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 * Validation of SM2 keys has been separated from the validation of regular EC
27 keys, allowing to improve the SM2 validation process to reject loaded private
28 keys that are not conforming to the SM2 ISO standard.
29 In particular, a private scalar `k` outside the range `1 <= k < n-1` is now
34 * Behavior of the `pkey` app is changed, when using the `-check` or `-pubcheck`
35 switches: a validation failure triggers an early exit, returning a failure
36 exit status to the parent process.
40 * Changed behavior of SSL_CTX_set_ciphersuites() and SSL_set_ciphersuites()
41 to ignore unknown ciphers.
45 * The `-cipher-commands` and `-digest-commands` options
46 of the command line utility `list` have been deprecated.
47 Instead use the `-cipher-algorithms` and `-digest-algorithms` options.
51 * Deprecated all the libcrypto and libssl error string loading
52 functions: ERR_load_ASN1_strings(), ERR_load_ASYNC_strings(),
53 ERR_load_BIO_strings(), ERR_load_BN_strings(), ERR_load_BUF_strings(),
54 ERR_load_CMS_strings(), ERR_load_COMP_strings(), ERR_load_CONF_strings(),
55 ERR_load_CRYPTO_strings(), ERR_load_CT_strings(), ERR_load_DH_strings(),
56 ERR_load_DSA_strings(), ERR_load_EC_strings(), ERR_load_ENGINE_strings(),
57 ERR_load_ERR_strings(), ERR_load_EVP_strings(), ERR_load_KDF_strings(),
58 ERR_load_OBJ_strings(), ERR_load_OCSP_strings(), ERR_load_PEM_strings(),
59 ERR_load_PKCS12_strings(), ERR_load_PKCS7_strings(), ERR_load_RAND_strings(),
60 ERR_load_RSA_strings(), ERR_load_OSSL_STORE_strings(), ERR_load_TS_strings(),
61 ERR_load_UI_strings(), ERR_load_X509_strings(), ERR_load_X509V3_strings().
63 Calling these functions is not necessary since OpenSSL 1.1.0, as OpenSSL
64 now loads error strings automatically.
68 * The functions SSL_CTX_set_tmp_dh_callback and SSL_set_tmp_dh_callback, as
69 well as the macros SSL_CTX_set_tmp_dh() and SSL_set_tmp_dh() have been
70 deprecated. These are used to set the Diffie-Hellman (DH) parameters that
71 are to be used by servers requiring ephemeral DH keys. Instead applications
72 should consider using the built-in DH parameters that are available by
73 calling SSL_CTX_set_dh_auto() or SSL_set_dh_auto(). If custom parameters are
74 necessary then applications can use the alternative functions
75 SSL_CTX_set0_tmp_dh_pkey() and SSL_set0_tmp_dh_pkey(). There is no direct
76 replacement for the "callback" functions. The callback was originally useful
77 in order to have different parameters for export and non-export ciphersuites.
78 Export ciphersuites are no longer supported by OpenSSL. Use of the callback
79 functions should be replaced by one of the other methods described above.
83 * The `-crypt` option to the `passwd` command line tool has been removed.
87 * The -C option to the `x509`, `dhparam`, `dsaparam`, and `ecparam` commands
92 * Add support for AES Key Wrap inverse ciphers to the EVP layer.
94 "AES-128-WRAP-INV", "AES-192-WRAP-INV", "AES-256-WRAP-INV",
95 "AES-128-WRAP-PAD-INV", "AES-192-WRAP-PAD-INV" and "AES-256-WRAP-PAD-INV".
96 The inverse ciphers use AES decryption for wrapping, and
97 AES encryption for unwrapping.
101 * Deprecated EVP_PKEY_set1_tls_encodedpoint() and
102 EVP_PKEY_get1_tls_encodedpoint(). These functions were previously used by
103 libssl to set or get an encoded public key in/from an EVP_PKEY object. With
104 OpenSSL 3.0 these are replaced by the more generic functions
105 EVP_PKEY_set1_encoded_public_key() and EVP_PKEY_get1_encoded_public_key().
106 The old versions have been converted to deprecated macros that just call the
111 * The security callback, which can be customised by application code, supports
112 the security operation SSL_SECOP_TMP_DH. This is defined to take an EVP_PKEY
113 in the "other" parameter. In most places this is what is passed. All these
114 places occur server side. However there was one client side call of this
115 security operation and it passed a DH object instead. This is incorrect
116 according to the definition of SSL_SECOP_TMP_DH, and is inconsistent with all
117 of the other locations. Therefore this client side call has been changed to
118 pass an EVP_PKEY instead.
122 * Add PKCS7_get_octet_string() and PKCS7_type_is_other() to the public
123 interface. Their functionality remains unchanged.
127 * Added new option for 'openssl list', '-providers', which will display the
128 list of loaded providers, their names, version and status. It optionally
129 displays their gettable parameters.
133 * Deprecated EVP_PKEY_set_alias_type(). This function was previously
134 needed as a workaround to recognise SM2 keys. With OpenSSL 3.0, this key
135 type is internally recognised so the workaround is no longer needed.
137 Functionality is still retained as it is, but will only work with
138 EVP_PKEYs with a legacy internal key.
142 * Deprecated `EVP_PKEY_CTX_set_rsa_keygen_pubexp()` and introduced
143 `EVP_PKEY_CTX_set1_rsa_keygen_pubexp()`, which is now preferred.
147 * Changed all "STACK" functions to be macros instead of inline functions. Macro
148 parameters are still checked for type safety at compile time via helper
153 * Remove the RAND_DRBG API
155 The RAND_DRBG API did not fit well into the new provider concept as
156 implemented by EVP_RAND and EVP_RAND_CTX. The main reason is that the
157 RAND_DRBG API is a mixture of 'front end' and 'back end' API calls
158 and some of its API calls are rather low-level. This holds in particular
159 for the callback mechanism (`RAND_DRBG_set_callbacks()`).
161 Adding a compatibility layer to continue supporting the RAND_DRBG API as
162 a legacy API for a regular deprecation period turned out to come at the
163 price of complicating the new provider API unnecessarily. Since the
164 RAND_DRBG API exists only since version 1.1.1, it was decided by the OMC
167 *Paul Dale and Matthias St. Pierre*
169 * Allow `SSL_set1_host()` and `SSL_add1_host()` to take IP literal addresses
170 as well as actual hostnames.
174 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
175 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
176 conversely, silently ignore DTLS protocol version bounds when configuring
177 TLS-based contexts. The commands can be repeated to set bounds of both
178 types. The same applies with the corresponding "min_protocol" and
179 "max_protocol" command-line switches, in case some application uses both TLS
182 SSL_CTX instances that are created for a fixed protocol version (e.g.
183 `TLSv1_server_method()`) also silently ignore version bounds. Previously
184 attempts to apply bounds to these protocol versions would result in an
185 error. Now only the "version-flexible" SSL_CTX instances are subject to
186 limits in configuration files in command-line options.
190 * Deprecated the `ENGINE` API. Engines should be replaced with providers
195 * Reworked the recorded ERR codes to make better space for system errors.
196 To distinguish them, the macro `ERR_SYSTEM_ERROR()` indicates if the
197 given code is a system error (true) or an OpenSSL error (false).
201 * Reworked the test perl framework to better allow parallel testing.
203 *Nicola Tuveri and David von Oheimb*
205 * Added ciphertext stealing algorithms AES-128-CBC-CTS, AES-192-CBC-CTS and
206 AES-256-CBC-CTS to the providers. CS1, CS2 and CS3 variants are supported.
210 * 'Configure' has been changed to figure out the configuration target if
211 none is given on the command line. Consequently, the 'config' script is
212 now only a mere wrapper. All documentation is changed to only mention
215 *Rich Salz and Richard Levitte*
217 * Added a library context `OSSL_LIB_CTX` that applications as well as
218 other libraries can use to form a separate context within which
219 libcrypto operations are performed.
221 There are two ways this can be used:
223 - Directly, by passing a library context to functions that take
224 such an argument, such as `EVP_CIPHER_fetch` and similar algorithm
226 - Indirectly, by creating a new library context and then assigning
227 it as the new default, with `OSSL_LIB_CTX_set0_default`.
229 All public OpenSSL functions that take an `OSSL_LIB_CTX` pointer,
230 apart from the functions directly related to `OSSL_LIB_CTX`, accept
231 NULL to indicate that the default library context should be used.
233 Library code that changes the default library context using
234 `OSSL_LIB_CTX_set0_default` should take care to restore it with a
235 second call before returning to the caller.
237 _(Note: the library context was initially called `OPENSSL_CTX` and
238 renamed to `OSSL_LIB_CTX` in version 3.0.0 alpha7.)_
242 * Handshake now fails if Extended Master Secret extension is dropped
247 * Dropped interactive mode from the `openssl` program. From now on,
248 running it without arguments is equivalent to `openssl help`.
252 * Renamed `EVP_PKEY_cmp()` to `EVP_PKEY_eq()` and
253 `EVP_PKEY_cmp_parameters()` to `EVP_PKEY_parameters_eq()`.
254 While the old function names have been retained for backward compatibility
255 they should not be used in new developments
256 because their return values are confusing: Unlike other `_cmp()` functions
257 they do not return 0 in case their arguments are equal.
261 * Deprecated `EC_METHOD_get_field_type()`. Applications should switch to
262 `EC_GROUP_get_field_type()`.
266 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
267 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
268 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
269 Applications should rely on the library automatically assigning a suitable
270 EC_METHOD internally upon EC_GROUP construction.
274 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
275 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
276 assigned internally without application intervention.
277 Users of EC_GROUP_new() should switch to a different suitable constructor.
281 * Add CAdES-BES signature verification support, mostly derived
282 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
284 *Filipe Raimundo da Silva*
286 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
290 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
291 functions are not widely used and now OpenSSL automatically perform this
292 conversion when needed.
296 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
297 EC_KEY_precompute_mult(). These functions are not widely used and
298 applications should instead switch to named curves which OpenSSL has
299 hardcoded lookup tables for.
303 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
304 should instead use the L<EC_POINT_mul(3)> function.
308 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
309 that are not applicable to the new provider model. Applications should
310 instead use EVP_default_properties_is_fips_enabled() and
311 EVP_default_properties_enable_fips().
315 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
316 is set, an unexpected EOF is ignored, it pretends a close notify was received
317 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
321 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
322 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
323 used and applications should instead use the
324 L<EC_POINT_set_affine_coordinates(3)> and
325 L<EC_POINT_get_affine_coordinates(3)> functions.
329 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
330 arrays to be more easily constructed via a series of utility functions.
331 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
332 the various push functions and finally convert to a passable OSSL_PARAM
333 array using OSSL_PARAM_BLD_to_param().
337 * The security strength of SHA1 and MD5 based signatures in TLS has been
338 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
339 working at the default security level of 1 and instead requires security
340 level 0. The security level can be changed either using the cipher string
341 with `@SECLEVEL`, or calling `SSL_CTX_set_security_level()`.
345 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
346 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
347 internal keys, if they correspond to one of those built in types.
351 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
352 contain a provider side internal key.
356 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
357 They are old functions that we don't use, and that you could disable with
358 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
362 * Project text documents not yet having a proper file name extension
363 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
364 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
368 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
369 have been converted to Markdown with the goal to produce documents
370 which not only look pretty when viewed online in the browser, but
371 remain well readable inside a plain text editor.
373 To achieve this goal, a 'minimalistic' Markdown style has been applied
374 which avoids formatting elements that interfere too much with the
375 reading flow in the text file. For example, it
377 * avoids [ATX headings][] and uses [setext headings][] instead
378 (which works for `<h1>` and `<h2>` headings only).
379 * avoids [inline links][] and uses [reference links][] instead.
380 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
382 [ATX headings]: https://github.github.com/gfm/#atx-headings
383 [setext headings]: https://github.github.com/gfm/#setext-headings
384 [inline links]: https://github.github.com/gfm/#inline-link
385 [reference links]: https://github.github.com/gfm/#reference-link
386 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
387 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
389 *Matthias St. Pierre*
391 * The test suite is changed to preserve results of each test recipe.
392 A new directory test-runs/ with subdirectories named like the
393 test recipes are created in the build tree for this purpose.
397 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
398 This adds `crypto/cmp/`, `crpyto/crmf/`, `apps/cmp.c`, and `test/cmp_*`.
399 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
401 *David von Oheimb, Martin Peylo*
403 * Generalized the HTTP client code from `crypto/ocsp/` into `crpyto/http/`.
404 The legacy OCSP-focused and only partly documented API is retained for
405 backward compatibility. See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
409 * Added `util/check-format.pl`, a tool for checking adherence to the
410 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
411 The checks performed are incomplete and yield some false positives.
412 Still the tool should be useful for detecting most typical glitches.
416 * `BIO_do_connect()` and `BIO_do_handshake()` have been extended:
417 If domain name resolution yields multiple IP addresses all of them are tried
418 after `connect()` failures.
422 * All of the low level RSA functions have been deprecated including:
424 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
425 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
426 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
427 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
428 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
429 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
430 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
431 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
432 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
433 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
434 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
435 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
436 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
437 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
438 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
439 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
440 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
441 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
442 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
443 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
444 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
445 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
446 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
447 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
448 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
449 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
450 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
451 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
453 Use of these low level functions has been informally discouraged for a long
454 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
455 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
456 L<EVP_PKEY_decrypt(3)>.
460 * X509 certificates signed using SHA1 are no longer allowed at security
462 In TLS/SSL the default security level is 1. It can be set either
463 using the cipher string with `@SECLEVEL`, or calling
464 `SSL_CTX_set_security_level()`. If the leaf certificate is signed with SHA-1,
465 a call to `SSL_CTX_use_certificate()` will fail if the security level is not
467 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
468 be set using `X509_VERIFY_PARAM_set_auth_level()` or using the `-auth_level`
469 options of the commands.
473 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
474 modified to use PKEY APIs. These commands are now in maintenance mode
475 and no new features will be added to them.
479 * The command line utility rsautl has been deprecated.
480 Instead use the pkeyutl program.
484 * The command line utilities genrsa and rsa have been modified to use PKEY
485 APIs. They now write PKCS#8 keys by default. These commands are now in
486 maintenance mode and no new features will be added to them.
490 * All of the low level DH functions have been deprecated including:
492 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
493 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
494 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
495 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
496 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
497 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
498 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
499 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
500 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
501 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
502 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
503 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
504 DH_meth_get_generate_params and DH_meth_set_generate_params.
506 Use of these low level functions has been informally discouraged for a long
507 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
508 and L<EVP_PKEY_derive(3)>.
510 Additionally functions that read and write DH objects such as d2i_DHparams,
511 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
512 functions have also been deprecated. Applications should instead use the
513 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
515 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
516 `EVP_PKEY_assign_DH()`, `EVP_PKEY_get0_DH()`, `EVP_PKEY_get1_DH()`, and
517 `EVP_PKEY_set1_DH()` are also deprecated.
518 Applications should instead either read or write an
519 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs.
520 Or load an EVP_PKEY directly from DH data using `EVP_PKEY_fromdata()`.
522 *Paul Dale and Matt Caswell*
524 * All of the low level DSA functions have been deprecated including:
526 DSA_new, DSA_free, DSA_up_ref, DSA_bits, DSA_get0_pqg, DSA_set0_pqg,
527 DSA_get0_key, DSA_set0_key, DSA_get0_p, DSA_get0_q, DSA_get0_g,
528 DSA_get0_pub_key, DSA_get0_priv_key, DSA_clear_flags, DSA_test_flags,
529 DSA_set_flags, DSA_do_sign, DSA_do_verify, DSA_OpenSSL,
530 DSA_set_default_method, DSA_get_default_method, DSA_set_method,
531 DSA_get_method, DSA_new_method, DSA_size, DSA_security_bits,
532 DSA_sign_setup, DSA_sign, DSA_verify, DSA_get_ex_new_index,
533 DSA_set_ex_data, DSA_get_ex_data, DSA_generate_parameters_ex,
534 DSA_generate_key, DSA_meth_new, DSA_get0_engine, DSA_meth_free,
535 DSA_meth_dup, DSA_meth_get0_name, DSA_meth_set1_name, DSA_meth_get_flags,
536 DSA_meth_set_flags, DSA_meth_get0_app_data, DSA_meth_set0_app_data,
537 DSA_meth_get_sign, DSA_meth_set_sign, DSA_meth_get_sign_setup,
538 DSA_meth_set_sign_setup, DSA_meth_get_verify, DSA_meth_set_verify,
539 DSA_meth_get_mod_exp, DSA_meth_set_mod_exp, DSA_meth_get_bn_mod_exp,
540 DSA_meth_set_bn_mod_exp, DSA_meth_get_init, DSA_meth_set_init,
541 DSA_meth_get_finish, DSA_meth_set_finish, DSA_meth_get_paramgen,
542 DSA_meth_set_paramgen, DSA_meth_get_keygen and DSA_meth_set_keygen.
544 Use of these low level functions has been informally discouraged for a long
545 time. Instead applications should use L<EVP_DigestSignInit_ex(3)>,
546 L<EVP_DigestSignUpdate(3)> and L<EVP_DigestSignFinal(3)>.
550 * Reworked the treatment of EC EVP_PKEYs with the SM2 curve to
551 automatically become EVP_PKEY_SM2 rather than EVP_PKEY_EC.
552 This means that applications don't have to look at the curve NID and
553 `EVP_PKEY_set_alias_type(pkey, EVP_PKEY_SM2)` to get SM2 computations.
554 However, they still can, that `EVP_PKEY_set_alias_type()` call acts as
555 a no-op when the EVP_PKEY is already of the given type.
557 Parameter and key generation is also reworked to make it possible
558 to generate EVP_PKEY_SM2 parameters and keys without having to go
559 through EVP_PKEY_EC generation and then change the EVP_PKEY type.
560 However, code that does the latter will still work as before.
564 * Deprecated low level ECDH and ECDSA functions. These include:
566 ECDH_compute_key, ECDSA_do_sign, ECDSA_do_sign_ex, ECDSA_do_verify,
567 ECDSA_sign_setup, ECDSA_sign, ECDSA_sign_ex, ECDSA_verify and
570 Use of these low level functions has been informally discouraged for a long
571 time. Instead applications should use the EVP_PKEY_derive(3),
572 EVP_DigestSign(3) and EVP_DigestVerify(3) functions.
576 * Deprecated the EC_KEY_METHOD functions. These include:
578 EC_KEY_METHOD_new, EC_KEY_METHOD_free, EC_KEY_METHOD_set_init,
579 EC_KEY_METHOD_set_keygen, EC_KEY_METHOD_set_compute_key,
580 EC_KEY_METHOD_set_sign, EC_KEY_METHOD_set_verify,
581 EC_KEY_METHOD_get_init, EC_KEY_METHOD_get_keygen,
582 EC_KEY_METHOD_get_compute_key, EC_KEY_METHOD_get_sign and
583 EC_KEY_METHOD_get_verify.
585 Instead applications and extension writers should use the OSSL_PROVIDER APIs.
589 * Deprecated EVP_PKEY_decrypt_old(), please use EVP_PKEY_decrypt_init()
590 and EVP_PKEY_decrypt() instead.
591 Deprecated EVP_PKEY_encrypt_old(), please use EVP_PKEY_encrypt_init()
592 and EVP_PKEY_encrypt() instead.
596 * Enhanced the documentation of EVP_PKEY_size(), EVP_PKEY_bits()
597 and EVP_PKEY_security_bits(). Especially EVP_PKEY_size() needed
598 a new formulation to include all the things it can be used for,
599 as well as words of caution.
603 * The SSL_CTX_set_tlsext_ticket_key_cb(3) function has been deprecated.
604 Instead used the new SSL_CTX_set_tlsext_ticket_key_evp_cb(3) function.
608 * All of the low level HMAC functions have been deprecated including:
610 HMAC, HMAC_size, HMAC_CTX_new, HMAC_CTX_reset, HMAC_CTX_free,
611 HMAC_Init_ex, HMAC_Update, HMAC_Final, HMAC_CTX_copy, HMAC_CTX_set_flags
614 Use of these low level functions has been informally discouraged for a long
615 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
616 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
617 and L<EVP_MAC_final(3)>.
621 * Over two thousand fixes were made to the documentation, including:
622 - Common options (such as -rand/-writerand, TLS version control, etc)
623 were refactored and point to newly-enhanced descriptions in openssl.pod.
624 - Added style conformance for all options (with help from Richard Levitte),
625 documented all reported missing options, added a CI build to check
626 that all options are documented and that no unimplemented options
628 - Documented some internals, such as all use of environment variables.
629 - Addressed all internal broken L<> references.
633 * All of the low level CMAC functions have been deprecated including:
635 CMAC_CTX_new, CMAC_CTX_cleanup, CMAC_CTX_free, CMAC_CTX_get0_cipher_ctx,
636 CMAC_CTX_copy, CMAC_Init, CMAC_Update, CMAC_Final and CMAC_resume.
638 Use of these low level functions has been informally discouraged for a long
639 time. Instead applications should use L<EVP_MAC_CTX_new(3)>,
640 L<EVP_MAC_CTX_free(3)>, L<EVP_MAC_init(3)>, L<EVP_MAC_update(3)>
641 and L<EVP_MAC_final(3)>.
645 * All of the low level MD2, MD4, MD5, MDC2, RIPEMD160, SHA1, SHA224, SHA256,
646 SHA384, SHA512 and Whirlpool digest functions have been deprecated.
649 MD2, MD2_options, MD2_Init, MD2_Update, MD2_Final, MD4, MD4_Init,
650 MD4_Update, MD4_Final, MD4_Transform, MD5, MD5_Init, MD5_Update,
651 MD5_Final, MD5_Transform, MDC2, MDC2_Init, MDC2_Update, MDC2_Final,
652 RIPEMD160, RIPEMD160_Init, RIPEMD160_Update, RIPEMD160_Final,
653 RIPEMD160_Transform, SHA1_Init, SHA1_Update, SHA1_Final, SHA1_Transform,
654 SHA224_Init, SHA224_Update, SHA224_Final, SHA224_Transform, SHA256_Init,
655 SHA256_Update, SHA256_Final, SHA256_Transform, SHA384, SHA384_Init,
656 SHA384_Update, SHA384_Final, SHA512, SHA512_Init, SHA512_Update,
657 SHA512_Final, SHA512_Transform, WHIRLPOOL, WHIRLPOOL_Init,
658 WHIRLPOOL_Update, WHIRLPOOL_BitUpdate and WHIRLPOOL_Final.
660 Use of these low level functions has been informally discouraged
661 for a long time. Applications should use the EVP_DigestInit_ex(3),
662 EVP_DigestUpdate(3) and EVP_DigestFinal_ex(3) functions instead.
666 * Corrected the documentation of the return values from the `EVP_DigestSign*`
667 set of functions. The documentation mentioned negative values for some
668 errors, but this was never the case, so the mention of negative values
671 Code that followed the documentation and thereby check with something
672 like `EVP_DigestSignInit(...) <= 0` will continue to work undisturbed.
676 * All of the low level cipher functions have been deprecated including:
678 AES_options, AES_set_encrypt_key, AES_set_decrypt_key, AES_encrypt,
679 AES_decrypt, AES_ecb_encrypt, AES_cbc_encrypt, AES_cfb128_encrypt,
680 AES_cfb1_encrypt, AES_cfb8_encrypt, AES_ofb128_encrypt,
681 AES_wrap_key, AES_unwrap_key, BF_set_key, BF_encrypt, BF_decrypt,
682 BF_ecb_encrypt, BF_cbc_encrypt, BF_cfb64_encrypt, BF_ofb64_encrypt,
683 BF_options, Camellia_set_key, Camellia_encrypt, Camellia_decrypt,
684 Camellia_ecb_encrypt, Camellia_cbc_encrypt, Camellia_cfb128_encrypt,
685 Camellia_cfb1_encrypt, Camellia_cfb8_encrypt, Camellia_ofb128_encrypt,
686 Camellia_ctr128_encrypt, CAST_set_key, CAST_encrypt, CAST_decrypt,
687 CAST_ecb_encrypt, CAST_cbc_encrypt, CAST_cfb64_encrypt,
688 CAST_ofb64_encrypt, DES_options, DES_encrypt1, DES_encrypt2,
689 DES_encrypt3, DES_decrypt3, DES_cbc_encrypt, DES_ncbc_encrypt,
690 DES_pcbc_encrypt, DES_xcbc_encrypt, DES_cfb_encrypt, DES_cfb64_encrypt,
691 DES_ecb_encrypt, DES_ofb_encrypt, DES_ofb64_encrypt, DES_random_key,
692 DES_set_odd_parity, DES_check_key_parity, DES_is_weak_key, DES_set_key,
693 DES_key_sched, DES_set_key_checked, DES_set_key_unchecked,
694 DES_string_to_key, DES_string_to_2keys, DES_fixup_key_parity,
695 DES_ecb2_encrypt, DES_ede2_cbc_encrypt, DES_ede2_cfb64_encrypt,
696 DES_ede2_ofb64_encrypt, DES_ecb3_encrypt, DES_ede3_cbc_encrypt,
697 DES_ede3_cfb64_encrypt, DES_ede3_cfb_encrypt, DES_ede3_ofb64_encrypt,
698 DES_cbc_cksum, DES_quad_cksum, IDEA_encrypt, IDEA_options,
699 IDEA_ecb_encrypt, IDEA_set_encrypt_key, IDEA_set_decrypt_key,
700 IDEA_cbc_encrypt, IDEA_cfb64_encrypt, IDEA_ofb64_encrypt, RC2_set_key,
701 RC2_encrypt, RC2_decrypt, RC2_ecb_encrypt, RC2_cbc_encrypt,
702 RC2_cfb64_encrypt, RC2_ofb64_encrypt, RC4, RC4_options, RC4_set_key,
703 RC5_32_set_key, RC5_32_encrypt, RC5_32_decrypt, RC5_32_ecb_encrypt,
704 RC5_32_cbc_encrypt, RC5_32_cfb64_encrypt, RC5_32_ofb64_encrypt,
705 SEED_set_key, SEED_encrypt, SEED_decrypt, SEED_ecb_encrypt,
706 SEED_cbc_encrypt, SEED_cfb128_encrypt and SEED_ofb128_encrypt.
708 Use of these low level functions has been informally discouraged for
709 a long time. Applications should use the high level EVP APIs, e.g.
710 EVP_EncryptInit_ex, EVP_EncryptUpdate, EVP_EncryptFinal_ex, and the
711 equivalently named decrypt functions instead.
713 *Matt Caswell and Paul Dale*
715 * Removed include/openssl/opensslconf.h.in and replaced it with
716 include/openssl/configuration.h.in, which differs in not including
717 <openssl/macros.h>. A short header include/openssl/opensslconf.h
718 was added to include both.
720 This allows internal hacks where one might need to modify the set
721 of configured macros, for example this if deprecated symbols are
722 still supposed to be available internally:
724 #include <openssl/configuration.h>
726 #undef OPENSSL_NO_DEPRECATED
727 #define OPENSSL_SUPPRESS_DEPRECATED
729 #include <openssl/macros.h>
731 This should not be used by applications that use the exported
732 symbols, as that will lead to linking errors.
736 * Fixed an overflow bug in the x64_64 Montgomery squaring procedure
737 used in exponentiation with 512-bit moduli. No EC algorithms are
738 affected. Analysis suggests that attacks against 2-prime RSA1024,
739 3-prime RSA1536, and DSA1024 as a result of this defect would be very
740 difficult to perform and are not believed likely. Attacks against DH512
741 are considered just feasible. However, for an attack the target would
742 have to re-use the DH512 private key, which is not recommended anyway.
743 Also applications directly using the low level API BN_mod_exp may be
744 affected if they use BN_FLG_CONSTTIME.
749 * Most memory-debug features have been deprecated, and the functionality
750 replaced with no-ops.
754 * Added documentation for the STACK API.
758 * Introduced a new method type and API, OSSL_ENCODER, to
759 represent generic encoders. An implementation is expected to
760 be able to encode an object associated with a given name (such
761 as an algorithm name for an asymmetric key) into forms given by
762 implementation properties.
764 Encoders are primarily used from inside libcrypto, through
765 calls to functions like EVP_PKEY_print_private(),
766 PEM_write_bio_PrivateKey() and similar.
768 Encoders are specified in such a way that they can be made to
769 directly handle the provider side portion of an object, if this
770 provider side part comes from the same provider as the encoder
771 itself, but can also be made to handle objects in parametrized
772 form (as an OSSL_PARAM array of data). This allows a provider to
773 offer generic encoders as a service for any other provider.
777 * Added a .pragma directive to the syntax of configuration files, to
778 allow varying behavior in a supported and predictable manner.
779 Currently added pragma:
783 This allows dollar signs to be a keyword character unless it's
784 followed by a opening brace or parenthesis. This is useful for
785 platforms where dollar signs are commonly used in names, such as
786 volume names and system directory names on VMS.
790 * Added functionality to create an EVP_PKEY from user data. This
791 is effectively the same as creating a RSA, DH or DSA object and
792 then assigning them to an EVP_PKEY, but directly using algorithm
793 agnostic EVP functions. A benefit is that this should be future
794 proof for public key algorithms to come.
798 * Change the interpretation of the '--api' configuration option to
799 mean that this is a desired API compatibility level with no
800 further meaning. The previous interpretation, that this would
801 also mean to remove all deprecated symbols up to and including
802 the given version, no requires that 'no-deprecated' is also used
803 in the configuration.
805 When building applications, the desired API compatibility level
806 can be set with the OPENSSL_API_COMPAT macro like before. For
807 API compatibility version below 3.0, the old style numerical
808 value is valid as before, such as -DOPENSSL_API_COMPAT=0x10100000L.
809 For version 3.0 and on, the value is expected to be the decimal
810 value calculated from the major and minor version like this:
812 MAJOR * 10000 + MINOR * 100
816 -DOPENSSL_API_COMPAT=30000 For 3.0
817 -DOPENSSL_API_COMPAT=30200 For 3.2
819 To hide declarations that are deprecated up to and including the
820 given API compatibility level, -DOPENSSL_NO_DEPRECATED must be
821 given when building the application as well.
825 * Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
826 access to certificate and CRL stores via URIs and OSSL_STORE
829 This adds the following functions:
831 - X509_LOOKUP_store()
832 - X509_STORE_load_file()
833 - X509_STORE_load_path()
834 - X509_STORE_load_store()
835 - SSL_add_store_cert_subjects_to_stack()
836 - SSL_CTX_set_default_verify_store()
837 - SSL_CTX_load_verify_file()
838 - SSL_CTX_load_verify_dir()
839 - SSL_CTX_load_verify_store()
843 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
844 The presence of this system service is determined at run-time.
848 * Added functionality to create an EVP_PKEY context based on data
849 for methods from providers. This takes an algorithm name and a
850 property query string and simply stores them, with the intent
851 that any operation that uses this context will use those strings
852 to fetch the needed methods implicitly, thereby making the port
853 of application written for pre-3.0 OpenSSL easier.
857 * The undocumented function NCONF_WIN32() has been deprecated; for
858 conversion details see the HISTORY section of doc/man5/config.pod
862 * Introduced the new functions EVP_DigestSignInit_ex() and
863 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
864 EVP_DigestVerifyUpdate() have been converted to functions. See the man
865 pages for further details.
869 * Over two thousand fixes were made to the documentation, including:
870 adding missing command flags, better style conformance, documentation
873 *Rich Salz, Richard Levitte*
875 * s390x assembly pack: add hardware-support for P-256, P-384, P-521,
876 X25519, X448, Ed25519 and Ed448.
880 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
885 * Deprecated the public definition of `ERR_STATE` as well as the function
886 `ERR_get_state()`. This is done in preparation of making `ERR_STATE` an
891 * Added ERR functionality to give callers access to the stored function
892 names that have replaced the older function code based functions.
894 New functions are ERR_peek_error_func(), ERR_peek_last_error_func(),
895 ERR_peek_error_data(), ERR_peek_last_error_data(), ERR_get_error_all(),
896 ERR_peek_error_all() and ERR_peek_last_error_all().
898 These functions have become deprecated: ERR_get_error_line(),
899 ERR_get_error_line_data(), ERR_peek_error_line_data(),
900 ERR_peek_last_error_line_data() and ERR_func_error_string().
902 Users are recommended to use ERR_get_error_all(), or to pick information
903 with ERR_peek functions and finish off with getting the error code by using
908 * Extended testing to be verbose for failing tests only. The make variables
909 VERBOSE_FAILURE or VF can be used to enable this:
911 $ make VF=1 test # Unix
912 $ mms /macro=(VF=1) test ! OpenVMS
913 $ nmake VF=1 test # Windows
917 * The `x509`, `req`, and `ca` commands now make sure that certificates they
918 generate are RFC 5280 compliant by default: For X.509 version 3 certs they ensure that
919 a subjectKeyIdentifier extension is included containing a hash value of the public key
920 and an authorityKeyIdentifier extension is included for not self-signed certs
921 containing a keyIdentifier field with the hash value identifying the signing key.
922 This is done unless some configuration overrides the new default behavior,
923 e.g. `authorityKeyIdentifier = none`.
927 * Added several checks to `X509_verify_cert()` according to requirements in
928 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
929 (which may be done by using the CLI option `-x509_strict`):
930 * The basicConstraints of CA certificates must be marked critical.
931 * CA certificates must explicitly include the keyUsage extension.
932 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
933 * The issuer name of any certificate must not be empty.
934 * The subject name of CA certs, certs with keyUsage crlSign,
935 and certs without subjectAlternativeName must not be empty.
936 * If a subjectAlternativeName extension is given it must not be empty.
937 * The signatureAlgorithm field and the cert signature must be consistent.
938 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
939 must not be marked critical.
940 * The authorityKeyIdentifier must be given for X.509v3 certs
941 unless they are self-signed.
942 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
946 * Certificate verification using `X509_verify_cert()` meanwhile rejects EC keys
947 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
951 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
952 used even when parsing explicit parameters, when loading a encoded key
953 or calling `EC_GROUP_new_from_ecpkparameters()`/
954 `EC_GROUP_new_from_ecparameters()`.
955 This prevents bypass of security hardening and performance gains,
956 especially for curves with specialized EC_METHODs.
957 By default, if a key encoded with explicit parameters is loaded and later
958 encoded, the output is still encoded with explicit parameters, even if
959 internally a "named" EC_GROUP is used for computation.
963 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
964 this change, EC_GROUP_set_generator would accept order and/or cofactor as
965 NULL. After this change, only the cofactor parameter can be NULL. It also
966 does some minimal sanity checks on the passed order.
971 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
972 An attack is simple, if the first CMS_recipientInfo is valid but the
973 second CMS_recipientInfo is chosen ciphertext. If the second
974 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
975 encryption key will be replaced by garbage, and the message cannot be
976 decoded, but if the RSA decryption fails, the correct encryption key is
977 used and the recipient will not notice the attack.
978 As a work around for this potential attack the length of the decrypted
979 key must be equal to the cipher default key length, in case the
980 certifiate is not given and all recipientInfo are tried out.
981 The old behaviour can be re-enabled in the CMS code by setting the
982 CMS_DEBUG_DECRYPT flag.
986 * Early start up entropy quality from the DEVRANDOM seed source has been
987 improved for older Linux systems. The RAND subsystem will wait for
988 /dev/random to be producing output before seeding from /dev/urandom.
989 The seeded state is stored for future library initialisations using
990 a system global shared memory segment. The shared memory identifier
991 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
992 the desired value. The default identifier is 114.
996 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
997 when primes for RSA keys are computed.
998 Since we previously always generated primes == 2 (mod 3) for RSA keys,
999 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1000 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
1001 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1002 This avoids possible fingerprinting of newly generated RSA modules.
1006 * Correct the extended master secret constant on EBCDIC systems. Without this
1007 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1008 negotiate EMS will fail. Unfortunately this also means that TLS connections
1009 between EBCDIC systems with this fix, and EBCDIC systems without this
1010 fix will fail if they negotiate EMS.
1014 * Changed the library initialisation so that the config file is now loaded
1015 by default. This was already the case for libssl. It now occurs for both
1016 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1017 `OPENSSL_init_crypto()` to suppress automatic loading of a config file.
1021 * Introduced new error raising macros, `ERR_raise()` and `ERR_raise_data()`,
1022 where the former acts as a replacement for `ERR_put_error()`, and the
1023 latter replaces the combination `ERR_put_error()` + `ERR_add_error_data()`.
1024 `ERR_raise_data()` adds more flexibility by taking a format string and
1025 an arbitrary number of arguments following it, to be processed with
1030 * Introduced a new function, `OSSL_PROVIDER_available()`, which can be used
1031 to check if a named provider is loaded and available. When called, it
1032 will also activate all fallback providers if such are still present.
1036 * Enforce a minimum DH modulus size of 512 bits.
1040 * Changed DH parameters to generate the order q subgroup instead of 2q.
1041 Previously generated DH parameters are still accepted by DH_check
1042 but DH_generate_key works around that by clearing bit 0 of the
1043 private key for those. This avoids leaking bit 0 of the private key.
1047 * Significantly reduce secure memory usage by the randomness pools.
1051 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1056 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1057 algorithms. An implementation of a key exchange algorithm can be obtained
1058 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1059 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1060 the older EVP_PKEY_derive_init() function. See the man pages for the new
1061 functions for further details.
1065 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1069 * Removed the function names from error messages and deprecated the
1072 * Removed NextStep support and the macro OPENSSL_UNISTD
1076 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1077 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1078 Also removed "export var as function" capability; we do not export
1079 variables, only functions.
1083 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1084 an error and 1 indicating success. In previous versions of OpenSSL this
1085 was a void type. If a key was set longer than the maximum possible this
1090 * Support SM2 signing and verification schemes with X509 certificate.
1094 * Use SHA256 as the default digest for TS query in the `ts` app.
1098 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1099 This checks that the salt length is at least 128 bits, the derived key
1100 length is at least 112 bits, and that the iteration count is at least 1000.
1101 For backwards compatibility these checks are disabled by default in the
1102 default provider, but are enabled by default in the fips provider.
1103 To enable or disable these checks use the control
1104 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1108 * Default cipher lists/suites are now available via a function, the
1109 #defines are deprecated.
1113 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1114 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1115 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1119 * Join the directories crypto/x509 and crypto/x509v3
1123 * Added command 'openssl kdf' that uses the EVP_KDF API.
