP_add1_basic_nonce()
12787 this is to allow responders to include a nonce in a response even if
12788 the request is nonce-less.
12792 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12793 skipped when using openssl x509 multiple times on a single input file,
12794 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12798 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12799 set string type: to handle setting ASN1_TIME structures. Fix ca
12800 utility to correctly initialize revocation date of CRLs.
12804 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12805 the clients preferred ciphersuites and rather use its own preferences.
12806 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12807 Internet Explorer by ensuring unchanged hash method during stepup.
12808 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12812 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12813 to aes and add a new 'exist' option to print out symbols that don't
12818 * Additional options to ocsp utility to allow flags to be set and
12819 additional certificates supplied.
12823 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12824 OCSP client a number of certificate to only verify the response
12829 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12830 handle the new API. Currently only ECB, CBC modes supported. Add new
12833 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12834 Encryption Standard (AES) Ciphersuites for Transport Layer
12835 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12836 not enabled by default and were not part of the "ALL" ciphersuite
12837 alias because they were not yet official; they could be
12838 explicitly requested by specifying the "AESdraft" ciphersuite
12839 group alias. In the final release of OpenSSL 0.9.7, the group
12840 alias is called "AES" and is part of "ALL".)
12842 *Ben Laurie, Steve Henson, Bodo Moeller*
12844 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12845 request to response.
12849 * Functions for OCSP responders. OCSP_request_onereq_count(),
12850 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12851 extract information from a certificate request. OCSP_response_create()
12852 creates a response and optionally adds a basic response structure.
12853 OCSP_basic_add1_status() adds a complete single response to a basic
12854 response and returns the OCSP_SINGLERESP structure just added (to allow
12855 extensions to be included for example). OCSP_basic_add1_cert() adds a
12856 certificate to a basic response and OCSP_basic_sign() signs a basic
12857 response with various flags. New helper functions ASN1_TIME_check()
12858 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12859 (converts ASN1_TIME to GeneralizedTime).
12863 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12864 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12865 structure from a certificate. X509_pubkey_digest() digests the public_key
12866 contents: this is used in various key identifiers.
12870 * Make sk_sort() tolerate a NULL argument.
12872 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12874 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12875 passed by the function are trusted implicitly. If any of them signed the
12876 response then it is assumed to be valid and is not verified.
12880 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12881 to data. This was previously part of the PKCS7 ASN1 code. This
12882 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12883 *Steve Henson, reported by Kenneth R. Robinette
12884 <support@securenetterm.com>*
12886 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12887 routines: without these tracing memory leaks is very painful.
12888 Fix leaks in PKCS12 and PKCS7 routines.
12892 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12893 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12894 effectively meant GeneralizedTime would never be used. Now it
12895 is initialised to -1 but X509_time_adj() now has to check the value
12896 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12897 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12898 *Steve Henson, reported by Kenneth R. Robinette
12899 <support@securenetterm.com>*
12901 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12902 result in a zero length in the ASN1_INTEGER structure which was
12903 not consistent with the structure when d2i_ASN1_INTEGER() was used
12904 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12905 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12906 where it did not print out a minus for negative ASN1_INTEGER.
12910 * Add summary printout to ocsp utility. The various functions which
12911 convert status values to strings have been renamed to:
12912 OCSP_response_status_str(), OCSP_cert_status_str() and
12913 OCSP_crl_reason_str() and are no longer static. New options
12914 to verify nonce values and to disable verification. OCSP response
12915 printout format cleaned up.
12919 * Add additional OCSP certificate checks. These are those specified
12920 in RFC2560. This consists of two separate checks: the CA of the
12921 certificate being checked must either be the OCSP signer certificate
12922 or the issuer of the OCSP signer certificate. In the latter case the
12923 OCSP signer certificate must contain the OCSP signing extended key
12924 usage. This check is performed by attempting to match the OCSP
12925 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12926 in the OCSP_CERTID structures of the response.
12930 * Initial OCSP certificate verification added to OCSP_basic_verify()
12931 and related routines. This uses the standard OpenSSL certificate
12932 verify routines to perform initial checks (just CA validity) and
12933 to obtain the certificate chain. Then additional checks will be
12934 performed on the chain. Currently the root CA is checked to see
12935 if it is explicitly trusted for OCSP signing. This is used to set
12936 a root CA as a global signing root: that is any certificate that
12937 chains to that CA is an acceptable OCSP signing certificate.
12941 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12942 extensions from a separate configuration file.
12943 As when reading extensions from the main configuration file,
12944 the '-extensions ...' option may be used for specifying the
12947 *Massimiliano Pala <madwolf@comune.modena.it>*
12949 * New OCSP utility. Allows OCSP requests to be generated or
12950 read. The request can be sent to a responder and the output
12951 parsed, outputted or printed in text form. Not complete yet:
12952 still needs to check the OCSP response validity.
12956 * New subcommands for 'openssl ca':
12957 `openssl ca -status <serial>` prints the status of the cert with
12958 the given serial number (according to the index file).
12959 `openssl ca -updatedb` updates the expiry status of certificates
12962 *Massimiliano Pala <madwolf@comune.modena.it>*
12964 * New '-newreq-nodes' command option to CA.pl. This is like
12965 '-newreq', but calls 'openssl req' with the '-nodes' option
12966 so that the resulting key is not encrypted.
12968 *Damien Miller <djm@mindrot.org>*
12970 * New configuration for the GNU Hurd.
12972 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12974 * Initial code to implement OCSP basic response verify. This
12975 is currently incomplete. Currently just finds the signer's
12976 certificate and verifies the signature on the response.
12980 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12981 value of OPENSSLDIR. This is available via the new '-d' option
12982 to 'openssl version', and is also included in 'openssl version -a'.
12986 * Allowing defining memory allocation callbacks that will be given
12987 file name and line number information in additional arguments
12988 (a `const char*` and an int). The basic functionality remains, as
12989 well as the original possibility to just replace malloc(),
12990 realloc() and free() by functions that do not know about these
12991 additional arguments. To register and find out the current
12992 settings for extended allocation functions, the following
12993 functions are provided:
12995 CRYPTO_set_mem_ex_functions
12996 CRYPTO_set_locked_mem_ex_functions
12997 CRYPTO_get_mem_ex_functions
12998 CRYPTO_get_locked_mem_ex_functions
13000 These work the same way as CRYPTO_set_mem_functions and friends.
13001 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
13002 extended allocation function is enabled.
13003 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
13004 a conventional allocation function is enabled.
13006 *Richard Levitte, Bodo Moeller*
13008 * Finish off removing the remaining LHASH function pointer casts.
13009 There should no longer be any prototype-casting required when using
13010 the LHASH abstraction, and any casts that remain are "bugs". See
13011 the callback types and macros at the head of lhash.h for details
13012 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
13016 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
13017 If /dev/[u]random devices are not available or do not return enough
13018 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
13020 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
13021 /etc/entropy will be queried once each in this sequence, querying stops
13022 when enough entropy was collected without querying more sockets.
13026 * Change the Unix RAND_poll() variant to be able to poll several
13027 random devices, as specified by DEVRANDOM, until a sufficient amount
13028 of data has been collected. We spend at most 10 ms on each file
13029 (select timeout) and read in non-blocking mode. DEVRANDOM now
13030 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
13031 (previously it was just the string "/dev/urandom"), so on typical
13032 platforms the 10 ms delay will never occur.
13033 Also separate out the Unix variant to its own file, rand_unix.c.
13034 For VMS, there's a currently-empty rand_vms.c.
13038 * Move OCSP client related routines to ocsp_cl.c. These
13039 provide utility functions which an application needing
13040 to issue a request to an OCSP responder and analyse the
13041 response will typically need: as opposed to those which an
13042 OCSP responder itself would need which will be added later.
13044 OCSP_request_sign() signs an OCSP request with an API similar
13045 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13046 response. OCSP_response_get1_basic() extracts basic response
13047 from response. OCSP_resp_find_status(): finds and extracts status
13048 information from an OCSP_CERTID structure (which will be created
13049 when the request structure is built). These are built from lower
13050 level functions which work on OCSP_SINGLERESP structures but
13051 won't normally be used unless the application wishes to examine
13052 extensions in the OCSP response for example.
13054 Replace nonce routines with a pair of functions.
13055 OCSP_request_add1_nonce() adds a nonce value and optionally
13056 generates a random value. OCSP_check_nonce() checks the
13057 validity of the nonce in an OCSP response.
13061 * Change function OCSP_request_add() to OCSP_request_add0_id().
13062 This doesn't copy the supplied OCSP_CERTID and avoids the
13063 need to free up the newly created id. Change return type
13064 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13065 This can then be used to add extensions to the request.
13066 Deleted OCSP_request_new(), since most of its functionality
13067 is now in OCSP_REQUEST_new() (and the case insensitive name
13068 clash) apart from the ability to set the request name which
13069 will be added elsewhere.
13073 * Update OCSP API. Remove obsolete extensions argument from
13074 various functions. Extensions are now handled using the new
13075 OCSP extension code. New simple OCSP HTTP function which
13076 can be used to send requests and parse the response.
13080 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13081 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13082 uses the special reorder version of SET OF to sort the attributes
13083 and reorder them to match the encoded order. This resolves a long
13084 standing problem: a verify on a PKCS7 structure just after signing
13085 it used to fail because the attribute order did not match the
13086 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13087 it uses the received order. This is necessary to tolerate some broken
13088 software that does not order SET OF. This is handled by encoding
13089 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13090 to produce the required SET OF.
13094 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13095 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13096 files to get correct declarations of the ASN.1 item variables.
13100 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13101 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13102 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13103 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13104 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13105 ASN1_ITEM and no wrapper functions.
13109 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13110 replace the old function pointer based I/O routines. Change most of
13111 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13115 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13116 lines, recognize more "algorithms" that can be deselected, and make
13117 it complain about algorithm deselection that isn't recognised.
13121 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13122 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13123 to use new functions. Add NO_ASN1_OLD which can be set to remove
13124 some old style ASN1 functions: this can be used to determine if old
13125 code will still work when these eventually go away.
13129 * New extension functions for OCSP structures, these follow the
13130 same conventions as certificates and CRLs.
13134 * New function X509V3_add1_i2d(). This automatically encodes and
13135 adds an extension. Its behaviour can be customised with various
13136 flags to append, replace or delete. Various wrappers added for
13137 certificates and CRLs.
13141 * Fix to avoid calling the underlying ASN1 print routine when
13142 an extension cannot be parsed. Correct a typo in the
13143 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13147 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13148 entries for variables.
13152 * Add functionality to `apps/openssl.c` for detecting locking
13153 problems: As the program is single-threaded, all we have
13154 to do is register a locking callback using an array for
13155 storing which locks are currently held by the program.
13159 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13160 SSL_get_ex_data_X509_STORE_idx(), which is used in
13161 ssl_verify_cert_chain() and thus can be called at any time
13162 during TLS/SSL handshakes so that thread-safety is essential.
13163 Unfortunately, the ex_data design is not at all suited
13164 for multi-threaded use, so it probably should be abolished.
13168 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13170 *Broadcom, tweaked and integrated by Geoff Thorpe*
13172 * Move common extension printing code to new function
13173 X509V3_print_extensions(). Reorganise OCSP print routines and
13174 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13178 * New function X509_signature_print() to remove duplication in some
13183 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13184 set (this was treated exactly the same as SET OF previously). This
13185 is used to reorder the STACK representing the structure to match the
13186 encoding. This will be used to get round a problem where a PKCS7
13187 structure which was signed could not be verified because the STACK
13188 order did not reflect the encoded order.
13192 * Reimplement the OCSP ASN1 module using the new code.
13196 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13197 for its ASN1 operations. The old style function pointers still exist
13198 for now but they will eventually go away.
13202 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13203 completely replaces the old ASN1 functionality with a table driven
13204 encoder and decoder which interprets an ASN1_ITEM structure describing
13205 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13206 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13207 has also been converted to the new form.
13211 * Change BN_mod_exp_recp so that negative moduli are tolerated
13212 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13213 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13214 for negative moduli.
13218 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13219 of not touching the result's sign bit.
13223 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13228 * Changed the LHASH code to use prototypes for callbacks, and created
13229 macros to declare and implement thin (optionally static) functions
13230 that provide type-safety and avoid function pointer casting for the
13231 type-specific callbacks.
13235 * Added Kerberos Cipher Suites to be used with TLS, as written in
13237 *Veers Staats <staatsvr@asc.hpc.mil>,
13238 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13240 * Reformat the FAQ so the different questions and answers can be divided
13241 in sections depending on the subject.
13245 * Have the zlib compression code load ZLIB.DLL dynamically under
13250 * New function BN_mod_sqrt for computing square roots modulo a prime
13251 (using the probabilistic Tonelli-Shanks algorithm unless
13252 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13253 be handled deterministically).
13255 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13257 * Make BN_mod_inverse faster by explicitly handling small quotients
13258 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13259 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13263 * New function BN_kronecker.
13267 * Fix BN_gcd so that it works on negative inputs; the result is
13268 positive unless both parameters are zero.
13269 Previously something reasonably close to an infinite loop was
13270 possible because numbers could be growing instead of shrinking
13271 in the implementation of Euclid's algorithm.
13275 * Fix BN_is_word() and BN_is_one() macros to take into account the
13276 sign of the number in question.
13278 Fix BN_is_word(a,w) to work correctly for w == 0.
13280 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13281 because its test if the absolute value of 'a' equals 'w'.
13282 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13283 it exists mostly for use in the implementations of BN_is_zero(),
13284 BN_is_one(), and BN_is_word().
13288 * New function BN_swap.
13292 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13293 the exponentiation functions are more likely to produce reasonable
13294 results on negative inputs.
13298 * Change BN_mod_mul so that the result is always non-negative.
13299 Previously, it could be negative if one of the factors was negative;
13300 I don't think anyone really wanted that behaviour.
13304 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13305 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13306 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13307 and add new functions:
13316 BN_mod_lshift1_quick
13318 BN_mod_lshift_quick
13320 These functions always generate non-negative results.
13322 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13323 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13325 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13326 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13327 be reduced modulo `m`.
13329 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13332 The following entry accidentally appeared in the CHANGES file
13333 distributed with OpenSSL 0.9.7. The modifications described in
13334 it do *not* apply to OpenSSL 0.9.7.
13336 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13337 was actually never needed) and in BN_mul(). The removal in BN_mul()
13338 required a small change in bn_mul_part_recursive() and the addition
13339 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13340 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13341 bn_sub_words() and bn_add_words() except they take arrays with
13347 * In 'openssl passwd', verify passwords read from the terminal
13348 unless the '-salt' option is used (which usually means that
13349 verification would just waste user's time since the resulting
13350 hash is going to be compared with some given password hash)
13351 or the new '-noverify' option is used.
13353 This is an incompatible change, but it does not affect
13354 non-interactive use of 'openssl passwd' (passwords on the command
13355 line, '-stdin' option, '-in ...' option) and thus should not
13356 cause any problems.
13360 * Remove all references to RSAref, since there's no more need for it.
13364 * Make DSO load along a path given through an environment variable
13365 (SHLIB_PATH) with shl_load().
13369 * Constify the ENGINE code as a result of BIGNUM constification.
13370 Also constify the RSA code and most things related to it. In a
13371 few places, most notable in the depth of the ASN.1 code, ugly
13372 casts back to non-const were required (to be solved at a later
13377 * Make it so the openssl application has all engines loaded by default.
13381 * Constify the BIGNUM routines a little more.
13385 * Add the following functions:
13387 ENGINE_load_cswift()
13389 ENGINE_load_atalla()
13390 ENGINE_load_nuron()
13391 ENGINE_load_builtin_engines()
13393 That way, an application can itself choose if external engines that
13394 are built-in in OpenSSL shall ever be used or not. The benefit is
13395 that applications won't have to be linked with libdl or other dso
13396 libraries unless it's really needed.
13398 Changed 'openssl engine' to load all engines on demand.
13399 Changed the engine header files to avoid the duplication of some
13400 declarations (they differed!).
13404 * 'openssl engine' can now list capabilities.
13408 * Better error reporting in 'openssl engine'.
13412 * Never call load_dh_param(NULL) in s_server.
13416 * Add engine application. It can currently list engines by name and
13417 identity, and test if they are actually available.
13421 * Improve RPM specification file by forcing symbolic linking and making
13422 sure the installed documentation is also owned by root.root.
13424 *Damien Miller <djm@mindrot.org>*
13426 * Give the OpenSSL applications more possibilities to make use of
13427 keys (public as well as private) handled by engines.
13431 * Add OCSP code that comes from CertCo.
13435 * Add VMS support for the Rijndael code.
13439 * Added untested support for Nuron crypto accelerator.
13443 * Add support for external cryptographic devices. This code was
13444 previously distributed separately as the "engine" branch.
13446 *Geoff Thorpe, Richard Levitte*
13448 * Rework the filename-translation in the DSO code. It is now possible to
13449 have far greater control over how a "name" is turned into a filename
13450 depending on the operating environment and any oddities about the
13451 different shared library filenames on each system.
13455 * Support threads on FreeBSD-elf in Configure.
13459 * Fix for SHA1 assembly problem with MASM: it produces
13460 warnings about corrupt line number information when assembling
13461 with debugging information. This is caused by the overlapping
13464 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13467 NCONF_get_number() has no error checking at all. As a replacement,
13468 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13469 promoted strongly. The old NCONF_get_number is kept around for
13470 binary backward compatibility.
13471 Make it possible for methods to load from something other than a BIO,
13472 by providing a function pointer that is given a name instead of a BIO.
13473 For example, this could be used to load configuration data from an
13478 * Fix for non blocking accept BIOs. Added new I/O special reason
13479 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13480 with non blocking I/O was not possible because no retry code was
13481 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13486 * Added the beginnings of Rijndael support.
13490 * Fix for bug in DirectoryString mask setting. Add support for
13491 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13492 to allow certificate printing to more controllable, additional
13493 'certopt' option to 'x509' to allow new printing options to be
13498 * Clean old EAY MD5 hack from e_os.h.
13502 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13504 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13505 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13507 *Joe Orton, Steve Henson*
13509 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13511 * Fix additional bug revealed by the NISCC test suite:
13513 Stop bug triggering large recursion when presented with
13514 certain ASN.1 tags ([CVE-2003-0851])
13518 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13520 * Fix various bugs revealed by running the NISCC test suite:
13522 Stop out of bounds reads in the ASN1 code when presented with
13523 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13525 If verify callback ignores invalid public key errors don't try to check
13526 certificate signature with the NULL public key.