1127 * Added command 'openssl mac' that uses the EVP_MAC API.
1131 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1132 as default directories. Also added the command 'openssl info'
1133 for scripting purposes.
1137 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1138 deprecated. These undocumented functions were never integrated into the EVP
1139 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1140 Bi-directional IGE mode. These modes were never formally standardised and
1141 usage of these functions is believed to be very small. In particular
1142 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1143 is ever used. The security implications are believed to be minimal, but
1144 this issue was never fixed for backwards compatibility reasons. New code
1145 should not use these modes.
1149 * Add prediction resistance to the DRBG reseeding process.
1153 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1154 mandated by IEEE Std 1619-2018.
1158 * Added newline escaping functionality to a filename when using openssl dgst.
1159 This output format is to replicate the output format found in the `*sum`
1160 checksum programs. This aims to preserve backward compatibility.
1162 *Matt Eaton, Richard Levitte, and Paul Dale*
1164 * Removed the heartbeat message in DTLS feature, as it has very
1165 little usage and doesn't seem to fulfill a valuable purpose.
1166 The configuration option is now deprecated.
1170 * Changed the output of 'openssl {digestname} < file' to display the
1171 digest name in its output.
1175 * Added a new generic trace API which provides support for enabling
1176 instrumentation through trace output. This feature is mainly intended
1177 as an aid for developers and is disabled by default. To utilize it,
1178 OpenSSL needs to be configured with the `enable-trace` option.
1180 If the tracing API is enabled, the application can activate trace output
1181 by registering BIOs as trace channels for a number of tracing and debugging
1184 The `openssl` program has been expanded to enable any of the types
1185 available via environment variables defined by the user, and serves as
1186 one possible example on how to use this functionality.
1188 *Richard Levitte & Matthias St. Pierre*
1190 * Added build tests for C++. These are generated files that only do one
1191 thing, to include one public OpenSSL head file each. This tests that
1192 the public header files can be usefully included in a C++ application.
1194 This test isn't enabled by default. It can be enabled with the option
1195 'enable-buildtest-c++'.
1199 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1203 * Add KMAC to EVP_MAC.
1207 * Added property based algorithm implementation selection framework to
1212 * Added SCA hardening for modular field inversion in EC_GROUP through
1213 a new dedicated field_inv() pointer in EC_METHOD.
1214 This also addresses a leakage affecting conversions from projective
1215 to affine coordinates.
1217 *Billy Bob Brumley, Nicola Tuveri*
1219 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1220 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1221 those algorithms that were already supported through the EVP_PKEY API
1222 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1223 and scrypt are now wrappers that call EVP_KDF.
1227 * Build devcrypto engine as a dynamic engine.
1229 *Eneas U de Queiroz*
1231 * Add keyed BLAKE2 to EVP_MAC.
1235 * Fix a bug in the computation of the endpoint-pair shared secret used
1236 by DTLS over SCTP. This breaks interoperability with older versions
1237 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1238 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1239 interoperability with such broken implementations. However, enabling
1240 this switch breaks interoperability with correct implementations.
1242 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1243 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1247 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1251 * Change the license to the Apache License v2.0.
1255 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1257 - Major releases (indicated by incrementing the MAJOR release number)
1258 may introduce incompatible API/ABI changes.
1259 - Minor releases (indicated by incrementing the MINOR release number)
1260 may introduce new features but retain API/ABI compatibility.
1261 - Patch releases (indicated by incrementing the PATCH number)
1262 are intended for bug fixes and other improvements of existing
1263 features only (like improving performance or adding documentation)
1264 and retain API/ABI compatibility.
1268 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1272 * Remove the 'dist' target and add a tarball building script. The
1273 'dist' target has fallen out of use, and it shouldn't be
1274 necessary to configure just to create a source distribution.
1278 * Recreate the OS390-Unix config target. It no longer relies on a
1279 special script like it did for OpenSSL pre-1.1.0.
1283 * Instead of having the source directories listed in Configure, add
1284 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1289 * Add GMAC to EVP_MAC.
1293 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1297 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1298 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1299 to facilitate the continued use of MACs through raw private keys in
1300 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1304 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1305 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1309 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1310 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1311 are retained for backwards compatibility.
1315 * AES-XTS mode now enforces that its two keys are different to mitigate
1316 the attacked described in "Efficient Instantiations of Tweakable
1317 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1318 Details of this attack can be obtained from:
1319 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1323 * Rename the object files, i.e. give them other names than in previous
1324 versions. Their names now include the name of the final product, as
1325 well as its type mnemonic (bin, lib, shlib).
1329 * Added new option for 'openssl list', '-objects', which will display the
1330 list of built in objects, i.e. OIDs with names.
1334 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1335 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1340 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1341 improves application performance by removing data copies and providing
1342 applications with zero-copy system calls such as sendfile and splice.
1346 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1347 option is set, openssl cleanses (zeroize) plaintext bytes from
1348 internal buffers after delivering them to the application. Note,
1349 the application is still responsible for cleansing other copies
1350 (e.g.: data received by SSL_read(3)).
1354 * `PKCS12_parse` now maintains the order of the parsed certificates
1355 when outputting them via `*ca` (rather than reversing it).
1359 * Deprecated pthread fork support methods. These were unused so no
1360 replacement is required.
1362 - OPENSSL_fork_prepare()
1363 - OPENSSL_fork_parent()
1364 - OPENSSL_fork_child()
1371 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1373 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1374 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1375 If an attacker can control both items being compared then this could lead
1376 to a possible denial of service attack. OpenSSL itself uses the
1377 GENERAL_NAME_cmp function for two purposes:
1378 1) Comparing CRL distribution point names between an available CRL and a
1379 CRL distribution point embedded in an X509 certificate
1380 2) When verifying that a timestamp response token signer matches the
1381 timestamp authority name (exposed via the API functions
1382 TS_RESP_verify_response and TS_RESP_verify_token)
1387 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1389 * Certificates with explicit curve parameters are now disallowed in
1390 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1394 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1395 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1396 conversely, silently ignore DTLS protocol version bounds when configuring
1397 TLS-based contexts. The commands can be repeated to set bounds of both
1398 types. The same applies with the corresponding "min_protocol" and
1399 "max_protocol" command-line switches, in case some application uses both TLS
1402 SSL_CTX instances that are created for a fixed protocol version (e.g.
1403 TLSv1_server_method()) also silently ignore version bounds. Previously
1404 attempts to apply bounds to these protocol versions would result in an
1405 error. Now only the "version-flexible" SSL_CTX instances are subject to
1406 limits in configuration files in command-line options.
1410 * Handshake now fails if Extended Master Secret extension is dropped
1415 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1417 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1419 * Fixed segmentation fault in SSL_check_chain()
1420 Server or client applications that call the SSL_check_chain() function
1421 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1422 dereference as a result of incorrect handling of the
1423 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1424 or unrecognised signature algorithm is received from the peer. This could
1425 be exploited by a malicious peer in a Denial of Service attack.
1430 * Added AES consttime code for no-asm configurations
1431 an optional constant time support for AES was added
1432 when building openssl for no-asm.
1433 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1434 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1435 At this time this feature is by default disabled.
1436 It will be enabled by default in 3.0.
1440 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1442 * Revert the change of EOF detection while reading in libssl to avoid
1443 regressions in applications depending on the current way of reporting
1444 the EOF. As the existing method is not fully accurate the change to
1445 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1446 branch and will be present in the 3.0 release.
1450 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1451 when primes for RSA keys are computed.
1452 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1453 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1454 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1455 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1456 This avoids possible fingerprinting of newly generated RSA modules.
1460 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1462 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1463 while reading in libssl then we would report an error back to the
1464 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1465 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1466 therefore give a hint as to what went wrong.
1470 * Check that ed25519 and ed448 are allowed by the security level. Previously
1471 signature algorithms not using an MD were not being checked that they were
1472 allowed by the security level.
1476 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1477 was not quite right. The behaviour was not consistent between resumption
1478 and normal handshakes, and also not quite consistent with historical
1479 behaviour. The behaviour in various scenarios has been clarified and
1480 it has been updated to make it match historical behaviour as closely as
1485 * *[VMS only]* The header files that the VMS compilers include automatically,
1486 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1487 that the C++ compiler doesn't understand. This is a shortcoming in the
1488 compiler, but can be worked around with `__cplusplus` guards.
1490 C++ applications that use OpenSSL libraries must be compiled using the
1491 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1492 functions. Otherwise, only functions with symbols of less than 31
1493 characters can be used, as the linker will not be able to successfully
1494 resolve symbols with longer names.
1498 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1499 The presence of this system service is determined at run-time.
1503 * Added newline escaping functionality to a filename when using openssl dgst.
1504 This output format is to replicate the output format found in the `*sum`
1505 checksum programs. This aims to preserve backward compatibility.
1507 *Matt Eaton, Richard Levitte, and Paul Dale*
1509 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1514 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1516 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1517 number generator (RNG). This was intended to include protection in the
1518 event of a fork() system call in order to ensure that the parent and child
1519 processes did not share the same RNG state. However this protection was not
1520 being used in the default case.
1522 A partial mitigation for this issue is that the output from a high
1523 precision timer is mixed into the RNG state so the likelihood of a parent
1524 and child process sharing state is significantly reduced.
1526 If an application already calls OPENSSL_init_crypto() explicitly using
1527 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1530 *Matthias St. Pierre*
1532 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1533 used even when parsing explicit parameters, when loading a encoded key
1534 or calling `EC_GROUP_new_from_ecpkparameters()`/
1535 `EC_GROUP_new_from_ecparameters()`.
1536 This prevents bypass of security hardening and performance gains,
1537 especially for curves with specialized EC_METHODs.
1538 By default, if a key encoded with explicit parameters is loaded and later
1539 encoded, the output is still encoded with explicit parameters, even if
1540 internally a "named" EC_GROUP is used for computation.
1544 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1545 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1546 NULL. After this change, only the cofactor parameter can be NULL. It also
1547 does some minimal sanity checks on the passed order.
1552 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1553 An attack is simple, if the first CMS_recipientInfo is valid but the
1554 second CMS_recipientInfo is chosen ciphertext. If the second
1555 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1556 encryption key will be replaced by garbage, and the message cannot be
1557 decoded, but if the RSA decryption fails, the correct encryption key is
1558 used and the recipient will not notice the attack.
1559 As a work around for this potential attack the length of the decrypted
1560 key must be equal to the cipher default key length, in case the
1561 certifiate is not given and all recipientInfo are tried out.
1562 The old behaviour can be re-enabled in the CMS code by setting the
1563 CMS_DEBUG_DECRYPT flag.
1568 * Early start up entropy quality from the DEVRANDOM seed source has been
1569 improved for older Linux systems. The RAND subsystem will wait for
1570 /dev/random to be producing output before seeding from /dev/urandom.
1571 The seeded state is stored for future library initialisations using
1572 a system global shared memory segment. The shared memory identifier
1573 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1574 the desired value. The default identifier is 114.
1578 * Correct the extended master secret constant on EBCDIC systems. Without this
1579 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1580 negotiate EMS will fail. Unfortunately this also means that TLS connections
1581 between EBCDIC systems with this fix, and EBCDIC systems without this
1582 fix will fail if they negotiate EMS.
1586 * Use Windows installation paths in the mingw builds
1588 Mingw isn't a POSIX environment per se, which means that Windows
1589 paths should be used for installation.
1594 * Changed DH_check to accept parameters with order q and 2q subgroups.
1595 With order 2q subgroups the bit 0 of the private key is not secret
1596 but DH_generate_key works around that by clearing bit 0 of the
1597 private key for those. This avoids leaking bit 0 of the private key.
1601 * Significantly reduce secure memory usage by the randomness pools.
1605 * Revert the DEVRANDOM_WAIT feature for Linux systems
1607 The DEVRANDOM_WAIT feature added a select() call to wait for the
1608 /dev/random device to become readable before reading from the
1609 /dev/urandom device.
1611 It turned out that this change had negative side effects on
1612 performance which were not acceptable. After some discussion it
1613 was decided to revert this feature and leave it up to the OS
1614 resp. the platform maintainer to ensure a proper initialization
1615 during early boot time.
1617 *Matthias St. Pierre*
1619 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1621 * Add build tests for C++. These are generated files that only do one
1622 thing, to include one public OpenSSL head file each. This tests that
1623 the public header files can be usefully included in a C++ application.
1625 This test isn't enabled by default. It can be enabled with the option
1626 'enable-buildtest-c++'.
1630 * Enable SHA3 pre-hashing for ECDSA and DSA.
1634 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1635 This changes the size when using the `genpkey` command when no size is given.
1636 It fixes an omission in earlier changes that changed all RSA, DSA and DH
1637 generation commands to use 2048 bits by default.
1641 * Reorganize the manual pages to consistently have RETURN VALUES,
1642 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1643 util/fix-doc-nits accordingly.
1645 *Paul Yang, Joshua Lock*
1647 * Add the missing accessor EVP_PKEY_get0_engine()
1651 * Have commands like `s_client` and `s_server` output the signature scheme
1652 along with other cipher suite parameters when debugging.
1656 * Make OPENSSL_config() error agnostic again.
1660 * Do the error handling in RSA decryption constant time.
1664 * Prevent over long nonces in ChaCha20-Poly1305.
1666 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1667 for every encryption operation. RFC 7539 specifies that the nonce value
1668 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1669 and front pads the nonce with 0 bytes if it is less than 12
1670 bytes. However it also incorrectly allows a nonce to be set of up to 16
1671 bytes. In this case only the last 12 bytes are significant and any
1672 additional leading bytes are ignored.
1674 It is a requirement of using this cipher that nonce values are
1675 unique. Messages encrypted using a reused nonce value are susceptible to
1676 serious confidentiality and integrity attacks. If an application changes
1677 the default nonce length to be longer than 12 bytes and then makes a
1678 change to the leading bytes of the nonce expecting the new value to be a
1679 new unique nonce then such an application could inadvertently encrypt
1680 messages with a reused nonce.
1682 Additionally the ignored bytes in a long nonce are not covered by the
1683 integrity guarantee of this cipher. Any application that relies on the
1684 integrity of these ignored leading bytes of a long nonce may be further
1685 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1686 is safe because no such use sets such a long nonce value. However user
1687 applications that use this cipher directly and set a non-default nonce
1688 length to be longer than 12 bytes may be vulnerable.
1690 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1696 * Add DEVRANDOM_WAIT feature for Linux systems
1698 On older Linux systems where the getrandom() system call is not available,
1699 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1700 Contrary to getrandom(), the /dev/urandom device will not block during
1701 early boot when the kernel CSPRNG has not been seeded yet.
1703 To mitigate this known weakness, use select() to wait for /dev/random to
1704 become readable before reading from /dev/urandom.
1706 * Ensure that SM2 only uses SM3 as digest algorithm
1710 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1712 * Change the info callback signals for the start and end of a post-handshake
1713 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1714 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1715 confused by this and assume that a TLSv1.2 renegotiation has started. This
1716 can break KeyUpdate handling. Instead we no longer signal the start and end
1717 of a post handshake message exchange (although the messages themselves are
1718 still signalled). This could break some applications that were expecting
1719 the old signals. However without this KeyUpdate is not usable for many
1724 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1726 * Timing vulnerability in DSA signature generation
1728 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1729 timing side channel attack. An attacker could use variations in the signing
1730 algorithm to recover the private key.
1732 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1737 * Timing vulnerability in ECDSA signature generation
1739 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1740 timing side channel attack. An attacker could use variations in the signing
1741 algorithm to recover the private key.
1743 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1748 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1749 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1750 of two gigabytes and the error handling improved.
1752 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1753 categorized as a normal bug, not a security issue, because the DRBG reseeds
1754 automatically and is fully functional even without additional randomness
1755 provided by the application.
1757 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1759 * Add a new ClientHello callback. Provides a callback interface that gives
1760 the application the ability to adjust the nascent SSL object at the
1761 earliest stage of ClientHello processing, immediately after extensions have
1762 been collected but before they have been processed. In particular, this
1763 callback can adjust the supported TLS versions in response to the contents
1768 * Add SM2 base algorithm support.
1772 * s390x assembly pack: add (improved) hardware-support for the following
1773 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1774 aes-cfb/cfb8, aes-ecb.
1778 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1779 parameter is no longer accepted, as it leads to a corrupt table. NULL
1780 pem_str is reserved for alias entries only.
1784 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1785 step for prime curves. The new implementation is based on formulae from
1786 differential addition-and-doubling in homogeneous projective coordinates
1787 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1788 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1789 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1790 to work in projective coordinates.
1792 *Billy Bob Brumley, Nicola Tuveri*
1794 * Change generating and checking of primes so that the error rate of not
1795 being prime depends on the intended use based on the size of the input.
1796 For larger primes this will result in more rounds of Miller-Rabin.
1797 The maximal error rate for primes with more than 1080 bits is lowered
1800 *Kurt Roeckx, Annie Yousar*
1802 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1806 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1807 moving between systems, and to avoid confusion when a Windows build is
1808 done with mingw vs with MSVC. For POSIX installs, there's still a
1809 symlink or copy named 'tsget' to avoid that confusion as well.
1813 * Revert blinding in ECDSA sign and instead make problematic addition
1814 length-invariant. Switch even to fixed-length Montgomery multiplication.
1818 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1819 step for binary curves. The new implementation is based on formulae from
1820 differential addition-and-doubling in mixed Lopez-Dahab projective
1821 coordinates, modified to independently blind the operands.
1823 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1825 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1826 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1827 EC_METHODs to implement their own specialized "ladder step", to take
1828 advantage of more favorable coordinate systems or more efficient
1829 differential addition-and-doubling algorithms.
1831 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1833 * Modified the random device based seed sources to keep the relevant
1834 file descriptors open rather than reopening them on each access.
1835 This allows such sources to operate in a chroot() jail without
1836 the associated device nodes being available. This behaviour can be
1837 controlled using RAND_keep_random_devices_open().
1841 * Numerous side-channel attack mitigations have been applied. This may have
1842 performance impacts for some algorithms for the benefit of improved
1843 security. Specific changes are noted in this change log by their respective
1848 * AIX shared library support overhaul. Switch to AIX "natural" way of
1849 handling shared libraries, which means collecting shared objects of
1850 different versions and bitnesses in one common archive. This allows to
1851 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1852 doesn't affect the way 3rd party applications are linked, only how
1853 multi-version installation is managed.
1857 * Make ec_group_do_inverse_ord() more robust and available to other
1858 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1859 mitigations are applied to the fallback BN_mod_inverse().
1860 When using this function rather than BN_mod_inverse() directly, new
1861 EC cryptosystem implementations are then safer-by-default.
1865 * Add coordinate blinding for EC_POINT and implement projective
1866 coordinate blinding for generic prime curves as a countermeasure to
1867 chosen point SCA attacks.
1869 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1871 * Add blinding to ECDSA and DSA signatures to protect against side channel
1872 attacks discovered by Keegan Ryan (NCC Group).
1876 * Enforce checking in the `pkeyutl` command to ensure that the input
1877 length does not exceed the maximum supported digest length when performing
1878 a sign, verify or verifyrecover operation.
1882 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1883 I/O in combination with something like select() or poll() will hang. This
1884 can be turned off again using SSL_CTX_clear_mode().
1885 Many applications do not properly handle non-application data records, and
1886 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1887 around the problems in those applications, but can also break some.
1888 It's recommended to read the manpages about SSL_read(), SSL_write(),
1889 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1890 SSL_CTX_set_read_ahead() again.
1894 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1895 now allow empty (zero character) pass phrases.
1899 * Apply blinding to binary field modular inversion and remove patent
1900 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1904 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1905 binary and prime elliptic curves.
1909 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1910 constant time fixed point multiplication.
1914 * Revise elliptic curve scalar multiplication with timing attack
1915 defenses: ec_wNAF_mul redirects to a constant time implementation
1916 when computing fixed point and variable point multiplication (which
1917 in OpenSSL are mostly used with secret scalars in keygen, sign,
1918 ECDH derive operations).
1919 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1922 * Updated CONTRIBUTING
1926 * Updated DRBG / RAND to request nonce and additional low entropy
1927 randomness from the system.
1929 *Matthias St. Pierre*
1931 * Updated 'openssl rehash' to use OpenSSL consistent default.
1935 * Moved the load of the ssl_conf module to libcrypto, which helps
1936 loading engines that libssl uses before libssl is initialised.
1940 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1944 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1946 *Ingo Schwarze, Rich Salz*
1948 * Added output of accepting IP address and port for 'openssl s_server'
1952 * Added a new API for TLSv1.3 ciphersuites:
1953 SSL_CTX_set_ciphersuites()
1954 SSL_set_ciphersuites()
1958 * Memory allocation failures consistently add an error to the error
1963 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1964 in libcrypto when run as setuid/setgid.
1968 * Load any config file by default when libssl is used.
1972 * Added new public header file <openssl/rand_drbg.h> and documentation
1973 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1975 *Matthias St. Pierre*
1977 * QNX support removed (cannot find contributors to get their approval
1978 for the license change).
1982 * TLSv1.3 replay protection for early data has been implemented. See the
1983 SSL_read_early_data() man page for further details.
1987 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1988 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1989 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1990 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1991 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1992 configuration has been separated out. See the ciphers man page or the
1993 SSL_CTX_set_ciphersuites() man page for more information.
1997 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1998 in responder mode now supports the new "-multi" option, which
1999 spawns the specified number of child processes to handle OCSP
2000 requests. The "-timeout" option now also limits the OCSP
2001 responder's patience to wait to receive the full client request
2002 on a newly accepted connection. Child processes are respawned
2003 as needed, and the CA index file is automatically reloaded
2004 when changed. This makes it possible to run the "ocsp" responder
2005 as a long-running service, making the OpenSSL CA somewhat more
2006 feature-complete. In this mode, most diagnostic messages logged
2007 after entering the event loop are logged via syslog(3) rather than
2012 * Added support for X448 and Ed448. Heavily based on original work by
2017 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2018 objects loaded. This adds the functions OSSL_STORE_expect() and
2019 OSSL_STORE_find() as well as needed tools to construct searches and
2020 get the search data out of them.
2024 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2025 version of OpenSSL should review their configuration settings to ensure
2026 that they are still appropriate for TLSv1.3. For further information see:
2027 <https://wiki.openssl.org/index.php/TLS1.3>
2031 * Grand redesign of the OpenSSL random generator
2033 The default RAND method now utilizes an AES-CTR DRBG according to
2034 NIST standard SP 800-90Ar1. The new random generator is essentially
2035 a port of the default random generator from the OpenSSL FIPS 2.0
2036 object module. It is a hybrid deterministic random bit generator
2037 using an AES-CTR bit stream and which seeds and reseeds itself
2038 automatically using trusted system entropy sources.
2040 Some of its new features are:
2041 - Support for multiple DRBG instances with seed chaining.
2042 - The default RAND method makes use of a DRBG.
2043 - There is a public and private DRBG instance.
2044 - The DRBG instances are fork-safe.
2045 - Keep all global DRBG instances on the secure heap if it is enabled.
2046 - The public and private DRBG instance are per thread for lock free
2049 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2051 * Changed Configure so it only says what it does and doesn't dump
2052 so much data. Instead, ./configdata.pm should be used as a script
2053 to display all sorts of configuration data.
2057 * Added processing of "make variables" to Configure.
2061 * Added SHA512/224 and SHA512/256 algorithm support.
2065 * The last traces of Netware support, first removed in 1.1.0, have
2070 * Get rid of Makefile.shared, and in the process, make the processing
2071 of certain files (rc.obj, or the .def/.map/.opt files produced from
2072 the ordinal files) more visible and hopefully easier to trace and
2073 debug (or make silent).
2077 * Make it possible to have environment variable assignments as
2078 arguments to config / Configure.
2082 * Add multi-prime RSA (RFC 8017) support.
2086 * Add SM3 implemented according to GB/T 32905-2016
2087 *Jack Lloyd <jack.lloyd@ribose.com>,*
2088 *Ronald Tse <ronald.tse@ribose.com>,*
2089 *Erick Borsboom <erick.borsboom@ribose.com>*
2091 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2092 as documented in RFC6066.
2093 Based on a patch from Tomasz Moń
2095 *Filipe Raimundo da Silva*
2097 * Add SM4 implemented according to GB/T 32907-2016.
2098 *Jack Lloyd <jack.lloyd@ribose.com>,*
2099 *Ronald Tse <ronald.tse@ribose.com>,*
2100 *Erick Borsboom <erick.borsboom@ribose.com>*
2102 * Reimplement -newreq-nodes and ERR_error_string_n; the
2103 original author does not agree with the license change.
2107 * Add ARIA AEAD TLS support.
2111 * Some macro definitions to support VS6 have been removed. Visual
2112 Studio 6 has not worked since 1.1.0
2116 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2117 without clearing the errors.
2121 * Add "atfork" functions. If building on a system that without
2122 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2123 requirements. The RAND facility now uses/requires this.
2131 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2132 not possible to disable entirely. However, it's still possible to
2133 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2136 To disable, configure with 'no-ui-console'. 'no-ui' is still
2137 possible to use as an alias. Check at compile time with the
2138 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2139 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2143 * Add a STORE module, which implements a uniform and URI based reader of
2144 stores that can contain keys, certificates, CRLs and numerous other
2145 objects. The main API is loosely based on a few stdio functions,
2146 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2147 OSSL_STORE_error and OSSL_STORE_close.
2148 The implementation uses backends called "loaders" to implement arbitrary
2149 URI schemes. There is one built in "loader" for the 'file' scheme.
2153 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2154 then adjusted to work on FreeBSD 8.4 as well.
2155 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2156 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2160 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2161 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2162 error code calls like this:
2164 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2166 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2167 that can be encoded in C. For the foreseeable future, this will only
2170 *Richard Levitte and Tim Hudson*
2172 * Removed BSD cryptodev engine.
2176 * Add a build target 'build_all_generated', to build all generated files
2177 and only that. This can be used to prepare everything that requires
2178 things like perl for a system that lacks perl and then move everything
2179 to that system and do the rest of the build there.
2183 * In the UI interface, make it possible to duplicate the user data. This
2184 can be used by engines that need to retain the data for a longer time
2185 than just the call where this user data is passed.
2189 * Ignore the '-named_curve auto' value for compatibility of applications
2192 *Tomas Mraz <tmraz@fedoraproject.org>*
2194 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2195 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2196 alerts across multiple records (some of which could be empty). In practice
2197 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2198 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2199 support this at all. Supporting it adds significant complexity to the
2200 record layer, and its removal is unlikely to cause interoperability
2205 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2206 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2207 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2212 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2213 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2215 *Richard Levitte, Andy Polyakov*
2217 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2222 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2223 platform rather than 'mingw'.
2227 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2228 success if they are asked to add an object which already exists
2229 in the store. This change cascades to other functions which load
2230 certificates and CRLs.
2234 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2235 facilitate stack unwinding even from assembly subroutines.
2239 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2240 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2244 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2245 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2246 which is the minimum version we support.
2250 * Certificate time validation (X509_cmp_time) enforces stricter
2251 compliance with RFC 5280. Fractional seconds and timezone offsets
2252 are no longer allowed.
2256 * Add support for ARIA
2260 * s_client will now send the Server Name Indication (SNI) extension by
2261 default unless the new "-noservername" option is used. The server name is
2262 based on the host provided to the "-connect" option unless overridden by
2263 using "-servername".
2267 * Add support for SipHash
2271 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2272 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2273 prevent issues where no progress is being made and the peer continually
2274 sends unrecognised record types, using up resources processing them.
2278 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2279 using the algorithm defined in
2280 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2284 * Heartbeat support has been removed; the ABI is changed for now.
2286 *Richard Levitte, Rich Salz*
2288 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2292 * The RSA "null" method, which was partially supported to avoid patent
2293 issues, has been replaced to always returns NULL.
2300 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2302 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2303 used even when parsing explicit parameters, when loading a encoded key
2304 or calling `EC_GROUP_new_from_ecpkparameters()`/
2305 `EC_GROUP_new_from_ecparameters()`.
2306 This prevents bypass of security hardening and performance gains,
2307 especially for curves with specialized EC_METHODs.
2308 By default, if a key encoded with explicit parameters is loaded and later
2309 encoded, the output is still encoded with explicit parameters, even if
2310 internally a "named" EC_GROUP is used for computation.
2314 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2315 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2316 NULL. After this change, only the cofactor parameter can be NULL. It also
2317 does some minimal sanity checks on the passed order.
2322 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2323 An attack is simple, if the first CMS_recipientInfo is valid but the
2324 second CMS_recipientInfo is chosen ciphertext. If the second
2325 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2326 encryption key will be replaced by garbage, and the message cannot be
2327 decoded, but if the RSA decryption fails, the correct encryption key is
2328 used and the recipient will not notice the attack.
2329 As a work around for this potential attack the length of the decrypted
2330 key must be equal to the cipher default key length, in case the
2331 certifiate is not given and all recipientInfo are tried out.
2332 The old behaviour can be re-enabled in the CMS code by setting the
2333 CMS_DEBUG_DECRYPT flag.
2338 * Use Windows installation paths in the mingw builds
2340 Mingw isn't a POSIX environment per se, which means that Windows
2341 paths should be used for installation.
2346 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2348 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2349 This changes the size when using the `genpkey` command when no size is given.
2350 It fixes an omission in earlier changes that changed all RSA, DSA and DH
2351 generation commands to use 2048 bits by default.
2355 * Prevent over long nonces in ChaCha20-Poly1305.
2357 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2358 for every encryption operation. RFC 7539 specifies that the nonce value
2359 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2360 and front pads the nonce with 0 bytes if it is less than 12
2361 bytes. However it also incorrectly allows a nonce to be set of up to 16
2362 bytes. In this case only the last 12 bytes are significant and any
2363 additional leading bytes are ignored.
2365 It is a requirement of using this cipher that nonce values are
2366 unique. Messages encrypted using a reused nonce value are susceptible to
2367 serious confidentiality and integrity attacks. If an application changes
2368 the default nonce length to be longer than 12 bytes and then makes a
2369 change to the leading bytes of the nonce expecting the new value to be a
2370 new unique nonce then such an application could inadvertently encrypt
2371 messages with a reused nonce.
2373 Additionally the ignored bytes in a long nonce are not covered by the
2374 integrity guarantee of this cipher. Any application that relies on the
2375 integrity of these ignored leading bytes of a long nonce may be further
2376 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2377 is safe because no such use sets such a long nonce value. However user
2378 applications that use this cipher directly and set a non-default nonce
2379 length to be longer than 12 bytes may be vulnerable.
2381 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2387 * Added SCA hardening for modular field inversion in EC_GROUP through
2388 a new dedicated field_inv() pointer in EC_METHOD.
2389 This also addresses a leakage affecting conversions from projective
2390 to affine coordinates.
2392 *Billy Bob Brumley, Nicola Tuveri*
2394 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2395 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2399 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2403 * Remove the 'dist' target and add a tarball building script. The
2404 'dist' target has fallen out of use, and it shouldn't be
2405 necessary to configure just to create a source distribution.
2409 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2411 * Timing vulnerability in DSA signature generation
2413 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2414 timing side channel attack. An attacker could use variations in the signing
2415 algorithm to recover the private key.
2417 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2422 * Timing vulnerability in ECDSA signature generation
2424 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2425 timing side channel attack. An attacker could use variations in the signing
2426 algorithm to recover the private key.
2428 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2433 * Add coordinate blinding for EC_POINT and implement projective
2434 coordinate blinding for generic prime curves as a countermeasure to
2435 chosen point SCA attacks.
2437 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2439 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2441 * Client DoS due to large DH parameter
2443 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2444 malicious server can send a very large prime value to the client. This will
2445 cause the client to spend an unreasonably long period of time generating a
2446 key for this prime resulting in a hang until the client has finished. This
2447 could be exploited in a Denial Of Service attack.
2449 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2454 * Cache timing vulnerability in RSA Key Generation
2456 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2457 a cache timing side channel attack. An attacker with sufficient access to
2458 mount cache timing attacks during the RSA key generation process could
2459 recover the private key.
2461 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2462 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2467 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2468 parameter is no longer accepted, as it leads to a corrupt table. NULL
2469 pem_str is reserved for alias entries only.
2473 * Revert blinding in ECDSA sign and instead make problematic addition
2474 length-invariant. Switch even to fixed-length Montgomery multiplication.
2478 * Change generating and checking of primes so that the error rate of not
2479 being prime depends on the intended use based on the size of the input.
2480 For larger primes this will result in more rounds of Miller-Rabin.
2481 The maximal error rate for primes with more than 1080 bits is lowered
2484 *Kurt Roeckx, Annie Yousar*
2486 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2490 * Add blinding to ECDSA and DSA signatures to protect against side channel
2491 attacks discovered by Keegan Ryan (NCC Group).
2495 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2496 now allow empty (zero character) pass phrases.
2500 * Certificate time validation (X509_cmp_time) enforces stricter
2501 compliance with RFC 5280. Fractional seconds and timezone offsets
2502 are no longer allowed.
2506 * Fixed a text canonicalisation bug in CMS
2508 Where a CMS detached signature is used with text content the text goes
2509 through a canonicalisation process first prior to signing or verifying a
2510 signature. This process strips trailing space at the end of lines, converts
2511 line terminators to CRLF and removes additional trailing line terminators
2512 at the end of a file. A bug in the canonicalisation process meant that
2513 some characters, such as form-feed, were incorrectly treated as whitespace
2514 and removed. This is contrary to the specification (RFC5485). This fix
2515 could mean that detached text data signed with an earlier version of
2516 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2517 signed with a fixed OpenSSL may fail to verify with an earlier version of
2518 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2519 and use the "-binary" flag (for the "cms" command line application) or set
2520 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2524 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2526 * Constructed ASN.1 types with a recursive definition could exceed the stack
2528 Constructed ASN.1 types with a recursive definition (such as can be found
2529 in PKCS7) could eventually exceed the stack given malicious input with
2530 excessive recursion. This could result in a Denial Of Service attack. There
2531 are no such structures used within SSL/TLS that come from untrusted sources
2532 so this is considered safe.
2534 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2540 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2542 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2543 effectively reduced to only comparing the least significant bit of each
2544 byte. This allows an attacker to forge messages that would be considered as
2545 authenticated in an amount of tries lower than that guaranteed by the
2546 security claims of the scheme. The module can only be compiled by the
2547 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2549 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2555 * Add a build target 'build_all_generated', to build all generated files
2556 and only that. This can be used to prepare everything that requires
2557 things like perl for a system that lacks perl and then move everything
2558 to that system and do the rest of the build there.
2562 * Backport SSL_OP_NO_RENGOTIATION
2564 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2565 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2566 changes this is no longer possible in 1.1.0. Therefore the new
2567 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2568 1.1.0 to provide equivalent functionality.
2570 Note that if an application built against 1.1.0h headers (or above) is run
2571 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2572 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2576 * Removed the OS390-Unix config target. It relied on a script that doesn't
2581 * rsaz_1024_mul_avx2 overflow bug on x86_64
2583 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2584 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2585 Analysis suggests that attacks against RSA and DSA as a result of this
2586 defect would be very difficult to perform and are not believed likely.
2587 Attacks against DH1024 are considered just feasible, because most of the
2588 work necessary to deduce information about a private key may be performed
2589 offline. The amount of resources required for such an attack would be
2590 significant. However, for an attack on TLS to be meaningful, the server
2591 would have to share the DH1024 private key among multiple clients, which is
2592 no longer an option since CVE-2016-0701.
2594 This only affects processors that support the AVX2 but not ADX extensions
2595 like Intel Haswell (4th generation).
2597 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2598 was originally found via the OSS-Fuzz project.
2603 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2605 * bn_sqrx8x_internal carry bug on x86_64
2607 There is a carry propagating bug in the x86_64 Montgomery squaring
2608 procedure. No EC algorithms are affected. Analysis suggests that attacks
2609 against RSA and DSA as a result of this defect would be very difficult to
2610 perform and are not believed likely. Attacks against DH are considered just
2611 feasible (although very difficult) because most of the work necessary to
2612 deduce information about a private key may be performed offline. The amount
2613 of resources required for such an attack would be very significant and
2614 likely only accessible to a limited number of attackers. An attacker would
2615 additionally need online access to an unpatched system using the target
2616 private key in a scenario with persistent DH parameters and a private
2617 key that is shared between multiple clients.
2619 This only affects processors that support the BMI1, BMI2 and ADX extensions
2620 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2622 This issue was reported to OpenSSL by the OSS-Fuzz project.
2627 * Malformed X.509 IPAddressFamily could cause OOB read
2629 If an X.509 certificate has a malformed IPAddressFamily extension,
2630 OpenSSL could do a one-byte buffer overread. The most likely result
2631 would be an erroneous display of the certificate in text format.
2633 This issue was reported to OpenSSL by the OSS-Fuzz project.
2638 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2640 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2641 platform rather than 'mingw'.
2645 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2646 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2647 which is the minimum version we support.
2651 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2653 * Encrypt-Then-Mac renegotiation crash
2655 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2656 negotiated where it was not in the original handshake (or vice-versa) then
2657 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2658 and servers are affected.
2660 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2665 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2667 * Truncated packet could crash via OOB read
2669 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2670 cipher is being used, then a truncated packet can cause that host to
2671 perform an out-of-bounds read, usually resulting in a crash.
2673 This issue was reported to OpenSSL by Robert Święcki of Google.
2678 * Bad (EC)DHE parameters cause a client crash
2680 If a malicious server supplies bad parameters for a DHE or ECDHE key
2681 exchange then this can result in the client attempting to dereference a
2682 NULL pointer leading to a client crash. This could be exploited in a Denial
2685 This issue was reported to OpenSSL by Guido Vranken.
2690 * BN_mod_exp may produce incorrect results on x86_64
2692 There is a carry propagating bug in the x86_64 Montgomery squaring
2693 procedure. No EC algorithms are affected. Analysis suggests that attacks
2694 against RSA and DSA as a result of this defect would be very difficult to
2695 perform and are not believed likely. Attacks against DH are considered just
2696 feasible (although very difficult) because most of the work necessary to
2697 deduce information about a private key may be performed offline. The amount
2698 of resources required for such an attack would be very significant and
2699 likely only accessible to a limited number of attackers. An attacker would
2700 additionally need online access to an unpatched system using the target
2701 private key in a scenario with persistent DH parameters and a private
2702 key that is shared between multiple clients. For example this can occur by
2703 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2704 similar to CVE-2015-3193 but must be treated as a separate problem.
2706 This issue was reported to OpenSSL by the OSS-Fuzz project.
2711 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2713 * ChaCha20/Poly1305 heap-buffer-overflow
2715 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2716 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2717 crash. This issue is not considered to be exploitable beyond a DoS.
2719 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2724 * CMS Null dereference
2726 Applications parsing invalid CMS structures can crash with a NULL pointer
2727 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2728 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2729 structure callback if an attempt is made to free certain invalid encodings.
2730 Only CHOICE structures using a callback which do not handle NULL value are
2733 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2738 * Montgomery multiplication may produce incorrect results
2740 There is a carry propagating bug in the Broadwell-specific Montgomery
2741 multiplication procedure that handles input lengths divisible by, but
2742 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2743 and DH private keys are impossible. This is because the subroutine in
2744 question is not used in operations with the private key itself and an input
2745 of the attacker's direct choice. Otherwise the bug can manifest itself as
2746 transient authentication and key negotiation failures or reproducible
2747 erroneous outcome of public-key operations with specially crafted input.
2748 Among EC algorithms only Brainpool P-512 curves are affected and one
2749 presumably can attack ECDH key negotiation. Impact was not analyzed in
2750 detail, because pre-requisites for attack are considered unlikely. Namely
2751 multiple clients have to choose the curve in question and the server has to
2752 share the private key among them, neither of which is default behaviour.
2753 Even then only clients that chose the curve will be affected.
2755 This issue was publicly reported as transient failures and was not
2756 initially recognized as a security issue. Thanks to Richard Morgan for
2757 providing reproducible case.
2762 * Removed automatic addition of RPATH in shared libraries and executables,
2763 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2767 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2769 * Fix Use After Free for large message sizes
2771 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2772 message larger than approx 16k is received then the underlying buffer to
2773 store the incoming message is reallocated and moved. Unfortunately a
2774 dangling pointer to the old location is left which results in an attempt to
2775 write to the previously freed location. This is likely to result in a
2776 crash, however it could potentially lead to execution of arbitrary code.
2778 This issue only affects OpenSSL 1.1.0a.
2780 This issue was reported to OpenSSL by Robert Święcki.
2785 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2787 * OCSP Status Request extension unbounded memory growth
2789 A malicious client can send an excessively large OCSP Status Request
2790 extension. If that client continually requests renegotiation, sending a
2791 large OCSP Status Request extension each time, then there will be unbounded
2792 memory growth on the server. This will eventually lead to a Denial Of
2793 Service attack through memory exhaustion. Servers with a default
2794 configuration are vulnerable even if they do not support OCSP. Builds using
2795 the "no-ocsp" build time option are not affected.
2797 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2802 * SSL_peek() hang on empty record
2804 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2805 sends an empty record. This could be exploited by a malicious peer in a
2806 Denial Of Service attack.
2808 This issue was reported to OpenSSL by Alex Gaynor.
2813 * Excessive allocation of memory in tls_get_message_header() and
2814 dtls1_preprocess_fragment()
2816 A (D)TLS message includes 3 bytes for its length in the header for the
2817 message. This would allow for messages up to 16Mb in length. Messages of
2818 this length are excessive and OpenSSL includes a check to ensure that a
2819 peer is sending reasonably sized messages in order to avoid too much memory
2820 being consumed to service a connection. A flaw in the logic of version
2821 1.1.0 means that memory for the message is allocated too early, prior to
2822 the excessive message length check. Due to way memory is allocated in
2823 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2824 to service a connection. This could lead to a Denial of Service through
2825 memory exhaustion. However, the excessive message length check still takes
2826 place, and this would cause the connection to immediately fail. Assuming
2827 that the application calls SSL_free() on the failed connection in a timely
2828 manner then the 21Mb of allocated memory will then be immediately freed
2829 again. Therefore the excessive memory allocation will be transitory in
2830 nature. This then means that there is only a security impact if:
2832 1) The application does not call SSL_free() in a timely manner in the event
2833 that the connection fails
2835 2) The application is working in a constrained environment where there is
2836 very little free memory
2838 3) The attacker initiates multiple connection attempts such that there are
2839 multiple connections in a state where memory has been allocated for the
2840 connection; SSL_free() has not yet been called; and there is insufficient
2841 memory to service the multiple requests.
2843 Except in the instance of (1) above any Denial Of Service is likely to be
2844 transitory because as soon as the connection fails the memory is
2845 subsequently freed again in the SSL_free() call. However there is an
2846 increased risk during this period of application crashes due to the lack of
2847 memory - which would then mean a more serious Denial of Service.
2849 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2850 (CVE-2016-6307 and CVE-2016-6308)
2854 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2855 had to be removed. Primary reason is that vendor assembler can't
2856 assemble our modules with -KPIC flag. As result it, assembly
2857 support, was not even available as option. But its lack means
2858 lack of side-channel resistant code, which is incompatible with
2859 security by todays standards. Fortunately gcc is readily available
2860 prepackaged option, which we firmly point at...
2864 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2866 * Windows command-line tool supports UTF-8 opt-in option for arguments
2867 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2868 (to any value) allows Windows user to access PKCS#12 file generated
2869 with Windows CryptoAPI and protected with non-ASCII password, as well
2870 as files generated under UTF-8 locale on Linux also protected with
2875 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2876 have been disabled by default and removed from DEFAULT, just like RC4.
2877 See the RC4 item below to re-enable both.
2881 * The method for finding the storage location for the Windows RAND seed file
2882 has changed. First we check %RANDFILE%. If that is not set then we check
2883 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2884 all else fails we fall back to C:\.
2888 * The EVP_EncryptUpdate() function has had its return type changed from void
2889 to int. A return of 0 indicates and error while a return of 1 indicates
2894 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2895 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2896 off the constant time implementation for RSA, DSA and DH have been made
2897 no-ops and deprecated.
2901 * Windows RAND implementation was simplified to only get entropy by
2902 calling CryptGenRandom(). Various other RAND-related tickets
2905 *Joseph Wylie Yandle, Rich Salz*
2907 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2908 and `OPENSSL_LH_`, respectively. The old names are available
2909 with API compatibility. They new names are now completely documented.
2913 * Unify TYPE_up_ref(obj) methods signature.
2914 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2915 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2916 int (instead of void) like all others TYPE_up_ref() methods.
2917 So now these methods also check the return value of CRYPTO_atomic_add(),
2918 and the validity of object reference counter.
2920 *fdasilvayy@gmail.com*
2922 * With Windows Visual Studio builds, the .pdb files are installed
2923 alongside the installed libraries and executables. For a static
2924 library installation, ossl_static.pdb is the associate compiler
2925 generated .pdb file to be used when linking programs.
2929 * Remove openssl.spec. Packaging files belong with the packagers.
2933 * Automatic Darwin/OSX configuration has had a refresh, it will now
2934 recognise x86_64 architectures automatically. You can still decide
2935 to build for a different bitness with the environment variable
2936 KERNEL_BITS (can be 32 or 64), for example:
2938 KERNEL_BITS=32 ./config
2942 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2943 256 bit AES and HMAC with SHA256.
2947 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2951 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2955 * To enable users to have their own config files and build file templates,
2956 Configure looks in the directory indicated by the environment variable
2957 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2958 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2959 name and is used as is.
2963 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2964 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2965 X509_CERT_FILE_CTX was removed.
2969 * "shared" builds are now the default. To create only static libraries use
2970 the "no-shared" Configure option.
2974 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2975 All of these option have not worked for some while and are fundamental
2980 * Make various cleanup routines no-ops and mark them as deprecated. Most
2981 global cleanup functions are no longer required because they are handled
2982 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2983 Explicitly de-initing can cause problems (e.g. where a library that uses
2984 OpenSSL de-inits, but an application is still using it). The affected
2985 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2986 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2987 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2988 COMP_zlib_cleanup().
2992 * --strict-warnings no longer enables runtime debugging options
2993 such as REF_DEBUG. Instead, debug options are automatically
2994 enabled with '--debug' builds.
2996 *Andy Polyakov, Emilia Käsper*
2998 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2999 have been moved out of the public header files. New functions for managing
3000 these have been added.
3004 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3005 objects have been moved out of the public header files. New
3006 functions for managing these have been added.
3010 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3011 have been moved out of the public header files. New functions for managing
3012 these have been added.
3016 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3017 moved out of the public header files. New functions for managing these
3022 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3026 * Removed the mk1mf build scripts.
3030 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3031 it is always safe to #include a header now.
3035 * Removed the aged BC-32 config and all its supporting scripts
3039 * Removed support for Ultrix, Netware, and OS/2.
3043 * Add support for HKDF.
3045 *Alessandro Ghedini*
3047 * Add support for blake2b and blake2s
3051 * Added support for "pipelining". Ciphers that have the
3052 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3053 encryptions/decryptions simultaneously. There are currently no built-in
3054 ciphers with this property but the expectation is that engines will be able
3055 to offer it to significantly improve throughput. Support has been extended
3056 into libssl so that multiple records for a single connection can be
3057 processed in one go (for >=TLS 1.1).
3061 * Added the AFALG engine. This is an async capable engine which is able to
3062 offload work to the Linux kernel. In this initial version it only supports
3063 AES128-CBC. The kernel must be version 4.1.0 or greater.
3067 * OpenSSL now uses a new threading API. It is no longer necessary to
3068 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3069 are two supported threading models: pthreads and windows threads. It is
3070 also possible to configure OpenSSL at compile time for "no-threads". The
3071 old threading API should no longer be used. The functions have been
3072 replaced with "no-op" compatibility macros.
3074 *Alessandro Ghedini, Matt Caswell*
3076 * Modify behavior of ALPN to invoke callback after SNI/servername
3077 callback, such that updates to the SSL_CTX affect ALPN.
3081 * Add SSL_CIPHER queries for authentication and key-exchange.
3085 * Changes to the DEFAULT cipherlist:
3086 - Prefer (EC)DHE handshakes over plain RSA.
3087 - Prefer AEAD ciphers over legacy ciphers.
3088 - Prefer ECDSA over RSA when both certificates are available.
3089 - Prefer TLSv1.2 ciphers/PRF.
3090 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3095 * Change the ECC default curve list to be this, in order: x25519,
3096 secp256r1, secp521r1, secp384r1.
3100 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3101 disabled by default. They can be re-enabled using the
3102 enable-weak-ssl-ciphers option to Configure.
3106 * If the server has ALPN configured, but supports no protocols that the
3107 client advertises, send a fatal "no_application_protocol" alert.
3108 This behaviour is SHALL in RFC 7301, though it isn't universally
3109 implemented by other servers.
3113 * Add X25519 support.
3114 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3115 for public and private key encoding using the format documented in
3116 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3117 key generation and key derivation.
3119 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3124 * Deprecate SRP_VBASE_get_by_user.
3125 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3126 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3127 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3128 seed, even if the seed is configured.
3130 Users should use SRP_VBASE_get1_by_user instead. Note that in
3131 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3132 also that even though configuring the SRP seed attempts to hide
3133 invalid usernames by continuing the handshake with fake
3134 credentials, this behaviour is not constant time and no strong
3135 guarantees are made that the handshake is indistinguishable from
3136 that of a valid user.
3140 * Configuration change; it's now possible to build dynamic engines
3141 without having to build shared libraries and vice versa. This
3142 only applies to the engines in `engines/`, those in `crypto/engine/`
3143 will always be built into libcrypto (i.e. "static").
3145 Building dynamic engines is enabled by default; to disable, use
3146 the configuration option "disable-dynamic-engine".
3148 The only requirements for building dynamic engines are the
3149 presence of the DSO module and building with position independent
3150 code, so they will also automatically be disabled if configuring
3151 with "disable-dso" or "disable-pic".
3153 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3154 are also taken away from openssl/opensslconf.h, as they are
3159 * Configuration change; if there is a known flag to compile
3160 position independent code, it will always be applied on the
3161 libcrypto and libssl object files, and never on the application
3162 object files. This means other libraries that use routines from
3163 libcrypto / libssl can be made into shared libraries regardless
3164 of how OpenSSL was configured.
3166 If this isn't desirable, the configuration options "disable-pic"
3167 or "no-pic" can be used to disable the use of PIC. This will
3168 also disable building shared libraries and dynamic engines.
3172 * Removed JPAKE code. It was experimental and has no wide use.
3176 * The INSTALL_PREFIX Makefile variable has been renamed to
3177 DESTDIR. That makes for less confusion on what this variable
3178 is for. Also, the configuration option --install_prefix is
3183 * Heartbeat for TLS has been removed and is disabled by default
3184 for DTLS; configure with enable-heartbeats. Code that uses the
3185 old #define's might need to be updated.
3187 *Emilia Käsper, Rich Salz*
3189 * Rename REF_CHECK to REF_DEBUG.
3193 * New "unified" build system
3195 The "unified" build system is aimed to be a common system for all
3196 platforms we support. With it comes new support for VMS.
3198 This system builds supports building in a different directory tree
3199 than the source tree. It produces one Makefile (for unix family
3200 or lookalikes), or one descrip.mms (for VMS).
3202 The source of information to make the Makefile / descrip.mms is
3203 small files called 'build.info', holding the necessary
3204 information for each directory with source to compile, and a
3205 template in Configurations, like unix-Makefile.tmpl or
3208 With this change, the library names were also renamed on Windows
3209 and on VMS. They now have names that are closer to the standard
3210 on Unix, and include the major version number, and in certain
3211 cases, the architecture they are built for. See "Notes on shared
3212 libraries" in INSTALL.
3214 We rely heavily on the perl module Text::Template.
3218 * Added support for auto-initialisation and de-initialisation of the library.
3219 OpenSSL no longer requires explicit init or deinit routines to be called,
3220 except in certain circumstances. See the OPENSSL_init_crypto() and
3221 OPENSSL_init_ssl() man pages for further information.
3225 * The arguments to the DTLSv1_listen function have changed. Specifically the
3226 "peer" argument is now expected to be a BIO_ADDR object.
3228 * Rewrite of BIO networking library. The BIO library lacked consistent
3229 support of IPv6, and adding it required some more extensive
3230 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3231 which hold all types of addresses and chains of address information.
3232 It also introduces a new API, with functions like BIO_socket,
3233 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3234 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3235 have been adapted accordingly.
3239 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3244 * CRIME protection: disable compression by default, even if OpenSSL is
3245 compiled with zlib enabled. Applications can still enable compression
3246 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3247 using the SSL_CONF library to configure compression.
3251 * The signature of the session callback configured with
3252 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3253 was explicitly marked as `const unsigned char*` instead of
3258 * Always DPURIFY. Remove the use of uninitialized memory in the
3259 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3263 * Removed many obsolete configuration items, including
3264 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3265 MD2_CHAR, MD2_INT, MD2_LONG
3267 IDEA_SHORT, IDEA_LONG
3268 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3270 *Rich Salz, with advice from Andy Polyakov*
3272 * Many BN internals have been moved to an internal header file.
3274 *Rich Salz with help from Andy Polyakov*
3276 * Configuration and writing out the results from it has changed.
3277 Files such as Makefile include/openssl/opensslconf.h and are now
3278 produced through general templates, such as Makefile.in and
3279 crypto/opensslconf.h.in and some help from the perl module
3282 Also, the center of configuration information is no longer
3283 Makefile. Instead, Configure produces a perl module in
3284 configdata.pm which holds most of the config data (in the hash
3285 table %config), the target data that comes from the target
3286 configuration in one of the `Configurations/*.conf` files (in
3291 * To clarify their intended purposes, the Configure options
3292 --prefix and --openssldir change their semantics, and become more
3293 straightforward and less interdependent.
3295 --prefix shall be used exclusively to give the location INSTALLTOP
3296 where programs, scripts, libraries, include files and manuals are
3297 going to be installed. The default is now /usr/local.
3299 --openssldir shall be used exclusively to give the default
3300 location OPENSSLDIR where certificates, private keys, CRLs are
3301 managed. This is also where the default openssl.cnf gets
3303 If the directory given with this option is a relative path, the
3304 values of both the --prefix value and the --openssldir value will
3305 be combined to become OPENSSLDIR.
3306 The default for --openssldir is INSTALLTOP/ssl.
3308 Anyone who uses --openssldir to specify where OpenSSL is to be
3309 installed MUST change to use --prefix instead.
3313 * The GOST engine was out of date and therefore it has been removed. An up
3314 to date GOST engine is now being maintained in an external repository.
3315 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3316 support for GOST ciphersuites (these are only activated if a GOST engine
3321 * EGD is no longer supported by default; use enable-egd when
3324 *Ben Kaduk and Rich Salz*
3326 * The distribution now has Makefile.in files, which are used to
3327 create Makefile's when Configure is run. *Configure must be run
3328 before trying to build now.*
3332 * The return value for SSL_CIPHER_description() for error conditions
3337 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3339 Obtaining and performing DNSSEC validation of TLSA records is
3340 the application's responsibility. The application provides
3341 the TLSA records of its choice to OpenSSL, and these are then
3342 used to authenticate the peer.
3344 The TLSA records need not even come from DNS. They can, for
3345 example, be used to implement local end-entity certificate or
3346 trust-anchor "pinning", where the "pin" data takes the form
3347 of TLSA records, which can augment or replace verification
3348 based on the usual WebPKI public certification authorities.
3352 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3353 continues to support deprecated interfaces in default builds.
3354 However, applications are strongly advised to compile their
3355 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3356 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3357 or the 1.1.0 releases.
3359 In environments in which all applications have been ported to
3360 not use any deprecated interfaces OpenSSL's Configure script
3361 should be used with the --api=1.1.0 option to entirely remove
3362 support for the deprecated features from the library and
3363 unconditionally disable them in the installed headers.
3364 Essentially the same effect can be achieved with the "no-deprecated"
3365 argument to Configure, except that this will always restrict
3366 the build to just the latest API, rather than a fixed API
3369 As applications are ported to future revisions of the API,
3370 they should update their compile-time OPENSSL_API_COMPAT define
3371 accordingly, but in most cases should be able to continue to
3372 compile with later releases.
3374 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3375 0x10000000L and 0x00908000L, respectively. However those
3376 versions did not support the OPENSSL_API_COMPAT feature, and
3377 so applications are not typically tested for explicit support
3378 of just the undeprecated features of either release.
3382 * Add support for setting the minimum and maximum supported protocol.
3383 It can bet set via the SSL_set_min_proto_version() and
3384 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3385 MaxProtocol. It's recommended to use the new APIs to disable
3386 protocols instead of disabling individual protocols using
3387 SSL_set_options() or SSL_CONF's Protocol. This change also
3388 removes support for disabling TLS 1.2 in the OpenSSL TLS
3389 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3393 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3397 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3398 and integrates ECDSA and ECDH functionality into EC. Implementations can
3399 now redirect key generation and no longer need to convert to or from
3402 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3403 include the ec.h header file instead.
3407 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3408 ciphers who are no longer supported and drops support the ephemeral RSA key
3409 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3413 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3414 opaque. For HMAC_CTX, the following constructors and destructors
3417 HMAC_CTX *HMAC_CTX_new(void);
3418 void HMAC_CTX_free(HMAC_CTX *ctx);
3420 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3421 destroy such methods has been added. See EVP_MD_meth_new(3) and
3422 EVP_CIPHER_meth_new(3) for documentation.
3425 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3426 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3427 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3428 an already created structure.
3429 2) For consistency with the majority of our object creators and
3430 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3431 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3432 for deprecated builds.
3436 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3437 cryptographic operations to be performed asynchronously as long as an
3438 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3439 further details. Libssl has also had this capability integrated with the
3440 introduction of the new mode SSL_MODE_ASYNC and associated error
3441 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3442 pages. This work was developed in partnership with Intel Corp.
3446 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3447 always enabled now. If you want to disable the support you should
3448 exclude it using the list of supported ciphers. This also means that the
3449 "-no_ecdhe" option has been removed from s_server.
3453 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3454 SSL_{CTX_}set1_curves() which can set a list.
3458 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3459 curve you want to support using SSL_{CTX_}set1_curves().
3463 * State machine rewrite. The state machine code has been significantly
3464 refactored in order to remove much duplication of code and solve issues
3465 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3466 further details). This change does have some associated API changes.
3467 Notably the SSL_state() function has been removed and replaced by
3468 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3469 SSL_set_state() has been removed altogether. The previous handshake states
3470 defined in ssl.h and ssl3.h have also been removed.
3474 * All instances of the string "ssleay" in the public API were replaced
3475 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3476 Some error codes related to internal RSA_eay API's were renamed.
3480 * The demo files in crypto/threads were moved to demo/threads.
3484 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3487 *Matt Caswell, Rich Salz*
3489 * New ASN.1 embed macro.
3491 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3492 structure is not allocated: it is part of the parent. That is instead of
3500 This reduces memory fragmentation and make it impossible to accidentally
3501 set a mandatory field to NULL.
3503 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3504 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3505 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3510 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3514 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3515 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3516 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3517 DES and RC4 ciphersuites.
3521 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3522 This changes the decoding behaviour for some invalid messages,
3523 though the change is mostly in the more lenient direction, and
3524 legacy behaviour is preserved as much as possible.
3528 * Fix no-stdio build.
3529 *David Woodhouse <David.Woodhouse@intel.com> and also*
3530 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3532 * New testing framework
3533 The testing framework has been largely rewritten and is now using
3534 perl and the perl modules Test::Harness and an extended variant of
3535 Test::More called OpenSSL::Test to do its work. All test scripts in
3536 test/ have been rewritten into test recipes, and all direct calls to
3537 executables in test/Makefile have become individual recipes using the
3538 simplified testing OpenSSL::Test::Simple.
3540 For documentation on our testing modules, do:
3542 perldoc test/testlib/OpenSSL/Test/Simple.pm
3543 perldoc test/testlib/OpenSSL/Test.pm
3547 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3548 are used; the latter aborts on memory leaks (usually checked on exit).
3549 Some undocumented "set malloc, etc., hooks" functions were removed
3550 and others were changed. All are now documented.
3554 * In DSA_generate_parameters_ex, if the provided seed is too short,
3557 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3559 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3560 from RFC4279, RFC4785, RFC5487, RFC5489.
3562 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3563 original RSA_PSK patch.
3567 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3568 era flag was never set throughout the codebase (only read). Also removed
3569 SSL3_FLAGS_POP_BUFFER which was only used if
3570 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3574 * Changed the default name options in the "ca", "crl", "req" and "x509"
3575 to be "oneline" instead of "compat".
3579 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3580 not aware of clients that still exhibit this bug, and the workaround
3581 hasn't been working properly for a while.
3585 * The return type of BIO_number_read() and BIO_number_written() as well as
3586 the corresponding num_read and num_write members in the BIO structure has
3587 changed from unsigned long to uint64_t. On platforms where an unsigned
3588 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3593 * Given the pervasive nature of TLS extensions it is inadvisable to run
3594 OpenSSL without support for them. It also means that maintaining
3595 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3596 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3600 * Removed support for the two export grade static DH ciphersuites
3601 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3602 were newly added (along with a number of other static DH ciphersuites) to
3603 1.0.2. However the two export ones have *never* worked since they were
3604 introduced. It seems strange in any case to be adding new export
3605 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3609 * Version negotiation has been rewritten. In particular SSLv23_method(),
3610 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3611 and turned into macros which simply call the new preferred function names
3612 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3613 should use the new names instead. Also as part of this change the ssl23.h
3614 header file has been removed.
3618 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3619 code and the associated standard is no longer considered fit-for-purpose.
3623 * RT2547 was closed. When generating a private key, try to make the
3624 output file readable only by the owner. This behavior change might
3625 be noticeable when interacting with other software.
3627 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3632 * Added HTTP GET support to the ocsp command.
3636 * Changed default digest for the dgst and enc commands from MD5 to
3641 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3645 * Added support for TLS extended master secret from
3646 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3647 initial patch which was a great help during development.
3651 * All libssl internal structures have been removed from the public header
3652 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3653 now redundant). Users should not attempt to access internal structures
3654 directly. Instead they should use the provided API functions.
3658 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3659 Access to deprecated functions can be re-enabled by running config with
3660 "enable-deprecated". In addition applications wishing to use deprecated
3661 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3662 will, by default, disable some transitive includes that previously existed
3663 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3667 * Added support for OCB mode. OpenSSL has been granted a patent license
3668 compatible with the OpenSSL license for use of OCB. Details are available
3669 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3670 for OCB can be removed by calling config with no-ocb.
3674 * SSLv2 support has been removed. It still supports receiving a SSLv2
3675 compatible client hello.
3679 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3680 done while fixing the error code for the key-too-small case.
3682 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3684 * CA.sh has been removed; use CA.pl instead.
3688 * Removed old DES API.
3692 * Remove various unsupported platforms:
3698 Sinix/ReliantUNIX RM400
3703 16-bit platforms such as WIN16
3707 * Clean up OPENSSL_NO_xxx #define's
3708 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3709 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3710 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3711 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3712 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3713 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3714 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3715 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3716 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3717 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3721 * Cleaned up dead code
3722 Remove all but one '#ifdef undef' which is to be looked at.
3726 * Clean up calling of xxx_free routines.
3727 Just like free(), fix most of the xxx_free routines to accept
3728 NULL. Remove the non-null checks from callers. Save much code.
3732 * Add secure heap for storage of private keys (when possible).
3733 Add BIO_s_secmem(), CBIGNUM, etc.
3734 Contributed by Akamai Technologies under our Corporate CLA.
3738 * Experimental support for a new, fast, unbiased prime candidate generator,
3739 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3741 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3743 * New output format NSS in the sess_id command line tool. This allows
3744 exporting the session id and the master key in NSS keylog format.
3746 *Martin Kaiser <martin@kaiser.cx>*
3748 * Harmonize version and its documentation. -f flag is used to display
3751 *mancha <mancha1@zoho.com>*
3753 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3754 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3756 *mancha <mancha1@zoho.com>*
3758 * Fix some double frees. These are not thought to be exploitable.
3760 *mancha <mancha1@zoho.com>*
3762 * A missing bounds check in the handling of the TLS heartbeat extension
3763 can be used to reveal up to 64k of memory to a connected client or
3766 Thanks for Neel Mehta of Google Security for discovering this bug and to
3767 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3768 preparing the fix ([CVE-2014-0160])
3770 *Adam Langley, Bodo Moeller*
3772 * Fix for the attack described in the paper "Recovering OpenSSL
3773 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3774 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3775 <http://eprint.iacr.org/2014/140>
3777 Thanks to Yuval Yarom and Naomi Benger for discovering this
3778 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3780 *Yuval Yarom and Naomi Benger*
3782 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3783 this fixes a limitation in previous versions of OpenSSL.
3787 * Experimental encrypt-then-mac support.
3789 Experimental support for encrypt then mac from
3790 draft-gutmann-tls-encrypt-then-mac-02.txt
3792 To enable it set the appropriate extension number (0x42 for the test
3793 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3795 For non-compliant peers (i.e. just about everything) this should have no
3798 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3802 * Add EVP support for key wrapping algorithms, to avoid problems with
3803 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3804 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3805 algorithms and include tests cases.
3809 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3814 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3815 MGF1 digest and OAEP label.
3819 * Make openssl verify return errors.
3821 *Chris Palmer <palmer@google.com> and Ben Laurie*
3823 * New function ASN1_TIME_diff to calculate the difference between two
3824 ASN1_TIME structures or one structure and the current time.
3828 * Update fips_test_suite to support multiple command line options. New
3829 test to induce all self test errors in sequence and check expected
3834 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3835 sign or verify all in one operation.
3839 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3840 test programs and fips_test_suite. Includes functionality to parse
3841 the minimal script output of fipsalgest.pl directly.
3845 * Add authorisation parameter to FIPS_module_mode_set().
3849 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3853 * Use separate DRBG fields for internal and external flags. New function
3854 FIPS_drbg_health_check() to perform on demand health checking. Add
3855 generation tests to fips_test_suite with reduced health check interval to
3856 demonstrate periodic health checking. Add "nodh" option to
3857 fips_test_suite to skip very slow DH test.
3861 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3866 * More extensive health check for DRBG checking many more failure modes.
3867 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3868 combination: call this in fips_test_suite.
3872 * Add support for canonical generation of DSA parameter 'g'. See
3875 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3876 POST to handle HMAC cases.
3880 * Add functions FIPS_module_version() and FIPS_module_version_text()
3881 to return numerical and string versions of the FIPS module number.
3885 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3886 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3887 outside the validated module in the FIPS capable OpenSSL.
3891 * Minor change to DRBG entropy callback semantics. In some cases
3892 there is no multiple of the block length between min_len and
3893 max_len. Allow the callback to return more than max_len bytes
3894 of entropy but discard any extra: it is the callback's responsibility
3895 to ensure that the extra data discarded does not impact the
3896 requested amount of entropy.
3900 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3901 information in FIPS186-3, SP800-57 and SP800-131A.
3905 * CCM support via EVP. Interface is very similar to GCM case except we
3906 must supply all data in one chunk (i.e. no update, final) and the
3907 message length must be supplied if AAD is used. Add algorithm test
3912 * Initial version of POST overhaul. Add POST callback to allow the status
3913 of POST to be monitored and/or failures induced. Modify fips_test_suite
3914 to use callback. Always run all selftests even if one fails.
3918 * XTS support including algorithm test driver in the fips_gcmtest program.
3919 Note: this does increase the maximum key length from 32 to 64 bytes but
3920 there should be no binary compatibility issues as existing applications
3921 will never use XTS mode.
3925 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3926 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3927 performs algorithm blocking for unapproved PRNG types. Also do not
3928 set PRNG type in FIPS_mode_set(): leave this to the application.
3929 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3930 the standard OpenSSL PRNG: set additional data to a date time vector.
3934 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3935 This shouldn't present any incompatibility problems because applications
3936 shouldn't be using these directly and any that are will need to rethink
3937 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3941 * Extensive self tests and health checking required by SP800-90 DRBG.
3942 Remove strength parameter from FIPS_drbg_instantiate and always
3943 instantiate at maximum supported strength.
3947 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3951 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3955 * New function DH_compute_key_padded() to compute a DH key and pad with
3956 leading zeroes if needed: this complies with SP800-56A et al.
3960 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3961 anything, incomplete, subject to change and largely untested at present.
3965 * Modify fipscanisteronly build option to only build the necessary object
3966 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3970 * Add experimental option FIPSSYMS to give all symbols in
3971 fipscanister.o and FIPS or fips prefix. This will avoid
3972 conflicts with future versions of OpenSSL. Add perl script
3973 util/fipsas.pl to preprocess assembly language source files
3974 and rename any affected symbols.
3978 * Add selftest checks and algorithm block of non-fips algorithms in
3979 FIPS mode. Remove DES2 from selftests.
3983 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3984 return internal method without any ENGINE dependencies. Add new
3985 tiny fips sign and verify functions.
3989 * New build option no-ec2m to disable characteristic 2 code.
3993 * New build option "fipscanisteronly". This only builds fipscanister.o
3994 and (currently) associated fips utilities. Uses the file Makefile.fips
3995 instead of Makefile.org as the prototype.
3999 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
4000 Update fips_gcmtest to use IV generator.
4004 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4005 setting output buffer to NULL. The `*Final` function must be
4006 called although it will not retrieve any additional data. The tag
4007 can be set or retrieved with a ctrl. The IV length is by default 12
4008 bytes (96 bits) but can be set to an alternative value. If the IV
4009 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4014 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4015 underlying do_cipher function handles all cipher semantics itself
4016 including padding and finalisation. This is useful if (for example)
4017 an ENGINE cipher handles block padding itself. The behaviour of
4018 do_cipher is subtly changed if this flag is set: the return value
4019 is the number of characters written to the output buffer (zero is
4020 no longer an error code) or a negative error code. Also if the
4021 input buffer is NULL and length 0 finalisation should be performed.
4025 * If a candidate issuer certificate is already part of the constructed
4026 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4030 * Improve forward-security support: add functions
4032 void SSL_CTX_set_not_resumable_session_callback(
4033 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4034 void SSL_set_not_resumable_session_callback(
4035 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4037 for use by SSL/TLS servers; the callback function will be called whenever a
4038 new session is created, and gets to decide whether the session may be
4039 cached to make it resumable (return 0) or not (return 1). (As by the
4040 SSL/TLS protocol specifications, the session_id sent by the server will be
4041 empty to indicate that the session is not resumable; also, the server will
4042 not generate RFC 4507 (RFC 5077) session tickets.)
4044 A simple reasonable callback implementation is to return is_forward_secure.
4045 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4046 by the SSL/TLS server library, indicating whether it can provide forward
4049 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4051 * New -verify_name option in command line utilities to set verification
4056 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4057 Add CMAC pkey methods.
4061 * Experimental renegotiation in s_server -www mode. If the client
4062 browses /reneg connection is renegotiated. If /renegcert it is
4063 renegotiated requesting a certificate.
4067 * Add an "external" session cache for debugging purposes to s_server. This
4068 should help trace issues which normally are only apparent in deployed
4069 multi-process servers.
4073 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4074 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4075 BIO_set_cipher() and some obscure PEM functions were changed so they
4076 can now return an error. The RAND changes required a change to the
4077 RAND_METHOD structure.
4081 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4082 a gcc attribute to warn if the result of a function is ignored. This
4083 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4084 whose return value is often ignored.
4088 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4089 These allow SCTs (signed certificate timestamps) to be requested and
4090 validated when establishing a connection.
4092 *Rob Percival <robpercival@google.com>*
4097 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4099 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4100 used even when parsing explicit parameters, when loading a encoded key
4101 or calling `EC_GROUP_new_from_ecpkparameters()`/
4102 `EC_GROUP_new_from_ecparameters()`.
4103 This prevents bypass of security hardening and performance gains,
4104 especially for curves with specialized EC_METHODs.
4105 By default, if a key encoded with explicit parameters is loaded and later
4106 encoded, the output is still encoded with explicit parameters, even if
4107 internally a "named" EC_GROUP is used for computation.
4111 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4112 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4113 NULL. After this change, only the cofactor parameter can be NULL. It also
4114 does some minimal sanity checks on the passed order.
4119 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4120 An attack is simple, if the first CMS_recipientInfo is valid but the
4121 second CMS_recipientInfo is chosen ciphertext. If the second
4122 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4123 encryption key will be replaced by garbage, and the message cannot be
4124 decoded, but if the RSA decryption fails, the correct encryption key is
4125 used and the recipient will not notice the attack.
4126 As a work around for this potential attack the length of the decrypted
4127 key must be equal to the cipher default key length, in case the
4128 certifiate is not given and all recipientInfo are tried out.
4129 The old behaviour can be re-enabled in the CMS code by setting the
4130 CMS_DEBUG_DECRYPT flag.
4135 * Document issue with installation paths in diverse Windows builds
4137 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4138 binaries and run-time config file.
4143 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4145 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4146 This changes the size when using the `genpkey` command when no size is given.
4147 It fixes an omission in earlier changes that changed all RSA, DSA and DH
4148 generation commands to use 2048 bits by default.
4152 * Add FIPS support for Android Arm 64-bit
4154 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4155 Module in Version 2.0.10. For some reason, the corresponding target
4156 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4157 built with FIPS support on Android Arm 64-bit. This omission has been
4160 *Matthias St. Pierre*
4162 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4164 * 0-byte record padding oracle
4166 If an application encounters a fatal protocol error and then calls
4167 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4168 then OpenSSL can respond differently to the calling application if a 0 byte
4169 record is received with invalid padding compared to if a 0 byte record is
4170 received with an invalid MAC. If the application then behaves differently
4171 based on that in a way that is detectable to the remote peer, then this
4172 amounts to a padding oracle that could be used to decrypt data.
4174 In order for this to be exploitable "non-stitched" ciphersuites must be in
4175 use. Stitched ciphersuites are optimised implementations of certain
4176 commonly used ciphersuites. Also the application must call SSL_shutdown()
4177 twice even if a protocol error has occurred (applications should not do
4178 this but some do anyway).
4180 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4181 Aviram, with additional investigation by Steven Collison and Andrew
4182 Hourselt. It was reported to OpenSSL on 10th December 2018.
4187 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4191 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4193 * Microarchitecture timing vulnerability in ECC scalar multiplication
4195 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4196 shown to be vulnerable to a microarchitecture timing side channel attack.
4197 An attacker with sufficient access to mount local timing attacks during
4198 ECDSA signature generation could recover the private key.
4200 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4201 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4207 * Timing vulnerability in DSA signature generation
4209 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4210 timing side channel attack. An attacker could use variations in the signing
4211 algorithm to recover the private key.
4213 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4218 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4219 Module, accidentally introduced while backporting security fixes from the
4220 development branch and hindering the use of ECC in FIPS mode.
4224 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4226 * Client DoS due to large DH parameter
4228 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4229 malicious server can send a very large prime value to the client. This will
4230 cause the client to spend an unreasonably long period of time generating a
4231 key for this prime resulting in a hang until the client has finished. This
4232 could be exploited in a Denial Of Service attack.
4234 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4239 * Cache timing vulnerability in RSA Key Generation
4241 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4242 a cache timing side channel attack. An attacker with sufficient access to
4243 mount cache timing attacks during the RSA key generation process could
4244 recover the private key.
4246 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4247 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4252 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4253 parameter is no longer accepted, as it leads to a corrupt table. NULL
4254 pem_str is reserved for alias entries only.
4258 * Revert blinding in ECDSA sign and instead make problematic addition
4259 length-invariant. Switch even to fixed-length Montgomery multiplication.
4263 * Change generating and checking of primes so that the error rate of not
4264 being prime depends on the intended use based on the size of the input.
4265 For larger primes this will result in more rounds of Miller-Rabin.
4266 The maximal error rate for primes with more than 1080 bits is lowered
4269 *Kurt Roeckx, Annie Yousar*
4271 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4275 * Add blinding to ECDSA and DSA signatures to protect against side channel
4276 attacks discovered by Keegan Ryan (NCC Group).
4280 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4281 now allow empty (zero character) pass phrases.
4285 * Certificate time validation (X509_cmp_time) enforces stricter
4286 compliance with RFC 5280. Fractional seconds and timezone offsets
4287 are no longer allowed.