13530 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13531 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13536 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13537 extra data after the compression methods not only for TLS 1.0
13538 but also for SSL 3.0 (as required by the specification).
13540 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13542 * Change X509_certificate_type() to mark the key as exported/exportable
13543 when it's 512 *bits* long, not 512 bytes.
13547 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13549 * Countermeasure against the Klima-Pokorny-Rosa extension of
13550 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13551 a protocol version number mismatch like a decryption error
13552 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13556 * Turn on RSA blinding by default in the default implementation
13557 to avoid a timing attack. Applications that don't want it can call
13558 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13559 They would be ill-advised to do so in most cases.
13561 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13563 * Change RSA blinding code so that it works when the PRNG is not
13564 seeded (in this case, the secret RSA exponent is abused as
13565 an unpredictable seed -- if it is not unpredictable, there
13566 is no point in blinding anyway). Make RSA blinding thread-safe
13567 by remembering the creator's thread ID in rsa->blinding and
13568 having all other threads use local one-time blinding factors
13569 (this requires more computation than sharing rsa->blinding, but
13570 avoids excessive locking; and if an RSA object is not shared
13571 between threads, blinding will still be very fast).
13575 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13577 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13578 via timing by performing a MAC computation even if incorrect
13579 block cipher padding has been found. This is a countermeasure
13580 against active attacks where the attacker has to distinguish
13581 between bad padding and a MAC verification error. ([CVE-2003-0078])
13583 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13584 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13585 Martin Vuagnoux (EPFL, Ilion)*
13587 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13589 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13590 memory from its contents. This is done with a counter that will
13591 place alternating values in each byte. This can be used to solve
13592 two issues: 1) the removal of calls to memset() by highly optimizing
13593 compilers, and 2) cleansing with other values than 0, since those can
13594 be read through on certain media, for example a swap space on disk.
13598 * Bugfix: client side session caching did not work with external caching,
13599 because the session->cipher setting was not restored when reloading
13600 from the external cache. This problem was masked, when
13601 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13602 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13606 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13607 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13609 *Zeev Lieber <zeev-l@yahoo.com>*
13611 * Undo an undocumented change introduced in 0.9.6e which caused
13612 repeated calls to OpenSSL_add_all_ciphers() and
13613 OpenSSL_add_all_digests() to be ignored, even after calling
13618 * Change the default configuration reader to deal with last line not
13619 being properly terminated.
13623 * Change X509_NAME_cmp() so it applies the special rules on handling
13624 DN values that are of type PrintableString, as well as RDNs of type
13625 emailAddress where the value has the type ia5String.
13627 *stefank@valicert.com via Richard Levitte*
13629 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13630 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13631 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13632 the bitwise-OR of the two for use by the majority of applications
13633 wanting this behaviour, and update the docs. The documented
13634 behaviour and actual behaviour were inconsistent and had been
13635 changing anyway, so this is more a bug-fix than a behavioural
13638 *Geoff Thorpe, diagnosed by Nadav Har'El*
13640 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13641 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13645 * Fix initialization code race conditions in
13646 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13647 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13648 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13649 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13650 ssl2_get_cipher_by_char(),
13651 ssl3_get_cipher_by_char().
13653 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13655 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13656 the cached sessions are flushed, as the remove_cb() might use ex_data
13657 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13658 (see [openssl.org #212]).
13660 *Geoff Thorpe, Lutz Jaenicke*
13662 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13663 length, instead of the encoding length to d2i_ASN1_OBJECT.
13667 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13669 * [In 0.9.6g-engine release:]
13670 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13672 *Lynn Gazis <lgazis@rainbow.com>*
13674 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13676 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13677 and get fix the header length calculation.
13678 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13679 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13681 * Use proper error handling instead of 'assertions' in buffer
13682 overflow checks added in 0.9.6e. This prevents DoS (the
13683 assertions could call abort()).
13685 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13687 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13689 * Add various sanity checks to asn1_get_length() to reject
13690 the ASN1 length bytes if they exceed sizeof(long), will appear
13691 negative or the content length exceeds the length of the
13694 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13696 * Fix cipher selection routines: ciphers without encryption had no flags
13697 for the cipher strength set and where therefore not handled correctly
13698 by the selection routines (PR #130).
13702 * Fix EVP_dsa_sha macro.
13707 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13708 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13709 that was added in OpenSSL 0.9.6d.
13711 As the countermeasure turned out to be incompatible with some
13712 broken SSL implementations, the new option is part of SSL_OP_ALL.
13713 SSL_OP_ALL is usually employed when compatibility with weird SSL
13714 implementations is desired (e.g. '-bugs' option to 's_client' and
13715 's_server'), so the new option is automatically set in many
13720 * Changes in security patch:
13722 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13723 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13724 Air Force Materiel Command, USAF, under agreement number
13727 * Add various sanity checks to asn1_get_length() to reject
13728 the ASN1 length bytes if they exceed sizeof(long), will appear
13729 negative or the content length exceeds the length of the
13730 supplied buffer. ([CVE-2002-0659])
13732 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13734 * Assertions for various potential buffer overflows, not known to
13735 happen in practice.
13737 *Ben Laurie (CHATS)*
13739 * Various temporary buffers to hold ASCII versions of integers were
13740 too small for 64 bit platforms. ([CVE-2002-0655])
13741 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13743 * Remote buffer overflow in SSL3 protocol - an attacker could
13744 supply an oversized session ID to a client. ([CVE-2002-0656])
13746 *Ben Laurie (CHATS)*
13748 * Remote buffer overflow in SSL2 protocol - an attacker could
13749 supply an oversized client master key. ([CVE-2002-0656])
13751 *Ben Laurie (CHATS)*
13753 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13755 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13756 encoded as NULL) with id-dsa-with-sha1.
13758 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13760 * Check various `X509_...()` return values in `apps/req.c`.
13762 *Nils Larsch <nla@trustcenter.de>*
13764 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13765 an end-of-file condition would erroneously be flagged, when the CRLF
13766 was just at the end of a processed block. The bug was discovered when
13767 processing data through a buffering memory BIO handing the data to a
13768 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13769 <ptsekov@syntrex.com> and Nedelcho Stanev.
13773 * Implement a countermeasure against a vulnerability recently found
13774 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13775 before application data chunks to avoid the use of known IVs
13776 with data potentially chosen by the attacker.
13780 * Fix length checks in ssl3_get_client_hello().
13784 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13785 to prevent ssl3_read_internal() from incorrectly assuming that
13786 ssl3_read_bytes() found application data while handshake
13787 processing was enabled when in fact s->s3->in_read_app_data was
13788 merely automatically cleared during the initial handshake.
13790 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13792 * Fix object definitions for Private and Enterprise: they were not
13793 recognized in their shortname (=lowercase) representation. Extend
13794 obj_dat.pl to issue an error when using undefined keywords instead
13795 of silently ignoring the problem (Svenning Sorensen
13796 <sss@sss.dnsalias.net>).
13800 * Fix DH_generate_parameters() so that it works for 'non-standard'
13801 generators, i.e. generators other than 2 and 5. (Previously, the
13802 code did not properly initialise the 'add' and 'rem' values to
13803 BN_generate_prime().)
13805 In the new general case, we do not insist that 'generator' is
13806 actually a primitive root: This requirement is rather pointless;
13807 a generator of the order-q subgroup is just as good, if not
13812 * Map new X509 verification errors to alerts. Discovered and submitted by
13813 Tom Wu <tom@arcot.com>.
13817 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13818 returning non-zero before the data has been completely received
13819 when using non-blocking I/O.
13821 *Bodo Moeller; problem pointed out by John Hughes*
13823 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13825 *Ben Laurie, Lutz Jaenicke*
13827 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13828 Yoram Zahavi <YoramZ@gilian.com>).
13832 * Add information about CygWin 1.3 and on, and preserve proper
13833 configuration for the versions before that.
13835 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13837 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13838 check whether we deal with a copy of a session and do not delete from
13839 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13840 <izhar@checkpoint.com>.
13844 * Do not store session data into the internal session cache, if it
13845 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13846 flag is set). Proposed by Aslam <aslam@funk.com>.
13850 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13855 * [In 0.9.6d-engine release:]
13856 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13858 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13860 * Add the configuration target linux-s390x.
13862 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13864 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13865 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13866 variable as an indication that a ClientHello message has been
13867 received. As the flag value will be lost between multiple
13868 invocations of ssl3_accept when using non-blocking I/O, the
13869 function may not be aware that a handshake has actually taken
13870 place, thus preventing a new session from being added to the
13873 To avoid this problem, we now set s->new_session to 2 instead of
13874 using a local variable.
13876 *Lutz Jaenicke, Bodo Moeller*
13878 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13879 if the SSL_R_LENGTH_MISMATCH error is detected.
13881 *Geoff Thorpe, Bodo Moeller*
13883 * New 'shared_ldflag' column in Configure platform table.
13887 * Fix EVP_CIPHER_mode macro.
13889 *"Dan S. Camper" <dan@bti.net>*
13891 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13892 type, we must throw them away by setting rr->length to 0.
13894 *D P Chang <dpc@qualys.com>*
13896 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13898 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13899 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13900 worked incorrectly for those cases where range = `10..._2` and
13901 `3*range` is two bits longer than range.)
13905 * Only add signing time to PKCS7 structures if it is not already
13910 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13911 OBJ_ld_ce should be OBJ_id_ce.
13912 Also some ip-pda OIDs in crypto/objects/objects.txt were
13913 incorrect (cf. RFC 3039).
13915 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13917 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13918 returns early because it has nothing to do.
13920 *Andy Schneider <andy.schneider@bjss.co.uk>*
13922 * [In 0.9.6c-engine release:]
13923 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13925 *Andy Schneider <andy.schneider@bjss.co.uk>*
13927 * [In 0.9.6c-engine release:]
13928 Add support for Cryptographic Appliance's keyserver technology.
13929 (Use engine 'keyclient')
13931 *Cryptographic Appliances and Geoff Thorpe*
13933 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13934 is called via tools/c89.sh because arguments have to be
13935 rearranged (all '-L' options must appear before the first object
13938 *Richard Shapiro <rshapiro@abinitio.com>*
13940 * [In 0.9.6c-engine release:]
13941 Add support for Broadcom crypto accelerator cards, backported
13944 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13946 * [In 0.9.6c-engine release:]
13947 Add support for SureWare crypto accelerator cards from
13948 Baltimore Technologies. (Use engine 'sureware')
13950 *Baltimore Technologies and Mark Cox*
13952 * [In 0.9.6c-engine release:]
13953 Add support for crypto accelerator cards from Accelerated
13954 Encryption Processing, www.aep.ie. (Use engine 'aep')
13956 *AEP Inc. and Mark Cox*
13958 * Add a configuration entry for gcc on UnixWare.
13960 *Gary Benson <gbenson@redhat.com>*
13962 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13963 messages are stored in a single piece (fixed-length part and
13964 variable-length part combined) and fix various bugs found on the way.
13968 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13969 instead. BIO_gethostbyname() does not know what timeouts are
13970 appropriate, so entries would stay in cache even when they have
13972 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13974 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13975 faced with a pathologically small ClientHello fragment that does
13976 not contain client_version: Instead of aborting with an error,
13977 simply choose the highest available protocol version (i.e.,
13978 TLS 1.0 unless it is disabled). In practice, ClientHello
13979 messages are never sent like this, but this change gives us
13980 strictly correct behaviour at least for TLS.
13984 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13985 never resets s->method to s->ctx->method when called from within
13986 one of the SSL handshake functions.
13988 *Bodo Moeller; problem pointed out by Niko Baric*
13990 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13991 (sent using the client's version number) if client_version is
13992 smaller than the protocol version in use. Also change
13993 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13994 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13995 the client will at least see that alert.
13999 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
14004 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
14005 client receives HelloRequest while in a handshake.
14007 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
14009 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
14010 should end in 'break', not 'goto end' which circumvents various
14011 cleanups done in state SSL_ST_OK. But session related stuff
14012 must be disabled for SSL_ST_OK in the case that we just sent a
14015 Also avoid some overhead by not calling ssl_init_wbio_buffer()
14016 before just sending a HelloRequest.
14018 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
14020 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
14021 reveal whether illegal block cipher padding was found or a MAC
14022 verification error occurred. (Neither SSLerr() codes nor alerts
14023 are directly visible to potential attackers, but the information
14024 may leak via logfiles.)
14026 Similar changes are not required for the SSL 2.0 implementation
14027 because the number of padding bytes is sent in clear for SSL 2.0,
14028 and the extra bytes are just ignored. However ssl/s2_pkt.c
14029 failed to verify that the purported number of padding bytes is in
14034 * Add OpenUNIX-8 support including shared libraries
14035 (Boyd Lynn Gerber <gerberb@zenez.com>).
14039 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
14040 'wristwatch attack' using huge encoding parameters (cf.
14041 James H. Manger's CRYPTO 2001 paper). Note that the
14042 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
14043 encoding parameters and hence was not vulnerable.
14047 * BN_sqr() bug fix.
14049 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14051 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14052 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14053 followed by modular reduction.
14055 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14057 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14058 equivalent based on BN_pseudo_rand() instead of BN_rand().
14062 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14063 This function was broken, as the check for a new client hello message
14064 to handle SGC did not allow these large messages.
14065 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14069 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14073 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14074 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14078 * Rework the configuration and shared library support for Tru64 Unix.
14079 The configuration part makes use of modern compiler features and
14080 still retains old compiler behavior for those that run older versions
14081 of the OS. The shared library support part includes a variant that
14082 uses the RPATH feature, and is available through the special
14083 configuration target "alpha-cc-rpath", which will never be selected
14086 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14088 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14089 with the same message size as in ssl3_get_certificate_request().
14090 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14091 messages might inadvertently be reject as too long.
14093 *Petr Lampa <lampa@fee.vutbr.cz>*
14095 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14099 * Modified SSL library such that the verify_callback that has been set
14100 specifically for an SSL object with SSL_set_verify() is actually being
14101 used. Before the change, a verify_callback set with this function was
14102 ignored and the verify_callback() set in the SSL_CTX at the time of
14103 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14104 to allow the necessary settings.
14108 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14109 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14110 done automatically (in contradiction to the requirements of the C
14111 standard). This made problems when used from OpenSSH.
14115 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14116 dh->length and always used
14118 BN_rand_range(priv_key, dh->p).
14120 BN_rand_range() is not necessary for Diffie-Hellman, and this
14121 specific range makes Diffie-Hellman unnecessarily inefficient if
14122 dh->length (recommended exponent length) is much smaller than the
14123 length of dh->p. We could use BN_rand_range() if the order of
14124 the subgroup was stored in the DH structure, but we only have
14129 BN_rand(priv_key, l, ...)
14131 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14138 RSA_eay_public_encrypt
14139 RSA_eay_private_decrypt
14140 RSA_eay_private_encrypt (signing)
14141 RSA_eay_public_decrypt (signature verification)
14143 (default implementations for RSA_public_encrypt,
14144 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14145 always reject numbers >= n.
14149 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14150 to synchronize access to 'locking_thread'. This is necessary on
14151 systems where access to 'locking_thread' (an 'unsigned long'
14152 variable) is not atomic.
14156 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14157 *before* setting the 'crypto_lock_rand' flag. The previous code had
14158 a race condition if 0 is a valid thread ID.
14160 *Travis Vitek <vitek@roguewave.com>*
14162 * Add support for shared libraries under Irix.
14164 *Albert Chin-A-Young <china@thewrittenword.com>*
14166 * Add configuration option to build on Linux on both big-endian and
14167 little-endian MIPS.
14169 *Ralf Baechle <ralf@uni-koblenz.de>*
14171 * Add the possibility to create shared libraries on HP-UX.
14175 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14177 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14178 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14179 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14180 PRNG state recovery was possible based on the output of
14181 one PRNG request appropriately sized to gain knowledge on
14182 'md' followed by enough consecutive 1-byte PRNG requests
14183 to traverse all of 'state'.
14185 1. When updating 'md_local' (the current thread's copy of 'md')
14186 during PRNG output generation, hash all of the previous
14187 'md_local' value, not just the half used for PRNG output.
14189 2. Make the number of bytes from 'state' included into the hash
14190 independent from the number of PRNG bytes requested.
14192 The first measure alone would be sufficient to avoid
14193 Markku-Juhani's attack. (Actually it had never occurred
14194 to me that the half of 'md_local' used for chaining was the
14195 half from which PRNG output bytes were taken -- I had always
14196 assumed that the secret half would be used.) The second
14197 measure makes sure that additional data from 'state' is never
14198 mixed into 'md_local' in small portions; this heuristically
14199 further strengthens the PRNG.
14203 * Fix crypto/bn/asm/mips3.s.
14207 * When only the key is given to "enc", the IV is undefined. Print out
14208 an error message in this case.
14212 * Handle special case when X509_NAME is empty in X509 printing routines.
14216 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14217 positive and less than q.
14221 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14222 used: it isn't thread safe and the add_lock_callback should handle
14225 *Paul Rose <Paul.Rose@bridge.com>*
14227 * Verify that incoming data obeys the block size in
14228 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14234 *Ulf Möller, Bodo Möller*
14236 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14237 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14238 when fixing the server behaviour for backwards-compatible 'client
14239 hello' messages. (Note that the attack is impractical against
14240 SSL 3.0 and TLS 1.0 anyway because length and version checking
14241 means that the probability of guessing a valid ciphertext is
14242 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14245 Before 0.9.5, the countermeasure (hide the error by generating a
14246 random 'decryption result') did not work properly because
14247 ERR_clear_error() was missing, meaning that SSL_get_error() would
14248 detect the supposedly ignored error.
14250 Both problems are now fixed.
14254 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14255 (previously it was 1024).
14259 * Fix for compatibility mode trust settings: ignore trust settings
14260 unless some valid trust or reject settings are present.
14264 * Fix for blowfish EVP: its a variable length cipher.
14268 * Fix various bugs related to DSA S/MIME verification. Handle missing
14269 parameters in DSA public key structures and return an error in the
14270 DSA routines if parameters are absent.
14274 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14275 in the current directory if neither $RANDFILE nor $HOME was set.
14276 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14277 caused some confusion to Windows users who haven't defined $HOME.