4291 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4293 * Constructed ASN.1 types with a recursive definition could exceed the stack
4295 Constructed ASN.1 types with a recursive definition (such as can be found
4296 in PKCS7) could eventually exceed the stack given malicious input with
4297 excessive recursion. This could result in a Denial Of Service attack. There
4298 are no such structures used within SSL/TLS that come from untrusted sources
4299 so this is considered safe.
4301 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4307 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4309 * Read/write after SSL object in error state
4311 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4312 mechanism. The intent was that if a fatal error occurred during a handshake
4313 then OpenSSL would move into the error state and would immediately fail if
4314 you attempted to continue the handshake. This works as designed for the
4315 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4316 SSL_connect()), however due to a bug it does not work correctly if
4317 SSL_read() or SSL_write() is called directly. In that scenario, if the
4318 handshake fails then a fatal error will be returned in the initial function
4319 call. If SSL_read()/SSL_write() is subsequently called by the application
4320 for the same SSL object then it will succeed and the data is passed without
4321 being decrypted/encrypted directly from the SSL/TLS record layer.
4323 In order to exploit this issue an application bug would have to be present
4324 that resulted in a call to SSL_read()/SSL_write() being issued after having
4325 already received a fatal error.
4327 This issue was reported to OpenSSL by David Benjamin (Google).
4332 * rsaz_1024_mul_avx2 overflow bug on x86_64
4334 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4335 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4336 Analysis suggests that attacks against RSA and DSA as a result of this
4337 defect would be very difficult to perform and are not believed likely.
4338 Attacks against DH1024 are considered just feasible, because most of the
4339 work necessary to deduce information about a private key may be performed
4340 offline. The amount of resources required for such an attack would be
4341 significant. However, for an attack on TLS to be meaningful, the server
4342 would have to share the DH1024 private key among multiple clients, which is
4343 no longer an option since CVE-2016-0701.
4345 This only affects processors that support the AVX2 but not ADX extensions
4346 like Intel Haswell (4th generation).
4348 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4349 was originally found via the OSS-Fuzz project.
4354 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4356 * bn_sqrx8x_internal carry bug on x86_64
4358 There is a carry propagating bug in the x86_64 Montgomery squaring
4359 procedure. No EC algorithms are affected. Analysis suggests that attacks
4360 against RSA and DSA as a result of this defect would be very difficult to
4361 perform and are not believed likely. Attacks against DH are considered just
4362 feasible (although very difficult) because most of the work necessary to
4363 deduce information about a private key may be performed offline. The amount
4364 of resources required for such an attack would be very significant and
4365 likely only accessible to a limited number of attackers. An attacker would
4366 additionally need online access to an unpatched system using the target
4367 private key in a scenario with persistent DH parameters and a private
4368 key that is shared between multiple clients.
4370 This only affects processors that support the BMI1, BMI2 and ADX extensions
4371 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4373 This issue was reported to OpenSSL by the OSS-Fuzz project.
4378 * Malformed X.509 IPAddressFamily could cause OOB read
4380 If an X.509 certificate has a malformed IPAddressFamily extension,
4381 OpenSSL could do a one-byte buffer overread. The most likely result
4382 would be an erroneous display of the certificate in text format.
4384 This issue was reported to OpenSSL by the OSS-Fuzz project.
4389 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4391 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4392 platform rather than 'mingw'.
4396 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4398 * Truncated packet could crash via OOB read
4400 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4401 cipher is being used, then a truncated packet can cause that host to
4402 perform an out-of-bounds read, usually resulting in a crash.
4404 This issue was reported to OpenSSL by Robert Święcki of Google.
4409 * BN_mod_exp may produce incorrect results on x86_64
4411 There is a carry propagating bug in the x86_64 Montgomery squaring
4412 procedure. No EC algorithms are affected. Analysis suggests that attacks
4413 against RSA and DSA as a result of this defect would be very difficult to
4414 perform and are not believed likely. Attacks against DH are considered just
4415 feasible (although very difficult) because most of the work necessary to
4416 deduce information about a private key may be performed offline. The amount
4417 of resources required for such an attack would be very significant and
4418 likely only accessible to a limited number of attackers. An attacker would
4419 additionally need online access to an unpatched system using the target
4420 private key in a scenario with persistent DH parameters and a private
4421 key that is shared between multiple clients. For example this can occur by
4422 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4423 similar to CVE-2015-3193 but must be treated as a separate problem.
4425 This issue was reported to OpenSSL by the OSS-Fuzz project.
4430 * Montgomery multiplication may produce incorrect results
4432 There is a carry propagating bug in the Broadwell-specific Montgomery
4433 multiplication procedure that handles input lengths divisible by, but
4434 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4435 and DH private keys are impossible. This is because the subroutine in
4436 question is not used in operations with the private key itself and an input
4437 of the attacker's direct choice. Otherwise the bug can manifest itself as
4438 transient authentication and key negotiation failures or reproducible
4439 erroneous outcome of public-key operations with specially crafted input.
4440 Among EC algorithms only Brainpool P-512 curves are affected and one
4441 presumably can attack ECDH key negotiation. Impact was not analyzed in
4442 detail, because pre-requisites for attack are considered unlikely. Namely
4443 multiple clients have to choose the curve in question and the server has to
4444 share the private key among them, neither of which is default behaviour.
4445 Even then only clients that chose the curve will be affected.
4447 This issue was publicly reported as transient failures and was not
4448 initially recognized as a security issue. Thanks to Richard Morgan for
4449 providing reproducible case.
4454 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4455 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4456 prevent issues where no progress is being made and the peer continually
4457 sends unrecognised record types, using up resources processing them.
4461 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4463 * Missing CRL sanity check
4465 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4466 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4467 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4469 This issue only affects the OpenSSL 1.0.2i
4474 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4476 * OCSP Status Request extension unbounded memory growth
4478 A malicious client can send an excessively large OCSP Status Request
4479 extension. If that client continually requests renegotiation, sending a
4480 large OCSP Status Request extension each time, then there will be unbounded
4481 memory growth on the server. This will eventually lead to a Denial Of
4482 Service attack through memory exhaustion. Servers with a default
4483 configuration are vulnerable even if they do not support OCSP. Builds using
4484 the "no-ocsp" build time option are not affected.
4486 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4491 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4494 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4500 * OOB write in MDC2_Update()
4502 An overflow can occur in MDC2_Update() either if called directly or
4503 through the EVP_DigestUpdate() function using MDC2. If an attacker
4504 is able to supply very large amounts of input data after a previous
4505 call to EVP_EncryptUpdate() with a partial block then a length check
4506 can overflow resulting in a heap corruption.
4508 The amount of data needed is comparable to SIZE_MAX which is impractical
4511 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4516 * Malformed SHA512 ticket DoS
4518 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4519 DoS attack where a malformed ticket will result in an OOB read which will
4522 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4523 a custom server callback and ticket lookup mechanism.
4525 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4530 * OOB write in BN_bn2dec()
4532 The function BN_bn2dec() does not check the return value of BN_div_word().
4533 This can cause an OOB write if an application uses this function with an
4534 overly large BIGNUM. This could be a problem if an overly large certificate
4535 or CRL is printed out from an untrusted source. TLS is not affected because
4536 record limits will reject an oversized certificate before it is parsed.
4538 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4543 * OOB read in TS_OBJ_print_bio()
4545 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4546 the total length the OID text representation would use and not the amount
4547 of data written. This will result in OOB reads when large OIDs are
4550 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4555 * Pointer arithmetic undefined behaviour
4557 Avoid some undefined pointer arithmetic
4559 A common idiom in the codebase is to check limits in the following manner:
4562 Where "p" points to some malloc'd data of SIZE bytes and
4565 "len" here could be from some externally supplied data (e.g. from a TLS
4568 The rules of C pointer arithmetic are such that "p + len" is only well
4569 defined where len <= SIZE. Therefore the above idiom is actually
4570 undefined behaviour.
4572 For example this could cause problems if some malloc implementation
4573 provides an address for "p" such that "p + len" actually overflows for
4574 values of len that are too big and therefore p + len < limit.
4576 This issue was reported to OpenSSL by Guido Vranken
4581 * Constant time flag not preserved in DSA signing
4583 Operations in the DSA signing algorithm should run in constant time in
4584 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4585 implementation means that a non-constant time codepath is followed for
4586 certain operations. This has been demonstrated through a cache-timing
4587 attack to be sufficient for an attacker to recover the private DSA key.
4589 This issue was reported by César Pereida (Aalto University), Billy Brumley
4590 (Tampere University of Technology), and Yuval Yarom (The University of
4591 Adelaide and NICTA).
4596 * DTLS buffered message DoS
4598 In a DTLS connection where handshake messages are delivered out-of-order
4599 those messages that OpenSSL is not yet ready to process will be buffered
4600 for later use. Under certain circumstances, a flaw in the logic means that
4601 those messages do not get removed from the buffer even though the handshake
4602 has been completed. An attacker could force up to approx. 15 messages to
4603 remain in the buffer when they are no longer required. These messages will
4604 be cleared when the DTLS connection is closed. The default maximum size for
4605 a message is 100k. Therefore the attacker could force an additional 1500k
4606 to be consumed per connection. By opening many simulataneous connections an
4607 attacker could cause a DoS attack through memory exhaustion.
4609 This issue was reported to OpenSSL by Quan Luo.
4614 * DTLS replay protection DoS
4616 A flaw in the DTLS replay attack protection mechanism means that records
4617 that arrive for future epochs update the replay protection "window" before
4618 the MAC for the record has been validated. This could be exploited by an
4619 attacker by sending a record for the next epoch (which does not have to
4620 decrypt or have a valid MAC), with a very large sequence number. This means
4621 that all subsequent legitimate packets are dropped causing a denial of
4622 service for a specific DTLS connection.
4624 This issue was reported to OpenSSL by the OCAP audit team.
4629 * Certificate message OOB reads
4631 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4632 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4633 theoretical DoS risk but this has not been observed in practice on common
4636 The messages affected are client certificate, client certificate request
4637 and server certificate. As a result the attack can only be performed
4638 against a client or a server which enables client authentication.
4640 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4645 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4647 * Prevent padding oracle in AES-NI CBC MAC check
4649 A MITM attacker can use a padding oracle attack to decrypt traffic
4650 when the connection uses an AES CBC cipher and the server support
4653 This issue was introduced as part of the fix for Lucky 13 padding
4654 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4655 constant time by making sure that always the same bytes are read and
4656 compared against either the MAC or padding bytes. But it no longer
4657 checked that there was enough data to have both the MAC and padding
4660 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4665 * Fix EVP_EncodeUpdate overflow
4667 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4668 Base64 encoding of binary data. If an attacker is able to supply very large
4669 amounts of input data then a length check can overflow resulting in a heap
4672 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4673 the `PEM_write_bio*` family of functions. These are mainly used within the
4674 OpenSSL command line applications, so any application which processes data
4675 from an untrusted source and outputs it as a PEM file should be considered
4676 vulnerable to this issue. User applications that call these APIs directly
4677 with large amounts of untrusted data may also be vulnerable.
4679 This issue was reported by Guido Vranken.
4684 * Fix EVP_EncryptUpdate overflow
4686 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4687 is able to supply very large amounts of input data after a previous call to
4688 EVP_EncryptUpdate() with a partial block then a length check can overflow
4689 resulting in a heap corruption. Following an analysis of all OpenSSL
4690 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4691 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4692 the first called function after an EVP_EncryptInit(), and therefore that
4693 specific call must be safe. The second form is where the length passed to
4694 EVP_EncryptUpdate() can be seen from the code to be some small value and
4695 therefore there is no possibility of an overflow. Since all instances are
4696 one of these two forms, it is believed that there can be no overflows in
4697 internal code due to this problem. It should be noted that
4698 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4699 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4700 of these calls have also been analysed too and it is believed there are no
4701 instances in internal usage where an overflow could occur.
4703 This issue was reported by Guido Vranken.
4708 * Prevent ASN.1 BIO excessive memory allocation
4710 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4711 a short invalid encoding can cause allocation of large amounts of memory
4712 potentially consuming excessive resources or exhausting memory.
4714 Any application parsing untrusted data through d2i BIO functions is
4715 affected. The memory based functions such as d2i_X509() are *not* affected.
4716 Since the memory based functions are used by the TLS library, TLS
4717 applications are not affected.
4719 This issue was reported by Brian Carpenter.
4726 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4727 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4728 in arbitrary stack data being returned in the buffer.
4730 This issue was reported by Guido Vranken.
4735 * Modify behavior of ALPN to invoke callback after SNI/servername
4736 callback, such that updates to the SSL_CTX affect ALPN.
4740 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4745 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4746 methods are enabled and ssl2 is disabled the methods return NULL.
4750 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4752 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4753 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4754 provide any "EXPORT" or "LOW" strength ciphers.
4758 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4759 is by default disabled at build-time. Builds that are not configured with
4760 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4761 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4762 will need to explicitly call either of:
4764 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4766 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4768 as appropriate. Even if either of those is used, or the application
4769 explicitly uses the version-specific SSLv2_method() or its client and
4770 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4771 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4772 ciphers, and SSLv2 56-bit DES are no longer available.
4777 * Fix a double-free in DSA code
4779 A double free bug was discovered when OpenSSL parses malformed DSA private
4780 keys and could lead to a DoS attack or memory corruption for applications
4781 that receive DSA private keys from untrusted sources. This scenario is
4784 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4790 * Disable SRP fake user seed to address a server memory leak.
4792 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4794 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4795 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4796 was changed to ignore the "fake user" SRP seed, even if the seed
4799 Users should use SRP_VBASE_get1_by_user instead. Note that in
4800 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4801 also that even though configuring the SRP seed attempts to hide
4802 invalid usernames by continuing the handshake with fake
4803 credentials, this behaviour is not constant time and no strong
4804 guarantees are made that the handshake is indistinguishable from
4805 that of a valid user.
4810 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4812 In the BN_hex2bn function the number of hex digits is calculated using an
4813 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4814 large values of `i` this can result in `bn_expand` not allocating any
4815 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4816 field as NULL leading to a subsequent NULL ptr deref. For very large values
4817 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4818 In this case memory is allocated to the internal BIGNUM data field, but it
4819 is insufficiently sized leading to heap corruption. A similar issue exists
4820 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4821 is ever called by user applications with very large untrusted hex/dec data.
4822 This is anticipated to be a rare occurrence.
4824 All OpenSSL internal usage of these functions use data that is not expected
4825 to be untrusted, e.g. config file data or application command line
4826 arguments. If user developed applications generate config file data based
4827 on untrusted data then it is possible that this could also lead to security
4828 consequences. This is also anticipated to be rare.
4830 This issue was reported to OpenSSL by Guido Vranken.
4835 * Fix memory issues in `BIO_*printf` functions
4837 The internal `fmtstr` function used in processing a "%s" format string in
4838 the `BIO_*printf` functions could overflow while calculating the length of a
4839 string and cause an OOB read when printing very long strings.
4841 Additionally the internal `doapr_outch` function can attempt to write to an
4842 OOB memory location (at an offset from the NULL pointer) in the event of a
4843 memory allocation failure. In 1.0.2 and below this could be caused where
4844 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4845 could be in processing a very long "%s" format string. Memory leaks can
4848 The first issue may mask the second issue dependent on compiler behaviour.
4849 These problems could enable attacks where large amounts of untrusted data
4850 is passed to the `BIO_*printf` functions. If applications use these functions
4851 in this way then they could be vulnerable. OpenSSL itself uses these
4852 functions when printing out human-readable dumps of ASN.1 data. Therefore
4853 applications that print this data could be vulnerable if the data is from
4854 untrusted sources. OpenSSL command line applications could also be
4855 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4856 as command line arguments.
4858 Libssl is not considered directly vulnerable. Additionally certificates etc
4859 received via remote connections via libssl are also unlikely to be able to
4860 trigger these issues because of message size limits enforced within libssl.
4862 This issue was reported to OpenSSL Guido Vranken.
4867 * Side channel attack on modular exponentiation
4869 A side-channel attack was found which makes use of cache-bank conflicts on
4870 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4871 of RSA keys. The ability to exploit this issue is limited as it relies on
4872 an attacker who has control of code in a thread running on the same
4873 hyper-threaded core as the victim thread which is performing decryptions.
4875 This issue was reported to OpenSSL by Yuval Yarom, The University of
4876 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4877 Nadia Heninger, University of Pennsylvania with more information at
4878 <http://cachebleed.info>.
4883 * Change the `req` command to generate a 2048-bit RSA/DSA key by default,
4884 if no keysize is specified with default_bits. This fixes an
4885 omission in an earlier change that changed all RSA/DSA key generation
4886 commands to use 2048 bits by default.
4890 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4892 * DH small subgroups
4894 Historically OpenSSL only ever generated DH parameters based on "safe"
4895 primes. More recently (in version 1.0.2) support was provided for
4896 generating X9.42 style parameter files such as those required for RFC 5114
4897 support. The primes used in such files may not be "safe". Where an
4898 application is using DH configured with parameters based on primes that are
4899 not "safe" then an attacker could use this fact to find a peer's private
4900 DH exponent. This attack requires that the attacker complete multiple
4901 handshakes in which the peer uses the same private DH exponent. For example
4902 this could be used to discover a TLS server's private DH exponent if it's
4903 reusing the private DH exponent or it's using a static DH ciphersuite.
4905 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4906 TLS. It is not on by default. If the option is not set then the server
4907 reuses the same private DH exponent for the life of the server process and
4908 would be vulnerable to this attack. It is believed that many popular
4909 applications do set this option and would therefore not be at risk.
4911 The fix for this issue adds an additional check where a "q" parameter is
4912 available (as is the case in X9.42 based parameters). This detects the
4913 only known attack, and is the only possible defense for static DH
4914 ciphersuites. This could have some performance impact.
4916 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4917 default and cannot be disabled. This could have some performance impact.
4919 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4924 * SSLv2 doesn't block disabled ciphers
4926 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4927 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4928 been disabled, provided that the SSLv2 protocol was not also disabled via
4931 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4932 and Sebastian Schinzel.
4937 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4939 * BN_mod_exp may produce incorrect results on x86_64
4941 There is a carry propagating bug in the x86_64 Montgomery squaring
4942 procedure. No EC algorithms are affected. Analysis suggests that attacks
4943 against RSA and DSA as a result of this defect would be very difficult to
4944 perform and are not believed likely. Attacks against DH are considered just
4945 feasible (although very difficult) because most of the work necessary to
4946 deduce information about a private key may be performed offline. The amount
4947 of resources required for such an attack would be very significant and
4948 likely only accessible to a limited number of attackers. An attacker would
4949 additionally need online access to an unpatched system using the target
4950 private key in a scenario with persistent DH parameters and a private
4951 key that is shared between multiple clients. For example this can occur by
4952 default in OpenSSL DHE based SSL/TLS ciphersuites.
4954 This issue was reported to OpenSSL by Hanno Böck.
4959 * Certificate verify crash with missing PSS parameter
4961 The signature verification routines will crash with a NULL pointer
4962 dereference if presented with an ASN.1 signature using the RSA PSS
4963 algorithm and absent mask generation function parameter. Since these
4964 routines are used to verify certificate signature algorithms this can be
4965 used to crash any certificate verification operation and exploited in a
4966 DoS attack. Any application which performs certificate verification is
4967 vulnerable including OpenSSL clients and servers which enable client
4970 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4975 * X509_ATTRIBUTE memory leak
4977 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4978 memory. This structure is used by the PKCS#7 and CMS routines so any
4979 application which reads PKCS#7 or CMS data from untrusted sources is
4980 affected. SSL/TLS is not affected.
4982 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4988 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4989 This changes the decoding behaviour for some invalid messages,
4990 though the change is mostly in the more lenient direction, and
4991 legacy behaviour is preserved as much as possible.
4995 * In DSA_generate_parameters_ex, if the provided seed is too short,
4998 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
5000 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
5002 * Alternate chains certificate forgery
5004 During certificate verification, OpenSSL will attempt to find an
5005 alternative certificate chain if the first attempt to build such a chain
5006 fails. An error in the implementation of this logic can mean that an
5007 attacker could cause certain checks on untrusted certificates to be
5008 bypassed, such as the CA flag, enabling them to use a valid leaf
5009 certificate to act as a CA and "issue" an invalid certificate.
5011 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5016 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5018 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5019 incompatibility in the handling of HMAC. The previous ABI has now been
5024 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5026 * Malformed ECParameters causes infinite loop
5028 When processing an ECParameters structure OpenSSL enters an infinite loop
5029 if the curve specified is over a specially malformed binary polynomial
5032 This can be used to perform denial of service against any
5033 system which processes public keys, certificate requests or
5034 certificates. This includes TLS clients and TLS servers with
5035 client authentication enabled.
5037 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5042 * Exploitable out-of-bounds read in X509_cmp_time
5044 X509_cmp_time does not properly check the length of the ASN1_TIME
5045 string and can read a few bytes out of bounds. In addition,
5046 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5049 An attacker can use this to craft malformed certificates and CRLs of
5050 various sizes and potentially cause a segmentation fault, resulting in
5051 a DoS on applications that verify certificates or CRLs. TLS clients
5052 that verify CRLs are affected. TLS clients and servers with client
5053 authentication enabled may be affected if they use custom verification
5056 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5057 independently by Hanno Böck.
5062 * PKCS7 crash with missing EnvelopedContent
5064 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5065 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5066 with missing content and trigger a NULL pointer dereference on parsing.
5068 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5069 structures from untrusted sources are affected. OpenSSL clients and
5070 servers are not affected.
5072 This issue was reported to OpenSSL by Michal Zalewski (Google).
5077 * CMS verify infinite loop with unknown hash function
5079 When verifying a signedData message the CMS code can enter an infinite loop
5080 if presented with an unknown hash function OID. This can be used to perform
5081 denial of service against any system which verifies signedData messages using
5083 This issue was reported to OpenSSL by Johannes Bauer.
5088 * Race condition handling NewSessionTicket
5090 If a NewSessionTicket is received by a multi-threaded client when attempting to
5091 reuse a previous ticket then a race condition can occur potentially leading to
5092 a double free of the ticket data.
5097 * Only support 256-bit or stronger elliptic curves with the
5098 'ecdh_auto' setting (server) or by default (client). Of supported
5099 curves, prefer P-256 (both).
5103 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5105 * ClientHello sigalgs DoS fix
5107 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5108 invalid signature algorithms extension a NULL pointer dereference will
5109 occur. This can be exploited in a DoS attack against the server.
5111 This issue was was reported to OpenSSL by David Ramos of Stanford
5115 *Stephen Henson and Matt Caswell*
5117 * Multiblock corrupted pointer fix
5119 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5120 feature only applies on 64 bit x86 architecture platforms that support AES
5121 NI instructions. A defect in the implementation of "multiblock" can cause
5122 OpenSSL's internal write buffer to become incorrectly set to NULL when
5123 using non-blocking IO. Typically, when the user application is using a
5124 socket BIO for writing, this will only result in a failed connection.
5125 However if some other BIO is used then it is likely that a segmentation
5126 fault will be triggered, thus enabling a potential DoS attack.
5128 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5133 * Segmentation fault in DTLSv1_listen fix
5135 The DTLSv1_listen function is intended to be stateless and processes the
5136 initial ClientHello from many peers. It is common for user code to loop
5137 over the call to DTLSv1_listen until a valid ClientHello is received with
5138 an associated cookie. A defect in the implementation of DTLSv1_listen means
5139 that state is preserved in the SSL object from one invocation to the next
5140 that can lead to a segmentation fault. Errors processing the initial
5141 ClientHello can trigger this scenario. An example of such an error could be
5142 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5145 This issue was reported to OpenSSL by Per Allansson.
5150 * Segmentation fault in ASN1_TYPE_cmp fix
5152 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5153 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5154 certificate signature algorithm consistency this can be used to crash any
5155 certificate verification operation and exploited in a DoS attack. Any
5156 application which performs certificate verification is vulnerable including
5157 OpenSSL clients and servers which enable client authentication.
5162 * Segmentation fault for invalid PSS parameters fix
5164 The signature verification routines will crash with a NULL pointer
5165 dereference if presented with an ASN.1 signature using the RSA PSS
5166 algorithm and invalid parameters. Since these routines are used to verify
5167 certificate signature algorithms this can be used to crash any
5168 certificate verification operation and exploited in a DoS attack. Any
5169 application which performs certificate verification is vulnerable including
5170 OpenSSL clients and servers which enable client authentication.
5172 This issue was was reported to OpenSSL by Brian Carpenter.
5177 * ASN.1 structure reuse memory corruption fix
5179 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5180 memory corruption via an invalid write. Such reuse is and has been
5181 strongly discouraged and is believed to be rare.
5183 Applications that parse structures containing CHOICE or ANY DEFINED BY
5184 components may be affected. Certificate parsing (d2i_X509 and related
5185 functions) are however not affected. OpenSSL clients and servers are
5191 * PKCS7 NULL pointer dereferences fix
5193 The PKCS#7 parsing code does not handle missing outer ContentInfo
5194 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5195 missing content and trigger a NULL pointer dereference on parsing.
5197 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5198 otherwise parse PKCS#7 structures from untrusted sources are
5199 affected. OpenSSL clients and servers are not affected.
5201 This issue was reported to OpenSSL by Michal Zalewski (Google).
5206 * DoS via reachable assert in SSLv2 servers fix
5208 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5209 servers that both support SSLv2 and enable export cipher suites by sending
5210 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5212 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5213 (OpenSSL development team).
5218 * Empty CKE with client auth and DHE fix
5220 If client auth is used then a server can seg fault in the event of a DHE
5221 ciphersuite being selected and a zero length ClientKeyExchange message
5222 being sent by the client. This could be exploited in a DoS attack.
5227 * Handshake with unseeded PRNG fix
5229 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5230 with an unseeded PRNG. The conditions are:
5231 - The client is on a platform where the PRNG has not been seeded
5232 automatically, and the user has not seeded manually
5233 - A protocol specific client method version has been used (i.e. not
5234 SSL_client_methodv23)
5235 - A ciphersuite is used that does not require additional random data from
5236 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5238 If the handshake succeeds then the client random that has been used will
5239 have been generated from a PRNG with insufficient entropy and therefore the
5240 output may be predictable.
5242 For example using the following command with an unseeded openssl will
5243 succeed on an unpatched platform:
5245 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5250 * Use After Free following d2i_ECPrivatekey error fix
5252 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5253 could cause a use after free condition. This, in turn, could cause a double
5254 free in several private key parsing functions (such as d2i_PrivateKey
5255 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5256 for applications that receive EC private keys from untrusted
5257 sources. This scenario is considered rare.
5259 This issue was discovered by the BoringSSL project and fixed in their
5265 * X509_to_X509_REQ NULL pointer deref fix
5267 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5268 the certificate key is invalid. This function is rarely used in practice.
5270 This issue was discovered by Brian Carpenter.
5275 * Removed the export ciphers from the DEFAULT ciphers
5279 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5281 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5282 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5283 So far those who have to target multiple platforms would compromise
5284 and argue that binary targeting say ARMv5 would still execute on
5285 ARMv8. "Universal" build resolves this compromise by providing
5286 near-optimal performance even on newer platforms.
5290 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5291 (other platforms pending).
5293 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5295 * Add support for the SignedCertificateTimestampList certificate and
5296 OCSP response extensions from RFC6962.
5300 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5301 for corner cases. (Certain input points at infinity could lead to
5302 bogus results, with non-infinity inputs mapped to infinity too.)
5306 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5307 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5308 common cases are optimized and there still is room for further
5309 improvements. Vector Permutation AES for Altivec is also added.
5313 * Add support for little-endian ppc64 Linux target.
5315 *Marcelo Cerri (IBM)*
5317 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5318 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5319 are optimized and there still is room for further improvements.
5320 Both 32- and 64-bit modes are supported.
5322 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5324 * Improved ARMv7 NEON support.
5328 * Support for SPARC Architecture 2011 crypto extensions, first
5329 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5330 SHA256/512, MD5, GHASH and modular exponentiation.
5332 *Andy Polyakov, David Miller*
5334 * Accelerated modular exponentiation for Intel processors, a.k.a.
5337 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5339 * Support for new and upcoming Intel processors, including AVX2,
5340 BMI and SHA ISA extensions. This includes additional "stitched"
5341 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5344 This work was sponsored by Intel Corp.
5348 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5349 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5350 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5354 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5355 this fixes a limitation in previous versions of OpenSSL.
5359 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5360 MGF1 digest and OAEP label.
5364 * Add EVP support for key wrapping algorithms, to avoid problems with
5365 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5366 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5367 algorithms and include tests cases.
5371 * Add functions to allocate and set the fields of an ECDSA_METHOD
5374 *Douglas E. Engert, Steve Henson*
5376 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5377 difference in days and seconds between two tm or ASN1_TIME structures.
5381 * Add -rev test option to s_server to just reverse order of characters
5382 received by client and send back to server. Also prints an abbreviated
5383 summary of the connection parameters.
5387 * New option -brief for s_client and s_server to print out a brief summary
5388 of connection parameters.
5392 * Add callbacks for arbitrary TLS extensions.
5394 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5396 * New option -crl_download in several openssl utilities to download CRLs
5397 from CRLDP extension in certificates.
5401 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5405 * New function X509_CRL_diff to generate a delta CRL from the difference
5406 of two full CRLs. Add support to "crl" utility.
5410 * New functions to set lookup_crls function and to retrieve
5411 X509_STORE from X509_STORE_CTX.
5415 * Print out deprecated issuer and subject unique ID fields in
5420 * Extend OCSP I/O functions so they can be used for simple general purpose
5421 HTTP as well as OCSP. New wrapper function which can be used to download
5422 CRLs using the OCSP API.
5426 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5430 * `SSL_CONF*` functions. These provide a common framework for application
5431 configuration using configuration files or command lines.
5435 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5436 message callback and prints the results. Needs compile time option
5437 "enable-ssl-trace". New options to s_client and s_server to enable
5442 * New ctrl and macro to retrieve supported points extensions.
5443 Print out extension in s_server and s_client.
5447 * New functions to retrieve certificate signature and signature
5452 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5457 * New Suite B modes for TLS code. These use and enforce the requirements
5458 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5459 only use Suite B curves. The Suite B modes can be set by using the
5460 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5464 * New chain verification flags for Suite B levels of security. Check
5465 algorithms are acceptable when flags are set in X509_verify_cert.
5469 * Make tls1_check_chain return a set of flags indicating checks passed
5470 by a certificate chain. Add additional tests to handle client
5471 certificates: checks for matching certificate type and issuer name
5476 * If an attempt is made to use a signature algorithm not in the peer
5477 preference list abort the handshake. If client has no suitable
5478 signature algorithms in response to a certificate request do not
5479 use the certificate.
5483 * If server EC tmp key is not in client preference list abort handshake.
5487 * Add support for certificate stores in CERT structure. This makes it
5488 possible to have different stores per SSL structure or one store in
5489 the parent SSL_CTX. Include distinct stores for certificate chain
5490 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5491 to build and store a certificate chain in CERT structure: returning
5492 an error if the chain cannot be built: this will allow applications
5493 to test if a chain is correctly configured.
5495 Note: if the CERT based stores are not set then the parent SSL_CTX
5496 store is used to retain compatibility with existing behaviour.
5500 * New function ssl_set_client_disabled to set a ciphersuite disabled
5501 mask based on the current session, check mask when sending client
5502 hello and checking the requested ciphersuite.
5506 * New ctrls to retrieve and set certificate types in a certificate
5507 request message. Print out received values in s_client. If certificate
5508 types is not set with custom values set sensible values based on
5509 supported signature algorithms.
5513 * Support for distinct client and server supported signature algorithms.
5517 * Add certificate callback. If set this is called whenever a certificate
5518 is required by client or server. An application can decide which
5519 certificate chain to present based on arbitrary criteria: for example
5520 supported signature algorithms. Add very simple example to s_server.
5521 This fixes many of the problems and restrictions of the existing client
5522 certificate callback: for example you can now clear an existing
5523 certificate and specify the whole chain.
5527 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5528 the certificate can be used for (if anything). Set valid_flags field
5529 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5530 to have similar checks in it.
5532 Add new "cert_flags" field to CERT structure and include a "strict mode".
5533 This enforces some TLS certificate requirements (such as only permitting
5534 certificate signature algorithms contained in the supported algorithms
5535 extension) which some implementations ignore: this option should be used
5536 with caution as it could cause interoperability issues.
5540 * Update and tidy signature algorithm extension processing. Work out
5541 shared signature algorithms based on preferences and peer algorithms
5542 and print them out in s_client and s_server. Abort handshake if no
5543 shared signature algorithms.
5547 * Add new functions to allow customised supported signature algorithms
5548 for SSL and SSL_CTX structures. Add options to s_client and s_server
5553 * New function SSL_certs_clear() to delete all references to certificates
5554 from an SSL structure. Before this once a certificate had been added
5555 it couldn't be removed.
5559 * Integrate hostname, email address and IP address checking with certificate
5560 verification. New verify options supporting checking in openssl utility.
5564 * Fixes and wildcard matching support to hostname and email checking
5565 functions. Add manual page.
5567 *Florian Weimer (Red Hat Product Security Team)*
5569 * New functions to check a hostname email or IP address against a
5570 certificate. Add options x509 utility to print results of checks against
5575 * Fix OCSP checking.
5577 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5579 * Initial experimental support for explicitly trusted non-root CAs.
5580 OpenSSL still tries to build a complete chain to a root but if an
5581 intermediate CA has a trust setting included that is used. The first
5582 setting is used: whether to trust (e.g., -addtrust option to the x509
5587 * Add -trusted_first option which attempts to find certificates in the
5588 trusted store even if an untrusted chain is also supplied.
5592 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5593 platform support for Linux and Android.
5597 * Support for linux-x32, ILP32 environment in x86_64 framework.
5601 * Experimental multi-implementation support for FIPS capable OpenSSL.
5602 When in FIPS mode the approved implementations are used as normal,
5603 when not in FIPS mode the internal unapproved versions are used instead.
5604 This means that the FIPS capable OpenSSL isn't forced to use the
5605 (often lower performance) FIPS implementations outside FIPS mode.
5609 * Transparently support X9.42 DH parameters when calling
5610 PEM_read_bio_DHparameters. This means existing applications can handle
5611 the new parameter format automatically.
5615 * Initial experimental support for X9.42 DH parameter format: mainly
5616 to support use of 'q' parameter for RFC5114 parameters.
5620 * Add DH parameters from RFC5114 including test data to dhtest.
5624 * Support for automatic EC temporary key parameter selection. If enabled
5625 the most preferred EC parameters are automatically used instead of
5626 hardcoded fixed parameters. Now a server just has to call:
5627 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5628 support ECDH and use the most appropriate parameters.
5632 * Enhance and tidy EC curve and point format TLS extension code. Use
5633 static structures instead of allocation if default values are used.
5634 New ctrls to set curves we wish to support and to retrieve shared curves.
5635 Print out shared curves in s_server. New options to s_server and s_client
5636 to set list of supported curves.
5640 * New ctrls to retrieve supported signature algorithms and
5641 supported curve values as an array of NIDs. Extend openssl utility
5642 to print out received values.
5646 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5647 between NIDs and the more common NIST names such as "P-256". Enhance
5648 ecparam utility and ECC method to recognise the NIST names for curves.
5652 * Enhance SSL/TLS certificate chain handling to support different
5653 chains for each certificate instead of one chain in the parent SSL_CTX.
5657 * Support for fixed DH ciphersuite client authentication: where both
5658 server and client use DH certificates with common parameters.
5662 * Support for fixed DH ciphersuites: those requiring DH server
5667 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5669 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5670 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5671 X509_CINF_get_signature were reverted post internal team review.
5676 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5678 * OCSP Status Request extension unbounded memory growth
5680 A malicious client can send an excessively large OCSP Status Request
5681 extension. If that client continually requests renegotiation, sending a
5682 large OCSP Status Request extension each time, then there will be unbounded
5683 memory growth on the server. This will eventually lead to a Denial Of
5684 Service attack through memory exhaustion. Servers with a default
5685 configuration are vulnerable even if they do not support OCSP. Builds using
5686 the "no-ocsp" build time option are not affected.
5688 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5693 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5696 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5702 * OOB write in MDC2_Update()
5704 An overflow can occur in MDC2_Update() either if called directly or
5705 through the EVP_DigestUpdate() function using MDC2. If an attacker
5706 is able to supply very large amounts of input data after a previous
5707 call to EVP_EncryptUpdate() with a partial block then a length check
5708 can overflow resulting in a heap corruption.
5710 The amount of data needed is comparable to SIZE_MAX which is impractical
5713 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5718 * Malformed SHA512 ticket DoS
5720 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5721 DoS attack where a malformed ticket will result in an OOB read which will
5724 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5725 a custom server callback and ticket lookup mechanism.
5727 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5732 * OOB write in BN_bn2dec()
5734 The function BN_bn2dec() does not check the return value of BN_div_word().
5735 This can cause an OOB write if an application uses this function with an
5736 overly large BIGNUM. This could be a problem if an overly large certificate
5737 or CRL is printed out from an untrusted source. TLS is not affected because
5738 record limits will reject an oversized certificate before it is parsed.
5740 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5745 * OOB read in TS_OBJ_print_bio()
5747 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5748 the total length the OID text representation would use and not the amount
5749 of data written. This will result in OOB reads when large OIDs are
5752 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5757 * Pointer arithmetic undefined behaviour
5759 Avoid some undefined pointer arithmetic
5761 A common idiom in the codebase is to check limits in the following manner:
5764 Where "p" points to some malloc'd data of SIZE bytes and
5767 "len" here could be from some externally supplied data (e.g. from a TLS
5770 The rules of C pointer arithmetic are such that "p + len" is only well
5771 defined where len <= SIZE. Therefore the above idiom is actually
5772 undefined behaviour.
5774 For example this could cause problems if some malloc implementation
5775 provides an address for "p" such that "p + len" actually overflows for
5776 values of len that are too big and therefore p + len < limit.
5778 This issue was reported to OpenSSL by Guido Vranken
5783 * Constant time flag not preserved in DSA signing
5785 Operations in the DSA signing algorithm should run in constant time in
5786 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5787 implementation means that a non-constant time codepath is followed for
5788 certain operations. This has been demonstrated through a cache-timing
5789 attack to be sufficient for an attacker to recover the private DSA key.