14278 Thus RAND_file_name() is changed again: e_os.h can define a
14279 DEFAULT_HOME, which will be used if $HOME is not set.
14280 For Windows, we use "C:"; on other platforms, we still require
14281 environment variables.
14283 * Move 'if (!initialized) RAND_poll()' into regions protected by
14284 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14285 having multiple threads call RAND_poll() concurrently.
14289 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14290 combination of a flag and a thread ID variable.
14291 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14292 flag), *other* threads can enter ssleay_add_bytes without obeying
14293 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14294 that they do not hold after the first thread unsets add_do_not_lock).
14298 * Change bctest again: '-x' expressions are not available in all
14299 versions of 'test'.
14303 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14305 * Fix a couple of memory leaks in PKCS7_dataDecode()
14307 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14309 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14310 the default extension for executables, if any. Also, make the perl
14311 scripts that use symlink() to test if it really exists and use "cp"
14312 if it doesn't. All this made OpenSSL compilable and installable in
14317 * Fix for asn1_GetSequence() for indefinite length constructed data.
14318 If SEQUENCE is length is indefinite just set c->slen to the total
14319 amount of data available.
14321 *Steve Henson, reported by shige@FreeBSD.org*
14323 *This change does not apply to 0.9.7.*
14325 * Change bctest to avoid here-documents inside command substitution
14326 (workaround for FreeBSD /bin/sh bug).
14327 For compatibility with Ultrix, avoid shell functions (introduced
14328 in the bctest version that searches along $PATH).
14332 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14333 with des_encrypt() defined on some operating systems, like Solaris
14338 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14339 On the Importance of Eliminating Errors in Cryptographic
14340 Computations, J. Cryptology 14 (2001) 2, 101-119,
14341 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14345 * MIPS assembler BIGNUM division bug fix.
14349 * Disabled incorrect Alpha assembler code.
14353 * Fix PKCS#7 decode routines so they correctly update the length
14354 after reading an EOC for the EXPLICIT tag.
14358 *This change does not apply to 0.9.7.*
14360 * Fix bug in PKCS#12 key generation routines. This was triggered
14361 if a 3DES key was generated with a 0 initial byte. Include
14362 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14363 (but broken) behaviour.
14367 * Enhance bctest to search for a working bc along $PATH and print
14370 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14372 * Fix memory leaks in err.c: free err_data string if necessary;
14373 don't write to the wrong index in ERR_set_error_data.
14377 * Implement ssl23_peek (analogous to ssl23_read), which previously
14382 * Replace rdtsc with `_emit` statements for VC++ version 5.
14384 *Jeremy Cooper <jeremy@baymoo.org>*
14386 * Make it possible to reuse SSLv2 sessions.
14390 * In copy_email() check for >= 0 as a return value for
14391 X509_NAME_get_index_by_NID() since 0 is a valid index.
14393 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14395 * Avoid coredump with unsupported or invalid public keys by checking if
14396 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14397 PKCS7_verify() fails with non detached data.
14401 * Don't use getenv in library functions when run as setuid/setgid.
14402 New function OPENSSL_issetugid().
14406 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14407 due to incorrect handling of multi-threading:
14409 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14411 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14413 3. Count how many times MemCheck_off() has been called so that
14414 nested use can be treated correctly. This also avoids
14415 inband-signalling in the previous code (which relied on the
14416 assumption that thread ID 0 is impossible).
14420 * Add "-rand" option also to s_client and s_server.
14424 * Fix CPU detection on Irix 6.x.
14425 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14426 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14428 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14433 *This change does not apply to 0.9.7.*
14435 * Use the cached encoding of an X509_NAME structure rather than
14436 copying it. This is apparently the reason for the libsafe "errors"
14437 but the code is actually correct.
14441 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14442 Bleichenbacher's DSA attack.
14443 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14444 to be set and top=0 forces the highest bit to be set; top=-1 is new
14445 and leaves the highest bit random.
14447 *Ulf Moeller, Bodo Moeller*
14449 * In the `NCONF_...`-based implementations for `CONF_...` queries
14450 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14451 a temporary CONF structure with the data component set to NULL
14452 (which gives segmentation faults in lh_retrieve).
14453 Instead, use NULL for the CONF pointer in CONF_get_string and
14454 CONF_get_number (which may use environment variables) and directly
14455 return NULL from CONF_get_section.
14459 * Fix potential buffer overrun for EBCDIC.
14463 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14464 keyUsage if basicConstraints absent for a CA.
14468 * Make SMIME_write_PKCS7() write mail header values with a format that
14469 is more generally accepted (no spaces before the semicolon), since
14470 some programs can't parse those values properly otherwise. Also make
14471 sure BIO's that break lines after each write do not create invalid
14476 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14477 macros previously used would not encode an empty SEQUENCE OF
14478 and break the signature.
14482 *This change does not apply to 0.9.7.*
14484 * Zero the premaster secret after deriving the master secret in
14489 * Add some EVP_add_digest_alias registrations (as found in
14490 OpenSSL_add_all_digests()) to SSL_library_init()
14491 aka OpenSSL_add_ssl_algorithms(). This provides improved
14492 compatibility with peers using X.509 certificates
14493 with unconventional AlgorithmIdentifier OIDs.
14497 * Fix for Irix with NO_ASM.
14499 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14501 * ./config script fixes.
14503 *Ulf Moeller, Richard Levitte*
14505 * Fix 'openssl passwd -1'.
14509 * Change PKCS12_key_gen_asc() so it can cope with non null
14510 terminated strings whose length is passed in the passlen
14511 parameter, for example from PEM callbacks. This was done
14512 by adding an extra length parameter to asc2uni().
14514 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14516 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14517 call failed, free the DSA structure.
14521 * Fix to uni2asc() to cope with zero length Unicode strings.
14522 These are present in some PKCS#12 files.
14526 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14527 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14528 when writing a 32767 byte record.
14530 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14532 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14533 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14535 (RSA objects have a reference count access to which is protected
14536 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14537 so they are meant to be shared between threads.)
14538 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14539 "Reddie, Steven" <Steven.Reddie@ca.com>*
14541 * Fix a deadlock in CRYPTO_mem_leaks().
14545 * Use better test patterns in bntest.
14549 * rand_win.c fix for Borland C.
14553 * BN_rshift bugfix for n == 0.
14557 * Add a 'bctest' script that checks for some known 'bc' bugs
14558 so that 'make test' does not abort just because 'bc' is broken.
14562 * Store verify_result within SSL_SESSION also for client side to
14563 avoid potential security hole. (Re-used sessions on the client side
14564 always resulted in verify_result==X509_V_OK, not using the original
14565 result of the server certificate verification.)
14569 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14570 SSL3_RT_APPLICATION_DATA, return 0.
14571 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14576 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14577 releases, have been re-implemented by renaming the previous
14578 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14579 and ssl3_read_internal, respectively, and adding 'peek' parameters
14580 to them. The new ssl[23]_{read,peek} functions are calls to
14581 ssl[23]_read_internal with the 'peek' flag set appropriately.
14582 A 'peek' parameter has also been added to ssl3_read_bytes, which
14583 does the actual work for ssl3_read_internal.
14587 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14588 the method-specific "init()" handler. Also clean up ex_data after
14589 calling the method-specific "finish()" handler. Previously, this was
14590 happening the other way round.
14594 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14595 The previous value, 12, was not always sufficient for BN_mod_exp().
14599 * Make sure that shared libraries get the internal name engine with
14600 the full version number and not just 0. This should mark the
14601 shared libraries as not backward compatible. Of course, this should
14602 be changed again when we can guarantee backward binary compatibility.
14606 * Fix typo in get_cert_by_subject() in by_dir.c
14608 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14610 * Rework the system to generate shared libraries:
14612 - Make note of the expected extension for the shared libraries and
14613 if there is a need for symbolic links from for example libcrypto.so.0
14614 to libcrypto.so.0.9.7. There is extended info in Configure for
14617 - Make as few rebuilds of the shared libraries as possible.
14619 - Still avoid linking the OpenSSL programs with the shared libraries.
14621 - When installing, install the shared libraries separately from the
14626 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14628 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14629 and not in SSL_clear because the latter is also used by the
14630 accept/connect functions; previously, the settings made by
14631 SSL_set_read_ahead would be lost during the handshake.
14633 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14635 * Correct util/mkdef.pl to be selective about disabled algorithms.
14636 Previously, it would create entries for disabled algorithms no
14641 * Added several new manual pages for SSL_* function.
14645 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14647 * In ssl23_get_client_hello, generate an error message when faced
14648 with an initial SSL 3.0/TLS record that is too small to contain the
14649 first two bytes of the ClientHello message, i.e. client_version.
14650 (Note that this is a pathologic case that probably has never happened
14651 in real life.) The previous approach was to use the version number
14652 from the record header as a substitute; but our protocol choice
14653 should not depend on that one because it is not authenticated
14654 by the Finished messages.
14658 * More robust randomness gathering functions for Windows.
14660 *Jeffrey Altman <jaltman@columbia.edu>*
14662 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14663 not set then we don't setup the error code for issuer check errors
14664 to avoid possibly overwriting other errors which the callback does
14665 handle. If an application does set the flag then we assume it knows
14666 what it is doing and can handle the new informational codes
14671 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14672 a general "ANY" type, as such it should be able to decode anything
14673 including tagged types. However it didn't check the class so it would
14674 wrongly interpret tagged types in the same way as their universal
14675 counterpart and unknown types were just rejected. Changed so that the
14676 tagged and unknown types are handled in the same way as a SEQUENCE:
14677 that is the encoding is stored intact. There is also a new type
14678 "V_ASN1_OTHER" which is used when the class is not universal, in this
14679 case we have no idea what the actual type is so we just lump them all
14684 * On VMS, stdout may very well lead to a file that is written to
14685 in a record-oriented fashion. That means that every write() will
14686 write a separate record, which will be read separately by the
14687 programs trying to read from it. This can be very confusing.
14689 The solution is to put a BIO filter in the way that will buffer
14690 text until a linefeed is reached, and then write everything a
14691 line at a time, so every record written will be an actual line,
14692 not chunks of lines and not (usually doesn't happen, but I've
14693 seen it once) several lines in one record. BIO_f_linebuffer() is
14696 Currently, it's a VMS-only method, because that's where it has
14697 been tested well enough.
14701 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14702 it can return incorrect results.
14703 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14704 but it was in 0.9.6-beta[12].)
14708 * Disable the check for content being present when verifying detached
14709 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14710 include zero length content when signing messages.
14714 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14715 BIO_ctrl (for BIO pairs).
14719 * Add DSO method for VMS.
14723 * Bug fix: Montgomery multiplication could produce results with the
14728 * Add RPM specification openssl.spec and modify it to build three
14729 packages. The default package contains applications, application
14730 documentation and run-time libraries. The devel package contains
14731 include files, static libraries and function documentation. The
14732 doc package contains the contents of the doc directory. The original
14733 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14737 * Add a large number of documentation files for many SSL routines.
14739 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14741 * Add a configuration entry for Sony News 4.
14743 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14745 * Don't set the two most significant bits to one when generating a
14746 random number < q in the DSA library.
14750 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14751 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14752 the underlying transport is blocking) if a handshake took place.
14753 (The default behaviour is needed by applications such as s_client
14754 and s_server that use select() to determine when to use SSL_read;
14755 but for applications that know in advance when to expect data, it
14756 just makes things more complicated.)
14760 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14765 * Add a few more EBCDIC conditionals that make `req` and `x509`
14766 work better on such systems.
14768 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14770 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14771 Update PKCS12_parse() so it copies the friendlyName and the
14772 keyid to the certificates aux info.
14776 * Fix bug in PKCS7_verify() which caused an infinite loop
14777 if there was more than one signature.
14779 *Sven Uszpelkat <su@celocom.de>*
14781 * Major change in util/mkdef.pl to include extra information
14782 about each symbol, as well as presenting variables as well
14783 as functions. This change means that there's n more need
14784 to rebuild the .num files when some algorithms are excluded.
14788 * Allow the verify time to be set by an application,
14789 rather than always using the current time.
14793 * Phase 2 verify code reorganisation. The certificate
14794 verify code now looks up an issuer certificate by a
14795 number of criteria: subject name, authority key id
14796 and key usage. It also verifies self signed certificates
14797 by the same criteria. The main comparison function is
14798 X509_check_issued() which performs these checks.
14800 Lot of changes were necessary in order to support this
14801 without completely rewriting the lookup code.
14803 Authority and subject key identifier are now cached.
14805 The LHASH 'certs' is X509_STORE has now been replaced
14806 by a STACK_OF(X509_OBJECT). This is mainly because an
14807 LHASH can't store or retrieve multiple objects with
14808 the same hash value.
14810 As a result various functions (which were all internal
14811 use only) have changed to handle the new X509_STORE
14812 structure. This will break anything that messed round
14813 with X509_STORE internally.
14815 The functions X509_STORE_add_cert() now checks for an
14816 exact match, rather than just subject name.
14818 The X509_STORE API doesn't directly support the retrieval
14819 of multiple certificates matching a given criteria, however
14820 this can be worked round by performing a lookup first
14821 (which will fill the cache with candidate certificates)
14822 and then examining the cache for matches. This is probably
14823 the best we can do without throwing out X509_LOOKUP
14824 entirely (maybe later...).
14826 The X509_VERIFY_CTX structure has been enhanced considerably.
14828 All certificate lookup operations now go via a get_issuer()
14829 callback. Although this currently uses an X509_STORE it
14830 can be replaced by custom lookups. This is a simple way
14831 to bypass the X509_STORE hackery necessary to make this
14832 work and makes it possible to use more efficient techniques
14833 in future. A very simple version which uses a simple
14834 STACK for its trusted certificate store is also provided
14835 using X509_STORE_CTX_trusted_stack().
14837 The verify_cb() and verify() callbacks now have equivalents
14838 in the X509_STORE_CTX structure.
14840 X509_STORE_CTX also has a 'flags' field which can be used
14841 to customise the verify behaviour.
14845 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14846 excludes S/MIME capabilities.
14850 * When a certificate request is read in keep a copy of the
14851 original encoding of the signed data and use it when outputting
14852 again. Signatures then use the original encoding rather than
14853 a decoded, encoded version which may cause problems if the
14854 request is improperly encoded.
14858 * For consistency with other BIO_puts implementations, call
14859 buffer_write(b, ...) directly in buffer_puts instead of calling
14862 In BIO_puts, increment b->num_write as in BIO_write.
14864 *Peter.Sylvester@EdelWeb.fr*
14866 * Fix BN_mul_word for the case where the word is 0. (We have to use
14867 BN_zero, we may not return a BIGNUM with an array consisting of
14868 words set to zero.)
14872 * Avoid calling abort() from within the library when problems are
14873 detected, except if preprocessor symbols have been defined
14874 (such as REF_CHECK, BN_DEBUG etc.).
14878 * New openssl application 'rsautl'. This utility can be
14879 used for low level RSA operations. DER public key
14880 BIO/fp routines also added.
14884 * New Configure entry and patches for compiling on QNX 4.
14886 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14888 * A demo state-machine implementation was sponsored by
14889 Nuron (<http://www.nuron.com/>) and is now available in
14890 demos/state_machine.
14894 * New options added to the 'dgst' utility for signature
14895 generation and verification.
14899 * Unrecognized PKCS#7 content types are now handled via a
14900 catch all ASN1_TYPE structure. This allows unsupported
14901 types to be stored as a "blob" and an application can
14902 encode and decode it manually.
14906 * Fix various signed/unsigned issues to make a_strex.c
14907 compile under VC++.
14909 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14911 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14912 length if passed a buffer. ASN1_INTEGER_to_BN failed
14913 if passed a NULL BN and its argument was negative.
14915 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14917 * Modification to PKCS#7 encoding routines to output definite
14918 length encoding. Since currently the whole structures are in
14919 memory there's not real point in using indefinite length
14920 constructed encoding. However if OpenSSL is compiled with
14921 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14925 * Added BIO_vprintf() and BIO_vsnprintf().
14929 * Added more prefixes to parse for in the strings written
14930 through a logging bio, to cover all the levels that are available
14931 through syslog. The prefixes are now:
14933 PANIC, EMERG, EMR => LOG_EMERG
14934 ALERT, ALR => LOG_ALERT
14935 CRIT, CRI => LOG_CRIT
14936 ERROR, ERR => LOG_ERR
14937 WARNING, WARN, WAR => LOG_WARNING
14938 NOTICE, NOTE, NOT => LOG_NOTICE
14939 INFO, INF => LOG_INFO
14940 DEBUG, DBG => LOG_DEBUG
14942 and as before, if none of those prefixes are present at the
14943 beginning of the string, LOG_ERR is chosen.
14945 On Win32, the `LOG_*` levels are mapped according to this:
14947 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14948 LOG_WARNING => EVENTLOG_WARNING_TYPE
14949 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14953 * Made it possible to reconfigure with just the configuration
14954 argument "reconf" or "reconfigure". The command line arguments
14955 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14956 and are retrieved from there when reconfiguring.
14962 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14964 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14968 * The obj_dat.pl script was messing up the sorting of object
14969 names. The reason was that it compared the quoted version
14970 of strings as a result "OCSP" > "OCSP Signing" because
14971 " > SPACE. Changed script to store unquoted versions of
14972 names and add quotes on output. It was also omitting some
14973 names from the lookup table if they were given a default
14974 value (that is if SN is missing it is given the same
14975 value as LN and vice versa), these are now added on the
14976 grounds that if an object has a name we should be able to
14977 look it up. Finally added warning output when duplicate
14978 short or long names are found.
14982 * Changes needed for Tandem NSK.
14984 *Scott Uroff <scott@xypro.com>*
14986 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14987 RSA_padding_check_SSLv23(), special padding was never detected
14988 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14989 version rollback attacks was not effective.
14991 In s23_clnt.c, don't use special rollback-attack detection padding
14992 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14993 client; similarly, in s23_srvr.c, don't do the rollback check if
14994 SSL 2.0 is the only protocol enabled in the server.
14998 * Make it possible to get hexdumps of unprintable data with 'openssl
14999 asn1parse'. By implication, the functions ASN1_parse_dump() and
15000 BIO_dump_indent() are added.
15004 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
15005 these print out strings and name structures based on various
15006 flags including RFC2253 support and proper handling of
15007 multibyte characters. Added options to the 'x509' utility
15008 to allow the various flags to be set.
15012 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
15013 Also change the functions X509_cmp_current_time() and
15014 X509_gmtime_adj() work with an ASN1_TIME structure,
15015 this will enable certificates using GeneralizedTime in validity
15016 dates to be checked.
15020 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
15021 negative public key encodings) on by default,
15022 NO_NEG_PUBKEY_BUG can be set to disable it.
15026 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
15027 content octets. An i2c_ASN1_OBJECT is unnecessary because
15028 the encoding can be trivially obtained from the structure.
15032 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
15033 not read locks (`CRYPTO_r_[un]lock`).
15037 * A first attempt at creating official support for shared
15038 libraries through configuration. I've kept it so the
15039 default is static libraries only, and the OpenSSL programs
15040 are always statically linked for now, but there are
15041 preparations for dynamic linking in place.
15042 This has been tested on Linux and Tru64.
15046 * Randomness polling function for Win9x, as described in:
15047 Peter Gutmann, Software Generation of Practically Strong
15052 * Fix so PRNG is seeded in req if using an already existing
15057 * New options to smime application. -inform and -outform
15058 allow alternative formats for the S/MIME message including
15059 PEM and DER. The -content option allows the content to be
15060 specified separately. This should allow things like Netscape
15061 form signing output easier to verify.
15065 * Fix the ASN1 encoding of tags using the 'long form'.
15069 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15070 STRING types. These convert content octets to and from the
15071 underlying type. The actual tag and length octets are
15072 already assumed to have been read in and checked. These
15073 are needed because all other string types have virtually
15074 identical handling apart from the tag. By having versions
15075 of the ASN1 functions that just operate on content octets
15076 IMPLICIT tagging can be handled properly. It also allows
15077 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15078 and ASN1_INTEGER are identical apart from the tag.
15082 * Change the handling of OID objects as follows:
15084 - New object identifiers are inserted in objects.txt, following
15085 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15086 - objects.pl is used to process obj_mac.num and create a new
15088 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15091 This is currently kind of a hack, and the perl code in objects.pl
15092 isn't very elegant, but it works as I intended. The simplest way
15093 to check that it worked correctly is to look in obj_dat.h and
15094 check the array nid_objs and make sure the objects haven't moved
15095 around (this is important!). Additions are OK, as well as
15096 consistent name changes.
15100 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15104 * Addition of the command line parameter '-rand file' to 'openssl req'.
15105 The given file adds to whatever has already been seeded into the
15106 random pool through the RANDFILE configuration file option or
15107 environment variable, or the default random state file.
15111 * mkstack.pl now sorts each macro group into lexical order.
15112 Previously the output order depended on the order the files
15113 appeared in the directory, resulting in needless rewriting
15118 * Patches to make OpenSSL compile under Win32 again. Mostly
15119 work arounds for the VC++ problem that it treats func() as
15120 func(void). Also stripped out the parts of mkdef.pl that
15121 added extra typesafe functions: these no longer exist.
15125 * Reorganisation of the stack code. The macros are now all
15126 collected in safestack.h . Each macro is defined in terms of
15127 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15128 DEBUG_SAFESTACK is now handled in terms of function casts,
15129 this has the advantage of retaining type safety without the
15130 use of additional functions. If DEBUG_SAFESTACK is not defined
15131 then the non typesafe macros are used instead. Also modified the
15132 mkstack.pl script to handle the new form. Needs testing to see
15133 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15134 the default if no major problems. Similar behaviour for ASN1_SET_OF
15135 and PKCS12_STACK_OF.
15139 * When some versions of IIS use the 'NET' form of private key the
15140 key derivation algorithm is different. Normally MD5(password) is
15141 used as a 128 bit RC4 key. In the modified case
15142 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15143 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15144 as the old Netscape_RSA functions except they have an additional
15145 'sgckey' parameter which uses the modified algorithm. Also added
15146 an -sgckey command line option to the rsa utility. Thanks to
15147 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15148 algorithm to openssl-dev.
15152 * The evp_local.h macros were using 'c.##kname' which resulted in
15153 invalid expansion on some systems (SCO 5.0.5 for example).
15154 Corrected to 'c.kname'.
15156 *Phillip Porch <root@theporch.com>*
15158 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15159 a STACK of email addresses from a certificate or request, these look
15160 in the subject name and the subject alternative name extensions and
15161 omit any duplicate addresses.
15165 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15166 This makes DSA verification about 2 % faster.
15170 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15171 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15172 plus overhead for 1024 bit moduli).
15173 This makes exponentiations about 0.5 % faster for 1024 bit
15174 exponents (as measured by "openssl speed rsa2048").
15178 * Rename memory handling macros to avoid conflicts with other
15180 Malloc => OPENSSL_malloc
15181 Malloc_locked => OPENSSL_malloc_locked
15182 Realloc => OPENSSL_realloc
15183 Free => OPENSSL_free
15187 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15188 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15192 * CygWin32 support.
15194 *John Jarvie <jjarvie@newsguy.com>*
15196 * The type-safe stack code has been rejigged. It is now only compiled
15197 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15198 by default all type-specific stack functions are "#define"d back to
15199 standard stack functions. This results in more streamlined output
15200 but retains the type-safety checking possibilities of the original
15205 * The STACK code has been cleaned up, and certain type declarations
15206 that didn't make a lot of sense have been brought in line. This has
15207 also involved a cleanup of sorts in safestack.h to more correctly
15208 map type-safe stack functions onto their plain stack counterparts.
15209 This work has also resulted in a variety of "const"ifications of
15210 lots of the code, especially `_cmp` operations which should normally
15211 be prototyped with "const" parameters anyway.
15215 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15216 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15217 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15218 where all of 'md' is used each time the PRNG is used, but 'state'
15219 is used only indexed by a cyclic counter. As entropy may not be
15220 well distributed from the beginning, 'md' is important as a
15221 chaining variable. However, the output function chains only half
15222 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15223 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15224 in all of 'state' being rewritten, with the new values depending
15225 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15229 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15230 the handshake is continued after ssl_verify_cert_chain();
15231 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15232 can lead to 'unexplainable' connection aborts later.
15234 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15236 * Major EVP API cipher revision.
15237 Add hooks for extra EVP features. This allows various cipher
15238 parameters to be set in the EVP interface. Support added for variable
15239 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15240 setting of RC2 and RC5 parameters.
15242 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15245 Remove lots of duplicated code from the EVP library. For example *every*
15246 cipher init() function handles the 'iv' in the same way according to the
15247 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15248 for CFB and OFB modes they zero ctx->num.
15250 New functionality allows removal of S/MIME code RC2 hack.
15252 Most of the routines have the same form and so can be declared in terms
15255 By shifting this to the top level EVP_CipherInit() it can be removed from
15256 all individual ciphers. If the cipher wants to handle IVs or keys
15257 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15260 Change lots of functions like EVP_EncryptUpdate() to now return a
15261 value: although software versions of the algorithms cannot fail
15262 any installed hardware versions can.
15266 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15267 this option is set, tolerate broken clients that send the negotiated
15268 protocol version number instead of the requested protocol version
15273 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15274 i.e. non-zero for export ciphersuites, zero otherwise.
15275 Previous versions had this flag inverted, inconsistent with
15276 rsa_tmp_cb (..._TMP_RSA_CB).
15278 *Bodo Moeller; problem reported by Amit Chopra*
15280 * Add missing DSA library text string. Work around for some IIS
15281 key files with invalid SEQUENCE encoding.
15285 * Add a document (doc/standards.txt) that list all kinds of standards
15286 and so on that are implemented in OpenSSL.
15290 * Enhance c_rehash script. Old version would mishandle certificates
15291 with the same subject name hash and wouldn't handle CRLs at all.
15292 Added -fingerprint option to crl utility, to support new c_rehash
15297 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15301 * Fix for SSL server purpose checking. Server checking was
15302 rejecting certificates which had extended key usage present
15303 but no ssl client purpose.
15305 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15307 * Make PKCS#12 code work with no password. The PKCS#12 spec
15308 is a little unclear about how a blank password is handled.
15309 Since the password in encoded as a BMPString with terminating
15310 double NULL a zero length password would end up as just the
15311 double NULL. However no password at all is different and is
15312 handled differently in the PKCS#12 key generation code. NS
15313 treats a blank password as zero length. MSIE treats it as no
15314 password on export: but it will try both on import. We now do
15315 the same: PKCS12_parse() tries zero length and no password if
15316 the password is set to "" or NULL (NULL is now a valid password:
15317 it wasn't before) as does the pkcs12 application.
15321 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15322 perror when PEM_read_bio_X509_REQ fails, the error message must
15323 be obtained from the error queue.
15327 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15328 it in ERR_remove_state if appropriate, and change ERR_get_state
15329 accordingly to avoid race conditions (this is necessary because
15330 thread_hash is no longer constant once set).
15334 * Bugfix for linux-elf makefile.one.
15338 * RSA_get_default_method() will now cause a default
15339 RSA_METHOD to be chosen if one doesn't exist already.
15340 Previously this was only set during a call to RSA_new()
15341 or RSA_new_method(NULL) meaning it was possible for
15342 RSA_get_default_method() to return NULL.
15346 * Added native name translation to the existing DSO code
15347 that will convert (if the flag to do so is set) filenames
15348 that are sufficiently small and have no path information
15349 into a canonical native form. Eg. "blah" converted to
15350 "libblah.so" or "blah.dll" etc.
15354 * New function ERR_error_string_n(e, buf, len) which is like
15355 ERR_error_string(e, buf), but writes at most 'len' bytes
15356 including the 0 terminator. For ERR_error_string_n, 'buf'
15359 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15361 * CONF library reworked to become more general. A new CONF
15362 configuration file reader "class" is implemented as well as a
15363 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15364 old `CONF_*` functions are still there, but are reimplemented to
15365 work in terms of the new functions. Also, a set of functions
15366 to handle the internal storage of the configuration data is
15367 provided to make it easier to write new configuration file
15368 reader "classes" (I can definitely see something reading a
15369 configuration file in XML format, for example), called `_CONF_*`,
15370 or "the configuration storage API"...
15372 The new configuration file reading functions are:
15374 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15375 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15377 NCONF_default, NCONF_WIN32
15379 NCONF_dump_fp, NCONF_dump_bio
15381 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15382 NCONF_new creates a new CONF object. This works in the same way
15383 as other interfaces in OpenSSL, like the BIO interface.
15384 `NCONF_dump_*` dump the internal storage of the configuration file,
15385 which is useful for debugging. All other functions take the same
15386 arguments as the old `CONF_*` functions with the exception of the
15387 first that must be a `CONF *` instead of a `LHASH *`.
15389 To make it easier to use the new classes with the old `CONF_*` functions,
15390 the function CONF_set_default_method is provided.
15394 * Add '-tls1' option to 'openssl ciphers', which was already
15395 mentioned in the documentation but had not been implemented.
15396 (This option is not yet really useful because even the additional
15397 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15401 * Initial DSO code added into libcrypto for letting OpenSSL (and
15402 OpenSSL-based applications) load shared libraries and bind to
15403 them in a portable way.
15405 *Geoff Thorpe, with contributions from Richard Levitte*
15407 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15409 * Make sure _lrotl and _lrotr are only used with MSVC.
15411 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15412 (the default implementation of RAND_status).
15414 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15415 to '-clrext' (= clear extensions), as intended and documented.
15416 *Bodo Moeller; inconsistency pointed out by Michael Attili
15417 <attili@amaxo.com>*
15419 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15420 was larger than the MD block size.
15422 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15424 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15425 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15426 using the passed key: if the passed key was a private key the result
15427 of X509_print(), for example, would be to print out all the private key
15432 * des_quad_cksum() byte order bug fix.
15433 *Ulf Möller, using the problem description in krb4-0.9.7, where
15434 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15436 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15439 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15441 * For easily testing in shell scripts whether some command
15442 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15443 returns with exit code 0 iff no command of the given name is available.
15444 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15445 the output goes to stdout and nothing is printed to stderr.
15446 Additional arguments are always ignored.
15448 Since for each cipher there is a command of the same name,
15449 the 'no-cipher' compilation switches can be tested this way.
15451 ('openssl no-XXX' is not able to detect pseudo-commands such
15452 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15456 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15460 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15461 is set; it will be thrown away anyway because each handshake creates
15463 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15464 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15465 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15466 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15470 * New s_client option -ign_eof: EOF at stdin is ignored, and
15471 'Q' and 'R' lose their special meanings (quit/renegotiate).
15472 This is part of what -quiet does; unlike -quiet, -ign_eof
15473 does not suppress any output.
15477 * Add compatibility options to the purpose and trust code. The
15478 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15479 accepts a certificate or CA, this was the previous behaviour,
15480 with all the associated security issues.
15482 X509_TRUST_COMPAT is the old trust behaviour: only and
15483 automatically trust self signed roots in certificate store. A
15484 new trust setting X509_TRUST_DEFAULT is used to specify that
15485 a purpose has no associated trust setting and it should instead
15486 use the value in the default purpose.
15490 * Fix the PKCS#8 DSA private key code so it decodes keys again
15491 and fix a memory leak.
15495 * In util/mkerr.pl (which implements 'make errors'), preserve
15496 reason strings from the previous version of the .c file, as
15497 the default to have only downcase letters (and digits) in
15498 automatically generated reasons codes is not always appropriate.
15502 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15503 using strerror. Previously, ERR_reason_error_string() returned
15504 library names as reason strings for SYSerr; but SYSerr is a special
15505 case where small numbers are errno values, not library numbers.
15509 * Add '-dsaparam' option to 'openssl dhparam' application. This
15510 converts DSA parameters into DH parameters. (When creating parameters,
15511 DSA_generate_parameters is used.)
15515 * Include 'length' (recommended exponent length) in C code generated
15516 by 'openssl dhparam -C'.
15520 * The second argument to set_label in perlasm was already being used
15521 so couldn't be used as a "file scope" flag. Moved to third argument
15526 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15527 instead of RAND_bytes for encryption IVs and salts.
15531 * Include RAND_status() into RAND_METHOD instead of implementing
15532 it only for md_rand.c Otherwise replacing the PRNG by calling
15533 RAND_set_rand_method would be impossible.
15537 * Don't let DSA_generate_key() enter an infinite loop if the random
15538 number generation fails.
15542 * New 'rand' application for creating pseudo-random output.
15546 * Added configuration support for Linux/IA64
15548 *Rolf Haberrecker <rolf@suse.de>*
15550 * Assembler module support for Mingw32.
15554 * Shared library support for HPUX (in shlib/).
15556 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15558 * Shared library support for Solaris gcc.
15560 *Lutz Behnke <behnke@trustcenter.de>*
15562 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15564 * PKCS7_encrypt() was adding text MIME headers twice because they
15565 were added manually and by SMIME_crlf_copy().
15569 * In bntest.c don't call BN_rand with zero bits argument.
15571 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15573 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15574 case was implemented. This caused BN_div_recp() to fail occasionally.
15578 * Add an optional second argument to the set_label() in the perl
15579 assembly language builder. If this argument exists and is set
15580 to 1 it signals that the assembler should use a symbol whose
15581 scope is the entire file, not just the current function. This
15582 is needed with MASM which uses the format label:: for this scope.
15584 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15586 * Change the ASN1 types so they are typedefs by default. Before
15587 almost all types were #define'd to ASN1_STRING which was causing
15588 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15593 * Change names of new functions to the new get1/get0 naming
15594 convention: After 'get1', the caller owns a reference count
15595 and has to call `..._free`; 'get0' returns a pointer to some
15596 data structure without incrementing reference counters.
15597 (Some of the existing 'get' functions increment a reference
15598 counter, some don't.)
15599 Similarly, 'set1' and 'add1' functions increase reference
15600 counters or duplicate objects.
15604 * Allow for the possibility of temp RSA key generation failure:
15605 the code used to assume it always worked and crashed on failure.
15609 * Fix potential buffer overrun problem in BIO_printf().
15610 *Ulf Möller, using public domain code by Patrick Powell; problem
15611 pointed out by David Sacerdote <das33@cornell.edu>*
15613 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15614 RAND_egd() and RAND_status(). In the command line application,
15615 the EGD socket can be specified like a seed file using RANDFILE
15620 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15621 Some CAs (e.g. Verisign) distribute certificates in this form.
15625 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15626 list to exclude them. This means that no special compilation option
15627 is needed to use anonymous DH: it just needs to be included in the
15632 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15633 EVP_MD_type. The old functionality is available in a new macro called
15634 EVP_MD_md(). Change code that uses it and update docs.
15638 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15639 where the `void *` argument is replaced by a function pointer argument.
15640 Previously `void *` was abused to point to functions, which works on
15641 many platforms, but is not correct. As these functions are usually
15642 called by macros defined in OpenSSL header files, most source code
15643 should work without changes.
15647 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15648 sections with information on -D... compiler switches used for
15649 compiling the library so that applications can see them. To enable
15650 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15651 must be defined. E.g.,
15652 #define OPENSSL_ALGORITHM_DEFINES
15653 #include <openssl/opensslconf.h>
15654 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15656 *Richard Levitte, Ulf and Bodo Möller*
15658 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15663 * Change the 'other' type in certificate aux info to a STACK_OF
15664 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15665 the required ASN1 format: arbitrary types determined by an OID.
15669 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15670 argument to 'req'. This is not because the function is newer or
15671 better than others it just uses the work 'NEW' in the certificate
15672 request header lines. Some software needs this.
15676 * Reorganise password command line arguments: now passwords can be
15677 obtained from various sources. Delete the PEM_cb function and make
15678 it the default behaviour: i.e. if the callback is NULL and the
15679 usrdata argument is not NULL interpret it as a null terminated pass
15680 phrase. If usrdata and the callback are NULL then the pass phrase
15681 is prompted for as usual.
15685 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15686 the support is automatically enabled. The resulting binaries will
15687 autodetect the card and use it if present.
15689 *Ben Laurie and Compaq Inc.*
15691 * Work around for Netscape hang bug. This sends certificate request
15692 and server done in one record. Since this is perfectly legal in the
15693 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15694 the bugs/SSLv3 entry for more info.
15698 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15702 * Add -rand argument to smime and pkcs12 applications and read/write
15707 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15711 * Add command line password options to the remaining applications.
15715 * Bug fix for BN_div_recp() for numerators with an even number of
15720 * More tests in bntest.c, and changed test_bn output.
15724 * ./config recognizes MacOS X now.
15728 * Bug fix for BN_div() when the first words of num and divisor are
15729 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15733 * Add support for various broken PKCS#8 formats, and command line
15734 options to produce them.