5791 This issue was reported by César Pereida (Aalto University), Billy Brumley
5792 (Tampere University of Technology), and Yuval Yarom (The University of
5793 Adelaide and NICTA).
5798 * DTLS buffered message DoS
5800 In a DTLS connection where handshake messages are delivered out-of-order
5801 those messages that OpenSSL is not yet ready to process will be buffered
5802 for later use. Under certain circumstances, a flaw in the logic means that
5803 those messages do not get removed from the buffer even though the handshake
5804 has been completed. An attacker could force up to approx. 15 messages to
5805 remain in the buffer when they are no longer required. These messages will
5806 be cleared when the DTLS connection is closed. The default maximum size for
5807 a message is 100k. Therefore the attacker could force an additional 1500k
5808 to be consumed per connection. By opening many simulataneous connections an
5809 attacker could cause a DoS attack through memory exhaustion.
5811 This issue was reported to OpenSSL by Quan Luo.
5816 * DTLS replay protection DoS
5818 A flaw in the DTLS replay attack protection mechanism means that records
5819 that arrive for future epochs update the replay protection "window" before
5820 the MAC for the record has been validated. This could be exploited by an
5821 attacker by sending a record for the next epoch (which does not have to
5822 decrypt or have a valid MAC), with a very large sequence number. This means
5823 that all subsequent legitimate packets are dropped causing a denial of
5824 service for a specific DTLS connection.
5826 This issue was reported to OpenSSL by the OCAP audit team.
5831 * Certificate message OOB reads
5833 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5834 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5835 theoretical DoS risk but this has not been observed in practice on common
5838 The messages affected are client certificate, client certificate request
5839 and server certificate. As a result the attack can only be performed
5840 against a client or a server which enables client authentication.
5842 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5847 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5849 * Prevent padding oracle in AES-NI CBC MAC check
5851 A MITM attacker can use a padding oracle attack to decrypt traffic
5852 when the connection uses an AES CBC cipher and the server support
5855 This issue was introduced as part of the fix for Lucky 13 padding
5856 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5857 constant time by making sure that always the same bytes are read and
5858 compared against either the MAC or padding bytes. But it no longer
5859 checked that there was enough data to have both the MAC and padding
5862 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5867 * Fix EVP_EncodeUpdate overflow
5869 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5870 Base64 encoding of binary data. If an attacker is able to supply very large
5871 amounts of input data then a length check can overflow resulting in a heap
5874 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5875 the `PEM_write_bio*` family of functions. These are mainly used within the
5876 OpenSSL command line applications, so any application which processes data
5877 from an untrusted source and outputs it as a PEM file should be considered
5878 vulnerable to this issue. User applications that call these APIs directly
5879 with large amounts of untrusted data may also be vulnerable.
5881 This issue was reported by Guido Vranken.
5886 * Fix EVP_EncryptUpdate overflow
5888 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5889 is able to supply very large amounts of input data after a previous call to
5890 EVP_EncryptUpdate() with a partial block then a length check can overflow
5891 resulting in a heap corruption. Following an analysis of all OpenSSL
5892 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5893 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5894 the first called function after an EVP_EncryptInit(), and therefore that
5895 specific call must be safe. The second form is where the length passed to
5896 EVP_EncryptUpdate() can be seen from the code to be some small value and
5897 therefore there is no possibility of an overflow. Since all instances are
5898 one of these two forms, it is believed that there can be no overflows in
5899 internal code due to this problem. It should be noted that
5900 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5901 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5902 of these calls have also been analysed too and it is believed there are no
5903 instances in internal usage where an overflow could occur.
5905 This issue was reported by Guido Vranken.
5910 * Prevent ASN.1 BIO excessive memory allocation
5912 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5913 a short invalid encoding can casuse allocation of large amounts of memory
5914 potentially consuming excessive resources or exhausting memory.
5916 Any application parsing untrusted data through d2i BIO functions is
5917 affected. The memory based functions such as d2i_X509() are *not* affected.
5918 Since the memory based functions are used by the TLS library, TLS
5919 applications are not affected.
5921 This issue was reported by Brian Carpenter.
5928 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5929 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5930 in arbitrary stack data being returned in the buffer.
5932 This issue was reported by Guido Vranken.
5937 * Modify behavior of ALPN to invoke callback after SNI/servername
5938 callback, such that updates to the SSL_CTX affect ALPN.
5942 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5947 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5948 methods are enabled and ssl2 is disabled the methods return NULL.
5952 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5954 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5955 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5956 provide any "EXPORT" or "LOW" strength ciphers.
5960 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5961 is by default disabled at build-time. Builds that are not configured with
5962 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5963 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5964 will need to explicitly call either of:
5966 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5968 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5970 as appropriate. Even if either of those is used, or the application
5971 explicitly uses the version-specific SSLv2_method() or its client and
5972 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5973 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5974 ciphers, and SSLv2 56-bit DES are no longer available.
5979 * Fix a double-free in DSA code
5981 A double free bug was discovered when OpenSSL parses malformed DSA private
5982 keys and could lead to a DoS attack or memory corruption for applications
5983 that receive DSA private keys from untrusted sources. This scenario is
5986 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5992 * Disable SRP fake user seed to address a server memory leak.
5994 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5996 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5997 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5998 was changed to ignore the "fake user" SRP seed, even if the seed
6001 Users should use SRP_VBASE_get1_by_user instead. Note that in
6002 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6003 also that even though configuring the SRP seed attempts to hide
6004 invalid usernames by continuing the handshake with fake
6005 credentials, this behaviour is not constant time and no strong
6006 guarantees are made that the handshake is indistinguishable from
6007 that of a valid user.
6012 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6014 In the BN_hex2bn function the number of hex digits is calculated using an
6015 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6016 large values of `i` this can result in `bn_expand` not allocating any
6017 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6018 field as NULL leading to a subsequent NULL ptr deref. For very large values
6019 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6020 In this case memory is allocated to the internal BIGNUM data field, but it
6021 is insufficiently sized leading to heap corruption. A similar issue exists
6022 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6023 is ever called by user applications with very large untrusted hex/dec data.
6024 This is anticipated to be a rare occurrence.
6026 All OpenSSL internal usage of these functions use data that is not expected
6027 to be untrusted, e.g. config file data or application command line
6028 arguments. If user developed applications generate config file data based
6029 on untrusted data then it is possible that this could also lead to security
6030 consequences. This is also anticipated to be rare.
6032 This issue was reported to OpenSSL by Guido Vranken.
6037 * Fix memory issues in `BIO_*printf` functions
6039 The internal `fmtstr` function used in processing a "%s" format string in
6040 the `BIO_*printf` functions could overflow while calculating the length of a
6041 string and cause an OOB read when printing very long strings.
6043 Additionally the internal `doapr_outch` function can attempt to write to an
6044 OOB memory location (at an offset from the NULL pointer) in the event of a
6045 memory allocation failure. In 1.0.2 and below this could be caused where
6046 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6047 could be in processing a very long "%s" format string. Memory leaks can
6050 The first issue may mask the second issue dependent on compiler behaviour.
6051 These problems could enable attacks where large amounts of untrusted data
6052 is passed to the `BIO_*printf` functions. If applications use these functions
6053 in this way then they could be vulnerable. OpenSSL itself uses these
6054 functions when printing out human-readable dumps of ASN.1 data. Therefore
6055 applications that print this data could be vulnerable if the data is from
6056 untrusted sources. OpenSSL command line applications could also be
6057 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6058 as command line arguments.
6060 Libssl is not considered directly vulnerable. Additionally certificates etc
6061 received via remote connections via libssl are also unlikely to be able to
6062 trigger these issues because of message size limits enforced within libssl.
6064 This issue was reported to OpenSSL Guido Vranken.
6069 * Side channel attack on modular exponentiation
6071 A side-channel attack was found which makes use of cache-bank conflicts on
6072 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6073 of RSA keys. The ability to exploit this issue is limited as it relies on
6074 an attacker who has control of code in a thread running on the same
6075 hyper-threaded core as the victim thread which is performing decryptions.
6077 This issue was reported to OpenSSL by Yuval Yarom, The University of
6078 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6079 Nadia Heninger, University of Pennsylvania with more information at
6080 <http://cachebleed.info>.
6085 * Change the req command to generate a 2048-bit RSA/DSA key by default,
6086 if no keysize is specified with default_bits. This fixes an
6087 omission in an earlier change that changed all RSA/DSA key generation
6088 commands to use 2048 bits by default.
6092 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6094 * Protection for DH small subgroup attacks
6096 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6097 switched on by default and cannot be disabled. This could have some
6102 * SSLv2 doesn't block disabled ciphers
6104 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6105 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6106 been disabled, provided that the SSLv2 protocol was not also disabled via
6109 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6110 and Sebastian Schinzel.
6115 * Reject DH handshakes with parameters shorter than 1024 bits.
6119 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6121 * Certificate verify crash with missing PSS parameter
6123 The signature verification routines will crash with a NULL pointer
6124 dereference if presented with an ASN.1 signature using the RSA PSS
6125 algorithm and absent mask generation function parameter. Since these
6126 routines are used to verify certificate signature algorithms this can be
6127 used to crash any certificate verification operation and exploited in a
6128 DoS attack. Any application which performs certificate verification is
6129 vulnerable including OpenSSL clients and servers which enable client
6132 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6137 * X509_ATTRIBUTE memory leak
6139 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6140 memory. This structure is used by the PKCS#7 and CMS routines so any
6141 application which reads PKCS#7 or CMS data from untrusted sources is
6142 affected. SSL/TLS is not affected.
6144 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6150 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6151 This changes the decoding behaviour for some invalid messages,
6152 though the change is mostly in the more lenient direction, and
6153 legacy behaviour is preserved as much as possible.
6157 * In DSA_generate_parameters_ex, if the provided seed is too short,
6158 use a random seed, as already documented.
6160 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6162 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6164 * Alternate chains certificate forgery
6166 During certificate verfification, OpenSSL will attempt to find an
6167 alternative certificate chain if the first attempt to build such a chain
6168 fails. An error in the implementation of this logic can mean that an
6169 attacker could cause certain checks on untrusted certificates to be
6170 bypassed, such as the CA flag, enabling them to use a valid leaf
6171 certificate to act as a CA and "issue" an invalid certificate.
6173 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6179 * Race condition handling PSK identify hint
6181 If PSK identity hints are received by a multi-threaded client then
6182 the values are wrongly updated in the parent SSL_CTX structure. This can
6183 result in a race condition potentially leading to a double free of the
6189 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6191 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6192 incompatibility in the handling of HMAC. The previous ABI has now been
6195 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6197 * Malformed ECParameters causes infinite loop
6199 When processing an ECParameters structure OpenSSL enters an infinite loop
6200 if the curve specified is over a specially malformed binary polynomial
6203 This can be used to perform denial of service against any
6204 system which processes public keys, certificate requests or
6205 certificates. This includes TLS clients and TLS servers with
6206 client authentication enabled.
6208 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6213 * Exploitable out-of-bounds read in X509_cmp_time
6215 X509_cmp_time does not properly check the length of the ASN1_TIME
6216 string and can read a few bytes out of bounds. In addition,
6217 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6220 An attacker can use this to craft malformed certificates and CRLs of
6221 various sizes and potentially cause a segmentation fault, resulting in
6222 a DoS on applications that verify certificates or CRLs. TLS clients
6223 that verify CRLs are affected. TLS clients and servers with client
6224 authentication enabled may be affected if they use custom verification
6227 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6228 independently by Hanno Böck.
6233 * PKCS7 crash with missing EnvelopedContent
6235 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6236 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6237 with missing content and trigger a NULL pointer dereference on parsing.
6239 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6240 structures from untrusted sources are affected. OpenSSL clients and
6241 servers are not affected.
6243 This issue was reported to OpenSSL by Michal Zalewski (Google).
6248 * CMS verify infinite loop with unknown hash function
6250 When verifying a signedData message the CMS code can enter an infinite loop
6251 if presented with an unknown hash function OID. This can be used to perform
6252 denial of service against any system which verifies signedData messages using
6254 This issue was reported to OpenSSL by Johannes Bauer.
6259 * Race condition handling NewSessionTicket
6261 If a NewSessionTicket is received by a multi-threaded client when attempting to
6262 reuse a previous ticket then a race condition can occur potentially leading to
6263 a double free of the ticket data.
6268 * Reject DH handshakes with parameters shorter than 768 bits.
6270 *Kurt Roeckx and Emilia Kasper*
6272 * dhparam: generate 2048-bit parameters by default.
6274 *Kurt Roeckx and Emilia Kasper*
6276 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6278 * Segmentation fault in ASN1_TYPE_cmp fix
6280 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6281 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6282 certificate signature algorithm consistency this can be used to crash any
6283 certificate verification operation and exploited in a DoS attack. Any
6284 application which performs certificate verification is vulnerable including
6285 OpenSSL clients and servers which enable client authentication.
6290 * ASN.1 structure reuse memory corruption fix
6292 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6293 memory corruption via an invalid write. Such reuse is and has been
6294 strongly discouraged and is believed to be rare.
6296 Applications that parse structures containing CHOICE or ANY DEFINED BY
6297 components may be affected. Certificate parsing (d2i_X509 and related
6298 functions) are however not affected. OpenSSL clients and servers are
6304 * PKCS7 NULL pointer dereferences fix
6306 The PKCS#7 parsing code does not handle missing outer ContentInfo
6307 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6308 missing content and trigger a NULL pointer dereference on parsing.
6310 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6311 otherwise parse PKCS#7 structures from untrusted sources are
6312 affected. OpenSSL clients and servers are not affected.
6314 This issue was reported to OpenSSL by Michal Zalewski (Google).
6319 * DoS via reachable assert in SSLv2 servers fix
6321 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6322 servers that both support SSLv2 and enable export cipher suites by sending
6323 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6325 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6326 (OpenSSL development team).
6331 * Use After Free following d2i_ECPrivatekey error fix
6333 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6334 could cause a use after free condition. This, in turn, could cause a double
6335 free in several private key parsing functions (such as d2i_PrivateKey
6336 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6337 for applications that receive EC private keys from untrusted
6338 sources. This scenario is considered rare.
6340 This issue was discovered by the BoringSSL project and fixed in their
6346 * X509_to_X509_REQ NULL pointer deref fix
6348 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6349 the certificate key is invalid. This function is rarely used in practice.
6351 This issue was discovered by Brian Carpenter.
6356 * Removed the export ciphers from the DEFAULT ciphers
6360 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6362 * Build fixes for the Windows and OpenVMS platforms
6364 *Matt Caswell and Richard Levitte*
6366 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6368 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6369 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6370 dereference. This could lead to a Denial Of Service attack. Thanks to
6371 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6376 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6377 dtls1_buffer_record function under certain conditions. In particular this
6378 could occur if an attacker sent repeated DTLS records with the same
6379 sequence number but for the next epoch. The memory leak could be exploited
6380 by an attacker in a Denial of Service attack through memory exhaustion.
6381 Thanks to Chris Mueller for reporting this issue.
6386 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6387 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6388 method would be set to NULL which could later result in a NULL pointer
6389 dereference. Thanks to Frank Schmirler for reporting this issue.
6394 * Abort handshake if server key exchange message is omitted for ephemeral
6397 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6398 reporting this issue.
6403 * Remove non-export ephemeral RSA code on client and server. This code
6404 violated the TLS standard by allowing the use of temporary RSA keys in
6405 non-export ciphersuites and could be used by a server to effectively
6406 downgrade the RSA key length used to a value smaller than the server
6407 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6408 INRIA or reporting this issue.
6413 * Fixed issue where DH client certificates are accepted without verification.
6414 An OpenSSL server will accept a DH certificate for client authentication
6415 without the certificate verify message. This effectively allows a client to
6416 authenticate without the use of a private key. This only affects servers
6417 which trust a client certificate authority which issues certificates
6418 containing DH keys: these are extremely rare and hardly ever encountered.
6419 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6425 * Ensure that the session ID context of an SSL is updated when its
6426 SSL_CTX is updated via SSL_set_SSL_CTX.
6428 The session ID context is typically set from the parent SSL_CTX,
6429 and can vary with the CTX.
6433 * Fix various certificate fingerprint issues.
6435 By using non-DER or invalid encodings outside the signed portion of a
6436 certificate the fingerprint can be changed without breaking the signature.
6437 Although no details of the signed portion of the certificate can be changed
6438 this can cause problems with some applications: e.g. those using the
6439 certificate fingerprint for blacklists.
6441 1. Reject signatures with non zero unused bits.
6443 If the BIT STRING containing the signature has non zero unused bits reject
6444 the signature. All current signature algorithms require zero unused bits.
6446 2. Check certificate algorithm consistency.
6448 Check the AlgorithmIdentifier inside TBS matches the one in the
6449 certificate signature. NB: this will result in signature failure
6450 errors for some broken certificates.
6452 Thanks to Konrad Kraszewski from Google for reporting this issue.
6454 3. Check DSA/ECDSA signatures use DER.
6456 Re-encode DSA/ECDSA signatures and compare with the original received
6457 signature. Return an error if there is a mismatch.
6459 This will reject various cases including garbage after signature
6460 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6461 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6462 (negative or with leading zeroes).
6464 Further analysis was conducted and fixes were developed by Stephen Henson
6465 of the OpenSSL core team.
6471 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6472 results on some platforms, including x86_64. This bug occurs at random
6473 with a very low probability, and is not known to be exploitable in any
6474 way, though its exact impact is difficult to determine. Thanks to Pieter
6475 Wuille (Blockstream) who reported this issue and also suggested an initial
6476 fix. Further analysis was conducted by the OpenSSL development team and
6477 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6478 the OpenSSL core team.
6483 * Do not resume sessions on the server if the negotiated protocol
6484 version does not match the session's version. Resuming with a different
6485 version, while not strictly forbidden by the RFC, is of questionable
6486 sanity and breaks all known clients.
6488 *David Benjamin, Emilia Käsper*
6490 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6491 early CCS messages during renegotiation. (Note that because
6492 renegotiation is encrypted, this early CCS was not exploitable.)
6496 * Tighten client-side session ticket handling during renegotiation:
6497 ensure that the client only accepts a session ticket if the server sends
6498 the extension anew in the ServerHello. Previously, a TLS client would
6499 reuse the old extension state and thus accept a session ticket if one was
6500 announced in the initial ServerHello.
6502 Similarly, ensure that the client requires a session ticket if one
6503 was advertised in the ServerHello. Previously, a TLS client would
6504 ignore a missing NewSessionTicket message.
6508 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6512 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6513 sends a carefully crafted handshake message, to cause OpenSSL to fail
6514 to free up to 64k of memory causing a memory leak. This could be
6515 exploited in a Denial Of Service attack. This issue affects OpenSSL
6516 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6517 whether SRTP is used or configured. Implementations of OpenSSL that
6518 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6520 The fix was developed by the OpenSSL team.
6525 * Session Ticket Memory Leak.
6527 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6528 integrity of that ticket is first verified. In the event of a session
6529 ticket integrity check failing, OpenSSL will fail to free memory
6530 causing a memory leak. By sending a large number of invalid session
6531 tickets an attacker could exploit this issue in a Denial Of Service
6537 * Build option no-ssl3 is incomplete.
6539 When OpenSSL is configured with "no-ssl3" as a build option, servers
6540 could accept and complete a SSL 3.0 handshake, and clients could be
6541 configured to send them.
6544 *Akamai and the OpenSSL team*
6546 * Add support for TLS_FALLBACK_SCSV.
6547 Client applications doing fallback retries should call
6548 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6551 *Adam Langley, Bodo Moeller*
6553 * Add additional DigestInfo checks.
6555 Re-encode DigestInto in DER and check against the original when
6556 verifying RSA signature: this will reject any improperly encoded
6557 DigestInfo structures.
6559 Note: this is a precautionary measure and no attacks are currently known.
6563 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6565 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6566 SRP code can be overrun an internal buffer. Add sanity check that
6567 g, A, B < N to SRP code.
6569 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6570 Group for discovering this issue.
6575 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6576 TLS 1.0 instead of higher protocol versions when the ClientHello message
6577 is badly fragmented. This allows a man-in-the-middle attacker to force a
6578 downgrade to TLS 1.0 even if both the server and the client support a
6579 higher protocol version, by modifying the client's TLS records.
6581 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6582 researching this issue.
6587 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6588 to a denial of service attack. A malicious server can crash the client
6589 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6590 ciphersuite and sending carefully crafted handshake messages.
6592 Thanks to Felix Gröbert (Google) for discovering and researching this
6598 * By sending carefully crafted DTLS packets an attacker could cause openssl
6599 to leak memory. This can be exploited through a Denial of Service attack.
6600 Thanks to Adam Langley for discovering and researching this issue.
6605 * An attacker can force openssl to consume large amounts of memory whilst
6606 processing DTLS handshake messages. This can be exploited through a
6607 Denial of Service attack.
6608 Thanks to Adam Langley for discovering and researching this issue.
6613 * An attacker can force an error condition which causes openssl to crash
6614 whilst processing DTLS packets due to memory being freed twice. This
6615 can be exploited through a Denial of Service attack.
6616 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6622 * If a multithreaded client connects to a malicious server using a resumed
6623 session and the server sends an ec point format extension it could write
6624 up to 255 bytes to freed memory.
6626 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6632 * A malicious server can crash an OpenSSL client with a null pointer
6633 dereference (read) by specifying an SRP ciphersuite even though it was not
6634 properly negotiated with the client. This can be exploited through a
6635 Denial of Service attack.
6637 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6638 discovering and researching this issue.
6643 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6644 X509_name_oneline, X509_name_print_ex et al. to leak some information
6645 from the stack. Applications may be affected if they echo pretty printing
6646 output to the attacker.
6648 Thanks to Ivan Fratric (Google) for discovering this issue.
6651 *Emilia Käsper, and Steve Henson*
6653 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6654 for corner cases. (Certain input points at infinity could lead to
6655 bogus results, with non-infinity inputs mapped to infinity too.)
6659 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6661 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6662 handshake can force the use of weak keying material in OpenSSL
6663 SSL/TLS clients and servers.
6665 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6666 researching this issue. ([CVE-2014-0224])
6668 *KIKUCHI Masashi, Steve Henson*
6670 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6671 OpenSSL DTLS client the code can be made to recurse eventually crashing
6674 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6677 *Imre Rad, Steve Henson*
6679 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6680 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6681 client or server. This is potentially exploitable to run arbitrary
6682 code on a vulnerable client or server.
6684 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6686 *Jüri Aedla, Steve Henson*
6688 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6689 are subject to a denial of service attack.
6691 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6692 this issue. ([CVE-2014-3470])
6694 *Felix Gröbert, Ivan Fratric, Steve Henson*
6696 * Harmonize version and its documentation. -f flag is used to display
6699 *mancha <mancha1@zoho.com>*
6701 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6702 in i2d_ECPrivateKey.
6704 *mancha <mancha1@zoho.com>*
6706 * Fix some double frees. These are not thought to be exploitable.
6708 *mancha <mancha1@zoho.com>*
6710 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6712 * A missing bounds check in the handling of the TLS heartbeat extension
6713 can be used to reveal up to 64k of memory to a connected client or
6716 Thanks for Neel Mehta of Google Security for discovering this bug and to
6717 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6718 preparing the fix ([CVE-2014-0160])
6720 *Adam Langley, Bodo Moeller*
6722 * Fix for the attack described in the paper "Recovering OpenSSL
6723 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6724 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6725 <http://eprint.iacr.org/2014/140>
6727 Thanks to Yuval Yarom and Naomi Benger for discovering this
6728 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6730 *Yuval Yarom and Naomi Benger*
6732 * TLS pad extension: draft-agl-tls-padding-03
6734 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6735 TLS client Hello record length value would otherwise be > 255 and
6736 less that 512 pad with a dummy extension containing zeroes so it
6737 is at least 512 bytes long.
6739 *Adam Langley, Steve Henson*
6741 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6743 * Fix for TLS record tampering bug. A carefully crafted invalid
6744 handshake could crash OpenSSL with a NULL pointer exception.
6745 Thanks to Anton Johansson for reporting this issues.
6748 * Keep original DTLS digest and encryption contexts in retransmission
6749 structures so we can use the previous session parameters if they need
6750 to be resent. ([CVE-2013-6450])
6754 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6755 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6756 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6757 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6758 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6759 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6761 *Rob Stradling, Adam Langley*
6763 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6765 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6766 supporting platforms or when small records were transferred.
6768 *Andy Polyakov, Steve Henson*
6770 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6772 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6774 This addresses the flaw in CBC record processing discovered by
6775 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6776 at: <http://www.isg.rhul.ac.uk/tls/>
6778 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6779 Security Group at Royal Holloway, University of London
6780 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6781 Emilia Käsper for the initial patch.
6784 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6786 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6787 ciphersuites which can be exploited in a denial of service attack.
6788 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6789 and detecting this bug and to Wolfgang Ettlinger
6790 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6795 * Return an error when checking OCSP signatures when key is NULL.
6796 This fixes a DoS attack. ([CVE-2013-0166])
6800 * Make openssl verify return errors.
6802 *Chris Palmer <palmer@google.com> and Ben Laurie*
6804 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6805 the right response is stapled. Also change SSL_get_certificate()
6806 so it returns the certificate actually sent.
6807 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6809 *Rob Stradling <rob.stradling@comodo.com>*
6811 * Fix possible deadlock when decoding public keys.
6815 * Don't use TLS 1.0 record version number in initial client hello
6820 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6822 * Sanity check record length before skipping explicit IV in TLS
6823 1.2, 1.1 and DTLS to fix DoS attack.
6825 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6826 fuzzing as a service testing platform.
6831 * Initialise tkeylen properly when encrypting CMS messages.
6832 Thanks to Solar Designer of Openwall for reporting this issue.
6836 * In FIPS mode don't try to use composite ciphers as they are not
6841 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6843 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6844 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6845 mean any application compiled against OpenSSL 1.0.0 headers setting
6846 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6847 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6848 0x10000000L Any application which was previously compiled against
6849 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6850 will need to be recompiled as a result. Letting be results in
6851 inability to disable specifically TLS 1.1 and in client context,
6852 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6856 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6857 disable just protocol X, but all protocols above X *if* there are
6858 protocols *below* X still enabled. In more practical terms it means
6859 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6860 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6861 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6866 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6868 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6869 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6870 in CRYPTO_realloc_clean.
6872 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6873 issue and to Adam Langley <agl@chromium.org> for fixing it.
6876 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6878 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6882 * Workarounds for some broken servers that "hang" if a client hello
6883 record length exceeds 255 bytes.
6885 1. Do not use record version number > TLS 1.0 in initial client
6886 hello: some (but not all) hanging servers will now work.
6887 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6888 the number of ciphers sent in the client hello. This should be
6889 set to an even number, such as 50, for example by passing:
6890 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6891 Most broken servers should now work.
6892 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6893 TLS 1.2 client support entirely.
6897 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6901 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6903 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6904 STRING form instead of a DigestInfo.
6908 * The format used for MDC2 RSA signatures is inconsistent between EVP
6909 and the RSA_sign/RSA_verify functions. This was made more apparent when
6910 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6911 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6912 the correct format in RSA_verify so both forms transparently work.
6916 * Some servers which support TLS 1.0 can choke if we initially indicate
6917 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6918 encrypted premaster secret. As a workaround use the maximum permitted
6919 client version in client hello, this should keep such servers happy
6920 and still work with previous versions of OpenSSL.
6924 * Add support for TLS/DTLS heartbeats.
6926 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6928 * Add support for SCTP.
6930 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6932 * Improved PRNG seeding for VOS.
6934 *Paul Green <Paul.Green@stratus.com>*
6936 * Extensive assembler packs updates, most notably:
6938 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6939 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6940 - x86_64: bit-sliced AES implementation;
6941 - ARM: NEON support, contemporary platforms optimizations;
6942 - s390x: z196 support;
6943 - `*`: GHASH and GF(2^m) multiplication implementations;
6947 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6948 (removal of unnecessary code)
6950 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6952 * Add TLS key material exporter from RFC 5705.
6956 * Add DTLS-SRTP negotiation from RFC 5764.
6960 * Add Next Protocol Negotiation,
6961 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6962 disabled with a no-npn flag to config or Configure. Code donated
6965 *Adam Langley <agl@google.com> and Ben Laurie*
6967 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6968 NIST-P256, NIST-P521, with constant-time single point multiplication on
6969 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6970 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6971 Code made available under Apache License version 2.0.
6973 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6974 line to include this in your build of OpenSSL, and run "make depend" (or
6975 "make update"). This enables the following EC_METHODs:
6977 EC_GFp_nistp224_method()
6978 EC_GFp_nistp256_method()
6979 EC_GFp_nistp521_method()
6981 EC_GROUP_new_by_curve_name() will automatically use these (while
6982 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6985 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6987 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6988 all platforms. Move ssize_t definition from e_os.h to the public
6989 header file e_os2.h as it now appears in public header file cms.h
6993 * New -sigopt option to the ca, req and x509 utilities. Additional
6994 signature parameters can be passed using this option and in
6999 * Add RSA PSS signing function. This will generate and set the
7000 appropriate AlgorithmIdentifiers for PSS based on those in the
7001 corresponding EVP_MD_CTX structure. No application support yet.
7005 * Support for companion algorithm specific ASN1 signing routines.
7006 New function ASN1_item_sign_ctx() signs a pre-initialised
7007 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7008 the appropriate parameters.
7012 * Add new algorithm specific ASN1 verification initialisation function
7013 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7014 handling will be the same no matter what EVP_PKEY_METHOD is used.
7015 Add a PSS handler to support verification of PSS signatures: checked
7016 against a number of sample certificates.
7020 * Add signature printing for PSS. Add PSS OIDs.
7022 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7024 * Add algorithm specific signature printing. An individual ASN1 method
7025 can now print out signatures instead of the standard hex dump.
7027 More complex signatures (e.g. PSS) can print out more meaningful
7028 information. Include DSA version that prints out the signature
7033 * Password based recipient info support for CMS library: implementing
7038 * Split password based encryption into PBES2 and PBKDF2 functions. This
7039 neatly separates the code into cipher and PBE sections and is required
7040 for some algorithms that split PBES2 into separate pieces (such as
7041 password based CMS).
7045 * Session-handling fixes:
7046 - Fix handling of connections that are resuming with a session ID,
7047 but also support Session Tickets.
7048 - Fix a bug that suppressed issuing of a new ticket if the client
7049 presented a ticket with an expired session.
7050 - Try to set the ticket lifetime hint to something reasonable.
7051 - Make tickets shorter by excluding irrelevant information.
7052 - On the client side, don't ignore renewed tickets.
7054 *Adam Langley, Bodo Moeller (Google)*
7056 * Fix PSK session representation.
7060 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7062 This work was sponsored by Intel.
7066 * Add GCM support to TLS library. Some custom code is needed to split
7067 the IV between the fixed (from PRF) and explicit (from TLS record)
7068 portions. This adds all GCM ciphersuites supported by RFC5288 and
7069 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7070 add a special AESGCM string for GCM only.
7074 * Expand range of ctrls for AES GCM. Permit setting invocation
7075 field on decrypt and retrieval of invocation field only on encrypt.
7079 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7080 As required by RFC5289 these ciphersuites cannot be used if for
7081 versions of TLS earlier than 1.2.
7085 * For FIPS capable OpenSSL interpret a NULL default public key method
7086 as unset and return the appropriate default but do *not* set the default.
7087 This means we can return the appropriate method in applications that
7088 switch between FIPS and non-FIPS modes.
7092 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7093 ENGINE is used then we cannot handle that in the FIPS module so we
7094 keep original code iff non-FIPS operations are allowed.
7098 * Add -attime option to openssl utilities.
7100 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7102 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7106 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7107 FIPS EC methods unconditionally for now.
7111 * New build option no-ec2m to disable characteristic 2 code.
7115 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7116 all cases can be covered as some introduce binary incompatibilities.
7120 * Redirect RSA operations to FIPS module including keygen,
7121 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7125 * Add similar low level API blocking to ciphers.
7129 * Low level digest APIs are not approved in FIPS mode: any attempt
7130 to use these will cause a fatal error. Applications that *really* want
7131 to use them can use the `private_*` version instead.
7135 * Redirect cipher operations to FIPS module for FIPS builds.
7139 * Redirect digest operations to FIPS module for FIPS builds.
7143 * Update build system to add "fips" flag which will link in fipscanister.o
7144 for static and shared library builds embedding a signature if needed.
7148 * Output TLS supported curves in preference order instead of numerical
7149 order. This is currently hardcoded for the highest order curves first.
7150 This should be configurable so applications can judge speed vs strength.
7154 * Add TLS v1.2 server support for client authentication.
7158 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7163 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7164 FIPS modules versions.
7168 * Add TLS v1.2 client side support for client authentication. Keep cache
7169 of handshake records longer as we don't know the hash algorithm to use
7170 until after the certificate request message is received.
7174 * Initial TLS v1.2 client support. Add a default signature algorithms
7175 extension including all the algorithms we support. Parse new signature
7176 format in client key exchange. Relax some ECC signing restrictions for
7177 TLS v1.2 as indicated in RFC5246.
7181 * Add server support for TLS v1.2 signature algorithms extension. Switch
7182 to new signature format when needed using client digest preference.
7183 All server ciphersuites should now work correctly in TLS v1.2. No client
7184 support yet and no support for client certificates.
7188 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7189 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7190 ciphersuites. At present only RSA key exchange ciphersuites work with
7191 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7192 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7193 and version checking.
7197 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7198 with this defined it will not be affected by any changes to ssl internal
7199 structures. Add several utility functions to allow openssl application
7200 to work with OPENSSL_NO_SSL_INTERN defined.
7204 * A long standing patch to add support for SRP from EdelWeb (Peter
7205 Sylvester and Christophe Renou) was integrated.
7206 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7207 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7210 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7214 * Permit abbreviated handshakes when renegotiating using the function
7215 SSL_renegotiate_abbreviated().
7217 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7219 * Add call to ENGINE_register_all_complete() to
7220 ENGINE_load_builtin_engines(), so some implementations get used
7221 automatically instead of needing explicit application support.
7225 * Add support for TLS key exporter as described in RFC5705.
7227 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7229 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7230 a few changes are required:
7232 Add SSL_OP_NO_TLSv1_1 flag.
7233 Add TLSv1_1 methods.
7234 Update version checking logic to handle version 1.1.
7235 Add explicit IV handling (ported from DTLS code).
7236 Add command line options to s_client/s_server.
7243 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7245 * X509_ATTRIBUTE memory leak
7247 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7248 memory. This structure is used by the PKCS#7 and CMS routines so any
7249 application which reads PKCS#7 or CMS data from untrusted sources is
7250 affected. SSL/TLS is not affected.
7252 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7258 * Race condition handling PSK identify hint
7260 If PSK identity hints are received by a multi-threaded client then
7261 the values are wrongly updated in the parent SSL_CTX structure. This can
7262 result in a race condition potentially leading to a double free of the
7268 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7270 * Malformed ECParameters causes infinite loop
7272 When processing an ECParameters structure OpenSSL enters an infinite loop
7273 if the curve specified is over a specially malformed binary polynomial
7276 This can be used to perform denial of service against any
7277 system which processes public keys, certificate requests or
7278 certificates. This includes TLS clients and TLS servers with
7279 client authentication enabled.
7281 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7286 * Exploitable out-of-bounds read in X509_cmp_time
7288 X509_cmp_time does not properly check the length of the ASN1_TIME
7289 string and can read a few bytes out of bounds. In addition,
7290 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7293 An attacker can use this to craft malformed certificates and CRLs of
7294 various sizes and potentially cause a segmentation fault, resulting in
7295 a DoS on applications that verify certificates or CRLs. TLS clients
7296 that verify CRLs are affected. TLS clients and servers with client
7297 authentication enabled may be affected if they use custom verification
7300 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7301 independently by Hanno Böck.
7306 * PKCS7 crash with missing EnvelopedContent
7308 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7309 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7310 with missing content and trigger a NULL pointer dereference on parsing.
7312 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7313 structures from untrusted sources are affected. OpenSSL clients and
7314 servers are not affected.
7316 This issue was reported to OpenSSL by Michal Zalewski (Google).
7321 * CMS verify infinite loop with unknown hash function
7323 When verifying a signedData message the CMS code can enter an infinite loop
7324 if presented with an unknown hash function OID. This can be used to perform
7325 denial of service against any system which verifies signedData messages using
7327 This issue was reported to OpenSSL by Johannes Bauer.
7332 * Race condition handling NewSessionTicket
7334 If a NewSessionTicket is received by a multi-threaded client when attempting to
7335 reuse a previous ticket then a race condition can occur potentially leading to
7336 a double free of the ticket data.
7341 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7343 * Segmentation fault in ASN1_TYPE_cmp fix
7345 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7346 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7347 certificate signature algorithm consistency this can be used to crash any
7348 certificate verification operation and exploited in a DoS attack. Any
7349 application which performs certificate verification is vulnerable including
7350 OpenSSL clients and servers which enable client authentication.
7355 * ASN.1 structure reuse memory corruption fix
7357 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7358 memory corruption via an invalid write. Such reuse is and has been
7359 strongly discouraged and is believed to be rare.
7361 Applications that parse structures containing CHOICE or ANY DEFINED BY
7362 components may be affected. Certificate parsing (d2i_X509 and related
7363 functions) are however not affected. OpenSSL clients and servers are
7369 * PKCS7 NULL pointer dereferences fix
7371 The PKCS#7 parsing code does not handle missing outer ContentInfo
7372 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7373 missing content and trigger a NULL pointer dereference on parsing.
7375 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7376 otherwise parse PKCS#7 structures from untrusted sources are
7377 affected. OpenSSL clients and servers are not affected.
7379 This issue was reported to OpenSSL by Michal Zalewski (Google).
7384 * DoS via reachable assert in SSLv2 servers fix
7386 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7387 servers that both support SSLv2 and enable export cipher suites by sending
7388 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7390 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7391 (OpenSSL development team).
7396 * Use After Free following d2i_ECPrivatekey error fix
7398 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7399 could cause a use after free condition. This, in turn, could cause a double
7400 free in several private key parsing functions (such as d2i_PrivateKey
7401 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7402 for applications that receive EC private keys from untrusted
7403 sources. This scenario is considered rare.
7405 This issue was discovered by the BoringSSL project and fixed in their
7411 * X509_to_X509_REQ NULL pointer deref fix
7413 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7414 the certificate key is invalid. This function is rarely used in practice.