15738 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15739 get temporary BIGNUMs from a BN_CTX.
15743 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15748 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15749 include a #define from the old name to the new. The original intent
15750 was that statically linked binaries could for example just call
15751 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15752 link with digests. This never worked because SSLeay_add_all_digests()
15753 and SSLeay_add_all_ciphers() were in the same source file so calling
15754 one would link with the other. They are now in separate source files.
15758 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15762 * Use a less unusual form of the Miller-Rabin primality test (it used
15763 a binary algorithm for exponentiation integrated into the Miller-Rabin
15764 loop, our standard modexp algorithms are faster).
15768 * Support for the EBCDIC character set completed.
15770 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15772 * Source code cleanups: use const where appropriate, eliminate casts,
15773 use `void *` instead of `char *` in lhash.
15777 * Bugfix: ssl3_send_server_key_exchange was not restartable
15778 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15779 this the server could overwrite ephemeral keys that the client
15784 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15785 using 50 iterations of the Rabin-Miller test.
15787 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15788 iterations of the Rabin-Miller test as required by the appendix
15789 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15790 As BN_is_prime_fasttest includes trial division, DSA parameter
15791 generation becomes much faster.
15793 This implies a change for the callback functions in DSA_is_prime
15794 and DSA_generate_parameters: The callback function is called once
15795 for each positive witness in the Rabin-Miller test, not just
15796 occasionally in the inner loop; and the parameters to the
15797 callback function now provide an iteration count for the outer
15798 loop rather than for the current invocation of the inner loop.
15799 DSA_generate_parameters additionally can call the callback
15800 function with an 'iteration count' of -1, meaning that a
15801 candidate has passed the trial division test (when q is generated
15802 from an application-provided seed, trial division is skipped).
15806 * New function BN_is_prime_fasttest that optionally does trial
15807 division before starting the Rabin-Miller test and has
15808 an additional BN_CTX * argument (whereas BN_is_prime always
15809 has to allocate at least one BN_CTX).
15810 'callback(1, -1, cb_arg)' is called when a number has passed the
15811 trial division stage.
15815 * Fix for bug in CRL encoding. The validity dates weren't being handled
15820 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15824 * New function BN_pseudo_rand().
15828 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15829 bignum version of BN_from_montgomery() with the working code from
15830 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15835 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15836 made it impossible to use the same SSL_SESSION data structure in
15837 SSL2 clients in multiple threads.
15841 * The return value of RAND_load_file() no longer counts bytes obtained
15842 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15843 to seed the PRNG (previously an explicit byte count was required).
15845 *Ulf Möller, Bodo Möller*
15847 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15848 used `char *` instead of `void *` and had casts all over the place.
15852 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15856 * Retain source code compatibility for BN_prime_checks macro:
15857 BN_is_prime(..., BN_prime_checks, ...) now uses
15858 BN_prime_checks_for_size to determine the appropriate number of
15859 Rabin-Miller iterations.
15863 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15864 DH_CHECK_P_NOT_SAFE_PRIME.
15865 (Check if this is true? OpenPGP calls them "strong".)
15869 * Merge the functionality of "dh" and "gendh" programs into a new program
15870 "dhparam". The old programs are retained for now but will handle DH keys
15871 (instead of parameters) in future.
15875 * Make the ciphers, s_server and s_client programs check the return values
15876 when a new cipher list is set.
15880 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15881 ciphers. Before when the 56bit ciphers were enabled the sorting was
15884 The syntax for the cipher sorting has been extended to support sorting by
15885 cipher-strength (using the strength_bits hard coded in the tables).
15886 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15888 Fix a bug in the cipher-command parser: when supplying a cipher command
15889 string with an "undefined" symbol (neither command nor alphanumeric
15890 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15891 an error is flagged.
15893 Due to the strength-sorting extension, the code of the
15894 ssl_create_cipher_list() function was completely rearranged. I hope that
15895 the readability was also increased :-)
15897 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15899 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15900 for the first serial number and places 2 in the serial number file. This
15901 avoids problems when the root CA is created with serial number zero and
15902 the first user certificate has the same issuer name and serial number
15907 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15908 the new code. Add documentation for this stuff.
15912 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15913 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15914 structures and behave in an analogous way to the X509v3 functions:
15915 they shouldn't be called directly but wrapper functions should be used
15918 So we also now have some wrapper functions that call the X509at functions
15919 when passed certificate requests. (TO DO: similar things can be done with
15920 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15921 things. Some of these need some d2i or i2d and print functionality
15922 because they handle more complex structures.)
15926 * Add missing #ifndefs that caused missing symbols when building libssl
15927 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15928 NO_RSA in `ssl/s2*.c`.
15930 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15932 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15933 has a return value which indicates the quality of the random data
15934 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15935 error queue. New function RAND_pseudo_bytes() generates output that is
15936 guaranteed to be unique but not unpredictable. RAND_add is like
15937 RAND_seed, but takes an extra argument for an entropy estimate
15938 (RAND_seed always assumes full entropy).
15942 * Do more iterations of Rabin-Miller probable prime test (specifically,
15943 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15944 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15945 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15946 false-positive rate of at most 2^-80 for random input.
15950 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15954 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15955 in the 0.9.5 release), this returns the chain
15956 from an X509_CTX structure with a dup of the stack and all
15957 the X509 reference counts upped: so the stack will exist
15958 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15961 Also make SSL_SESSION_print() print out the verify return
15966 * Add manpage for the pkcs12 command. Also change the default
15967 behaviour so MAC iteration counts are used unless the new
15968 -nomaciter option is used. This improves file security and
15969 only older versions of MSIE (4.0 for example) need it.
15973 * Honor the no-xxx Configure options when creating .DEF files.
15977 * Add PKCS#10 attributes to field table: challengePassword,
15978 unstructuredName and unstructuredAddress. These are taken from
15979 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15980 international characters are used.
15982 More changes to X509_ATTRIBUTE code: allow the setting of types
15983 based on strings. Remove the 'loc' parameter when adding
15984 attributes because these will be a SET OF encoding which is sorted
15989 * Initial changes to the 'req' utility to allow request generation
15990 automation. This will allow an application to just generate a template
15991 file containing all the field values and have req construct the
15994 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15995 used all over the place including certificate requests and PKCS#7
15996 structures. They are currently handled manually where necessary with
15997 some primitive wrappers for PKCS#7. The new functions behave in a
15998 manner analogous to the X509 extension functions: they allow
15999 attributes to be looked up by NID and added.
16001 Later something similar to the X509V3 code would be desirable to
16002 automatically handle the encoding, decoding and printing of the
16003 more complex types. The string types like challengePassword can
16004 be handled by the string table functions.
16006 Also modified the multi byte string table handling. Now there is
16007 a 'global mask' which masks out certain types. The table itself
16008 can use the flag STABLE_NO_MASK to ignore the mask setting: this
16009 is useful when for example there is only one permissible type
16010 (as in countryName) and using the mask might result in no valid
16015 * Clean up 'Finished' handling, and add functions SSL_get_finished and
16016 SSL_get_peer_finished to allow applications to obtain the latest
16017 Finished messages sent to the peer or expected from the peer,
16018 respectively. (SSL_get_peer_finished is usually the Finished message
16019 actually received from the peer, otherwise the protocol will be aborted.)
16021 As the Finished message are message digests of the complete handshake
16022 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
16023 be used for external authentication procedures when the authentication
16024 provided by SSL/TLS is not desired or is not enough.
16028 * Enhanced support for Alpha Linux is added. Now ./config checks if
16029 the host supports BWX extension and if Compaq C is present on the
16030 $PATH. Just exploiting of the BWX extension results in 20-30%
16031 performance kick for some algorithms, e.g. DES and RC4 to mention
16032 a couple. Compaq C in turn generates ~20% faster code for MD5 and
16037 * Add support for MS "fast SGC". This is arguably a violation of the
16038 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
16039 weak crypto and after checking the certificate is SGC a second one
16040 with strong crypto. MS SGC stops the first handshake after receiving
16041 the server certificate message and sends a second client hello. Since
16042 a server will typically do all the time consuming operations before
16043 expecting any further messages from the client (server key exchange
16044 is the most expensive) there is little difference between the two.
16046 To get OpenSSL to support MS SGC we have to permit a second client
16047 hello message after we have sent server done. In addition we have to
16048 reset the MAC if we do get this second client hello.
16052 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16053 if a DER encoded private key is RSA or DSA traditional format. Changed
16054 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16055 format DER encoded private key. Newer code should use PKCS#8 format which
16056 has the key type encoded in the ASN1 structure. Added DER private key
16057 support to pkcs8 application.
16061 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16062 ciphersuites has been selected (as required by the SSL 3/TLS 1
16063 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16064 is set, we interpret this as a request to violate the specification
16065 (the worst that can happen is a handshake failure, and 'correct'
16066 behaviour would result in a handshake failure anyway).
16070 * In SSL_CTX_add_session, take into account that there might be multiple
16071 SSL_SESSION structures with the same session ID (e.g. when two threads
16072 concurrently obtain them from an external cache).
16073 The internal cache can handle only one SSL_SESSION with a given ID,
16074 so if there's a conflict, we now throw out the old one to achieve
16079 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16080 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16081 some routines that use cipher OIDs: some ciphers do not have OIDs
16082 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16087 * Simplify the trust setting structure and code. Now we just have
16088 two sequences of OIDs for trusted and rejected settings. These will
16089 typically have values the same as the extended key usage extension
16090 and any application specific purposes.
16092 The trust checking code now has a default behaviour: it will just
16093 check for an object with the same NID as the passed id. Functions can
16094 be provided to override either the default behaviour or the behaviour
16095 for a given id. SSL client, server and email already have functions
16096 in place for compatibility: they check the NID and also return "trusted"
16097 if the certificate is self signed.
16101 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16102 traditional format into an EVP_PKEY structure.
16106 * Add a password callback function PEM_cb() which either prompts for
16107 a password if usr_data is NULL or otherwise assumes it is a null
16108 terminated password. Allow passwords to be passed on command line
16109 environment or config files in a few more utilities.
16113 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16114 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16115 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16116 Update documentation.
16120 * Support for ASN1 "NULL" type. This could be handled before by using
16121 ASN1_TYPE but there wasn't any function that would try to read a NULL
16122 and produce an error if it couldn't. For compatibility we also have
16123 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16124 don't allocate anything because they don't need to.
16128 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16131 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16133 * Rebuild of the memory allocation routines used by OpenSSL code and
16134 possibly others as well. The purpose is to make an interface that
16135 provide hooks so anyone can build a separate set of allocation and
16136 deallocation routines to be used by OpenSSL, for example memory
16137 pool implementations, or something else, which was previously hard
16138 since Malloc(), Realloc() and Free() were defined as macros having
16139 the values malloc, realloc and free, respectively (except for Win32
16140 compilations). The same is provided for memory debugging code.
16141 OpenSSL already comes with functionality to find memory leaks, but
16142 this gives people a chance to debug other memory problems.
16144 With these changes, a new set of functions and macros have appeared:
16146 CRYPTO_set_mem_debug_functions() [F]
16147 CRYPTO_get_mem_debug_functions() [F]
16148 CRYPTO_dbg_set_options() [F]
16149 CRYPTO_dbg_get_options() [F]
16150 CRYPTO_malloc_debug_init() [M]
16152 The memory debug functions are NULL by default, unless the library
16153 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16154 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16155 gives the standard debugging functions that come with OpenSSL) or
16156 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16157 provided by the library user) must be used. When the standard
16158 debugging functions are used, CRYPTO_dbg_set_options can be used to
16159 request additional information:
16160 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16161 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16163 Also, things like CRYPTO_set_mem_functions will always give the
16164 expected result (the new set of functions is used for allocation
16165 and deallocation) at all times, regardless of platform and compiler
16168 To finish it up, some functions that were never use in any other
16169 way than through macros have a new API and new semantic:
16171 CRYPTO_dbg_malloc()
16172 CRYPTO_dbg_realloc()
16175 All macros of value have retained their old syntax.
16177 *Richard Levitte and Bodo Moeller*
16179 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16180 ordering of SMIMECapabilities wasn't in "strength order" and there
16181 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16186 * Some ASN1 types with illegal zero length encoding (INTEGER,
16187 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16189 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16191 * Merge in my S/MIME library for OpenSSL. This provides a simple
16192 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16193 functionality to handle multipart/signed properly) and a utility
16194 called 'smime' to call all this stuff. This is based on code I
16195 originally wrote for Celo who have kindly allowed it to be
16196 included in OpenSSL.
16200 * Add variants des_set_key_checked and des_set_key_unchecked of
16201 des_set_key (aka des_key_sched). Global variable des_check_key
16202 decides which of these is called by des_set_key; this way
16203 des_check_key behaves as it always did, but applications and
16204 the library itself, which was buggy for des_check_key == 1,
16205 have a cleaner way to pick the version they need.
16209 * New function PKCS12_newpass() which changes the password of a
16214 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16215 dynamic mix. In both cases the ids can be used as an index into the
16216 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16217 functions so they accept a list of the field values and the
16218 application doesn't need to directly manipulate the X509_TRUST
16223 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16228 * Modify the way the V3 extension code looks up extensions. This now
16229 works in a similar way to the object code: we have some "standard"
16230 extensions in a static table which is searched with OBJ_bsearch()
16231 and the application can add dynamic ones if needed. The file
16232 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16233 updated whenever a new extension is added to the core code and kept
16234 in ext_nid order. There is a simple program 'tabtest.c' which checks
16235 this. New extensions are not added too often so this file can readily
16236 be maintained manually.
16238 There are two big advantages in doing things this way. The extensions
16239 can be looked up immediately and no longer need to be "added" using
16240 X509V3_add_standard_extensions(): this function now does nothing.
16241 Side note: I get *lots* of email saying the extension code doesn't
16242 work because people forget to call this function.
16243 Also no dynamic allocation is done unless new extensions are added:
16244 so if we don't add custom extensions there is no need to call
16245 X509V3_EXT_cleanup().
16249 * Modify enc utility's salting as follows: make salting the default. Add a
16250 magic header, so unsalted files fail gracefully instead of just decrypting
16251 to garbage. This is because not salting is a big security hole, so people
16252 should be discouraged from doing it.
16256 * Fixes and enhancements to the 'x509' utility. It allowed a message
16257 digest to be passed on the command line but it only used this
16258 parameter when signing a certificate. Modified so all relevant
16259 operations are affected by the digest parameter including the
16260 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16261 DSA key was used because it didn't fix the digest.
16265 * Initial certificate chain verify code. Currently tests the untrusted
16266 certificates for consistency with the verify purpose (which is set
16267 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16269 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16270 this is because it will reject chains with invalid extensions whereas
16271 every previous version of OpenSSL and SSLeay made no checks at all.
16273 Trust code: checks the root CA for the relevant trust settings. Trust
16274 settings have an initial value consistent with the verify purpose: e.g.
16275 if the verify purpose is for SSL client use it expects the CA to be
16276 trusted for SSL client use. However the default value can be changed to
16277 permit custom trust settings: one example of this would be to only trust
16278 certificates from a specific "secure" set of CAs.
16280 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16281 which should be used for version portability: especially since the
16282 verify structure is likely to change more often now.
16284 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16285 to set them. If not set then assume SSL clients will verify SSL servers
16288 Two new options to the verify program: -untrusted allows a set of
16289 untrusted certificates to be passed in and -purpose which sets the
16290 intended purpose of the certificate. If a purpose is set then the
16291 new chain verify code is used to check extension consistency.
16295 * Support for the authority information access extension.
16299 * Modify RSA and DSA PEM read routines to transparently handle
16300 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16301 public keys in a format compatible with certificate
16302 SubjectPublicKeyInfo structures. Unfortunately there were already
16303 functions called *_PublicKey_* which used various odd formats so
16304 these are retained for compatibility: however the DSA variants were
16305 never in a public release so they have been deleted. Changed dsa/rsa
16306 utilities to handle the new format: note no releases ever handled public
16307 keys so we should be OK.
16309 The primary motivation for this change is to avoid the same fiasco
16310 that dogs private keys: there are several incompatible private key
16311 formats some of which are standard and some OpenSSL specific and
16312 require various evil hacks to allow partial transparent handling and
16313 even then it doesn't work with DER formats. Given the option anything
16314 other than PKCS#8 should be dumped: but the other formats have to
16315 stay in the name of compatibility.
16317 With public keys and the benefit of hindsight one standard format
16318 is used which works with EVP_PKEY, RSA or DSA structures: though
16319 it clearly returns an error if you try to read the wrong kind of key.
16321 Added a -pubkey option to the 'x509' utility to output the public key.
16322 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16323 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16324 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16325 that do the same as the `EVP_PKEY_assign_*()` except they up the
16326 reference count of the added key (they don't "swallow" the
16331 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16332 CRLs would fail if the file contained no certificates or no CRLs:
16333 added a new function to read in both types and return the number
16334 read: this means that if none are read it will be an error. The
16335 DER versions of the certificate and CRL reader would always fail
16336 because it isn't possible to mix certificates and CRLs in DER format
16337 without choking one or the other routine. Changed this to just read
16338 a certificate: this is the best we can do. Also modified the code
16339 in `apps/verify.c` to take notice of return codes: it was previously
16340 attempting to read in certificates from NULL pointers and ignoring
16341 any errors: this is one reason why the cert and CRL reader seemed
16342 to work. It doesn't check return codes from the default certificate
16343 routines: these may well fail if the certificates aren't installed.
16347 * Code to support otherName option in GeneralName.
16351 * First update to verify code. Change the verify utility
16352 so it warns if it is passed a self signed certificate:
16353 for consistency with the normal behaviour. X509_verify
16354 has been modified to it will now verify a self signed
16355 certificate if *exactly* the same certificate appears
16356 in the store: it was previously impossible to trust a
16357 single self signed certificate. This means that:
16358 openssl verify ss.pem
16359 now gives a warning about a self signed certificate but
16360 openssl verify -CAfile ss.pem ss.pem
16365 * For servers, store verify_result in SSL_SESSION data structure
16366 (and add it to external session representation).
16367 This is needed when client certificate verifications fails,
16368 but an application-provided verification callback (set by
16369 SSL_CTX_set_cert_verify_callback) allows accepting the session
16370 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16371 but returns 1): When the session is reused, we have to set
16372 ssl->verify_result to the appropriate error code to avoid
16375 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16377 * Fix a bug in the new PKCS#7 code: it didn't consider the
16378 case in PKCS7_dataInit() where the signed PKCS7 structure
16379 didn't contain any existing data because it was being created.