7416 This issue was discovered by Brian Carpenter.
7421 * Removed the export ciphers from the DEFAULT ciphers
7425 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7427 * Build fixes for the Windows and OpenVMS platforms
7429 *Matt Caswell and Richard Levitte*
7431 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7433 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7434 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7435 dereference. This could lead to a Denial Of Service attack. Thanks to
7436 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7441 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7442 dtls1_buffer_record function under certain conditions. In particular this
7443 could occur if an attacker sent repeated DTLS records with the same
7444 sequence number but for the next epoch. The memory leak could be exploited
7445 by an attacker in a Denial of Service attack through memory exhaustion.
7446 Thanks to Chris Mueller for reporting this issue.
7451 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7452 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7453 method would be set to NULL which could later result in a NULL pointer
7454 dereference. Thanks to Frank Schmirler for reporting this issue.
7459 * Abort handshake if server key exchange message is omitted for ephemeral
7462 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7463 reporting this issue.
7468 * Remove non-export ephemeral RSA code on client and server. This code
7469 violated the TLS standard by allowing the use of temporary RSA keys in
7470 non-export ciphersuites and could be used by a server to effectively
7471 downgrade the RSA key length used to a value smaller than the server
7472 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7473 INRIA or reporting this issue.
7478 * Fixed issue where DH client certificates are accepted without verification.
7479 An OpenSSL server will accept a DH certificate for client authentication
7480 without the certificate verify message. This effectively allows a client to
7481 authenticate without the use of a private key. This only affects servers
7482 which trust a client certificate authority which issues certificates
7483 containing DH keys: these are extremely rare and hardly ever encountered.
7484 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7490 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7491 results on some platforms, including x86_64. This bug occurs at random
7492 with a very low probability, and is not known to be exploitable in any
7493 way, though its exact impact is difficult to determine. Thanks to Pieter
7494 Wuille (Blockstream) who reported this issue and also suggested an initial
7495 fix. Further analysis was conducted by the OpenSSL development team and
7496 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7497 the OpenSSL core team.
7502 * Fix various certificate fingerprint issues.
7504 By using non-DER or invalid encodings outside the signed portion of a
7505 certificate the fingerprint can be changed without breaking the signature.
7506 Although no details of the signed portion of the certificate can be changed
7507 this can cause problems with some applications: e.g. those using the
7508 certificate fingerprint for blacklists.
7510 1. Reject signatures with non zero unused bits.
7512 If the BIT STRING containing the signature has non zero unused bits reject
7513 the signature. All current signature algorithms require zero unused bits.
7515 2. Check certificate algorithm consistency.
7517 Check the AlgorithmIdentifier inside TBS matches the one in the
7518 certificate signature. NB: this will result in signature failure
7519 errors for some broken certificates.
7521 Thanks to Konrad Kraszewski from Google for reporting this issue.
7523 3. Check DSA/ECDSA signatures use DER.
7525 Reencode DSA/ECDSA signatures and compare with the original received
7526 signature. Return an error if there is a mismatch.
7528 This will reject various cases including garbage after signature
7529 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7530 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7531 (negative or with leading zeroes).
7533 Further analysis was conducted and fixes were developed by Stephen Henson
7534 of the OpenSSL core team.
7540 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7542 * Session Ticket Memory Leak.
7544 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7545 integrity of that ticket is first verified. In the event of a session
7546 ticket integrity check failing, OpenSSL will fail to free memory
7547 causing a memory leak. By sending a large number of invalid session
7548 tickets an attacker could exploit this issue in a Denial Of Service
7554 * Build option no-ssl3 is incomplete.
7556 When OpenSSL is configured with "no-ssl3" as a build option, servers
7557 could accept and complete a SSL 3.0 handshake, and clients could be
7558 configured to send them.
7561 *Akamai and the OpenSSL team*
7563 * Add support for TLS_FALLBACK_SCSV.
7564 Client applications doing fallback retries should call
7565 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7568 *Adam Langley, Bodo Moeller*
7570 * Add additional DigestInfo checks.
7572 Reencode DigestInto in DER and check against the original when
7573 verifying RSA signature: this will reject any improperly encoded
7574 DigestInfo structures.
7576 Note: this is a precautionary measure and no attacks are currently known.
7580 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7582 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7583 to a denial of service attack. A malicious server can crash the client
7584 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7585 ciphersuite and sending carefully crafted handshake messages.
7587 Thanks to Felix Gröbert (Google) for discovering and researching this
7593 * By sending carefully crafted DTLS packets an attacker could cause openssl
7594 to leak memory. This can be exploited through a Denial of Service attack.
7595 Thanks to Adam Langley for discovering and researching this issue.
7600 * An attacker can force openssl to consume large amounts of memory whilst
7601 processing DTLS handshake messages. This can be exploited through a
7602 Denial of Service attack.
7603 Thanks to Adam Langley for discovering and researching this issue.
7608 * An attacker can force an error condition which causes openssl to crash
7609 whilst processing DTLS packets due to memory being freed twice. This
7610 can be exploited through a Denial of Service attack.
7611 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7617 * If a multithreaded client connects to a malicious server using a resumed
7618 session and the server sends an ec point format extension it could write
7619 up to 255 bytes to freed memory.
7621 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7627 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7628 X509_name_oneline, X509_name_print_ex et al. to leak some information
7629 from the stack. Applications may be affected if they echo pretty printing
7630 output to the attacker.
7632 Thanks to Ivan Fratric (Google) for discovering this issue.
7635 *Emilia Käsper, and Steve Henson*
7637 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7638 for corner cases. (Certain input points at infinity could lead to
7639 bogus results, with non-infinity inputs mapped to infinity too.)
7643 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7645 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7646 handshake can force the use of weak keying material in OpenSSL
7647 SSL/TLS clients and servers.
7649 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7650 researching this issue. ([CVE-2014-0224])
7652 *KIKUCHI Masashi, Steve Henson*
7654 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7655 OpenSSL DTLS client the code can be made to recurse eventually crashing
7658 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7661 *Imre Rad, Steve Henson*
7663 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7664 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7665 client or server. This is potentially exploitable to run arbitrary
7666 code on a vulnerable client or server.
7668 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7670 *Jüri Aedla, Steve Henson*
7672 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7673 are subject to a denial of service attack.
7675 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7676 this issue. ([CVE-2014-3470])
7678 *Felix Gröbert, Ivan Fratric, Steve Henson*
7680 * Harmonize version and its documentation. -f flag is used to display
7683 *mancha <mancha1@zoho.com>*
7685 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7686 in i2d_ECPrivateKey.
7688 *mancha <mancha1@zoho.com>*
7690 * Fix some double frees. These are not thought to be exploitable.
7692 *mancha <mancha1@zoho.com>*
7694 * Fix for the attack described in the paper "Recovering OpenSSL
7695 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7696 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7697 <http://eprint.iacr.org/2014/140>
7699 Thanks to Yuval Yarom and Naomi Benger for discovering this
7700 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7702 *Yuval Yarom and Naomi Benger*
7704 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7706 * Keep original DTLS digest and encryption contexts in retransmission
7707 structures so we can use the previous session parameters if they need
7708 to be resent. ([CVE-2013-6450])
7712 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7713 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7714 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7715 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7716 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7717 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7719 *Rob Stradling, Adam Langley*
7721 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7723 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7725 This addresses the flaw in CBC record processing discovered by
7726 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7727 at: <http://www.isg.rhul.ac.uk/tls/>
7729 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7730 Security Group at Royal Holloway, University of London
7731 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7732 Emilia Käsper for the initial patch.
7735 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7737 * Return an error when checking OCSP signatures when key is NULL.
7738 This fixes a DoS attack. ([CVE-2013-0166])
7742 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7743 the right response is stapled. Also change SSL_get_certificate()
7744 so it returns the certificate actually sent.
7745 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7746 (This is a backport)
7748 *Rob Stradling <rob.stradling@comodo.com>*
7750 * Fix possible deadlock when decoding public keys.
7754 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7756 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7759 * Sanity check record length before skipping explicit IV in DTLS
7762 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7763 fuzzing as a service testing platform.
7768 * Initialise tkeylen properly when encrypting CMS messages.
7769 Thanks to Solar Designer of Openwall for reporting this issue.
7773 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7775 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7776 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7777 in CRYPTO_realloc_clean.
7779 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7780 issue and to Adam Langley <agl@chromium.org> for fixing it.
7783 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7785 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7787 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7788 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7789 content decryption and always return the same error. Note: this attack
7790 needs on average 2^20 messages so it only affects automated senders. The
7791 old behaviour can be re-enabled in the CMS code by setting the
7792 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7793 an MMA defence is not necessary.
7794 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7795 this issue. ([CVE-2012-0884])
7799 * Fix CVE-2011-4619: make sure we really are receiving a
7800 client hello before rejecting multiple SGC restarts. Thanks to
7801 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7805 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7807 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7808 Thanks to Antonio Martin, Enterprise Secure Access Research and
7809 Development, Cisco Systems, Inc. for discovering this bug and
7810 preparing a fix. ([CVE-2012-0050])
7814 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7816 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7817 of the Vaudenay padding oracle attack on CBC mode encryption
7818 which enables an efficient plaintext recovery attack against
7819 the OpenSSL implementation of DTLS. Their attack exploits timing
7820 differences arising during decryption processing. A research
7821 paper describing this attack can be found at:
7822 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7823 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7824 Security Group at Royal Holloway, University of London
7825 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7826 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7827 for preparing the fix. ([CVE-2011-4108])
7829 *Robin Seggelmann, Michael Tuexen*
7831 * Clear bytes used for block padding of SSL 3.0 records.
7834 *Adam Langley (Google)*
7836 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7837 Kadianakis <desnacked@gmail.com> for discovering this issue and
7838 Adam Langley for preparing the fix. ([CVE-2011-4619])
7840 *Adam Langley (Google)*
7842 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7844 *Andrey Kulikov <amdeich@gmail.com>*
7846 * Prevent malformed RFC3779 data triggering an assertion failure.
7847 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7848 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7850 *Rob Austein <sra@hactrn.net>*
7852 * Improved PRNG seeding for VOS.
7854 *Paul Green <Paul.Green@stratus.com>*
7856 * Fix ssl_ciph.c set-up race.
7858 *Adam Langley (Google)*
7860 * Fix spurious failures in ecdsatest.c.
7862 *Emilia Käsper (Google)*
7864 * Fix the BIO_f_buffer() implementation (which was mixing different
7865 interpretations of the `..._len` fields).
7867 *Adam Langley (Google)*
7869 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7870 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7871 threads won't reuse the same blinding coefficients.
7873 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7874 lock to call BN_BLINDING_invert_ex, and avoids one use of
7875 BN_BLINDING_update for each BN_BLINDING structure (previously,
7876 the last update always remained unused).
7878 *Emilia Käsper (Google)*
7880 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7882 *Bob Buckholz (Google)*
7884 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7886 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7887 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7889 *Kaspar Brand <ossl@velox.ch>*
7891 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7892 for multi-threaded use of ECDH. ([CVE-2011-3210])
7894 *Adam Langley (Google)*
7896 * Fix x509_name_ex_d2i memory leak on bad inputs.
7900 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7901 signature public key algorithm by using OID xref utilities instead.
7902 Before this you could only use some ECC ciphersuites with SHA1 only.
7906 * Add protection against ECDSA timing attacks as mentioned in the paper
7907 by Billy Bob Brumley and Nicola Tuveri, see:
7908 <http://eprint.iacr.org/2011/232.pdf>
7910 *Billy Bob Brumley and Nicola Tuveri*
7912 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7914 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7916 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7918 * Fix bug in string printing code: if *any* escaping is enabled we must
7919 escape the escape character (backslash) or the resulting string is
7924 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7926 * Disable code workaround for ancient and obsolete Netscape browsers
7927 and servers: an attacker can use it in a ciphersuite downgrade attack.
7928 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7932 * Fixed J-PAKE implementation error, originally discovered by
7933 Sebastien Martini, further info and confirmation from Stefan
7934 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7938 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7940 * Fix extension code to avoid race conditions which can result in a buffer
7941 overrun vulnerability: resumed sessions must not be modified as they can
7942 be shared by multiple threads. CVE-2010-3864
7946 * Fix WIN32 build system to correctly link an ENGINE directory into
7951 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7953 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7956 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7958 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7960 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7961 context. The operation can be customised via the ctrl mechanism in
7962 case ENGINEs want to include additional functionality.
7966 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7970 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7971 output hashes compatible with older versions of OpenSSL.
7973 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7975 * Fix compression algorithm handling: if resuming a session use the
7976 compression algorithm of the resumed session instead of determining
7977 it from client hello again. Don't allow server to change algorithm.
7981 * Add load_crls() function to commands tidying load_certs() too. Add option
7982 to verify utility to allow additional CRLs to be included.
7986 * Update OCSP request code to permit adding custom headers to the request:
7987 some responders need this.
7991 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7994 *Julia Lawall <julia@diku.dk>*
7996 * Update verify callback code in `apps/s_cb.c` and `apps/verify.c`, it
7997 needlessly dereferenced structures, used obsolete functions and
7998 didn't handle all updated verify codes correctly.
8002 * Disable MD2 in the default configuration.
8006 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8007 indicate the initial BIO being pushed or popped. This makes it possible
8008 to determine whether the BIO is the one explicitly called or as a result
8009 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8010 it handles reference counts correctly and doesn't zero out the I/O bio
8011 when it is not being explicitly popped. WARNING: applications which
8012 included workarounds for the old buggy behaviour will need to be modified
8013 or they could free up already freed BIOs.
8017 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8018 renaming to all platforms (within the 0.9.8 branch, this was
8019 done conditionally on Netware platforms to avoid a name clash).
8021 *Guenter <lists@gknw.net>*
8023 * Add ECDHE and PSK support to DTLS.
8025 *Michael Tuexen <tuexen@fh-muenster.de>*
8027 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8032 * Add "missing" function EVP_MD_flags() (without this the only way to
8033 retrieve a digest flags is by accessing the structure directly. Update
8034 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8035 or cipher is registered as in the "from" argument. Print out all
8036 registered digests in the dgst usage message instead of manually
8037 attempting to work them out.
8041 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8042 this allows the use of compression and extensions. Change default cipher
8043 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8044 by default unless an application cipher string requests it.
8048 * Alter match criteria in PKCS12_parse(). It used to try to use local
8049 key ids to find matching certificates and keys but some PKCS#12 files
8050 don't follow the (somewhat unwritten) rules and this strategy fails.
8051 Now just gather all certificates together and the first private key
8052 then look for the first certificate that matches the key.
8056 * Support use of registered digest and cipher names for dgst and cipher
8057 commands instead of having to add each one as a special case. So now
8064 openssl dgst -sha256 foo
8066 and this works for ENGINE based algorithms too.
8070 * Update Gost ENGINE to support parameter files.
8072 *Victor B. Wagner <vitus@cryptocom.ru>*
8074 * Support GeneralizedTime in ca utility.
8076 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8078 * Enhance the hash format used for certificate directory links. The new
8079 form uses the canonical encoding (meaning equivalent names will work
8080 even if they aren't identical) and uses SHA1 instead of MD5. This form
8081 is incompatible with the older format and as a result c_rehash should
8082 be used to rebuild symbolic links.
8086 * Make PKCS#8 the default write format for private keys, replacing the
8087 traditional format. This form is standardised, more secure and doesn't
8088 include an implicit MD5 dependency.
8092 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8093 committed to OpenSSL should pass this lot as a minimum.
8097 * Add session ticket override functionality for use by EAP-FAST.
8099 *Jouni Malinen <j@w1.fi>*
8101 * Modify HMAC functions to return a value. Since these can be implemented
8102 in an ENGINE errors can occur.
8106 * Type-checked OBJ_bsearch_ex.
8110 * Type-checked OBJ_bsearch. Also some constification necessitated
8111 by type-checking. Still to come: TXT_DB, bsearch(?),
8112 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8117 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8118 seconds to a tm structure directly, instead of going through OS
8119 specific date routines. This avoids any issues with OS routines such
8120 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8121 and X509_time_adj_ex() to cover the extended range. The existing
8122 X509_time_adj() is still usable and will no longer have any date issues.
8126 * Delta CRL support. New use deltas option which will attempt to locate
8127 and search any appropriate delta CRLs available.
8129 This work was sponsored by Google.
8133 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8134 code and add additional score elements. Validate alternate CRL paths
8135 as part of the CRL checking and indicate a new error "CRL path validation
8136 error" in this case. Applications wanting additional details can use
8137 the verify callback and check the new "parent" field. If this is not
8138 NULL CRL path validation is taking place. Existing applications won't
8139 see this because it requires extended CRL support which is off by
8142 This work was sponsored by Google.
8146 * Support for freshest CRL extension.
8148 This work was sponsored by Google.
8152 * Initial indirect CRL support. Currently only supported in the CRLs
8153 passed directly and not via lookup. Process certificate issuer
8154 CRL entry extension and lookup CRL entries by bother issuer name
8155 and serial number. Check and process CRL issuer entry in IDP extension.
8157 This work was sponsored by Google.
8161 * Add support for distinct certificate and CRL paths. The CRL issuer
8162 certificate is validated separately in this case. Only enabled if
8163 an extended CRL support flag is set: this flag will enable additional
8164 CRL functionality in future.
8166 This work was sponsored by Google.
8170 * Add support for policy mappings extension.
8172 This work was sponsored by Google.
8176 * Fixes to pathlength constraint, self issued certificate handling,
8177 policy processing to align with RFC3280 and PKITS tests.
8179 This work was sponsored by Google.
8183 * Support for name constraints certificate extension. DN, email, DNS
8184 and URI types are currently supported.
8186 This work was sponsored by Google.
8190 * To cater for systems that provide a pointer-based thread ID rather
8191 than numeric, deprecate the current numeric thread ID mechanism and
8192 replace it with a structure and associated callback type. This
8193 mechanism allows a numeric "hash" to be extracted from a thread ID in
8194 either case, and on platforms where pointers are larger than 'long',
8195 mixing is done to help ensure the numeric 'hash' is usable even if it
8196 can't be guaranteed unique. The default mechanism is to use "&errno"
8197 as a pointer-based thread ID to distinguish between threads.
8199 Applications that want to provide their own thread IDs should now use
8200 CRYPTO_THREADID_set_callback() to register a callback that will call
8201 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8203 Note that ERR_remove_state() is now deprecated, because it is tied
8204 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8205 to free the current thread's error state should be replaced by
8206 ERR_remove_thread_state(NULL).
8208 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8209 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8210 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8211 application was previously providing a numeric thread callback that
8212 was inappropriate for distinguishing threads, then uniqueness might
8213 have been obtained with &errno that happened immediately in the
8214 intermediate development versions of OpenSSL; this is no longer the
8215 case, the numeric thread callback will now override the automatic use
8218 *Geoff Thorpe, with help from Bodo Moeller*
8220 * Initial support for different CRL issuing certificates. This covers a
8221 simple case where the self issued certificates in the chain exist and
8222 the real CRL issuer is higher in the existing chain.
8224 This work was sponsored by Google.
8228 * Removed effectively defunct crypto/store from the build.
8232 * Revamp of STACK to provide stronger type-checking. Still to come:
8233 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8234 ASN1_STRING, CONF_VALUE.
8238 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8239 RAM on SSL connections. This option can save about 34k per idle SSL.
8243 * Revamp of LHASH to provide stronger type-checking. Still to come:
8244 STACK, TXT_DB, bsearch, qsort.
8248 * Initial support for Cryptographic Message Syntax (aka CMS) based
8249 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8250 support for data, signedData, compressedData, digestedData and
8251 encryptedData, envelopedData types included. Scripts to check against
8252 RFC4134 examples draft and interop and consistency checks of many
8253 content types and variants.
8257 * Add options to enc utility to support use of zlib compression BIO.
8261 * Extend mk1mf to support importing of options and assembly language
8262 files from Configure script, currently only included in VC-WIN32.
8263 The assembly language rules can now optionally generate the source
8264 files from the associated perl scripts.
8268 * Implement remaining functionality needed to support GOST ciphersuites.
8269 Interop testing has been performed using CryptoPro implementations.
8271 *Victor B. Wagner <vitus@cryptocom.ru>*
8273 * s390x assembler pack.
8277 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8282 * Implement Opaque PRF Input TLS extension as specified in
8283 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8284 official specification yet and no extension type assignment by
8285 IANA exists, this extension (for now) will have to be explicitly
8286 enabled when building OpenSSL by providing the extension number
8287 to use. For example, specify an option
8289 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8291 to the "config" or "Configure" script to enable the extension,
8292 assuming extension number 0x9527 (which is a completely arbitrary
8293 and unofficial assignment based on the MD5 hash of the Internet
8294 Draft). Note that by doing so, you potentially lose
8295 interoperability with other TLS implementations since these might
8296 be using the same extension number for other purposes.
8298 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8299 opaque PRF input value to use in the handshake. This will create
8300 an internal copy of the length-'len' string at 'src', and will
8301 return non-zero for success.
8303 To get more control and flexibility, provide a callback function
8306 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8307 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8311 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8314 Callback function 'cb' will be called in handshakes, and is
8315 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8316 Argument 'arg' is for application purposes (the value as given to
8317 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8318 be provided to the callback function). The callback function
8319 has to return non-zero to report success: usually 1 to use opaque
8320 PRF input just if possible, or 2 to enforce use of the opaque PRF
8321 input. In the latter case, the library will abort the handshake
8322 if opaque PRF input is not successfully negotiated.
8324 Arguments 'peerinput' and 'len' given to the callback function
8325 will always be NULL and 0 in the case of a client. A server will
8326 see the client's opaque PRF input through these variables if
8327 available (NULL and 0 otherwise). Note that if the server
8328 provides an opaque PRF input, the length must be the same as the
8329 length of the client's opaque PRF input.
8331 Note that the callback function will only be called when creating
8332 a new session (session resumption can resume whatever was
8333 previously negotiated), and will not be called in SSL 2.0
8334 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8335 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8336 for applications that need to enforce opaque PRF input.
8340 * Update ssl code to support digests other than SHA1+MD5 for handshake
8343 *Victor B. Wagner <vitus@cryptocom.ru>*
8345 * Add RFC4507 support to OpenSSL. This includes the corrections in
8346 RFC4507bis. The encrypted ticket format is an encrypted encoded
8347 SSL_SESSION structure, that way new session features are automatically
8350 If a client application caches session in an SSL_SESSION structure
8351 support is transparent because tickets are now stored in the encoded
8354 The SSL_CTX structure automatically generates keys for ticket
8355 protection in servers so again support should be possible
8356 with no application modification.
8358 If a client or server wishes to disable RFC4507 support then the option
8359 SSL_OP_NO_TICKET can be set.
8361 Add a TLS extension debugging callback to allow the contents of any client
8362 or server extensions to be examined.
8364 This work was sponsored by Google.
8368 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8369 OpenSSL should now compile cleanly on gcc 4.2
8371 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8373 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8374 support including streaming MAC support: this is required for GOST
8375 ciphersuite support.
8377 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8379 * Add option -stream to use PKCS#7 streaming in smime utility. New
8380 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8381 to output in BER and PEM format.
8385 * Experimental support for use of HMAC via EVP_PKEY interface. This
8386 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8387 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8388 ENGINE support for HMAC keys which are unextractable. New -mac and
8389 -macopt options to dgst utility.
8393 * New option -sigopt to dgst utility. Update dgst to use
8394 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8395 alternative signing parameters such as X9.31 or PSS in the dgst
8400 * Change ssl_cipher_apply_rule(), the internal function that does
8401 the work each time a ciphersuite string requests enabling
8402 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8403 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8404 the order of disabled ciphersuites such that those ciphersuites
8405 that most recently went from enabled to disabled not only stay
8406 in order with respect to each other, but also have higher priority
8407 than other disabled ciphersuites the next time ciphersuites are
8410 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8411 the same ciphersuites as with "HIGH" alone, but in a specific
8412 order where the PSK ciphersuites come first (since they are the
8413 most recently disabled ciphersuites when "HIGH" is parsed).
8415 Also, change ssl_create_cipher_list() (using this new
8416 functionality) such that between otherwise identical
8417 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8422 * Change ssl_create_cipher_list() so that it automatically
8423 arranges the ciphersuites in reasonable order before starting
8424 to process the rule string. Thus, the definition for "DEFAULT"
8425 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8426 remains equivalent to `"AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH"`.
8427 This makes it much easier to arrive at a reasonable default order
8428 in applications for which anonymous ciphers are OK (meaning
8429 that you can't actually use DEFAULT).
8431 *Bodo Moeller; suggested by Victor Duchovni*
8433 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8434 processing) into multiple integers instead of setting
8435 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8436 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8437 (These masks as well as the individual bit definitions are hidden
8438 away into the non-exported interface ssl/ssl_locl.h, so this
8439 change to the definition of the SSL_CIPHER structure shouldn't
8440 affect applications.) This give us more bits for each of these
8441 categories, so there is no longer a need to coagulate AES128 and
8442 AES256 into a single algorithm bit, and to coagulate Camellia128
8443 and Camellia256 into a single algorithm bit, which has led to all
8446 Thus, among other things, the kludge introduced in 0.9.7m and
8447 0.9.8e for masking out AES256 independently of AES128 or masking
8448 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8450 With the change, we also introduce new ciphersuite aliases that
8451 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8456 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8457 Use the leftmost N bytes of the signature input if the input is
8458 larger than the prime q (with N being the size in bytes of q).
8462 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8463 it yet and it is largely untested.
8467 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8471 * Initial incomplete changes to avoid need for function casts in OpenSSL
8472 some compilers (gcc 4.2 and later) reject their use. Safestack is
8473 reimplemented. Update ASN1 to avoid use of legacy functions.
8477 * Win32/64 targets are linked with Winsock2.
8481 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8482 to external functions. This can be used to increase CRL handling
8483 efficiency especially when CRLs are very large by (for example) storing
8484 the CRL revoked certificates in a database.
8488 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8489 new CRLs added to a directory can be used. New command line option
8490 -verify_return_error to s_client and s_server. This causes real errors
8491 to be returned by the verify callback instead of carrying on no matter
8492 what. This reflects the way a "real world" verify callback would behave.
8496 * GOST engine, supporting several GOST algorithms and public key formats.
8497 Kindly donated by Cryptocom.
8501 * Partial support for Issuing Distribution Point CRL extension. CRLs
8502 partitioned by DP are handled but no indirect CRL or reason partitioning
8503 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8504 selected via a scoring technique which handles IDP and AKID in CRLs.
8508 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8509 will ultimately be used for all verify operations: this will remove the
8510 X509_STORE dependency on certificate verification and allow alternative
8511 lookup methods. X509_STORE based implementations of these two callbacks.
8515 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8516 Modify get_crl() to find a valid (unexpired) CRL if possible.
8520 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8521 this would be called X509_CRL_cmp() but that name is already used by
8522 a function that just compares CRL issuer names. Cache several CRL
8523 extensions in X509_CRL structure and cache CRLDP in X509.
8527 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8528 this maps equivalent X509_NAME structures into a consistent structure.
8529 Name comparison can then be performed rapidly using memcmp().
8533 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8538 * Allow digests to supply their own micalg string for S/MIME type using
8539 the ctrl EVP_MD_CTRL_MICALG.
8543 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8544 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8545 ctrl. It can then customise the structure before and/or after signing
8550 * New function OBJ_add_sigid() to allow application defined signature OIDs
8551 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8552 to free up any added signature OIDs.
8556 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8557 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8558 digest and cipher tables. New options added to openssl utility:
8559 list-message-digest-algorithms and list-cipher-algorithms.
8563 * Change the array representation of binary polynomials: the list
8564 of degrees of non-zero coefficients is now terminated with -1.
8565 Previously it was terminated with 0, which was also part of the
8566 value; thus, the array representation was not applicable to
8567 polynomials where t^0 has coefficient zero. This change makes
8568 the array representation useful in a more general context.
8572 * Various modifications and fixes to SSL/TLS cipher string
8573 handling. For ECC, the code now distinguishes between fixed ECDH
8574 with RSA certificates on the one hand and with ECDSA certificates
8575 on the other hand, since these are separate ciphersuites. The
8576 unused code for Fortezza ciphersuites has been removed.
8578 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8579 (not "ECDHE"). For consistency with the code for DH
8580 certificates, use of ECDH certificates is now considered ECDH
8581 authentication, not RSA or ECDSA authentication (the latter is
8582 merely the CA's signing algorithm and not actively used in the
8585 The temporary ciphersuite alias "ECCdraft" is no longer
8586 available, and ECC ciphersuites are no longer excluded from "ALL"
8587 and "DEFAULT". The following aliases now exist for RFC 4492
8588 ciphersuites, most of these by analogy with the DH case:
8590 kECDHr - ECDH cert, signed with RSA
8591 kECDHe - ECDH cert, signed with ECDSA
8592 kECDH - ECDH cert (signed with either RSA or ECDSA)
8593 kEECDH - ephemeral ECDH
8594 ECDH - ECDH cert or ephemeral ECDH
8600 AECDH - anonymous ECDH
8601 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8605 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8606 Use correct micalg parameters depending on digest(s) in signed message.
8610 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8611 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8615 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8616 an engine to register a method. Add ENGINE lookups for methods and
8617 functional reference processing.
8621 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8622 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8627 * New -resign option to smime utility. This adds one or more signers
8628 to an existing PKCS#7 signedData structure. Also -md option to use an
8629 alternative message digest algorithm for signing.
8633 * Tidy up PKCS#7 routines and add new functions to make it easier to
8634 create PKCS7 structures containing multiple signers. Update smime
8635 application to support multiple signers.
8639 * New -macalg option to pkcs12 utility to allow setting of an alternative
8644 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8645 Reorganize PBE internals to lookup from a static table using NIDs,
8646 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8647 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8648 PRF which will be automatically used with PBES2.
8652 * Replace the algorithm specific calls to generate keys in "req" with the
8657 * Update PKCS#7 enveloped data routines to use new API. This is now
8658 supported by any public key method supporting the encrypt operation. A
8659 ctrl is added to allow the public key algorithm to examine or modify
8660 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8665 * Add a ctrl to asn1 method to allow a public key algorithm to express
8666 a default digest type to use. In most cases this will be SHA1 but some
8667 algorithms (such as GOST) need to specify an alternative digest. The
8668 return value indicates how strong the preference is 1 means optional and
8669 2 is mandatory (that is it is the only supported type). Modify
8670 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8671 use the default md. Update openssl utilities to use the default digest
8672 type for signing if it is not explicitly indicated.
8676 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8677 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8678 signing method from the key type. This effectively removes the link
8679 between digests and public key types.
8683 * Add an OID cross reference table and utility functions. Its purpose is to
8684 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8685 rsaEncryption. This will allow some of the algorithm specific hackery
8686 needed to use the correct OID to be removed.
8690 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8691 structures for PKCS7_sign(). They are now set up by the relevant public
8696 * Add provisional EC pkey method with support for ECDSA and ECDH.
8700 * Add support for key derivation (agreement) in the API, DH method and
8705 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8706 public and private key formats. As a side effect these add additional
8707 command line functionality not previously available: DSA signatures can be
8708 generated and verified using pkeyutl and DH key support and generation in
8715 *Oliver Tappe <zooey@hirschkaefer.de>*
8717 * New make target "install_html_docs" installs HTML renditions of the
8720 *Oliver Tappe <zooey@hirschkaefer.de>*
8722 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8723 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8724 support key and parameter generation and add initial key generation
8725 functionality for RSA.
8729 * Add functions for main EVP_PKEY_method operations. The undocumented
8730 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8731 `EVP_PKEY_{encrypt,decrypt}_old`.
8735 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8736 key API, doesn't do much yet.
8740 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8741 public key algorithms. New option to openssl utility:
8742 "list-public-key-algorithms" to print out info.
8746 * Implement the Supported Elliptic Curves Extension for
8747 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8751 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8752 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8756 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8757 utilities such as rsa, dsa, dsaparam etc except they process any key
8762 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8763 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8764 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8769 * Initial support for pluggable public key ASN1.
8770 De-spaghettify the public key ASN1 handling. Move public and private
8771 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8772 algorithm specific handling to a single module within the relevant
8773 algorithm directory. Add functions to allow (near) opaque processing
8774 of public and private key structures.
8778 * Implement the Supported Point Formats Extension for
8779 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8783 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8784 for the psk identity [hint] and the psk callback functions to the
8785 SSL_SESSION, SSL and SSL_CTX structure.
8788 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8792 SSL_CTX_use_psk_identity_hint
8793 SSL_get_psk_identity_hint
8794 SSL_get_psk_identity
8795 SSL_use_psk_identity_hint
8797 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8799 * Add RFC 3161 compliant time stamp request creation, response generation
8800 and response verification functionality.
8802 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8804 * Add initial support for TLS extensions, specifically for the server_name
8805 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8806 have new members for a host name. The SSL data structure has an
8807 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8808 stored in that context to allow for session resumption, even after the
8809 SSL has been switched to a new SSL_CTX in reaction to a client's
8810 server_name extension.
8812 New functions (subject to change):
8814 SSL_get_servername()
8815 SSL_get_servername_type()
8818 New CTRL codes and macros (subject to change):
8820 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8821 - SSL_CTX_set_tlsext_servername_callback()
8822 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8823 - SSL_CTX_set_tlsext_servername_arg()
8824 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8826 openssl s_client has a new '-servername ...' option.
8828 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8829 '-key2 ...', '-servername_fatal' (subject to change). This allows
8830 testing the HostName extension for a specific single host name ('-cert'
8831 and '-key' remain fallbacks for handshakes without HostName
8832 negotiation). If the unrecognized_name alert has to be sent, this by
8833 default is a warning; it becomes fatal with the '-servername_fatal'
8836 *Peter Sylvester, Remy Allais, Christophe Renou*
8838 * Whirlpool hash implementation is added.
8842 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8843 bn(64,32). Because of instruction set limitations it doesn't have
8844 any negative impact on performance. This was done mostly in order
8845 to make it possible to share assembler modules, such as bn_mul_mont
8846 implementations, between 32- and 64-bit builds without hassle.
8850 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8851 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8856 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8857 dedicated Montgomery multiplication procedure, is introduced.
8858 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8859 "64-bit" performance on certain 32-bit targets.
8863 * New option SSL_OP_NO_COMP to disable use of compression selectively
8864 in SSL structures. New SSL ctrl to set maximum send fragment size.
8865 Save memory by setting the I/O buffer sizes dynamically instead of
8866 using the maximum available value.
8870 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8871 in addition to the text details.
8875 * Very, very preliminary EXPERIMENTAL support for printing of general
8876 ASN1 structures. This currently produces rather ugly output and doesn't
8877 handle several customised structures at all.
8881 * Integrated support for PVK file format and some related formats such
8882 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8883 these in the 'rsa' and 'dsa' utilities.
8887 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8891 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8892 place for the (very old) "NETSCAPE" format certificates which are now
8893 handled using new ASN1 code equivalents.
8897 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8898 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8899 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8903 * Modify CRL distribution points extension code to print out previously
8904 unsupported fields. Enhance extension setting code to allow setting of
8909 * Add print and set support for Issuing Distribution Point CRL extension.
8913 * Change 'Configure' script to enable Camellia by default.
8920 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8922 * When rejecting SSL/TLS records due to an incorrect version number, never
8923 update s->server with a new major version number. As of
8924 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8925 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8926 the previous behavior could result in a read attempt at NULL when
8927 receiving specific incorrect SSL/TLS records once record payload
8928 protection is active. ([CVE-2010-0740])
8930 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8932 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8933 could be crashed if the relevant tables were not present (e.g. chrooted).
8935 *Tomas Hoger <thoger@redhat.com>*
8937 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8939 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8941 *Martin Olsson, Neel Mehta*
8943 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8944 accommodate for stack sorting, always a write lock!).
8948 * On some versions of WIN32 Heap32Next is very slow. This can cause
8949 excessive delays in the RAND_poll(): over a minute. As a workaround
8950 include a time check in the inner Heap32Next loop too.
8954 * The code that handled flushing of data in SSL/TLS originally used the
8955 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8956 the problem outlined in PR#1949. The fix suggested there however can
8957 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8958 of Apache). So instead simplify the code to flush unconditionally.
8959 This should be fine since flushing with no data to flush is a no op.
8963 * Handle TLS versions 2.0 and later properly and correctly use the
8964 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8965 off ancient servers have a habit of sticking around for a while...
8969 * Modify compression code so it frees up structures without using the
8970 ex_data callbacks. This works around a problem where some applications
8971 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8972 restarting) then use compression (e.g. SSL with compression) later.
8973 This results in significant per-connection memory leaks and
8974 has caused some security issues including CVE-2008-1678 and
8979 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8980 change when encrypting or decrypting.
8984 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8985 connect and renegotiate with servers which do not support RI.
8986 Until RI is more widely deployed this option is enabled by default.
8990 * Add "missing" ssl ctrls to clear options and mode.
8994 * If client attempts to renegotiate and doesn't support RI respond with
8995 a no_renegotiation alert as required by RFC5746. Some renegotiating
8996 TLS clients will continue a connection gracefully when they receive
8997 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8998 waiting for a server hello which it will never receive. Now we treat a
8999 received no_renegotiation alert as a fatal error. This is because
9000 applications requesting a renegotiation might well expect it to succeed
9001 and would have no code in place to handle the server denying it so the
9002 only safe thing to do is to terminate the connection.
9006 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9007 peer supports secure renegotiation and 0 otherwise. Print out peer
9008 renegotiation support in s_client/s_server.
9012 * Replace the highly broken and deprecated SPKAC certification method with
9013 the updated NID creation version. This should correctly handle UTF8.
9017 * Implement RFC5746. Re-enable renegotiation but require the extension
9018 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9019 turns out to be a bad idea. It has been replaced by
9020 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9021 SSL_CTX_set_options(). This is really not recommended unless you
9022 know what you are doing.
9024 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9026 * Fixes to stateless session resumption handling. Use initial_ctx when
9027 issuing and attempting to decrypt tickets in case it has changed during
9028 servername handling. Use a non-zero length session ID when attempting
9029 stateless session resumption: this makes it possible to determine if
9030 a resumption has occurred immediately after receiving server hello
9031 (several places in OpenSSL subtly assume this) instead of later in
9036 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9037 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9038 fixes for a few places where the return code is not checked
9041 *Julia Lawall <julia@diku.dk>*
9043 * Add --strict-warnings option to Configure script to include devteam
9044 warnings in other configurations.