16381 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16383 * Add a salt to the key derivation routines in enc.c. This
16384 forms the first 8 bytes of the encrypted file. Also add a
16385 -S option to allow a salt to be input on the command line.
16389 * New function X509_cmp(). Oddly enough there wasn't a function
16390 to compare two certificates. We do this by working out the SHA1
16391 hash and comparing that. X509_cmp() will be needed by the trust
16396 * SSL_get1_session() is like SSL_get_session(), but increments
16397 the reference count in the SSL_SESSION returned.
16399 *Geoff Thorpe <geoff@eu.c2.net>*
16401 * Fix for 'req': it was adding a null to request attributes.
16402 Also change the X509_LOOKUP and X509_INFO code to handle
16403 certificate auxiliary information.
16407 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16412 * Add the possibility to add extra information to the memory leak
16413 detecting output, to form tracebacks, showing from where each
16414 allocation was originated: CRYPTO_push_info("constant string") adds
16415 the string plus current file name and line number to a per-thread
16416 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16417 is like calling CYRPTO_pop_info() until the stack is empty.
16418 Also updated memory leak detection code to be multi-thread-safe.
16422 * Add options -text and -noout to pkcs7 utility and delete the
16423 encryption options which never did anything. Update docs.
16427 * Add options to some of the utilities to allow the pass phrase
16428 to be included on either the command line (not recommended on
16429 OSes like Unix) or read from the environment. Update the
16430 manpages and fix a few bugs.
16434 * Add a few manpages for some of the openssl commands.
16438 * Fix the -revoke option in ca. It was freeing up memory twice,
16439 leaking and not finding already revoked certificates.
16443 * Extensive changes to support certificate auxiliary information.
16444 This involves the use of X509_CERT_AUX structure and X509_AUX
16445 functions. An X509_AUX function such as PEM_read_X509_AUX()
16446 can still read in a certificate file in the usual way but it
16447 will also read in any additional "auxiliary information". By
16448 doing things this way a fair degree of compatibility can be
16449 retained: existing certificates can have this information added
16450 using the new 'x509' options.
16452 Current auxiliary information includes an "alias" and some trust
16453 settings. The trust settings will ultimately be used in enhanced
16454 certificate chain verification routines: currently a certificate
16455 can only be trusted if it is self signed and then it is trusted
16460 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16461 The problem was that one of the replacement routines had not been working
16462 since SSLeay releases. For now the offending routine has been replaced
16463 with non-optimised assembler. Even so, this now gives around 95%
16464 performance improvement for 1024 bit RSA signs.
16468 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16469 handling. Most clients have the effective key size in bits equal to
16470 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16471 A few however don't do this and instead use the size of the decrypted key
16472 to determine the RC2 key length and the AlgorithmIdentifier to determine
16473 the effective key length. In this case the effective key length can still
16474 be 40 bits but the key length can be 168 bits for example. This is fixed
16475 by manually forcing an RC2 key into the EVP_PKEY structure because the
16476 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16477 the key length and effective key length are equal.
16481 * Add a bunch of functions that should simplify the creation of
16482 X509_NAME structures. Now you should be able to do:
16483 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16484 and have it automatically work out the correct field type and fill in
16485 the structures. The more adventurous can try:
16486 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16487 and it will (hopefully) work out the correct multibyte encoding.
16491 * Change the 'req' utility to use the new field handling and multibyte
16492 copy routines. Before the DN field creation was handled in an ad hoc
16493 way in req, ca, and x509 which was rather broken and didn't support
16494 BMPStrings or UTF8Strings. Since some software doesn't implement
16495 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16496 using the dirstring_type option. See the new comment in the default
16497 openssl.cnf for more info.
16501 * Make crypto/rand/md_rand.c more robust:
16502 - Assure unique random numbers after fork().
16503 - Make sure that concurrent threads access the global counter and
16504 md serializably so that we never lose entropy in them
16505 or use exactly the same state in multiple threads.
16506 Access to the large state is not always serializable because
16507 the additional locking could be a performance killer, and
16508 md should be large enough anyway.
16512 * New file `apps/app_rand.c` with commonly needed functionality
16513 for handling the random seed file.
16515 Use the random seed file in some applications that previously did not:
16517 dsaparam -genkey (which also ignored its '-rand' option),
16520 x509 (when signing).
16521 Except on systems with /dev/urandom, it is crucial to have a random
16522 seed file at least for key creation, DSA signing, and for DH exchanges;
16523 for RSA signatures we could do without one.
16525 gendh and gendsa (unlike genrsa) used to read only the first byte
16526 of each file listed in the '-rand' option. The function as previously
16527 found in genrsa is now in app_rand.c and is used by all programs
16528 that support '-rand'.
16532 * In RAND_write_file, use mode 0600 for creating files;
16533 don't just chmod when it may be too late.
16537 * Report an error from X509_STORE_load_locations
16538 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16542 * New function ASN1_mbstring_copy() this copies a string in either
16543 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16544 into an ASN1_STRING type. A mask of permissible types is passed
16545 and it chooses the "minimal" type to use or an error if not type
16550 * Add function equivalents to the various macros in asn1.h. The old
16551 macros are retained with an `M_` prefix. Code inside the library can
16552 use the `M_` macros. External code (including the openssl utility)
16553 should *NOT* in order to be "shared library friendly".
16557 * Add various functions that can check a certificate's extensions
16558 to see if it usable for various purposes such as SSL client,
16559 server or S/MIME and CAs of these types. This is currently
16560 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16561 verification. Also added a -purpose flag to x509 utility to
16562 print out all the purposes.
16566 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16571 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16572 for, obtain and decode and extension and obtain its critical flag.
16573 This allows all the necessary extension code to be handled in a
16574 single function call.
16578 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16579 platforms. See crypto/rc4/rc4_enc.c for further details.
16583 * New -noout option to asn1parse. This causes no output to be produced
16584 its main use is when combined with -strparse and -out to extract data
16585 from a file (which may not be in ASN.1 format).
16589 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16590 when producing the local key id.
16592 *Richard Levitte <levitte@stacken.kth.se>*
16594 * New option -dhparam in s_server. This allows a DH parameter file to be
16595 stated explicitly. If it is not stated then it tries the first server
16596 certificate file. The previous behaviour hard coded the filename
16601 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16602 a public key to be input or output. For example:
16603 openssl rsa -in key.pem -pubout -out pubkey.pem
16604 Also added necessary DSA public key functions to handle this.
16608 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16609 in the message. This was handled by allowing
16610 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16612 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16614 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16615 to the end of the strings whereas this didn't. This would cause problems
16616 if strings read with d2i_ASN1_bytes() were later modified.
16618 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16620 * Fix for base64 decode bug. When a base64 bio reads only one line of
16621 data and it contains EOF it will end up returning an error. This is
16622 caused by input 46 bytes long. The cause is due to the way base64
16623 BIOs find the start of base64 encoded data. They do this by trying a
16624 trial decode on each line until they find one that works. When they
16625 do a flag is set and it starts again knowing it can pass all the
16626 data directly through the decoder. Unfortunately it doesn't reset
16627 the context it uses. This means that if EOF is reached an attempt
16628 is made to pass two EOFs through the context and this causes the
16629 resulting error. This can also cause other problems as well. As is
16630 usual with these problems it takes *ages* to find and the fix is
16631 trivial: move one line.
16633 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16635 * Ugly workaround to get s_client and s_server working under Windows. The
16636 old code wouldn't work because it needed to select() on sockets and the
16637 tty (for keypresses and to see if data could be written). Win32 only
16638 supports select() on sockets so we select() with a 1s timeout on the
16639 sockets and then see if any characters are waiting to be read, if none
16640 are present then we retry, we also assume we can always write data to
16641 the tty. This isn't nice because the code then blocks until we've
16642 received a complete line of data and it is effectively polling the
16643 keyboard at 1s intervals: however it's quite a bit better than not
16644 working at all :-) A dedicated Windows application might handle this
16645 with an event loop for example.
16649 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16650 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16651 will be called when RSA_sign() and RSA_verify() are used. This is useful
16652 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16653 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16654 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16655 This necessitated the support of an extra signature type NID_md5_sha1
16656 for SSL signatures and modifications to the SSL library to use it instead
16657 of calling RSA_public_decrypt() and RSA_private_encrypt().
16661 * Add new -verify -CAfile and -CApath options to the crl program, these
16662 will lookup a CRL issuers certificate and verify the signature in a
16663 similar way to the verify program. Tidy up the crl program so it
16664 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16665 less strict. It will now permit CRL extensions even if it is not
16666 a V2 CRL: this will allow it to tolerate some broken CRLs.
16670 * Initialize all non-automatic variables each time one of the openssl
16671 sub-programs is started (this is necessary as they may be started
16672 multiple times from the "OpenSSL>" prompt).
16674 *Lennart Bang, Bodo Moeller*
16676 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16677 removing all other RSA functionality (this is what NO_RSA does). This
16678 is so (for example) those in the US can disable those operations covered
16679 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16684 * Non-copying interface to BIO pairs.
16685 (still largely untested)
16689 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16690 ASCII string. This was handled independently in various places before.
16694 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16695 UTF8 strings a character at a time.
16699 * Use client_version from client hello to select the protocol
16700 (s23_srvr.c) and for RSA client key exchange verification
16701 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16705 * Add various utility functions to handle SPKACs, these were previously
16706 handled by poking round in the structure internals. Added new function
16707 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16708 print, verify and generate SPKACs. Based on an original idea from
16709 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16713 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16717 * Allow the config file extension section to be overwritten on the
16718 command line. Based on an original idea from Massimiliano Pala
16719 <madwolf@comune.modena.it>. The new option is called -extensions
16720 and can be applied to ca, req and x509. Also -reqexts to override
16721 the request extensions in req and -crlexts to override the crl extensions
16726 * Add new feature to the SPKAC handling in ca. Now you can include
16727 the same field multiple times by preceding it by "XXXX." for example:
16730 this is the same syntax as used in the req config file.
16734 * Allow certificate extensions to be added to certificate requests. These
16735 are specified in a 'req_extensions' option of the req section of the
16736 config file. They can be printed out with the -text option to req but
16737 are otherwise ignored at present.
16741 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16742 data read consists of only the final block it would not decrypted because
16743 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16744 A misplaced 'break' also meant the decrypted final block might not be
16745 copied until the next read.
16749 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16750 a few extra parameters to the DH structure: these will be useful if
16751 for example we want the value of 'q' or implement X9.42 DH.
16755 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16756 provides hooks that allow the default DSA functions or functions on a
16757 "per key" basis to be replaced. This allows hardware acceleration and
16758 hardware key storage to be handled without major modification to the
16759 library. Also added low level modexp hooks and CRYPTO_EX structure and
16760 associated functions.
16764 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16765 as "read only": it can't be written to and the buffer it points to will
16766 not be freed. Reading from a read only BIO is much more efficient than
16767 a normal memory BIO. This was added because there are several times when
16768 an area of memory needs to be read from a BIO. The previous method was
16769 to create a memory BIO and write the data to it, this results in two
16770 copies of the data and an O(n^2) reading algorithm. There is a new
16771 function BIO_new_mem_buf() which creates a read only memory BIO from
16772 an area of memory. Also modified the PKCS#7 routines to use read only
16777 * Bugfix: ssl23_get_client_hello did not work properly when called in
16778 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16779 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16780 but a retry condition occurred while trying to read the rest.
16784 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16785 NID_pkcs7_encrypted by default: this was wrong since this should almost
16786 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16787 the encrypted data type: this is a more sensible place to put it and it
16788 allows the PKCS#12 code to be tidied up that duplicated this
16793 * Changed obj_dat.pl script so it takes its input and output files on
16794 the command line. This should avoid shell escape redirection problems
16799 * Initial support for certificate extension requests, these are included
16800 in things like Xenroll certificate requests. Included functions to allow
16801 extensions to be obtained and added.
16805 * -crlf option to s_client and s_server for sending newlines as
16806 CRLF (as required by many protocols).
16810 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16812 * Install libRSAglue.a when OpenSSL is built with RSAref.
16814 *Ralf S. Engelschall*
16816 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16818 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16820 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16825 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16826 DH parameters/keys (q is lost during that conversion, but the resulting
16827 DH parameters contain its length).
16829 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16830 much faster than DH_generate_parameters (which creates parameters
16831 where `p = 2*q + 1`), and also the smaller q makes DH computations
16832 much more efficient (160-bit exponentiation instead of 1024-bit
16833 exponentiation); so this provides a convenient way to support DHE
16834 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16835 utter importance to use
16836 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16838 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16839 when such DH parameters are used, because otherwise small subgroup
16840 attacks may become possible!
16844 * Avoid memory leak in i2d_DHparams.
16848 * Allow the -k option to be used more than once in the enc program:
16849 this allows the same encrypted message to be read by multiple recipients.
16853 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16854 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16855 it will always use the numerical form of the OID, even if it has a short
16860 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16861 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16862 otherwise bn_mod_exp was called. In the case of hardware keys for example
16863 no private key components need be present and it might store extra data
16864 in the RSA structure, which cannot be accessed from bn_mod_exp.
16865 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16866 private key operations.
16870 * Added support for SPARC Linux.
16874 * pem_password_cb function type incompatibly changed from
16875 typedef int pem_password_cb(char *buf, int size, int rwflag);
16877 ....(char *buf, int size, int rwflag, void *userdata);
16878 so that applications can pass data to their callbacks:
16879 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16880 additional void * argument, which is just handed through whenever
16881 the password callback is called.
16883 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16885 New function SSL_CTX_set_default_passwd_cb_userdata.
16887 Compatibility note: As many C implementations push function arguments
16888 onto the stack in reverse order, the new library version is likely to
16889 interoperate with programs that have been compiled with the old
16890 pem_password_cb definition (PEM_whatever takes some data that
16891 happens to be on the stack as its last argument, and the callback
16892 just ignores this garbage); but there is no guarantee whatsoever that
16895 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16896 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16897 problems not only on Windows, but also on some Unix platforms.
16898 To avoid problematic command lines, these definitions are now in an
16899 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16900 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16904 * MIPS III/IV assembler module is reimplemented.
16908 * More DES library cleanups: remove references to srand/rand and
16909 delete an unused file.
16913 * Add support for the free Netwide assembler (NASM) under Win32,
16914 since not many people have MASM (ml) and it can be hard to obtain.
16915 This is currently experimental but it seems to work OK and pass all
16916 the tests. Check out INSTALL.W32 for info.
16920 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16921 without temporary keys kept an extra copy of the server key,
16922 and connections with temporary keys did not free everything in case
16927 * New function RSA_check_key and new openssl rsa option -check
16928 for verifying the consistency of RSA keys.
16930 *Ulf Moeller, Bodo Moeller*
16932 * Various changes to make Win32 compile work:
16933 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16934 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16935 comparison" warnings.
16936 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16940 * Add a debugging option to PKCS#5 v2 key generation function: when
16941 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16942 derived keys are printed to stderr.
16946 * Copy the flags in ASN1_STRING_dup().
16948 *Roman E. Pavlov <pre@mo.msk.ru>*
16950 * The x509 application mishandled signing requests containing DSA
16951 keys when the signing key was also DSA and the parameters didn't match.
16953 It was supposed to omit the parameters when they matched the signing key:
16954 the verifying software was then supposed to automatically use the CA's
16955 parameters if they were absent from the end user certificate.
16957 Omitting parameters is no longer recommended. The test was also
16958 the wrong way round! This was probably due to unusual behaviour in
16959 EVP_cmp_parameters() which returns 1 if the parameters match.
16960 This meant that parameters were omitted when they *didn't* match and
16961 the certificate was useless. Certificates signed with 'ca' didn't have
16964 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16966 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16967 The interface is as follows:
16968 Applications can use
16969 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16970 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16971 "off" is now the default.
16972 The library internally uses
16973 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16974 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16975 to disable memory-checking temporarily.
16977 Some inconsistent states that previously were possible (and were
16978 even the default) are now avoided.
16980 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16981 with each memory chunk allocated; this is occasionally more helpful
16982 than just having a counter.
16984 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16986 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16991 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16992 which largely parallels "options", but is for changing API behaviour,
16993 whereas "options" are about protocol behaviour.
16994 Initial "mode" flags are:
16996 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16997 a single record has been written.
16998 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16999 retries use the same buffer location.
17000 (But all of the contents must be
17005 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
17008 * Fix problems with no-hmac etc.
17010 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
17012 * New functions RSA_get_default_method(), RSA_set_method() and
17013 RSA_get_method(). These allows replacement of RSA_METHODs without having
17014 to mess around with the internals of an RSA structure.
17018 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
17019 Also really enable memory leak checks in openssl.c and in some
17022 *Chad C. Mulligan, Bodo Moeller*
17024 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
17025 up the length of negative integers. This has now been simplified to just
17026 store the length when it is first determined and use it later, rather
17027 than trying to keep track of where data is copied and updating it to
17029 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
17031 * Add a new function PKCS7_signatureVerify. This allows the verification
17032 of a PKCS#7 signature but with the signing certificate passed to the
17033 function itself. This contrasts with PKCS7_dataVerify which assumes the
17034 certificate is present in the PKCS#7 structure. This isn't always the
17035 case: certificates can be omitted from a PKCS#7 structure and be
17036 distributed by "out of band" means (such as a certificate database).
17040 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
17041 function prototypes in pem.h, also change util/mkdef.pl to add the
17042 necessary function names.
17046 * mk1mf.pl (used by Windows builds) did not properly read the
17047 options set by Configure in the top level Makefile, and Configure
17048 was not even able to write more than one option correctly.
17049 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17053 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17054 file to be loaded from a BIO or FILE pointer. The BIO version will
17055 for example allow memory BIOs to contain config info.
17059 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17060 Whoever hopes to achieve shared-library compatibility across versions
17061 must use this, not the compile-time macro.
17062 (Exercise 0.9.4: Which is the minimum library version required by
17064 Note: All this applies only to multi-threaded programs, others don't
17069 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17070 through a BIO pair triggered the default case, i.e.
17071 SSLerr(...,SSL_R_UNKNOWN_STATE).
17075 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17076 can use the SSL library even if none of the specific BIOs is
17081 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17082 for the encoded length.
17084 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17086 * Add initial documentation of the X509V3 functions.