9048 * Add support for --libdir option and LIBDIR variable in makefiles. This
9049 makes it possible to install openssl libraries in locations which
9050 have names other than "lib", for example "/usr/lib64" which some
9053 *Steve Henson, based on patch from Jeremy Utley*
9055 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9056 X690 8.9.12 and can produce some misleading textual output of OIDs.
9058 *Steve Henson, reported by Dan Kaminsky*
9060 * Delete MD2 from algorithm tables. This follows the recommendation in
9061 several standards that it is not used in new applications due to
9062 several cryptographic weaknesses. For binary compatibility reasons
9063 the MD2 API is still compiled in by default.
9067 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9072 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9073 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9076 *Guenter <lists@gknw.net>*
9078 * Fix the server certificate chain building code to use X509_verify_cert(),
9079 it used to have an ad-hoc builder which was unable to cope with anything
9080 other than a simple chain.
9082 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9084 * Don't check self signed certificate signatures in X509_verify_cert()
9085 by default (a flag can override this): it just wastes time without
9086 adding any security. As a useful side effect self signed root CAs
9087 with non-FIPS digests are now usable in FIPS mode.
9091 * In dtls1_process_out_of_seq_message() the check if the current message
9092 is already buffered was missing. For every new message was memory
9093 allocated, allowing an attacker to perform an denial of service attack
9094 with sending out of seq handshake messages until there is no memory
9095 left. Additionally every future message was buffered, even if the
9096 sequence number made no sense and would be part of another handshake.
9097 So only messages with sequence numbers less than 10 in advance will be
9098 buffered. ([CVE-2009-1378])
9100 *Robin Seggelmann, discovered by Daniel Mentz*
9102 * Records are buffered if they arrive with a future epoch to be
9103 processed after finishing the corresponding handshake. There is
9104 currently no limitation to this buffer allowing an attacker to perform
9105 a DOS attack with sending records with future epochs until there is no
9106 memory left. This patch adds the pqueue_size() function to determine
9107 the size of a buffer and limits the record buffer to 100 entries.
9110 *Robin Seggelmann, discovered by Daniel Mentz*
9112 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9113 parent structure is freed. ([CVE-2009-1379])
9117 * Handle non-blocking I/O properly in SSL_shutdown() call.
9119 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9121 * Add `2.5.4.*` OIDs
9123 *Ilya O. <vrghost@gmail.com>*
9125 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9127 * Disable renegotiation completely - this fixes a severe security
9128 problem ([CVE-2009-3555]) at the cost of breaking all
9129 renegotiation. Renegotiation can be re-enabled by setting
9130 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9131 run-time. This is really not recommended unless you know what
9136 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9138 * Don't set val to NULL when freeing up structures, it is freed up by
9139 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9140 zeroing past the valid field. ([CVE-2009-0789])
9142 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9144 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9145 checked correctly. This would allow some invalid signed attributes to
9146 appear to verify correctly. ([CVE-2009-0591])
9148 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9150 * Reject UniversalString and BMPString types with invalid lengths. This
9151 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9152 a legal length. ([CVE-2009-0590])
9156 * Set S/MIME signing as the default purpose rather than setting it
9157 unconditionally. This allows applications to override it at the store
9162 * Permit restricted recursion of ASN1 strings. This is needed in practice
9163 to handle some structures.
9167 * Improve efficiency of mem_gets: don't search whole buffer each time
9170 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9172 * New -hex option for openssl rand.
9176 * Print out UTF8String and NumericString when parsing ASN1.
9180 * Support NumericString type for name components.
9184 * Allow CC in the environment to override the automatically chosen
9185 compiler. Note that nothing is done to ensure flags work with the
9190 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9192 * Properly check EVP_VerifyFinal() and similar return values
9195 *Ben Laurie, Bodo Moeller, Google Security Team*
9197 * Enable TLS extensions by default.
9201 * Allow the CHIL engine to be loaded, whether the application is
9202 multithreaded or not. (This does not release the developer from the
9203 obligation to set up the dynamic locking callbacks.)
9205 *Sander Temme <sander@temme.net>*
9207 * Use correct exit code if there is an error in dgst command.
9209 *Steve Henson; problem pointed out by Roland Dirlewanger*
9211 * Tweak Configure so that you need to say "experimental-jpake" to enable
9212 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9216 * Add experimental JPAKE support, including demo authentication in
9217 s_client and s_server.
9221 * Set the comparison function in v3_addr_canonize().
9223 *Rob Austein <sra@hactrn.net>*
9225 * Add support for XMPP STARTTLS in s_client.
9227 *Philip Paeps <philip@freebsd.org>*
9229 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9230 to ensure that even with this option, only ciphersuites in the
9231 server's preference list will be accepted. (Note that the option
9232 applies only when resuming a session, so the earlier behavior was
9233 just about the algorithm choice for symmetric cryptography.)
9237 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9239 * Fix NULL pointer dereference if a DTLS server received
9240 ChangeCipherSpec as first record ([CVE-2009-1386]).
9244 * Fix a state transition in s3_srvr.c and d1_srvr.c
9245 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9249 * The fix in 0.9.8c that supposedly got rid of unsafe
9250 double-checked locking was incomplete for RSA blinding,
9251 addressing just one layer of what turns out to have been
9252 doubly unsafe triple-checked locking.
9254 So now fix this for real by retiring the MONT_HELPER macro
9255 in crypto/rsa/rsa_eay.c.
9257 *Bodo Moeller; problem pointed out by Marius Schilder*
9259 * Various precautionary measures:
9261 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9263 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9264 (NB: This would require knowledge of the secret session ticket key
9265 to exploit, in which case you'd be SOL either way.)
9267 - Change bn_nist.c so that it will properly handle input BIGNUMs
9268 outside the expected range.
9270 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9273 *Neel Mehta, Bodo Moeller*
9275 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9276 the load fails. Useful for distros.
9278 *Ben Laurie and the FreeBSD team*
9280 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9284 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9288 * Expand ENGINE to support engine supplied SSL client certificate functions.
9290 This work was sponsored by Logica.
9294 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9295 keystores. Support for SSL/TLS client authentication too.
9296 Not compiled unless enable-capieng specified to Configure.
9298 This work was sponsored by Logica.
9302 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9303 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9304 attribute creation routines such as certificate requests and PKCS#12
9309 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9311 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9312 handshake which could lead to a client crash as found using the
9313 Codenomicon TLS test suite ([CVE-2008-1672])
9315 *Steve Henson, Mark Cox*
9317 * Fix double free in TLS server name extensions which could lead to
9318 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9322 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9324 Clear the error queue to ensure that error entries left from
9325 older function calls do not interfere with the correct operation.
9327 *Lutz Jaenicke, Erik de Castro Lopo*
9329 * Remove root CA certificates of commercial CAs:
9331 The OpenSSL project does not recommend any specific CA and does not
9332 have any policy with respect to including or excluding any CA.
9333 Therefore it does not make any sense to ship an arbitrary selection
9334 of root CA certificates with the OpenSSL software.
9338 * RSA OAEP patches to fix two separate invalid memory reads.
9339 The first one involves inputs when 'lzero' is greater than
9340 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9341 before the beginning of from). The second one involves inputs where
9342 the 'db' section contains nothing but zeroes (there is a one-byte
9343 invalid read after the end of 'db').
9345 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9347 * Partial backport from 0.9.9-dev:
9349 Introduce bn_mul_mont (dedicated Montgomery multiplication
9350 procedure) as a candidate for BIGNUM assembler implementation.
9351 While 0.9.9-dev uses assembler for various architectures, only
9352 x86_64 is available by default here in the 0.9.8 branch, and
9353 32-bit x86 is available through a compile-time setting.
9355 To try the 32-bit x86 assembler implementation, use Configure
9356 option "enable-montasm" (which exists only for this backport).
9358 As "enable-montasm" for 32-bit x86 disclaims code stability
9359 anyway, in this constellation we activate additional code
9360 backported from 0.9.9-dev for further performance improvements,
9361 namely BN_from_montgomery_word. (To enable this otherwise,
9362 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9364 *Andy Polyakov (backport partially by Bodo Moeller)*
9366 * Add TLS session ticket callback. This allows an application to set
9367 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9368 values. This is useful for key rollover for example where several key
9369 sets may exist with different names.
9373 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9374 This was broken until now in 0.9.8 releases, such that the only way
9375 a registered ENGINE could be used (assuming it initialises
9376 successfully on the host) was to explicitly set it as the default
9377 for the relevant algorithms. This is in contradiction with 0.9.7
9378 behaviour and the documentation. With this fix, when an ENGINE is
9379 registered into a given algorithm's table of implementations, the
9380 'uptodate' flag is reset so that auto-discovery will be used next
9381 time a new context for that algorithm attempts to select an
9384 *Ian Lister (tweaked by Geoff Thorpe)*
9386 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9387 implementation in the following ways:
9389 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9392 Lack of BER streaming support means one pass streaming processing is
9393 only supported if data is detached: setting the streaming flag is
9394 ignored for embedded content.
9396 CMS support is disabled by default and must be explicitly enabled
9397 with the enable-cms configuration option.
9401 * Update the GMP engine glue to do direct copies between BIGNUM and
9402 mpz_t when openssl and GMP use the same limb size. Otherwise the
9403 existing "conversion via a text string export" trick is still used.
9405 *Paul Sheer <paulsheer@gmail.com>*
9407 * Zlib compression BIO. This is a filter BIO which compressed and
9408 uncompresses any data passed through it.
9412 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9413 RFC3394 compatible AES key wrapping.
9417 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9418 sets string data without copying. X509_ALGOR_set0() and
9419 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9420 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9421 from an X509_ATTRIBUTE structure optionally checking it occurs only
9422 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9427 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9428 to get the expected BN_FLG_CONSTTIME behavior.
9430 *Bodo Moeller (Google)*
9434 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9435 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9436 - added some more tests to do_tests.pl
9437 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9438 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9439 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9440 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9441 - various changes to netware.pl to enable gcc-cross builds on Win32
9443 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9444 - various changes to fix missing prototype warnings
9445 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9446 - added AES, WHIRLPOOL and CPUID assembler code to build files
9447 - added missing AES assembler make rules to mk1mf.pl
9448 - fixed order of includes in `apps/ocsp.c` so that `e_os.h` settings apply
9450 *Guenter Knauf <eflash@gmx.net>*
9452 * Implement certificate status request TLS extension defined in RFC3546.
9453 A client can set the appropriate parameters and receive the encoded
9454 OCSP response via a callback. A server can query the supplied parameters
9455 and set the encoded OCSP response in the callback. Add simplified examples
9456 to s_client and s_server.
9460 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9463 + Binary incompatibility of ssl_ctx_st structure
9464 + DTLS interoperation with non-compliant servers
9465 + Don't call get_session_cb() without proposed session
9466 + Fix ia64 assembler code
9468 *Andy Polyakov, Steve Henson*
9470 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9472 * DTLS Handshake overhaul. There were longstanding issues with
9473 OpenSSL DTLS implementation, which were making it impossible for
9474 RFC 4347 compliant client to communicate with OpenSSL server.
9475 Unfortunately just fixing these incompatibilities would "cut off"
9476 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9477 server keeps tolerating non RFC compliant syntax. The opposite is
9478 not true, 0.9.8f client can not communicate with earlier server.
9479 This update even addresses CVE-2007-4995.
9483 * Changes to avoid need for function casts in OpenSSL: some compilers
9484 (gcc 4.2 and later) reject their use.
9485 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9488 * Add RFC4507 support to OpenSSL. This includes the corrections in
9489 RFC4507bis. The encrypted ticket format is an encrypted encoded
9490 SSL_SESSION structure, that way new session features are automatically
9493 If a client application caches session in an SSL_SESSION structure
9494 support is transparent because tickets are now stored in the encoded
9497 The SSL_CTX structure automatically generates keys for ticket
9498 protection in servers so again support should be possible
9499 with no application modification.
9501 If a client or server wishes to disable RFC4507 support then the option
9502 SSL_OP_NO_TICKET can be set.
9504 Add a TLS extension debugging callback to allow the contents of any client
9505 or server extensions to be examined.
9507 This work was sponsored by Google.
9511 * Add initial support for TLS extensions, specifically for the server_name
9512 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9513 have new members for a host name. The SSL data structure has an
9514 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9515 stored in that context to allow for session resumption, even after the
9516 SSL has been switched to a new SSL_CTX in reaction to a client's
9517 server_name extension.
9519 New functions (subject to change):
9521 SSL_get_servername()
9522 SSL_get_servername_type()
9525 New CTRL codes and macros (subject to change):
9527 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9528 - SSL_CTX_set_tlsext_servername_callback()
9529 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9530 - SSL_CTX_set_tlsext_servername_arg()
9531 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9533 openssl s_client has a new '-servername ...' option.
9535 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9536 '-key2 ...', '-servername_fatal' (subject to change). This allows
9537 testing the HostName extension for a specific single host name ('-cert'
9538 and '-key' remain fallbacks for handshakes without HostName
9539 negotiation). If the unrecognized_name alert has to be sent, this by
9540 default is a warning; it becomes fatal with the '-servername_fatal'
9543 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9545 * Add AES and SSE2 assembly language support to VC++ build.
9549 * Mitigate attack on final subtraction in Montgomery reduction.
9553 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9554 (which previously caused an internal error).
9558 * Squeeze another 10% out of IGE mode when in != out.
9562 * AES IGE mode speedup.
9564 *Dean Gaudet (Google)*
9566 * Add the Korean symmetric 128-bit cipher SEED (see
9567 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9568 add SEED ciphersuites from RFC 4162:
9570 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9571 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9572 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9573 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9575 To minimize changes between patchlevels in the OpenSSL 0.9.8
9576 series, SEED remains excluded from compilation unless OpenSSL
9577 is configured with 'enable-seed'.
9579 *KISA, Bodo Moeller*
9581 * Mitigate branch prediction attacks, which can be practical if a
9582 single processor is shared, allowing a spy process to extract
9583 information. For detailed background information, see
9584 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9585 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9586 and Necessary Software Countermeasures"). The core of the change
9587 are new versions BN_div_no_branch() and
9588 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9589 respectively, which are slower, but avoid the security-relevant
9590 conditional branches. These are automatically called by BN_div()
9591 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9592 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9593 remove a conditional branch.
9595 BN_FLG_CONSTTIME is the new name for the previous
9596 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9597 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9598 in the exponent causes BN_mod_exp_mont() to use the alternative
9599 implementation in BN_mod_exp_mont_consttime().) The old name
9600 remains as a deprecated alias.
9602 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9603 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9604 constant-time implementations for more than just exponentiation.
9605 Here too the old name is kept as a deprecated alias.
9607 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9608 the BN_BLINDING structure gets an independent copy of the
9609 modulus. This means that the previous `BIGNUM *m` argument to
9610 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9611 essentially becomes `const BIGNUM *m`, although we can't actually
9612 change this in the header file before 0.9.9. It allows
9613 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9614 enable BN_FLG_CONSTTIME.
9616 *Matthew D Wood (Intel Corp)*
9618 * In the SSL/TLS server implementation, be strict about session ID
9619 context matching (which matters if an application uses a single
9620 external cache for different purposes). Previously,
9621 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9622 set. This did ensure strict client verification, but meant that,
9623 with applications using a single external cache for quite
9624 different requirements, clients could circumvent ciphersuite
9625 restrictions for a given session ID context by starting a session
9626 in a different context.
9630 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9631 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9632 authentication-only ciphersuites.
9636 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9637 not complete and could lead to a possible single byte overflow
9638 ([CVE-2007-5135]) [Ben Laurie]
9640 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9642 * Since AES128 and AES256 (and similarly Camellia128 and
9643 Camellia256) share a single mask bit in the logic of
9644 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9645 kludge to work properly if AES128 is available and AES256 isn't
9646 (or if Camellia128 is available and Camellia256 isn't).
9650 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9651 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9652 When a point or a seed is encoded in a BIT STRING, we need to
9653 prevent the removal of trailing zero bits to get the proper DER
9654 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9655 of a NamedBitList, for which trailing 0 bits need to be removed.)
9659 * Have SSL/TLS server implementation tolerate "mismatched" record
9660 protocol version while receiving ClientHello even if the
9661 ClientHello is fragmented. (The server can't insist on the
9662 particular protocol version it has chosen before the ServerHello
9663 message has informed the client about his choice.)
9667 * Add RFC 3779 support.
9669 *Rob Austein for ARIN, Ben Laurie*
9671 * Load error codes if they are not already present instead of using a
9672 static variable. This allows them to be cleanly unloaded and reloaded.
9673 Improve header file function name parsing.
9677 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9678 or CAPABILITY handshake as required by RFCs.
9682 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9684 * Introduce limits to prevent malicious keys being able to
9685 cause a denial of service. ([CVE-2006-2940])
9687 *Steve Henson, Bodo Moeller*
9689 * Fix ASN.1 parsing of certain invalid structures that can result
9690 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9692 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9693 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9695 * Fix SSL client code which could crash if connecting to a
9696 malicious SSLv2 server. ([CVE-2006-4343])
9698 *Tavis Ormandy and Will Drewry, Google Security Team*
9700 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9701 match only those. Before that, "AES256-SHA" would be interpreted
9702 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9703 the same strength classification in 0.9.7h) as we currently only
9704 have a single AES bit in the ciphersuite description bitmap.
9705 That change, however, also applied to ciphersuite strings such as
9706 "RC4-MD5" that intentionally matched multiple ciphersuites --
9707 namely, SSL 2.0 ciphersuites in addition to the more common ones
9708 from SSL 3.0/TLS 1.0.
9710 So we change the selection algorithm again: Naming an explicit
9711 ciphersuite selects this one ciphersuite, and any other similar
9712 ciphersuite (same bitmap) from *other* protocol versions.
9713 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9714 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9716 Since SSL 2.0 does not have any ciphersuites for which the
9717 128/256 bit distinction would be relevant, this works for now.
9718 The proper fix will be to use different bits for AES128 and
9719 AES256, which would have avoided the problems from the beginning;
9720 however, bits are scarce, so we can only do this in a new release
9721 (not just a patchlevel) when we can change the SSL_CIPHER
9722 definition to split the single 'unsigned long mask' bitmap into
9723 multiple values to extend the available space.
9727 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9729 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9730 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9732 * Add AES IGE and biIGE modes.
9736 * Change the Unix randomness entropy gathering to use poll() when
9737 possible instead of select(), since the latter has some
9738 undesirable limitations.
9740 *Darryl Miles via Richard Levitte and Bodo Moeller*
9742 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9743 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9744 cannot be implicitly activated as part of, e.g., the "AES" alias.
9745 However, please upgrade to OpenSSL 0.9.9[-dev] for
9746 non-experimental use of the ECC ciphersuites to get TLS extension
9747 support, which is required for curve and point format negotiation
9748 to avoid potential handshake problems.
9752 * Disable rogue ciphersuites:
9754 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9755 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9756 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9758 The latter two were purportedly from
9759 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9762 Also deactivate the remaining ciphersuites from
9763 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9764 unofficial, and the ID has long expired.
9768 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9769 dual-core machines) and other potential thread-safety issues.
9773 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9774 versions), which is now available for royalty-free use
9775 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9776 Also, add Camellia TLS ciphersuites from RFC 4132.
9778 To minimize changes between patchlevels in the OpenSSL 0.9.8
9779 series, Camellia remains excluded from compilation unless OpenSSL
9780 is configured with 'enable-camellia'.
9784 * Disable the padding bug check when compression is in use. The padding
9785 bug check assumes the first packet is of even length, this is not
9786 necessarily true if compression is enabled and can result in false
9787 positives causing handshake failure. The actual bug test is ancient
9788 code so it is hoped that implementations will either have fixed it by
9789 now or any which still have the bug do not support compression.
9793 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9795 * When applying a cipher rule check to see if string match is an explicit
9796 cipher suite and only match that one cipher suite if it is.
9800 * Link in manifests for VC++ if needed.
9802 *Austin Ziegler <halostatue@gmail.com>*
9804 * Update support for ECC-based TLS ciphersuites according to
9805 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9806 TLS extensions, which are supported starting with the 0.9.9
9807 branch, not in the OpenSSL 0.9.8 branch).
9811 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9812 opaque EVP_CIPHER_CTX handling.
9816 * Fixes and enhancements to zlib compression code. We now only use
9817 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9818 to conform with the standards mentioned here:
9819 <http://www.zlib.net/DLL_FAQ.txt>
9820 Static zlib linking now works on Windows and the new --with-zlib-include
9821 --with-zlib-lib options to Configure can be used to supply the location
9822 of the headers and library. Gracefully handle case where zlib library
9827 * Several fixes and enhancements to the OID generation code. The old code
9828 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9829 handle numbers larger than ULONG_MAX, truncated printing and had a
9830 non standard OBJ_obj2txt() behaviour.
9834 * Add support for building of engines under engine/ as shared libraries
9835 under VC++ build system.
9839 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9840 Hopefully, we will not see any false combination of paths any more.
9844 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9846 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9847 (part of SSL_OP_ALL). This option used to disable the
9848 countermeasure against man-in-the-middle protocol-version
9849 rollback in the SSL 2.0 server implementation, which is a bad
9850 idea. ([CVE-2005-2969])
9852 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9853 for Information Security, National Institute of Advanced Industrial
9854 Science and Technology [AIST], Japan)*
9856 * Add two function to clear and return the verify parameter flags.
9860 * Keep cipherlists sorted in the source instead of sorting them at
9861 runtime, thus removing the need for a lock.
9865 * Avoid some small subgroup attacks in Diffie-Hellman.
9867 *Nick Mathewson and Ben Laurie*
9869 * Add functions for well-known primes.
9873 * Extended Windows CE support.
9875 *Satoshi Nakamura and Andy Polyakov*
9877 * Initialize SSL_METHOD structures at compile time instead of during
9878 runtime, thus removing the need for a lock.
9882 * Make PKCS7_decrypt() work even if no certificate is supplied by
9883 attempting to decrypt each encrypted key in turn. Add support to
9888 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9890 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9893 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9897 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9898 key into the same file any more.
9902 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9906 * Add -utf8 command line and config file option to 'ca'.
9908 *Stefan <stf@udoma.org*
9910 * Removed the macro des_crypt(), as it seems to conflict with some
9911 libraries. Use DES_crypt().
9915 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9916 involves renaming the source and generated shared-libs for
9917 both. The engines will accept the corrected or legacy ids
9918 ('ncipher' and '4758_cca' respectively) when binding. NB,
9919 this only applies when building 'shared'.
9921 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9923 * Add attribute functions to EVP_PKEY structure. Modify
9924 PKCS12_create() to recognize a CSP name attribute and
9925 use it. Make -CSP option work again in pkcs12 utility.
9929 * Add new functionality to the bn blinding code:
9930 - automatic re-creation of the BN_BLINDING parameters after
9931 a fixed number of uses (currently 32)
9932 - add new function for parameter creation
9933 - introduce flags to control the update behaviour of the
9934 BN_BLINDING parameters
9935 - hide BN_BLINDING structure
9936 Add a second BN_BLINDING slot to the RSA structure to improve
9937 performance when a single RSA object is shared among several
9942 * Add support for DTLS.
9944 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9946 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9947 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9951 * Remove buggy and incomplete DH cert support from
9952 ssl/ssl_rsa.c and ssl/s3_both.c
9956 * Use SHA-1 instead of MD5 as the default digest algorithm for
9957 the `apps/openssl` commands.
9961 * Compile clean with "-Wall -Wmissing-prototypes
9962 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9963 DEBUG_SAFESTACK must also be set.
9967 * Change ./Configure so that certain algorithms can be disabled by default.
9968 The new counterpiece to "no-xxx" is "enable-xxx".
9970 The patented RC5 and MDC2 algorithms will now be disabled unless
9971 "enable-rc5" and "enable-mdc2", respectively, are specified.
9973 (IDEA remains enabled despite being patented. This is because IDEA
9974 is frequently required for interoperability, and there is no license
9975 fee for non-commercial use. As before, "no-idea" can be used to
9976 avoid this algorithm.)
9980 * Add processing of proxy certificates (see RFC 3820). This work was
9981 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9982 EGEE (Enabling Grids for E-science in Europe).
9986 * RC4 performance overhaul on modern architectures/implementations, such
9987 as Intel P4, IA-64 and AMD64.
9991 * New utility extract-section.pl. This can be used specify an alternative
9992 section number in a pod file instead of having to treat each file as
9993 a separate case in Makefile. This can be done by adding two lines to the
9996 =for comment openssl_section:XXX
9998 The blank line is mandatory.
10002 * New arguments -certform, -keyform and -pass for s_client and s_server
10003 to allow alternative format key and certificate files and passphrase
10008 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10009 update associated structures and add various utility functions.
10011 Add new policy related verify parameters, include policy checking in
10012 standard verify code. Enhance 'smime' application with extra parameters
10013 to support policy checking and print out.
10017 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10018 Nehemiah processors. These extensions support AES encryption in hardware
10019 as well as RNG (though RNG support is currently disabled).
10021 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10023 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10027 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10029 *Andy Polyakov and a number of other people*
10031 * Improved PowerPC platform support. Most notably BIGNUM assembler
10032 implementation contributed by IBM.
10034 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10036 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10037 exponent rather than 'unsigned long'. There is a corresponding change to
10038 the new 'rsa_keygen' element of the RSA_METHOD structure.
10040 *Jelte Jansen, Geoff Thorpe*
10042 * Functionality for creating the initial serial number file is now
10043 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10045 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10046 number file to 1, which is bound to cause problems. To avoid
10047 the problems while respecting compatibility between different 0.9.7
10048 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10049 CA.pl for serial number initialization. With the new release 0.9.8,
10050 we can fix the problem directly in the 'ca' utility.)
10054 * Reduced header interdependencies by declaring more opaque objects in
10055 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10056 give fewer recursive includes, which could break lazy source code - so
10057 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10058 developers should define this symbol when building and using openssl to
10059 ensure they track the recommended behaviour, interfaces, [etc], but
10060 backwards-compatible behaviour prevails when this isn't defined.
10064 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10068 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10069 This will generate a random key of the appropriate length based on the
10070 cipher context. The EVP_CIPHER can provide its own random key generation
10071 routine to support keys of a specific form. This is used in the des and
10072 3des routines to generate a key of the correct parity. Update S/MIME
10073 code to use new functions and hence generate correct parity DES keys.
10074 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10075 valid (weak or incorrect parity).
10079 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10080 as looking them up. This is useful when the verified structure may contain
10081 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10082 present unless the new PKCS7_NO_CRL flag is asserted.
10086 * Extend ASN1 oid configuration module. It now additionally accepts the
10089 shortName = some long name, 1.2.3.4
10093 * Reimplemented the BN_CTX implementation. There is now no more static
10094 limitation on the number of variables it can handle nor the depth of the
10095 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10096 information can now expand as required, and rather than having a single
10097 static array of bignums, BN_CTX now uses a linked-list of such arrays
10098 allowing it to expand on demand whilst maintaining the usefulness of
10099 BN_CTX's "bundling".
10103 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10104 to allow all RSA operations to function using a single BN_CTX.
10108 * Preliminary support for certificate policy evaluation and checking. This
10109 is initially intended to pass the tests outlined in "Conformance Testing
10110 of Relying Party Client Certificate Path Processing Logic" v1.07.
10114 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10115 remained unused and not that useful. A variety of other little bignum
10116 tweaks and fixes have also been made continuing on from the audit (see
10121 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10122 associated ASN1, EVP and SSL functions and old ASN1 macros.
10126 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10127 and this should never fail. So the return value from the use of
10128 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10129 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10133 * BN_CTX_get() should return zero-valued bignums, providing the same
10134 initialised value as BN_new().
10136 *Geoff Thorpe, suggested by Ulf Möller*
10138 * Support for inhibitAnyPolicy certificate extension.
10142 * An audit of the BIGNUM code is underway, for which debugging code is
10143 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10144 is considered valid when processing BIGNUMs, and causes execution to
10145 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10146 further steps are taken to deliberately pollute unused data in BIGNUM
10147 structures to try and expose faulty code further on. For now, openssl will
10148 (in its default mode of operation) continue to tolerate the inconsistent
10149 forms that it has tolerated in the past, but authors and packagers should
10150 consider trying openssl and their own applications when compiled with
10151 these debugging symbols defined. It will help highlight potential bugs in
10152 their own code, and will improve the test coverage for OpenSSL itself. At
10153 some point, these tighter rules will become openssl's default to improve
10154 maintainability, though the assert()s and other overheads will remain only
10155 in debugging configurations. See bn.h for more details.
10157 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10159 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10160 that can only be obtained through BN_CTX_new() (which implicitly
10161 initialises it). The presence of this function only made it possible
10162 to overwrite an existing structure (and cause memory leaks).
10166 * Because of the callback-based approach for implementing LHASH as a
10167 template type, lh_insert() adds opaque objects to hash-tables and
10168 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10169 to clean up those corresponding objects before destroying the hash table
10170 (and losing the object pointers). So some over-zealous constifications in
10171 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10172 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10173 prototyped to have "const" restrictions on the object pointers they are
10174 given (and so aren't required to cast them away any more).
10178 * The tmdiff.h API was so ugly and minimal that our own timing utility
10179 (speed) prefers to use its own implementation. The two implementations
10180 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10181 its object type properly exposed (MS_TM) instead of casting to/from
10182 `char *`. This may still change yet if someone realises MS_TM and
10184 aren't necessarily the greatest nomenclatures - but this is what was used
10185 internally to the implementation so I've used that for now.
10189 * Ensure that deprecated functions do not get compiled when
10190 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10191 the self-tests were still using deprecated key-generation functions so
10192 these have been updated also.
10196 * Reorganise PKCS#7 code to separate the digest location functionality
10197 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10198 New function PKCS7_set_digest() to set the digest type for PKCS#7
10199 digestedData type. Add additional code to correctly generate the
10200 digestedData type and add support for this type in PKCS7 initialization
10205 * New function PKCS7_set0_type_other() this initializes a PKCS7
10206 structure of type "other".
10210 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10211 sure the loop does correctly stop and breaking ("division by zero")
10212 modulus operations are not performed. The (pre-generated) prime
10213 table crypto/bn/bn_prime.h was already correct, but it could not be
10214 re-generated on some platforms because of the "division by zero"
10215 situation in the script.
10217 *Ralf S. Engelschall*
10219 * Update support for ECC-based TLS ciphersuites according to
10220 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10221 SHA-1 now is only used for "small" curves (where the
10222 representation of a field element takes up to 24 bytes); for
10223 larger curves, the field element resulting from ECDH is directly
10224 used as premaster secret.
10226 *Douglas Stebila (Sun Microsystems Laboratories)*
10228 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10229 curve secp160r1 to the tests.
10231 *Douglas Stebila (Sun Microsystems Laboratories)*
10233 * Add the possibility to load symbols globally with DSO.
10235 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10237 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10238 control of the error stack.
10242 * Add support for STORE in ENGINE.
10246 * Add the STORE type. The intention is to provide a common interface
10247 to certificate and key stores, be they simple file-based stores, or
10248 HSM-type store, or LDAP stores, or...
10249 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10253 * Add a generic structure called OPENSSL_ITEM. This can be used to
10254 pass a list of arguments to any function as well as provide a way
10255 for a function to pass data back to the caller.
10259 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10260 works like BUF_strdup() but can be used to duplicate a portion of
10261 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10266 * Add the function sk_find_ex() which works like sk_find(), but will
10267 return an index to an element even if an exact match couldn't be
10268 found. The index is guaranteed to point at the element where the
10269 searched-for key would be inserted to preserve sorting order.
10273 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10274 takes an extra flags argument for optional functionality. Currently,
10275 the following flags are defined:
10277 OBJ_BSEARCH_VALUE_ON_NOMATCH
10278 This one gets OBJ_bsearch_ex() to return a pointer to the first
10279 element where the comparing function returns a negative or zero
10282 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10283 This one gets OBJ_bsearch_ex() to return a pointer to the first
10284 element where the comparing function returns zero. This is useful
10285 if there are more than one element where the comparing function
10290 * Make it possible to create self-signed certificates with 'openssl ca'
10291 in such a way that the self-signed certificate becomes part of the
10292 CA database and uses the same mechanisms for serial number generation
10293 as all other certificate signing. The new flag '-selfsign' enables
10294 this functionality. Adapt CA.sh and CA.pl.in.
10298 * Add functionality to check the public key of a certificate request
10299 against a given private. This is useful to check that a certificate
10300 request can be signed by that key (self-signing).
10304 * Make it possible to have multiple active certificates with the same
10305 subject in the CA index file. This is done only if the keyword
10306 'unique_subject' is set to 'no' in the main CA section (default
10307 if 'CA_default') of the configuration file. The value is saved
10308 with the database itself in a separate index attribute file,
10309 named like the index file with '.attr' appended to the name.
10313 * Generate multi-valued AVAs using '+' notation in config files for
10318 * Support for nameConstraints certificate extension.
10322 * Support for policyConstraints certificate extension.
10326 * Support for policyMappings certificate extension.
10330 * Make sure the default DSA_METHOD implementation only uses its
10331 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10332 and change its own handlers to be NULL so as to remove unnecessary
10333 indirection. This lets alternative implementations fallback to the
10334 default implementation more easily.
10338 * Support for directoryName in GeneralName related extensions
10343 * Make it possible to link applications using Makefile.shared.
10344 Make that possible even when linking against static libraries!
10348 * Support for single pass processing for S/MIME signing. This now
10349 means that S/MIME signing can be done from a pipe, in addition
10350 cleartext signing (multipart/signed type) is effectively streaming
10351 and the signed data does not need to be all held in memory.
10353 This is done with a new flag PKCS7_STREAM. When this flag is set
10354 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10355 is done after the data is output (and digests calculated) in
10356 SMIME_write_PKCS7().
10360 * Add full support for -rpath/-R, both in shared libraries and
10361 applications, at least on the platforms where it's known how
10366 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10367 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10368 will now compute a table of multiples of the generator that
10369 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10370 faster (notably in the case of a single point multiplication,
10371 scalar * generator).
10373 *Nils Larsch, Bodo Moeller*
10375 * IPv6 support for certificate extensions. The various extensions
10376 which use the IP:a.b.c.d can now take IPv6 addresses using the
10377 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10382 * Added an ENGINE that implements RSA by performing private key
10383 exponentiations with the GMP library. The conversions to and from
10384 GMP's mpz_t format aren't optimised nor are any montgomery forms
10385 cached, and on x86 it appears OpenSSL's own performance has caught up.
10386 However there are likely to be other architectures where GMP could
10387 provide a boost. This ENGINE is not built in by default, but it can be
10388 specified at Configure time and should be accompanied by the necessary
10389 linker additions, eg;
10390 ./config -DOPENSSL_USE_GMP -lgmp
10394 * "openssl engine" will not display ENGINE/DSO load failure errors when
10395 testing availability of engines with "-t" - the old behaviour is
10396 produced by increasing the feature's verbosity with "-tt".
10400 * ECDSA routines: under certain error conditions uninitialized BN objects
10401 could be freed. Solution: make sure initialization is performed early
10402 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10407 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10408 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10409 software implementations. For DSA and DH, parameter generation can
10410 also be overridden by providing the appropriate method callbacks.
10414 * Change the "progress" mechanism used in key-generation and
10415 primality testing to functions that take a new BN_GENCB pointer in
10416 place of callback/argument pairs. The new API functions have `_ex`
10417 postfixes and the older functions are reimplemented as wrappers for
10418 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10419 declarations of the old functions to help (graceful) attempts to
10420 migrate to the new functions. Also, the new key-generation API
10421 functions operate on a caller-supplied key-structure and return
10422 success/failure rather than returning a key or NULL - this is to
10423 help make "keygen" another member function of RSA_METHOD etc.
10425 Example for using the new callback interface:
10427 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10428 void *my_arg = ...;
10431 BN_GENCB_set(&my_cb, my_callback, my_arg);
10433 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10434 /* For the meaning of a, b in calls to my_callback(), see the
10435 * documentation of the function that calls the callback.
10436 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10437 * my_callback should return 1 if it wants BN_is_prime_ex()
10438 * to continue, or 0 to stop.
10443 * Change the ZLIB compression method to be stateful, and make it
10444 available to TLS with the number defined in
10445 draft-ietf-tls-compression-04.txt.
10449 * Add the ASN.1 structures and functions for CertificatePair, which
10450 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10452 CertificatePair ::= SEQUENCE {
10453 forward [0] Certificate OPTIONAL,
10454 reverse [1] Certificate OPTIONAL,
10455 -- at least one of the pair shall be present -- }
10457 Also implement the PEM functions to read and write certificate
10458 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10460 This needed to be defined, mostly for the sake of the LDAP
10461 attribute crossCertificatePair, but may prove useful elsewhere as
10466 * Make it possible to inhibit symlinking of shared libraries in
10467 Makefile.shared, for Cygwin's sake.
10471 * Extend the BIGNUM API by creating a function
10472 void BN_set_negative(BIGNUM *a, int neg);
10473 and a macro that behave like
10474 int BN_is_negative(const BIGNUM *a);
10476 to avoid the need to access 'a->neg' directly in applications.
10480 * Implement fast modular reduction for pseudo-Mersenne primes
10481 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10482 EC_GROUP_new_curve_GFp() will now automatically use this
10485 *Nils Larsch <nla@trustcenter.de>*
10487 * Add new lock type (CRYPTO_LOCK_BN).
10491 * Change the ENGINE framework to automatically load engines
10492 dynamically from specific directories unless they could be
10493 found to already be built in or loaded. Move all the
10494 current engines except for the cryptodev one to a new
10495 directory engines/.
10496 The engines in engines/ are built as shared libraries if
10497 the "shared" options was given to ./Configure or ./config.
10498 Otherwise, they are inserted in libcrypto.a.
10499 /usr/local/ssl/engines is the default directory for dynamic
10500 engines, but that can be overridden at configure time through
10501 the usual use of --prefix and/or --openssldir, and at run
10502 time with the environment variable OPENSSL_ENGINES.
10504 *Geoff Thorpe and Richard Levitte*
10506 * Add Makefile.shared, a helper makefile to build shared
10507 libraries. Adapt Makefile.org.
10511 * Add version info to Win32 DLLs.