17090 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17091 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17092 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17093 secure PKCS#8 private key format with a high iteration count.
17097 * Fix determination of Perl interpreter: A perl or perl5
17098 *directory* in $PATH was also accepted as the interpreter.
17100 *Ralf S. Engelschall*
17102 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17103 wrong with it but it was very old and did things like calling
17104 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17105 unusual formatting.
17109 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17110 to use the new extension code.
17114 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17115 with macros. This should make it easier to change their form, add extra
17116 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17121 * Add to configuration table a new entry that can specify an alternative
17122 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17123 according to Mark Crispin <MRC@Panda.COM>.
17127 * DES CBC did not update the IV. Weird.
17131 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17132 Changing the behaviour of the former might break existing programs --
17133 where IV updating is needed, des_ncbc_encrypt can be used.
17136 * When bntest is run from "make test" it drives bc to check its
17137 calculations, as well as internally checking them. If an internal check
17138 fails, it needs to cause bc to give a non-zero result or make test carries
17139 on without noticing the failure. Fixed.
17143 * DES library cleanups.
17147 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17148 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17149 ciphers. NOTE: although the key derivation function has been verified
17150 against some published test vectors it has not been extensively tested
17151 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17156 * Instead of "mkdir -p", which is not fully portable, use new
17157 Perl script "util/mkdir-p.pl".
17161 * Rewrite the way password based encryption (PBE) is handled. It used to
17162 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17163 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17164 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17165 the 'parameter' field of the AlgorithmIdentifier is passed to the
17166 underlying key generation function so it must do its own ASN1 parsing.
17167 This has also changed the EVP_PBE_CipherInit() function which now has a
17168 'parameter' argument instead of literal salt and iteration count values
17169 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17173 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17174 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17175 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17176 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17177 value was just used as a "magic string" and not used directly its
17178 value doesn't matter.
17182 * Introduce some semblance of const correctness to BN. Shame C doesn't
17187 * "linux-sparc64" configuration (ultrapenguin).
17189 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17190 "linux-sparc" configuration.
17192 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17194 * config now generates no-xxx options for missing ciphers.
17198 * Support the EBCDIC character set (work in progress).
17199 File ebcdic.c not yet included because it has a different license.
17201 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17203 * Support BS2000/OSD-POSIX.
17205 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17207 * Make callbacks for key generation use `void *` instead of `char *`.
17211 * Make S/MIME samples compile (not yet tested).
17215 * Additional typesafe stacks.
17219 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17223 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17225 * New configuration variant "sco5-gcc".
17227 * Updated some demos.
17229 *Sean O Riordain, Wade Scholine*
17231 * Add missing BIO_free at exit of pkcs12 application.
17235 * Fix memory leak in conf.c.
17239 * Updates for Win32 to assembler version of MD5.
17243 * Set #! path to perl in `apps/der_chop` to where we found it
17244 instead of using a fixed path.
17248 * SHA library changes for irix64-mips4-cc.
17252 * Improvements for VMS support.
17256 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17258 * Bignum library bug fix. IRIX 6 passes "make test" now!
17259 This also avoids the problems with SC4.2 and unpatched SC5.
17261 *Andy Polyakov <appro@fy.chalmers.se>*
17263 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17264 These are required because of the typesafe stack would otherwise break
17265 existing code. If old code used a structure member which used to be STACK
17266 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17267 sk_num or sk_value it would produce an error because the num, data members
17268 are not present in STACK_OF. Now it just produces a warning. sk_set
17269 replaces the old method of assigning a value to sk_value
17270 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17271 that does this will no longer work (and should use sk_set instead) but
17272 this could be regarded as a "questionable" behaviour anyway.
17276 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17277 correctly handle encrypted S/MIME data.
17281 * Change type of various DES function arguments from des_cblock
17282 (which means, in function argument declarations, pointer to char)
17283 to des_cblock * (meaning pointer to array with 8 char elements),
17284 which allows the compiler to do more typechecking; it was like
17285 that back in SSLeay, but with lots of ugly casts.
17287 Introduce new type const_des_cblock.
17291 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17292 problems: find RecipientInfo structure that matches recipient certificate
17293 and initialise the ASN1 structures properly based on passed cipher.
17297 * Belatedly make the BN tests actually check the results.
17301 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17302 to and from BNs: it was completely broken. New compilation option
17303 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17304 key elements as negative integers.
17308 * Reorganize and speed up MD5.
17310 *Andy Polyakov <appro@fy.chalmers.se>*
17314 *Richard Levitte <richard@levitte.org>*
17316 * New option -out to asn1parse to allow the parsed structure to be
17317 output to a file. This is most useful when combined with the -strparse
17318 option to examine the output of things like OCTET STRINGS.
17322 * Make SSL library a little more fool-proof by not requiring any longer
17323 that `SSL_set_{accept,connect}_state` be called before
17324 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17325 in many applications because usually everything *appeared* to work as
17326 intended anyway -- now it really works as intended).
17330 * Move openssl.cnf out of lib/.
17334 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17335 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17336 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17338 *Ralf S. Engelschall*
17340 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17341 handle PKCS#7 enveloped data properly.
17343 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17345 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17346 copying pointers. The cert_st handling is changed by this in
17347 various ways (and thus what used to be known as ctx->default_cert
17348 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17349 any longer when s->cert does not give us what we need).
17350 ssl_cert_instantiate becomes obsolete by this change.
17351 As soon as we've got the new code right (possibly it already is?),
17352 we have solved a couple of bugs of the earlier code where s->cert
17353 was used as if it could not have been shared with other SSL structures.
17355 Note that using the SSL API in certain dirty ways now will result
17356 in different behaviour than observed with earlier library versions:
17357 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17358 does not influence s as it used to.
17360 In order to clean up things more thoroughly, inside SSL_SESSION
17361 we don't use CERT any longer, but a new structure SESS_CERT
17362 that holds per-session data (if available); currently, this is
17363 the peer's certificate chain and, for clients, the server's certificate
17364 and temporary key. CERT holds only those values that can have
17365 meaningful defaults in an SSL_CTX.
17369 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17370 from the internal representation. Various PKCS#7 fixes: remove some
17371 evil casts and set the enc_dig_alg field properly based on the signing
17376 * Allow PKCS#12 password to be set from the command line or the
17377 environment. Let 'ca' get its config file name from the environment
17378 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17383 * Allow certificate policies extension to use an IA5STRING for the
17384 organization field. This is contrary to the PKIX definition but
17385 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17390 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17391 without disallowing inline assembler and the like for non-pedantic builds.
17395 * Support Borland C++ builder.
17397 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17403 * SHA-1 cleanups and performance enhancements.
17405 *Andy Polyakov <appro@fy.chalmers.se>*
17407 * Sparc v8plus assembler for the bignum library.
17409 *Andy Polyakov <appro@fy.chalmers.se>*
17411 * Accept any -xxx and +xxx compiler options in Configure.
17415 * Update HPUX configuration.
17419 * Add missing `sk_<type>_unshift()` function to safestack.h
17421 *Ralf S. Engelschall*
17423 * New function SSL_CTX_use_certificate_chain_file that sets the
17424 "extra_cert"s in addition to the certificate. (This makes sense
17425 only for "PEM" format files, as chains as a whole are not
17430 * Support verify_depth from the SSL API.
17431 x509_vfy.c had what can be considered an off-by-one-error:
17432 Its depth (which was not part of the external interface)
17433 was actually counting the number of certificates in a chain;
17434 now it really counts the depth.
17438 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17439 instead of X509err, which often resulted in confusing error
17440 messages since the error codes are not globally unique
17441 (e.g. an alleged error in ssl3_accept when a certificate
17442 didn't match the private key).
17444 * New function SSL_CTX_set_session_id_context that allows to set a default
17445 value (so that you don't need SSL_set_session_id_context for each
17446 connection using the SSL_CTX).
17450 * OAEP decoding bug fix.
17454 * Support INSTALL_PREFIX for package builders, as proposed by
17459 * New Configure options "threads" and "no-threads". For systems
17460 where the proper compiler options are known (currently Solaris
17461 and Linux), "threads" is the default.
17465 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17469 * Install various scripts to $(OPENSSLDIR)/misc, not to
17470 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17471 such as /usr/local/bin.
17475 * "make linux-shared" to build shared libraries.
17477 *Niels Poppe <niels@netbox.org>*
17479 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17483 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17484 extension adding in x509 utility.
17488 * Remove NOPROTO sections and error code comments.
17492 * Partial rewrite of the DEF file generator to now parse the ANSI
17497 * New Configure options --prefix=DIR and --openssldir=DIR.
17501 * Complete rewrite of the error code script(s). It is all now handled
17502 by one script at the top level which handles error code gathering,
17503 header rewriting and C source file generation. It should be much better
17504 than the old method: it now uses a modified version of Ulf's parser to
17505 read the ANSI prototypes in all header files (thus the old K&R definitions
17506 aren't needed for error creation any more) and do a better job of
17507 translating function codes into names. The old 'ASN1 error code embedded
17508 in a comment' is no longer necessary and it doesn't use .err files which
17509 have now been deleted. Also the error code call doesn't have to appear all
17510 on one line (which resulted in some large lines...).
17514 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17518 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17519 0 (which usually indicates a closed connection), but continue reading.
17523 * Fix some race conditions.
17527 * Add support for CRL distribution points extension. Add Certificate
17528 Policies and CRL distribution points documentation.
17532 * Move the autogenerated header file parts to crypto/opensslconf.h.
17536 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17537 8 of keying material. Merlin has also confirmed interop with this fix
17538 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17540 *Merlin Hughes <merlin@baltimore.ie>*
17542 * Fix lots of warnings.
17544 *Richard Levitte <levitte@stacken.kth.se>*
17546 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17547 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17549 *Richard Levitte <levitte@stacken.kth.se>*
17551 * Fix problems with sizeof(long) == 8.
17553 *Andy Polyakov <appro@fy.chalmers.se>*
17555 * Change functions to ANSI C.
17559 * Fix typos in error codes.
17561 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17563 * Remove defunct assembler files from Configure.
17567 * SPARC v8 assembler BIGNUM implementation.
17569 *Andy Polyakov <appro@fy.chalmers.se>*
17571 * Support for Certificate Policies extension: both print and set.
17572 Various additions to support the r2i method this uses.
17576 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17577 return a const string when you are expecting an allocated buffer.
17581 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17582 types DirectoryString and DisplayText.
17586 * Add code to allow r2i extensions to access the configuration database,
17587 add an LHASH database driver and add several ctx helper functions.
17591 * Fix an evil bug in bn_expand2() which caused various BN functions to
17592 fail when they extended the size of a BIGNUM.
17596 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17597 support typesafe stack.
17601 * Fix typo in SSL_[gs]et_options().
17603 *Nils Frostberg <nils@medcom.se>*
17605 * Delete various functions and files that belonged to the (now obsolete)
17606 old X509V3 handling code.
17610 * New Configure option "rsaref".
17614 * Don't auto-generate pem.h.
17618 * Introduce type-safe ASN.1 SETs.
17622 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17624 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17626 * Introduce type-safe STACKs. This will almost certainly break lots of code
17627 that links with OpenSSL (well at least cause lots of warnings), but fear
17628 not: the conversion is trivial, and it eliminates loads of evil casts. A
17629 few STACKed things have been converted already. Feel free to convert more.
17630 In the fullness of time, I'll do away with the STACK type altogether.
17634 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17635 specified in `<certfile>` by updating the entry in the index.txt file.
17636 This way one no longer has to edit the index.txt file manually for
17637 revoking a certificate. The -revoke option does the gory details now.
17639 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17641 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17642 `-text` option at all and this way the `-noout -text` combination was
17643 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17645 *Ralf S. Engelschall*
17647 * Make sure a corresponding plain text error message exists for the
17648 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17649 verify callback function determined that a certificate was revoked.
17651 *Ralf S. Engelschall*
17653 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17654 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17655 all available ciphers including rc5, which was forgotten until now.
17656 In order to let the testing shell script know which algorithms
17657 are available, a new (up to now undocumented) command
17658 `openssl list-cipher-commands` is used.
17662 * Bugfix: s_client occasionally would sleep in select() when
17663 it should have checked SSL_pending() first.
17667 * New functions DSA_do_sign and DSA_do_verify to provide access to
17668 the raw DSA values prior to ASN.1 encoding.
17672 * Tweaks to Configure
17674 *Niels Poppe <niels@netbox.org>*
17676 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17681 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17685 * New config option to avoid instructions that are illegal on the 80386.
17686 The default code is faster, but requires at least a 486.
17690 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17691 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17692 same as SSL2_VERSION anyway.
17696 * New "-showcerts" option for s_client.
17700 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17701 application. Various cleanups and fixes.
17705 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17706 modify error routines to work internally. Add error codes and PBE init
17707 to library startup routines.
17711 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17712 packing functions to asn1 and evp. Changed function names and error
17713 codes along the way.
17717 * PKCS12 integration: and so it begins... First of several patches to
17718 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17719 objects to objects.h
17723 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17724 and display support for Thawte strong extranet extension.
17728 * Add LinuxPPC support.
17730 *Jeff Dubrule <igor@pobox.org>*
17732 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17733 bn_div_words in alpha.s.
17735 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17737 * Make sure the RSA OAEP test is skipped under -DRSAref because
17738 OAEP isn't supported when OpenSSL is built with RSAref.
17740 *Ulf Moeller <ulf@fitug.de>*
17742 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17743 so they no longer are missing under -DNOPROTO.
17745 *Soren S. Jorvang <soren@t.dk>*
17747 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17749 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17750 doesn't work when the session is reused. Coming soon!
17754 * Fix a security hole, that allows sessions to be reused in the wrong
17755 context thus bypassing client cert protection! All software that uses
17756 client certs and session caches in multiple contexts NEEDS PATCHING to
17757 allow session reuse! A fuller solution is in the works.
17759 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17761 * Some more source tree cleanups (removed obsolete files
17762 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17763 permission on "config" script to be executable) and a fix for the INSTALL
17766 *Ulf Moeller <ulf@fitug.de>*
17768 * Remove some legacy and erroneous uses of malloc, free instead of
17771 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17773 * Make rsa_oaep_test return non-zero on error.
17775 *Ulf Moeller <ulf@fitug.de>*
17777 * Add support for native Solaris shared libraries. Configure
17778 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17779 if someone would make that last step automatic.
17781 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17783 * ctx_size was not built with the right compiler during "make links". Fixed.
17787 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17788 except NULL ciphers". This means the default cipher list will no longer
17789 enable NULL ciphers. They need to be specifically enabled e.g. with
17790 the string "DEFAULT:eNULL".
17794 * Fix to RSA private encryption routines: if p < q then it would
17795 occasionally produce an invalid result. This will only happen with
17796 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17800 * Be less restrictive and allow also `perl util/perlpath.pl
17801 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17802 because this way one can also use an interpreter named `perl5` (which is
17803 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17804 installed as `perl`).
17806 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17808 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17810 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17812 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17813 advapi32.lib to Win32 build and change the pem test comparison
17814 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17815 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17816 and crypto/des/ede_cbcm_enc.c.
17820 * DES quad checksum was broken on big-endian architectures. Fixed.
17824 * Comment out two functions in bio.h that aren't implemented. Fix up the
17825 Win32 test batch file so it (might) work again. The Win32 test batch file
17826 is horrible: I feel ill....
17830 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17831 in e_os.h. Audit of header files to check ANSI and non ANSI
17832 sections: 10 functions were absent from non ANSI section and not exported
17833 from Windows DLLs. Fixed up libeay.num for new functions.
17837 * Make `openssl version` output lines consistent.
17839 *Ralf S. Engelschall*
17841 * Fix Win32 symbol export lists for BIO functions: Added
17842 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17843 to ms/libeay{16,32}.def.
17845 *Ralf S. Engelschall*
17847 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17848 fine under Unix and passes some trivial tests I've now added. But the
17849 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17850 added to make sure no one expects that this stuff really works in the
17851 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17852 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17855 *Ralf S. Engelschall*
17857 * Fix the generation of two part addresses in perl.
17859 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17861 * Add config entry for Linux on MIPS.
17863 *John Tobey <jtobey@channel1.com>*
17865 * Make links whenever Configure is run, unless we are on Windoze.
17869 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17870 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17875 * Add a useful kludge to allow package maintainers to specify compiler and
17876 other platforms details on the command line without having to patch the
17877 Configure script every time: One now can use
17878 `perl Configure <id>:<details>`,
17879 i.e. platform ids are allowed to have details appended
17880 to them (separated by colons). This is treated as there would be a static
17881 pre-configured entry in Configure's %table under key `<id>` with value
17882 `<details>` and `perl Configure <id>` is called. So, when you want to
17883 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17884 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17885 now, which overrides the FreeBSD-elf entry on-the-fly.
17887 *Ralf S. Engelschall*
17889 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17893 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17894 on the `perl Configure ...` command line. This way one can compile
17895 OpenSSL libraries with Position Independent Code (PIC) which is needed
17896 for linking it into DSOs.
17898 *Ralf S. Engelschall*
17900 * Remarkably, export ciphers were totally broken and no-one had noticed!
17905 * Cleaned up the LICENSE document: The official contact for any license
17906 questions now is the OpenSSL core team under openssl-core@openssl.org.
17907 And add a paragraph about the dual-license situation to make sure people
17908 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17909 to the OpenSSL toolkit.
17911 *Ralf S. Engelschall*
17913 * General source tree makefile cleanups: Made `making xxx in yyy...`
17914 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17915 Additionally cleaned up the `make links` target: Remove unnecessary
17916 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17917 to speed processing and no longer clutter the display with confusing
17918 stuff. Instead only the actually done links are displayed.
17920 *Ralf S. Engelschall*
17922 * Permit null encryption ciphersuites, used for authentication only. It used
17923 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17924 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17929 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17930 signed attributes when verifying signatures (this would break them),
17931 the detached data encoding was wrong and public keys obtained using
17932 X509_get_pubkey() weren't freed.
17936 * Add text documentation for the BUFFER functions. Also added a work around
17937 to a Win95 console bug. This was triggered by the password read stuff: the
17938 last character typed gets carried over to the next fread(). If you were
17939 generating a new cert request using 'req' for example then the last
17940 character of the passphrase would be CR which would then enter the first
17945 * Added the new 'Includes OpenSSL Cryptography Software' button as
17946 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17947 button and can be used by applications based on OpenSSL to show the
17948 relationship to the OpenSSL project.
17950 *Ralf S. Engelschall*
17952 * Remove confusing variables in function signatures in files
17953 ssl/ssl_lib.c and ssl/ssl.h.
17955 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17957 * Don't install bss_file.c under PREFIX/include/
17959 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17961 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17962 functions that return function pointers and has support for NT specific
17963 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17964 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17965 unsigned to signed types: this was killing the Win32 compile.
17969 * Add new certificate file to stack functions,
17970 SSL_add_dir_cert_subjects_to_stack() and
17971 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17972 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17973 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17974 This means that Apache-SSL and similar packages don't have to mess around
17975 to add as many CAs as they want to the preferred list.
17979 * Experiment with doxygen documentation. Currently only partially applied to
17981 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17982 openssl.doxy as the configuration file.
17986 * Get rid of remaining C++-style comments which strict C compilers hate.
17988 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17990 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17991 compiled in by default: it has problems with large keys.
17995 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17996 DH private keys and/or callback functions which directly correspond to
17997 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17998 is needed for applications which have to configure certificates on a
17999 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
18001 For the RSA certificate situation is makes no difference, but
18002 for the DSA certificate situation this fixes the "no shared cipher"
18003 problem where the OpenSSL cipher selection procedure failed because the
18004 temporary keys were not overtaken from the context and the API provided
18005 no way to reconfigure them.
18006 The new functions now let applications reconfigure the stuff and they
18007 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
18008 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
18009 non-public-API function ssl_cert_instantiate() is used as a helper
18010 function and also to reduce code redundancy inside ssl_rsa.c.
18012 *Ralf S. Engelschall*
18014 * Move s_server -dcert and -dkey options out of the undocumented feature
18015 area because they are useful for the DSA situation and should be
18016 recognized by the users.
18018 *Ralf S. Engelschall*
18020 * Fix the cipher decision scheme for export ciphers: the export bits are
18021 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
18022 SSL_EXP_MASK. So, the original variable has to be used instead of the
18023 already masked variable.
18025 *Richard Levitte <levitte@stacken.kth.se>*
18027 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
18029 *Richard Levitte <levitte@stacken.kth.se>*
18031 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
18032 from `int` to `unsigned int` because it is a length and initialized by
18033 EVP_DigestFinal() which expects an `unsigned int *`.
18035 *Richard Levitte <levitte@stacken.kth.se>*
18037 * Don't hard-code path to Perl interpreter on shebang line of Configure
18038 script. Instead use the usual Shell->Perl transition trick.
18040 *Ralf S. Engelschall*
18042 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
18043 (in addition to RSA certificates) to match the behaviour of `openssl dsa
18044 -noout -modulus` as it's already the case for `openssl rsa -noout
18045 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18046 currently the public key is printed (a decision which was already done by
18047 `openssl dsa -modulus` in the past) which serves a similar purpose.
18048 Additionally the NO_RSA no longer completely removes the whole -modulus
18049 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18052 *Ralf S. Engelschall*
18054 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18055 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18057 *Arne Ansper <arne@ats.cyber.ee>*
18059 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18060 to be added. Now both 'req' and 'ca' can use new objects defined in the
18065 * Add cool BIO that does syslog (or event log on NT).
18067 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18069 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18070 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18071 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18072 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18076 * Add preliminary config info for new extension code.
18080 * Make RSA_NO_PADDING really use no padding.
18082 *Ulf Moeller <ulf@fitug.de>*
18084 * Generate errors when private/public key check is done.
18088 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18089 for some CRL extensions and new objects added.
18093 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18094 key usage extension and fuller support for authority key id.
18098 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18099 padding method for RSA, which is recommended for new applications in PKCS
18100 #1 v2.0 (RFC 2437, October 1998).
18101 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18102 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18103 against Bleichbacher's attack on RSA.
18104 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18107 * Updates to the new SSL compression code
18109 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18111 * Fix so that the version number in the master secret, when passed
18112 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18113 (because the server will not accept higher), that the version number
18114 is 0x03,0x01, not 0x03,0x00
18116 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18118 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18119 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18120 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18124 * Support for RAW extensions where an arbitrary extension can be
18125 created by including its DER encoding. See `apps/openssl.cnf` for
18130 * Make sure latest Perl versions don't interpret some generated C array
18131 code as Perl array code in the crypto/err/err_genc.pl script.
18133 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18135 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18136 not many people have the assembler. Various Win32 compilation fixes and
18137 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18138 build instructions.
18142 * Modify configure script 'Configure' to automatically create crypto/date.h
18143 file under Win32 and also build pem.h from pem.org. New script
18144 util/mkfiles.pl to create the MINFO file on environments that can't do a
18145 'make files': perl util/mkfiles.pl >MINFO should work.
18149 * Major rework of DES function declarations, in the pursuit of correctness
18150 and purity. As a result, many evil casts evaporated, and some weirdness,
18151 too. You may find this causes warnings in your code. Zapping your evil
18152 casts will probably fix them. Mostly.
18156 * Fix for a typo in asn1.h. Bug fix to object creation script
18157 obj_dat.pl. It considered a zero in an object definition to mean
18158 "end of object": none of the objects in objects.h have any zeros
18159 so it wasn't spotted.
18161 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18163 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18164 Masking (CBCM). In the absence of test vectors, the best I have been able
18165 to do is check that the decrypt undoes the encrypt, so far. Send me test
18166 vectors if you have them.
18170 * Correct calculation of key length for export ciphers (too much space was
18171 allocated for null ciphers). This has not been tested!
18175 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18176 message is now correct (it understands "crypto" and "ssl" on its
18177 command line). There is also now an "update" option. This will update
18178 the util/ssleay.num and util/libeay.num files with any new functions.
18180 perl util/mkdef.pl crypto ssl update
18181 it will update them.
18185 * Overhauled the Perl interface:
18186 - ported BN stuff to OpenSSL's different BN library
18187 - made the perl/ source tree CVS-aware
18188 - renamed the package from SSLeay to OpenSSL (the files still contain
18189 their history because I've copied them in the repository)
18190 - removed obsolete files (the test scripts will be replaced
18191 by better Test::Harness variants in the future)
18193 *Ralf S. Engelschall*
18195 * First cut for a very conservative source tree cleanup:
18196 1. merge various obsolete readme texts into doc/ssleay.txt
18197 where we collect the old documents and readme texts.
18198 2. remove the first part of files where I'm already sure that we no
18199 longer need them because of three reasons: either they are just temporary
18200 files which were left by Eric or they are preserved original files where
18201 I've verified that the diff is also available in the CVS via "cvs diff
18202 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18203 the crypto/md/ stuff).
18205 *Ralf S. Engelschall*
18207 * More extension code. Incomplete support for subject and issuer alt
18208 name, issuer and authority key id. Change the i2v function parameters
18209 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18210 what that's for :-) Fix to ASN1 macro which messed up
18211 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18215 * Preliminary support for ENUMERATED type. This is largely copied from the
18220 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18222 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18224 * Make sure `make rehash` target really finds the `openssl` program.
18226 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18228 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18229 like to hear about it if this slows down other processors.
18233 * Add CygWin32 platform information to Configure script.
18235 *Alan Batie <batie@aahz.jf.intel.com>*
18237 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18239 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18241 * New program nseq to manipulate netscape certificate sequences
18245 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18250 * Fixes to BN code. Previously the default was to define BN_RECURSION
18251 but the BN code had some problems that would cause failures when
18252 doing certificate verification and some other functions.
18254 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18256 * Add ASN1 and PEM code to support netscape certificate sequences.
18260 * Add ASN1 and PEM code to support netscape certificate sequences.
18264 * Add several PKIX and private extended key usage OIDs.
18268 * Modify the 'ca' program to handle the new extension code. Modify
18269 openssl.cnf for new extension format, add comments.
18273 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18274 and add a sample to openssl.cnf so req -x509 now adds appropriate
18279 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18280 error code, add initial support to X509_print() and x509 application.
18284 * Takes a deep breath and start adding X509 V3 extension support code. Add
18285 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18286 stuff is currently isolated and isn't even compiled yet.
18290 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18291 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18292 Removed the versions check from X509 routines when loading extensions:
18293 this allows certain broken certificates that don't set the version
18294 properly to be processed.
18298 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18299 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18300 can still be regenerated with "make depend".
18304 * Spelling mistake in C version of CAST-128.
18306 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18308 * Changes to the error generation code. The perl script err-code.pl
18309 now reads in the old error codes and retains the old numbers, only
18310 adding new ones if necessary. It also only changes the .err files if new
18311 codes are added. The makefiles have been modified to only insert errors
18312 when needed (to avoid needlessly modifying header files). This is done
18313 by only inserting errors if the .err file is newer than the auto generated
18314 C file. To rebuild all the error codes from scratch (the old behaviour)
18315 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18316 or delete all the .err files.
18320 * CAST-128 was incorrectly implemented for short keys. The C version has
18321 been fixed, but is untested. The assembler versions are also fixed, but
18322 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18323 to regenerate it if needed.
18324 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18325 Hagino <itojun@kame.net>*
18327 * File was opened incorrectly in randfile.c.
18329 *Ulf Möller <ulf@fitug.de>*
18331 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18332 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18333 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18334 al: it's just almost always a UTCTime. Note this patch adds new error
18335 codes so do a "make errors" if there are problems.
18339 * Correct Linux 1 recognition in config.
18341 *Ulf Möller <ulf@fitug.de>*
18343 * Remove pointless MD5 hash when using DSA keys in ca.
18345 *Anonymous <nobody@replay.com>*
18347 * Generate an error if given an empty string as a cert directory. Also
18348 generate an error if handed NULL (previously returned 0 to indicate an
18349 error, but didn't set one).
18351 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18353 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18357 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18358 parameters. This was causing a warning which killed off the Win32 compile.
18362 * Remove C++ style comments from crypto/bn/bn_local.h.
18364 *Neil Costigan <neil.costigan@celocom.com>*
18366 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18367 based on a text string, looking up short and long names and finally
18368 "dot" format. The "dot" format stuff didn't work. Added new function
18369 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18370 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18371 OID is not part of the table.
18375 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18376 X509_LOOKUP_by_alias().
18380 * Sort openssl functions by name.
18384 * Get the `gendsa` command working and add it to the `list` command. Remove
18385 encryption from sample DSA keys (in case anyone is interested the password
18390 * Make *all* `*_free` functions accept a NULL pointer.
18392 *Frans Heymans <fheymans@isaserver.be>*
18394 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18397 *Anonymous <nobody@replay.com>*
18399 * s_server should send the CAfile as acceptable CAs, not its own cert.
18401 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18403 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18405 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18407 * Temp key "for export" tests were wrong in s3_srvr.c.
18409 *Anonymous <nobody@replay.com>*
18411 * Add prototype for temp key callback functions
18412 SSL_CTX_set_tmp_{rsa,dh}_callback().
18416 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18417 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18421 * X509_name_add_entry() freed the wrong thing after an error.
18423 *Arne Ansper <arne@ats.cyber.ee>*
18425 * rsa_eay.c would attempt to free a NULL context.
18427 *Arne Ansper <arne@ats.cyber.ee>*
18429 * BIO_s_socket() had a broken should_retry() on Windoze.
18431 *Arne Ansper <arne@ats.cyber.ee>*
18433 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18435 *Arne Ansper <arne@ats.cyber.ee>*
18437 * Make sure the already existing X509_STORE->depth variable is initialized
18438 in X509_STORE_new(), but document the fact that this variable is still
18439 unused in the certificate verification process.
18441 *Ralf S. Engelschall*
18443 * Fix the various library and `apps/` files to free up pkeys obtained from
18444 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18448 * Fix reference counting in X509_PUBKEY_get(). This makes
18449 demos/maurice/example2.c work, amongst others, probably.
18451 *Steve Henson and Ben Laurie*
18453 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18454 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18455 are no longer created. This way we have a single and consistent command
18456 line interface `openssl <command>`, similar to `cvs <command>`.
18458 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18460 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18461 BIT STRING wrapper always have zero unused bits.
18465 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18469 * Make the top-level INSTALL documentation easier to understand.
18473 * Makefiles updated to exit if an error occurs in a sub-directory
18474 make (including if user presses ^C) [Paul Sutton]
18476 * Make Montgomery context stuff explicit in RSA data structure.
18480 * Fix build order of pem and err to allow for generated pem.h.
18484 * Fix renumbering bug in X509_NAME_delete_entry().
18488 * Enhanced the err-ins.pl script so it makes the error library number
18489 global and can add a library name. This is needed for external ASN1 and
18490 other error libraries.
18494 * Fixed sk_insert which never worked properly.
18498 * Fix ASN1 macros so they can handle indefinite length constructed
18499 EXPLICIT tags. Some non standard certificates use these: they can now
18504 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18505 into a single doc/ssleay.txt bundle. This way the information is still
18506 preserved but no longer messes up this directory. Now it's new room for
18507 the new set of documentation files.
18509 *Ralf S. Engelschall*
18511 * SETs were incorrectly DER encoded. This was a major pain, because they
18512 shared code with SEQUENCEs, which aren't coded the same. This means that
18513 almost everything to do with SETs or SEQUENCEs has either changed name or
18514 number of arguments.
18516 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18518 * Fix test data to work with the above.
18522 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18523 was already fixed by Eric for 0.9.1 it seems.
18525 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18527 * Autodetect FreeBSD3.
18531 * Fix various bugs in Configure. This affects the following platforms:
18535 unixware-2.0-pentium
18540 * Eliminate generated files from CVS. Reorder tests to regenerate files
18541 before they are needed.
18545 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18549 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18551 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18552 changed SSLeay to OpenSSL in version strings.
18554 *Ralf S. Engelschall*
18556 * Some fixups to the top-level documents.
18560 * Fixed the nasty bug where rsaref.h was not found under compile-time
18561 because the symlink to include/ was missing.
18563 *Ralf S. Engelschall*
18565 * Incorporated the popular no-RSA/DSA-only patches
18566 which allow to compile a RSA-free SSLeay.
18568 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18570 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18571 when "ssleay" is still not found.
18573 *Ralf S. Engelschall*
18575 * Added more platforms to Configure: Cray T3E, HPUX 11,
18577 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18579 * Updated the README file.
18581 *Ralf S. Engelschall*
18583 * Added various .cvsignore files in the CVS repository subdirs
18584 to make a "cvs update" really silent.
18586 *Ralf S. Engelschall*
18588 * Recompiled the error-definition header files and added
18589 missing symbols to the Win32 linker tables.
18591 *Ralf S. Engelschall*
18593 * Cleaned up the top-level documents;
18594 o new files: CHANGES and LICENSE
18595 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18596 o merged COPYRIGHT into LICENSE
18597 o removed obsolete TODO file
18598 o renamed MICROSOFT to INSTALL.W32
18600 *Ralf S. Engelschall*
18602 * Removed dummy files from the 0.9.1b source tree:
18604 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18605 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18606 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18607 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18608 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18611 *Ralf S. Engelschall*
18613 * Added various platform portability fixes.
18617 * The Genesis of the OpenSSL rpject:
18618 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18619 Young and Tim J. Hudson created while they were working for C2Net until
18622 *The OpenSSL Project*
18624 ### Changes between 0.9.0b and 0.9.1b [not released]
18626 * Updated a few CA certificates under certs/
18630 * Changed some BIGNUM api stuff.
18634 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18635 DGUX x86, Linux Alpha, etc.
18639 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18640 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18645 * Add -strparse option to asn1pars program which parses nested
18648 *Dr Stephen Henson <shenson@bigfoot.com>*
18650 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18654 * DSA fix for "ca" program.
18658 * Added "-genkey" option to "dsaparam" program.
18662 * Added RIPE MD160 (rmd160) message digest.
18666 * Added -a (all) option to "ssleay version" command.
18670 * Added PLATFORM define which is the id given to Configure.
18674 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18678 * Extended the ASN.1 parser routines.
18682 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18686 * Added a BN_CTX to the BN library.
18690 * Fixed the weak key values in DES library
18694 * Changed API in EVP library for cipher aliases.
18698 * Added support for RC2/64bit cipher.
18702 * Converted the lhash library to the crypto/mem.c functions.
18706 * Added more recognized ASN.1 object ids.
18710 * Added more RSA padding checks for SSL/TLS.
18714 * Added BIO proxy/filter functionality.
18718 * Added extra_certs to SSL_CTX which can be used
18719 send extra CA certificates to the client in the CA cert chain sending
18720 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18724 * Now Fortezza is denied in the authentication phase because
18725 this is key exchange mechanism is not supported by SSLeay at all.
18729 * Additional PKCS1 checks.
18733 * Support the string "TLSv1" for all TLS v1 ciphers.
18737 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18738 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18742 * Fixed a few memory leaks.
18746 * Fixed various code and comment typos.
18750 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18751 bytes sent in the client random.
18753 *Edward Bishop <ebishop@spyglass.com>*
18757 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18758 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18759 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18760 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18761 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18762 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18763 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18764 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18765 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18766 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18767 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18768 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18769 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18770 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18771 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18772 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18773 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18774 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18775 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18776 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18777 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18778 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18779 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18780 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18781 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18782 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18783 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18784 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18785 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18786 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18787 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18788 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18789 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18790 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18791 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18792 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18793 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18794 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18795 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18796 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18797 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18798 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18799 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18800 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18801 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18802 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18803 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18804 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18805 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18806 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18807 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18808 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18809 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18810 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18811 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18812 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18813 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18814 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18815 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18816 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18817 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18818 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18819 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18820 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18821 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18822 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18823 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18824 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18825 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18826 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18827 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18828 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18829 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18830 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18831 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18832 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18833 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18834 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18835 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18836 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18837 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18838 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18839 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18840 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18841 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18842 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18843 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18844 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18845 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18846 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18847 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18848 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18849 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18850 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18851 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18852 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18853 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18854 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18855 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18856 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18857 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18858 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18859 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18860 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18861 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18862 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18863 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18864 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18865 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18866 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18867 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18868 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18869 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18870 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18871 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18872 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18873 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18874 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18875 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18876 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18877 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18878 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18879 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18880 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18881 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18882 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18883 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18884 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18885 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18886 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18887 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18888 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18889 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18890 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18891 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18892 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18893 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18894 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18895 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18896 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18897 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18898 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18899 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18900 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18901 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18902 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18903 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18904 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18905 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18906 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18907 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18908 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18909 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18910 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18911 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18912 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18913 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18914 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18915 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18916 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18917 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18918 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655