10513 *Peter 'Luna' Runestig" <peter@runestig.com>*
10515 * Add new 'medium level' PKCS#12 API. Certificates and keys
10516 can be added using this API to created arbitrary PKCS#12
10517 files while avoiding the low level API.
10519 New options to PKCS12_create(), key or cert can be NULL and
10520 will then be omitted from the output file. The encryption
10521 algorithm NIDs can be set to -1 for no encryption, the mac
10522 iteration count can be set to 0 to omit the mac.
10524 Enhance pkcs12 utility by making the -nokeys and -nocerts
10525 options work when creating a PKCS#12 file. New option -nomac
10526 to omit the mac, NONE can be set for an encryption algorithm.
10527 New code is modified to use the enhanced PKCS12_create()
10528 instead of the low level API.
10532 * Extend ASN1 encoder to support indefinite length constructed
10533 encoding. This can output sequences tags and octet strings in
10534 this form. Modify pk7_asn1.c to support indefinite length
10535 encoding. This is experimental and needs additional code to
10536 be useful, such as an ASN1 bio and some enhanced streaming
10539 Extend template encode functionality so that tagging is passed
10540 down to the template encoder.
10544 * Let 'openssl req' fail if an argument to '-newkey' is not
10545 recognized instead of using RSA as a default.
10549 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10550 As these are not official, they are not included in "ALL";
10551 the "ECCdraft" ciphersuite group alias can be used to select them.
10553 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10555 * Add ECDH engine support.
10557 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10559 * Add ECDH in new directory crypto/ecdh/.
10561 *Douglas Stebila (Sun Microsystems Laboratories)*
10563 * Let BN_rand_range() abort with an error after 100 iterations
10564 without success (which indicates a broken PRNG).
10568 * Change BN_mod_sqrt() so that it verifies that the input value
10569 is really the square of the return value. (Previously,
10570 BN_mod_sqrt would show GIGO behaviour.)
10574 * Add named elliptic curves over binary fields from X9.62, SECG,
10575 and WAP/WTLS; add OIDs that were still missing.
10577 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10579 * Extend the EC library for elliptic curves over binary fields
10580 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10583 EC_GF2m_simple_method
10587 EC_GROUP_new_curve_GF2m
10588 EC_GROUP_set_curve_GF2m
10589 EC_GROUP_get_curve_GF2m
10590 EC_POINT_set_affine_coordinates_GF2m
10591 EC_POINT_get_affine_coordinates_GF2m
10592 EC_POINT_set_compressed_coordinates_GF2m
10594 Point compression for binary fields is disabled by default for
10595 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10598 As binary polynomials are represented as BIGNUMs, various members
10599 of the EC_GROUP and EC_POINT data structures can be shared
10600 between the implementations for prime fields and binary fields;
10601 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10602 are essentially identical to their `..._GFp` counterparts.
10603 (For simplicity, the `..._GFp` prefix has been dropped from
10604 various internal method names.)
10606 An internal 'field_div' method (similar to 'field_mul' and
10607 'field_sqr') has been added; this is used only for binary fields.
10609 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10611 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10612 through methods ('mul', 'precompute_mult').
10614 The generic implementations (now internally called 'ec_wNAF_mul'
10615 and 'ec_wNAF_precomputed_mult') remain the default if these
10616 methods are undefined.
10618 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10620 * New function EC_GROUP_get_degree, which is defined through
10621 EC_METHOD. For curves over prime fields, this returns the bit
10622 length of the modulus.
10624 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10626 * New functions EC_GROUP_dup, EC_POINT_dup.
10627 (These simply call ..._new and ..._copy).
10629 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10631 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10632 Polynomials are represented as BIGNUMs (where the sign bit is not
10633 used) in the following functions [macros]:
10636 BN_GF2m_sub [= BN_GF2m_add]
10637 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10638 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10639 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10641 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10642 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10643 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10644 BN_GF2m_cmp [= BN_ucmp]
10646 (Note that only the 'mod' functions are actually for fields GF(2^m).
10647 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10649 For some functions, an the irreducible polynomial defining a
10650 field can be given as an 'unsigned int[]' with strictly
10651 decreasing elements giving the indices of those bits that are set;
10652 i.e., p[] represents the polynomial
10653 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10655 p[0] > p[1] > ... > p[k] = 0.
10656 This applies to the following functions:
10659 BN_GF2m_mod_mul_arr
10660 BN_GF2m_mod_sqr_arr
10661 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10662 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10663 BN_GF2m_mod_exp_arr
10664 BN_GF2m_mod_sqrt_arr
10665 BN_GF2m_mod_solve_quad_arr
10669 Conversion can be performed by the following functions:
10674 bntest.c has additional tests for binary polynomial arithmetic.
10676 Two implementations for BN_GF2m_mod_div() are available.
10677 The default algorithm simply uses BN_GF2m_mod_inv() and
10678 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10679 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10680 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10682 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10684 * Add new error code 'ERR_R_DISABLED' that can be used when some
10685 functionality is disabled at compile-time.
10687 *Douglas Stebila <douglas.stebila@sun.com>*
10689 * Change default behaviour of 'openssl asn1parse' so that more
10690 information is visible when viewing, e.g., a certificate:
10692 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10693 mode the content of non-printable OCTET STRINGs is output in a
10694 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10695 avoid the appearance of a printable string.
10697 *Nils Larsch <nla@trustcenter.de>*
10699 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10701 EC_GROUP_set_asn1_flag()
10702 EC_GROUP_get_asn1_flag()
10703 EC_GROUP_set_point_conversion_form()
10704 EC_GROUP_get_point_conversion_form()
10705 These control ASN1 encoding details:
10706 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10707 has been set to OPENSSL_EC_NAMED_CURVE.
10708 - Points are encoded in uncompressed form by default; options for
10709 asn1_for are as for point2oct, namely
10710 POINT_CONVERSION_COMPRESSED
10711 POINT_CONVERSION_UNCOMPRESSED
10712 POINT_CONVERSION_HYBRID
10714 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10716 EC_GROUP_set_seed()
10717 EC_GROUP_get0_seed()
10718 EC_GROUP_get_seed_len()
10719 This is used only for ASN1 purposes (so far).
10721 *Nils Larsch <nla@trustcenter.de>*
10723 * Add 'field_type' member to EC_METHOD, which holds the NID
10724 of the appropriate field type OID. The new function
10725 EC_METHOD_get_field_type() returns this value.
10727 *Nils Larsch <nla@trustcenter.de>*
10730 EC_POINT_point2bn()
10731 EC_POINT_bn2point()
10732 EC_POINT_point2hex()
10733 EC_POINT_hex2point()
10734 providing useful interfaces to EC_POINT_point2oct() and
10735 EC_POINT_oct2point().
10737 *Nils Larsch <nla@trustcenter.de>*
10739 * Change internals of the EC library so that the functions
10740 EC_GROUP_set_generator()
10741 EC_GROUP_get_generator()
10742 EC_GROUP_get_order()
10743 EC_GROUP_get_cofactor()
10744 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10745 to methods, which would lead to unnecessary code duplication when
10746 adding different types of curves.
10748 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10750 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10751 arithmetic, and such that modified wNAFs are generated
10752 (which avoid length expansion in many cases).
10756 * Add a function EC_GROUP_check_discriminant() (defined via
10757 EC_METHOD) that verifies that the curve discriminant is non-zero.
10759 Add a function EC_GROUP_check() that makes some sanity tests
10760 on a EC_GROUP, its generator and order. This includes
10761 EC_GROUP_check_discriminant().
10763 *Nils Larsch <nla@trustcenter.de>*
10765 * Add ECDSA in new directory crypto/ecdsa/.
10767 Add applications 'openssl ecparam' and 'openssl ecdsa'
10768 (these are based on 'openssl dsaparam' and 'openssl dsa').
10770 ECDSA support is also included in various other files across the
10771 library. Most notably,
10772 - 'openssl req' now has a '-newkey ecdsa:file' option;
10773 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10774 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10775 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10776 them suitable for ECDSA where domain parameters must be
10777 extracted before the specific public key;
10778 - ECDSA engine support has been added.
10780 *Nils Larsch <nla@trustcenter.de>*
10782 * Include some named elliptic curves, and add OIDs from X9.62,
10783 SECG, and WAP/WTLS. Each curve can be obtained from the new
10785 EC_GROUP_new_by_curve_name(),
10786 and the list of available named curves can be obtained with
10787 EC_get_builtin_curves().
10788 Also add a 'curve_name' member to EC_GROUP objects, which can be
10790 EC_GROUP_set_curve_name()
10791 EC_GROUP_get_curve_name()
10793 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10795 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10796 was actually never needed) and in BN_mul(). The removal in BN_mul()
10797 required a small change in bn_mul_part_recursive() and the addition
10798 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10799 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10800 bn_sub_words() and bn_add_words() except they take arrays with
10805 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10807 * Cleanse PEM buffers before freeing them since they may contain
10810 *Benjamin Bennett <ben@psc.edu>*
10812 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10813 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10814 authentication-only ciphersuites.
10818 * Since AES128 and AES256 share a single mask bit in the logic of
10819 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10820 kludge to work properly if AES128 is available and AES256 isn't.
10824 * Expand security boundary to match 1.1.1 module.
10828 * Remove redundant features: hash file source, editing of test vectors
10829 modify fipsld to use external fips_premain.c signature.
10833 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10834 run algorithm test programs.
10838 * Make algorithm test programs more tolerant of whitespace.
10842 * Have SSL/TLS server implementation tolerate "mismatched" record
10843 protocol version while receiving ClientHello even if the
10844 ClientHello is fragmented. (The server can't insist on the
10845 particular protocol version it has chosen before the ServerHello
10846 message has informed the client about his choice.)
10850 * Load error codes if they are not already present instead of using a
10851 static variable. This allows them to be cleanly unloaded and reloaded.
10855 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10857 * Introduce limits to prevent malicious keys being able to
10858 cause a denial of service. ([CVE-2006-2940])
10860 *Steve Henson, Bodo Moeller*
10862 * Fix ASN.1 parsing of certain invalid structures that can result
10863 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10865 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10866 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10868 * Fix SSL client code which could crash if connecting to a
10869 malicious SSLv2 server. ([CVE-2006-4343])
10871 *Tavis Ormandy and Will Drewry, Google Security Team*
10873 * Change ciphersuite string processing so that an explicit
10874 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10875 will no longer include "AES128-SHA"), and any other similar
10876 ciphersuite (same bitmap) from *other* protocol versions (so that
10877 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10878 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10879 changes from 0.9.8b and 0.9.8d.
10883 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10885 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10886 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10888 * Change the Unix randomness entropy gathering to use poll() when
10889 possible instead of select(), since the latter has some
10890 undesirable limitations.
10892 *Darryl Miles via Richard Levitte and Bodo Moeller*
10894 * Disable rogue ciphersuites:
10896 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10897 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10898 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10900 The latter two were purportedly from
10901 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10904 Also deactivate the remaining ciphersuites from
10905 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10906 unofficial, and the ID has long expired.
10910 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10911 dual-core machines) and other potential thread-safety issues.
10915 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10917 * Adapt fipsld and the build system to link against the validated FIPS
10918 module in FIPS mode.
10922 * Fixes for VC++ 2005 build under Windows.
10926 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10927 from a Windows bash shell such as MSYS. It is autodetected from the
10928 "config" script when run from a VC++ environment. Modify standard VC++
10929 build to use fipscanister.o from the GNU make build.
10933 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10935 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10936 The value now differs depending on if you build for FIPS or not.
10937 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10938 safely run with a non-FIPSed libcrypto, as it may crash because of
10939 the difference induced by this change.
10943 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10945 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10946 (part of SSL_OP_ALL). This option used to disable the
10947 countermeasure against man-in-the-middle protocol-version
10948 rollback in the SSL 2.0 server implementation, which is a bad
10949 idea. ([CVE-2005-2969])
10951 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10952 for Information Security, National Institute of Advanced Industrial
10953 Science and Technology [AIST, Japan)]*
10955 * Minimal support for X9.31 signatures and PSS padding modes. This is
10956 mainly for FIPS compliance and not fully integrated at this stage.
10960 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10961 the exponentiation using a fixed-length exponent. (Otherwise,
10962 the information leaked through timing could expose the secret key
10963 after many signatures; cf. Bleichenbacher's attack on DSA with
10968 * Make a new fixed-window mod_exp implementation the default for
10969 RSA, DSA, and DH private-key operations so that the sequence of
10970 squares and multiplies and the memory access pattern are
10971 independent of the particular secret key. This will mitigate
10972 cache-timing and potential related attacks.
10974 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10975 and this is automatically used by BN_mod_exp_mont() if the new flag
10976 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10977 will use this BN flag for private exponents unless the flag
10978 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10979 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10981 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10983 * Change the client implementation for SSLv23_method() and
10984 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10985 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10986 (Previously, the SSL 2.0 backwards compatible Client Hello
10987 message format would be used even with SSL_OP_NO_SSLv2.)
10991 * Add support for smime-type MIME parameter in S/MIME messages which some
10996 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10997 a threadsafe manner. Modify rsa code to use new function and add calls
10998 to dsa and dh code (which had race conditions before).
11002 * Include the fixed error library code in the C error file definitions
11003 instead of fixing them up at runtime. This keeps the error code
11004 structures constant.
11008 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11010 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11013 * Fixes for newer kerberos headers. NB: the casts are needed because
11014 the 'length' field is signed on one version and unsigned on another
11015 with no (?) obvious way to tell the difference, without these VC++
11016 complains. Also the "definition" of FAR (blank) is no longer included
11017 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11018 some needed definitions.
11022 * Undo Cygwin change.
11026 * Added support for proxy certificates according to RFC 3820.
11027 Because they may be a security thread to unaware applications,
11028 they must be explicitly allowed in run-time. See
11029 docs/HOWTO/proxy_certificates.txt for further information.
11033 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11035 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11036 server and client random values. Previously
11037 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11038 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11040 This change has negligible security impact because:
11042 1. Server and client random values still have 24 bytes of pseudo random
11045 2. Server and client random values are sent in the clear in the initial
11048 3. The master secret is derived using the premaster secret (48 bytes in
11049 size for static RSA ciphersuites) as well as client server and random
11052 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11055 *Stephen Henson, reported by UK NISCC*
11057 * Use Windows randomness collection on Cygwin.
11061 * Fix hang in EGD/PRNGD query when communication socket is closed
11062 prematurely by EGD/PRNGD.
11064 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11066 * Prompt for pass phrases when appropriate for PKCS12 input format.
11070 * Back-port of selected performance improvements from development
11071 branch, as well as improved support for PowerPC platforms.
11075 * Add lots of checks for memory allocation failure, error codes to indicate
11076 failure and freeing up memory if a failure occurs.
11078 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11080 * Add new -passin argument to dgst.
11084 * Perform some character comparisons of different types in X509_NAME_cmp:
11085 this is needed for some certificates that re-encode DNs into UTF8Strings
11086 (in violation of RFC3280) and can't or won't issue name rollover
11091 * Make an explicit check during certificate validation to see that
11092 the CA setting in each certificate on the chain is correct. As a
11093 side effect always do the following basic checks on extensions,
11094 not just when there's an associated purpose to the check:
11096 - if there is an unhandled critical extension (unless the user
11097 has chosen to ignore this fault)
11098 - if the path length has been exceeded (if one is set at all)
11099 - that certain extensions fit the associated purpose (if one has
11104 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11106 * Avoid a race condition when CRLs are checked in a multi threaded
11107 environment. This would happen due to the reordering of the revoked
11108 entries during signature checking and serial number lookup. Now the
11109 encoding is cached and the serial number sort performed under a lock.
11110 Add new STACK function sk_is_sorted().
11114 * Add Delta CRL to the extension code.
11118 * Various fixes to s3_pkt.c so alerts are sent properly.
11120 *David Holmes <d.holmes@f5.com>*
11122 * Reduce the chances of duplicate issuer name and serial numbers (in
11123 violation of RFC3280) using the OpenSSL certificate creation utilities.
11124 This is done by creating a random 64 bit value for the initial serial
11125 number when a serial number file is created or when a self signed
11126 certificate is created using 'openssl req -x509'. The initial serial
11127 number file is created using 'openssl x509 -next_serial' in CA.pl
11128 rather than being initialized to 1.
11132 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11134 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11135 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11137 *Joe Orton, Steve Henson*
11139 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11142 *Joe Orton, Steve Henson*
11144 * Make it possible to have multiple active certificates with the same
11145 subject in the CA index file. This is done only if the keyword
11146 'unique_subject' is set to 'no' in the main CA section (default
11147 if 'CA_default') of the configuration file. The value is saved
11148 with the database itself in a separate index attribute file,
11149 named like the index file with '.attr' appended to the name.
11153 * X509 verify fixes. Disable broken certificate workarounds when
11154 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11155 keyUsage extension present. Don't accept CRLs with unhandled critical
11156 extensions: since verify currently doesn't process CRL extensions this
11157 rejects a CRL with *any* critical extensions. Add new verify error codes
11162 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11163 A clarification of RFC2560 will require the use of OCTET STRINGs and
11164 some implementations cannot handle the current raw format. Since OpenSSL
11165 copies and compares OCSP nonces as opaque blobs without any attempt at
11166 parsing them this should not create any compatibility issues.
11170 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11171 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11172 this HMAC (and other) operations are several times slower than OpenSSL
11177 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11179 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11181 * Use the correct content when signing type "other".
11185 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11187 * Fix various bugs revealed by running the NISCC test suite:
11189 Stop out of bounds reads in the ASN1 code when presented with
11190 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11192 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11194 If verify callback ignores invalid public key errors don't try to check
11195 certificate signature with the NULL public key.
11199 * New -ignore_err option in ocsp application to stop the server
11200 exiting on the first error in a request.
11204 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11205 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11210 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11211 extra data after the compression methods not only for TLS 1.0
11212 but also for SSL 3.0 (as required by the specification).
11214 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11216 * Change X509_certificate_type() to mark the key as exported/exportable
11217 when it's 512 *bits* long, not 512 bytes.
11221 * Change AES_cbc_encrypt() so it outputs exact multiple of
11222 blocks during encryption.
11226 * Various fixes to base64 BIO and non blocking I/O. On write
11227 flushes were not handled properly if the BIO retried. On read
11228 data was not being buffered properly and had various logic bugs.
11229 This also affects blocking I/O when the data being decoded is a
11234 * Various S/MIME bugfixes and compatibility changes:
11235 output correct application/pkcs7 MIME type if
11236 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11237 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11238 of files as .eml work). Correctly handle very long lines in MIME
11243 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11245 * Countermeasure against the Klima-Pokorny-Rosa extension of
11246 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11247 a protocol version number mismatch like a decryption error
11248 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11252 * Turn on RSA blinding by default in the default implementation
11253 to avoid a timing attack. Applications that don't want it can call
11254 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11255 They would be ill-advised to do so in most cases.
11257 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11259 * Change RSA blinding code so that it works when the PRNG is not
11260 seeded (in this case, the secret RSA exponent is abused as
11261 an unpredictable seed -- if it is not unpredictable, there
11262 is no point in blinding anyway). Make RSA blinding thread-safe
11263 by remembering the creator's thread ID in rsa->blinding and
11264 having all other threads use local one-time blinding factors
11265 (this requires more computation than sharing rsa->blinding, but
11266 avoids excessive locking; and if an RSA object is not shared
11267 between threads, blinding will still be very fast).
11271 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11272 ENGINE as defaults for all supported algorithms irrespective of
11273 the 'flags' parameter. 'flags' is now honoured, so applications
11274 should make sure they are passing it correctly.
11278 * Target "mingw" now allows native Windows code to be generated in
11279 the Cygwin environment as well as with the MinGW compiler.
11283 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11285 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11286 via timing by performing a MAC computation even if incorrect
11287 block cipher padding has been found. This is a countermeasure
11288 against active attacks where the attacker has to distinguish
11289 between bad padding and a MAC verification error. ([CVE-2003-0078])
11291 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11292 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11293 Martin Vuagnoux (EPFL, Ilion)*
11295 * Make the no-err option work as intended. The intention with no-err
11296 is not to have the whole error stack handling routines removed from
11297 libcrypto, it's only intended to remove all the function name and
11298 reason texts, thereby removing some of the footprint that may not
11299 be interesting if those errors aren't displayed anyway.
11301 NOTE: it's still possible for any application or module to have its
11302 own set of error texts inserted. The routines are there, just not
11303 used by default when no-err is given.
11307 * Add support for FreeBSD on IA64.
11309 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11311 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11312 Kerberos function mit_des_cbc_cksum(). Before this change,
11313 the value returned by DES_cbc_cksum() was like the one from
11314 mit_des_cbc_cksum(), except the bytes were swapped.
11316 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11318 * Allow an application to disable the automatic SSL chain building.
11319 Before this a rather primitive chain build was always performed in
11320 ssl3_output_cert_chain(): an application had no way to send the
11321 correct chain if the automatic operation produced an incorrect result.
11323 Now the chain builder is disabled if either:
11325 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11327 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11329 The reasoning behind this is that an application would not want the
11330 auto chain building to take place if extra chain certificates are
11331 present and it might also want a means of sending no additional
11332 certificates (for example the chain has two certificates and the
11337 * Add the possibility to build without the ENGINE framework.
11339 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11341 * Under Win32 gmtime() can return NULL: check return value in
11342 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11346 * DSA routines: under certain error conditions uninitialized BN objects
11347 could be freed. Solution: make sure initialization is performed early
11348 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11349 Nils Larsch <nla@trustcenter.de> via PR#459)
11353 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11354 checked on reconnect on the client side, therefore session resumption
11355 could still fail with a "ssl session id is different" error. This
11356 behaviour is masked when SSL_OP_ALL is used due to
11357 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11358 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11359 followup to PR #377.
11363 * IA-32 assembler support enhancements: unified ELF targets, support
11364 for SCO/Caldera platforms, fix for Cygwin shared build.
11368 * Add support for FreeBSD on sparc64. As a consequence, support for
11369 FreeBSD on non-x86 processors is separate from x86 processors on
11370 the config script, much like the NetBSD support.
11372 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11374 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11376 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11379 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11380 code (06) was taken as the first octet of the session ID and the last
11381 octet was ignored consequently. As a result SSLv2 client side session
11382 caching could not have worked due to the session ID mismatch between
11384 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11389 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11390 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11395 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11396 seems that in spite of existing for more than a year, many application
11397 author have done nothing to provide the necessary callbacks, which
11398 means that this particular engine will not work properly anywhere.
11399 This is a very unfortunate situation which forces us, in the name
11400 of usability, to give the hw_ncipher.c a static lock, which is part
11402 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11403 appear in 0.9.8 or later. We EXPECT application authors to have
11404 dealt properly with this when 0.9.8 is released (unless we actually
11405 make such changes in the libcrypto locking code that changes will
11406 have to be made anyway).
11410 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11411 octets have been read, EOF or an error occurs. Without this change
11412 some truncated ASN1 structures will not produce an error.
11416 * Disable Heimdal support, since it hasn't been fully implemented.
11417 Still give the possibility to force the use of Heimdal, but with
11418 warnings and a request that patches get sent to openssl-dev.
11422 * Add the VC-CE target, introduce the WINCE sysname, and add
11423 INSTALL.WCE and appropriate conditionals to make it build.
11425 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11427 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11428 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11429 edit numbers of the version.
11431 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11433 * Introduce safe string copy and catenation functions
11434 (BUF_strlcpy() and BUF_strlcat()).
11436 *Ben Laurie (CHATS) and Richard Levitte*
11438 * Avoid using fixed-size buffers for one-line DNs.
11440 *Ben Laurie (CHATS)*
11442 * Add BUF_MEM_grow_clean() to avoid information leakage when
11443 resizing buffers containing secrets, and use where appropriate.
11445 *Ben Laurie (CHATS)*
11447 * Avoid using fixed size buffers for configuration file location.
11449 *Ben Laurie (CHATS)*
11451 * Avoid filename truncation for various CA files.
11453 *Ben Laurie (CHATS)*
11455 * Use sizeof in preference to magic numbers.
11457 *Ben Laurie (CHATS)*
11459 * Avoid filename truncation in cert requests.
11461 *Ben Laurie (CHATS)*
11463 * Add assertions to check for (supposedly impossible) buffer
11466 *Ben Laurie (CHATS)*
11468 * Don't cache truncated DNS entries in the local cache (this could
11469 potentially lead to a spoofing attack).
11471 *Ben Laurie (CHATS)*
11473 * Fix various buffers to be large enough for hex/decimal
11474 representations in a platform independent manner.
11476 *Ben Laurie (CHATS)*
11478 * Add CRYPTO_realloc_clean() to avoid information leakage when
11479 resizing buffers containing secrets, and use where appropriate.
11481 *Ben Laurie (CHATS)*
11483 * Add BIO_indent() to avoid much slightly worrying code to do
11486 *Ben Laurie (CHATS)*
11488 * Convert sprintf()/BIO_puts() to BIO_printf().
11490 *Ben Laurie (CHATS)*
11492 * buffer_gets() could terminate with the buffer only half
11495 *Ben Laurie (CHATS)*
11497 * Add assertions to prevent user-supplied crypto functions from
11498 overflowing internal buffers by having large block sizes, etc.
11500 *Ben Laurie (CHATS)*
11502 * New OPENSSL_assert() macro (similar to assert(), but enabled
11505 *Ben Laurie (CHATS)*
11507 * Eliminate unused copy of key in RC4.
11509 *Ben Laurie (CHATS)*
11511 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11513 *Ben Laurie (CHATS)*
11515 * Fix off-by-one error in EGD path.
11517 *Ben Laurie (CHATS)*
11519 * If RANDFILE path is too long, ignore instead of truncating.
11521 *Ben Laurie (CHATS)*
11523 * Eliminate unused and incorrectly sized X.509 structure
11526 *Ben Laurie (CHATS)*
11528 * Eliminate unused and dangerous function knumber().
11530 *Ben Laurie (CHATS)*
11532 * Eliminate unused and dangerous structure, KSSL_ERR.
11534 *Ben Laurie (CHATS)*
11536 * Protect against overlong session ID context length in an encoded
11537 session object. Since these are local, this does not appear to be
11540 *Ben Laurie (CHATS)*
11542 * Change from security patch (see 0.9.6e below) that did not affect
11543 the 0.9.6 release series:
11545 Remote buffer overflow in SSL3 protocol - an attacker could
11546 supply an oversized master key in Kerberos-enabled versions.
11549 *Ben Laurie (CHATS)*
11551 * Change the SSL kerb5 codes to match RFC 2712.
11555 * Make -nameopt work fully for req and add -reqopt switch.
11557 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11559 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11561 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11563 * Make sure tests can be performed even if the corresponding algorithms
11564 have been removed entirely. This was also the last step to make
11565 OpenSSL compilable with DJGPP under all reasonable conditions.
11567 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11569 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11570 to allow version independent disabling of normally unselected ciphers,
11571 which may be activated as a side-effect of selecting a single cipher.
11573 (E.g., cipher list string "RSA" enables ciphersuites that are left
11574 out of "ALL" because they do not provide symmetric encryption.
11575 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11577 *Lutz Jaenicke, Bodo Moeller*
11579 * Add appropriate support for separate platform-dependent build
11580 directories. The recommended way to make a platform-dependent
11581 build directory is the following (tested on Linux), maybe with
11584 # Place yourself outside of the OpenSSL source tree. In
11585 # this example, the environment variable OPENSSL_SOURCE
11586 # is assumed to contain the absolute OpenSSL source directory.
11587 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11588 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11589 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11590 mkdir -p `dirname $F`
11591 ln -s $OPENSSL_SOURCE/$F $F
11594 To be absolutely sure not to disturb the source tree, a "make clean"
11595 is a good thing. If it isn't successful, don't worry about it,
11596 it probably means the source directory is very clean.
11600 * Make sure any ENGINE control commands make local copies of string
11601 pointers passed to them whenever necessary. Otherwise it is possible
11602 the caller may have overwritten (or deallocated) the original string
11603 data when a later ENGINE operation tries to use the stored values.
11605 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11607 * Improve diagnostics in file reading and command-line digests.
11609 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11611 * Add AES modes CFB and OFB to the object database. Correct an
11612 error in AES-CFB decryption.
11616 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11617 allows existing EVP_CIPHER_CTX structures to be reused after
11618 calling `EVP_*Final()`. This behaviour is used by encryption
11619 BIOs and some applications. This has the side effect that
11620 applications must explicitly clean up cipher contexts with
11621 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11625 * Check the values of dna and dnb in bn_mul_recursive before calling
11626 bn_mul_comba (a non zero value means the a or b arrays do not contain
11627 n2 elements) and fallback to bn_mul_normal if either is not zero.
11631 * Fix escaping of non-ASCII characters when using the -subj option
11632 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11636 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11637 form for "surname", serialNumber has no short form.
11638 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11639 therefore remove "mail" short name for "internet 7".
11640 The OID for unique identifiers in X509 certificates is
11641 x500UniqueIdentifier, not uniqueIdentifier.
11642 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11646 * Add an "init" command to the ENGINE config module and auto initialize
11647 ENGINEs. Without any "init" command the ENGINE will be initialized
11648 after all ctrl commands have been executed on it. If init=1 the
11649 ENGINE is initialized at that point (ctrls before that point are run
11650 on the uninitialized ENGINE and after on the initialized one). If
11651 init=0 then the ENGINE will not be initialized at all.
11655 * Fix the 'app_verify_callback' interface so that the user-defined
11656 argument is actually passed to the callback: In the
11657 SSL_CTX_set_cert_verify_callback() prototype, the callback
11658 declaration has been changed from
11661 int (*cb)(X509_STORE_CTX *,void *);
11662 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11663 i=s->ctx->app_verify_callback(&ctx)
11664 has been changed into
11665 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11667 To update applications using SSL_CTX_set_cert_verify_callback(),
11668 a dummy argument can be added to their callback functions.
11670 *D. K. Smetters <smetters@parc.xerox.com>*
11672 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11674 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11676 * Add and OPENSSL_LOAD_CONF define which will cause
11677 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11678 This allows older applications to transparently support certain
11679 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11680 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11681 load the config file and OPENSSL_add_all_algorithms_conf() which will
11682 always load it have also been added.
11686 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11687 Adjust NIDs and EVP layer.
11689 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11691 * Config modules support in openssl utility.
11693 Most commands now load modules from the config file,
11694 though in a few (such as version) this isn't done
11695 because it couldn't be used for anything.
11697 In the case of ca and req the config file used is
11698 the same as the utility itself: that is the -config
11699 command line option can be used to specify an
11704 * Move default behaviour from OPENSSL_config(). If appname is NULL
11705 use "openssl_conf" if filename is NULL use default openssl config file.
11709 * Add an argument to OPENSSL_config() to allow the use of an alternative
11710 config section name. Add a new flag to tolerate a missing config file
11711 and move code to CONF_modules_load_file().
11715 * Support for crypto accelerator cards from Accelerated Encryption
11716 Processing, www.aep.ie. (Use engine 'aep')
11717 The support was copied from 0.9.6c [engine] and adapted/corrected
11718 to work with the new engine framework.
11720 *AEP Inc. and Richard Levitte*
11722 * Support for SureWare crypto accelerator cards from Baltimore
11723 Technologies. (Use engine 'sureware')
11724 The support was copied from 0.9.6c [engine] and adapted
11725 to work with the new engine framework.
11729 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11730 make the newer ENGINE framework commands for the CHIL engine work.
11732 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11734 * Make it possible to produce shared libraries on ReliantUNIX.
11736 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11738 * Add the configuration target debug-linux-ppro.
11739 Make 'openssl rsa' use the general key loading routines
11740 implemented in `apps.c`, and make those routines able to
11741 handle the key format FORMAT_NETSCAPE and the variant
11744 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11746 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11748 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11750 * Add -keyform to rsautl, and document -engine.
11752 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11754 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11755 BIO_R_NO_SUCH_FILE error code rather than the generic
11756 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11760 * Add new functions
11761 ERR_peek_last_error
11762 ERR_peek_last_error_line
11763 ERR_peek_last_error_line_data.
11764 These are similar to
11766 ERR_peek_error_line
11767 ERR_peek_error_line_data,
11768 but report on the latest error recorded rather than the first one
11769 still in the error queue.
11771 *Ben Laurie, Bodo Moeller*
11773 * default_algorithms option in ENGINE config module. This allows things
11775 default_algorithms = ALL
11776 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11780 * Preliminary ENGINE config module.
11784 * New experimental application configuration code.
11788 * Change the AES code to follow the same name structure as all other
11789 symmetric ciphers, and behave the same way. Move everything to
11790 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11792 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11794 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11796 *Ben Laurie and Theo de Raadt*
11798 * Add option to output public keys in req command.
11800 *Massimiliano Pala madwolf@openca.org*
11802 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11803 (up to about 10% better than before for P-192 and P-224).
11807 * New functions/macros
11809 SSL_CTX_set_msg_callback(ctx, cb)
11810 SSL_CTX_set_msg_callback_arg(ctx, arg)
11811 SSL_set_msg_callback(ssl, cb)
11812 SSL_set_msg_callback_arg(ssl, arg)
11814 to request calling a callback function
11816 void cb(int write_p, int version, int content_type,
11817 const void *buf, size_t len, SSL *ssl, void *arg)
11819 whenever a protocol message has been completely received
11820 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11821 protocol version according to which the SSL library interprets
11822 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11823 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11824 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11825 specification (change_cipher_spec(20), alert(21), handshake(22)).
11826 'buf' and 'len' point to the actual message, 'ssl' to the
11827 SSL object, and 'arg' is the application-defined value set by
11828 SSL[_CTX]_set_msg_callback_arg().
11830 'openssl s_client' and 'openssl s_server' have new '-msg' options
11831 to enable a callback that displays all protocol messages.
11835 * Change the shared library support so shared libraries are built as
11836 soon as the corresponding static library is finished, and thereby get
11837 openssl and the test programs linked against the shared library.
11838 This still only happens when the keyword "shard" has been given to
11839 the configuration scripts.
11841 NOTE: shared library support is still an experimental thing, and
11842 backward binary compatibility is still not guaranteed.
11844 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11846 * Add support for Subject Information Access extension.
11848 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11850 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11851 additional bytes when new memory had to be allocated, not just
11852 when reusing an existing buffer.
11856 * New command line and configuration option 'utf8' for the req command.
11857 This allows field values to be specified as UTF8 strings.
11861 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11862 runs for the former and machine-readable output for the latter.
11866 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11867 of the e-mail address in the DN (i.e., it will go into a certificate
11868 extension only). The new configuration file option 'email_in_dn = no'
11869 has the same effect.
11871 *Massimiliano Pala madwolf@openca.org*
11873 * Change all functions with names starting with `des_` to be starting
11874 with `DES_` instead. Add wrappers that are compatible with libdes,
11875 but are named `_ossl_old_des_*`. Finally, add macros that map the
11876 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11877 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11878 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11881 Since we provide two compatibility mappings, the user needs to
11882 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11883 compatibility is desired. The default (i.e., when that macro
11884 isn't defined) is OpenSSL 0.9.6c compatibility.
11886 There are also macros that enable and disable the support of old
11887 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11888 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11889 are defined, the default will apply: to support the old des routines.
11891 In either case, one must include openssl/des.h to get the correct
11892 definitions. Do not try to just include openssl/des_old.h, that
11895 NOTE: This is a major break of an old API into a new one. Software
11896 authors are encouraged to switch to the `DES_` style functions. Some
11897 time in the future, des_old.h and the libdes compatibility functions
11898 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11899 default), and then completely removed.
11903 * Test for certificates which contain unsupported critical extensions.
11904 If such a certificate is found during a verify operation it is
11905 rejected by default: this behaviour can be overridden by either
11906 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11907 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11908 X509_supported_extension() has also been added which returns 1 if a
11909 particular extension is supported.
11913 * Modify the behaviour of EVP cipher functions in similar way to digests
11914 to retain compatibility with existing code.
11918 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11919 compatibility with existing code. In particular the 'ctx' parameter does
11920 not have to be to be initialized before the call to EVP_DigestInit() and
11921 it is tidied up after a call to EVP_DigestFinal(). New function
11922 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11923 EVP_MD_CTX_copy() changed to not require the destination to be
11924 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11925 requires the destination to be valid.
11927 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11928 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11932 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11933 so that complete 'Handshake' protocol structures are kept in memory
11934 instead of overwriting 'msg_type' and 'length' with 'body' data.
11938 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11940 *Massimo Santin via Richard Levitte*
11942 * Major restructuring to the underlying ENGINE code. This includes
11943 reduction of linker bloat, separation of pure "ENGINE" manipulation
11944 (initialisation, etc) from functionality dealing with implementations
11945 of specific crypto interfaces. This change also introduces integrated
11946 support for symmetric ciphers and digest implementations - so ENGINEs
11947 can now accelerate these by providing EVP_CIPHER and EVP_MD
11948 implementations of their own. This is detailed in
11949 [crypto/engine/README.md](crypto/engine/README.md)
11950 as it couldn't be adequately described here. However, there are a few
11951 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11952 were changed in the original introduction of ENGINE code have now
11953 reverted back - the hooking from this code to ENGINE is now a good
11954 deal more passive and at run-time, operations deal directly with
11955 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11956 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11957 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11958 they were not being used by the framework as there is no concept of a
11959 BIGNUM_METHOD and they could not be generalised to the new
11960 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11961 ENGINE_cpy() has been removed as it cannot be consistently defined in
11966 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11970 * Change mkdef.pl to sort symbols that get the same entry number,
11971 and make sure the automatically generated functions `ERR_load_*`
11972 become part of libeay.num as well.
11976 * New function SSL_renegotiate_pending(). This returns true once
11977 renegotiation has been requested (either SSL_renegotiate() call
11978 or HelloRequest/ClientHello received from the peer) and becomes
11979 false once a handshake has been completed.
11980 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11981 sends a HelloRequest, but does not ensure that a handshake takes
11982 place. SSL_renegotiate_pending() is useful for checking if the
11983 client has followed the request.)
11987 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11988 By default, clients may request session resumption even during
11989 renegotiation (if session ID contexts permit); with this option,
11990 session resumption is possible only in the first handshake.
11992 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11993 more bits available for options that should not be part of
11994 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11998 * Add some demos for certificate and certificate request creation.
12002 * Make maximum certificate chain size accepted from the peer application
12003 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12004 "Douglas E. Engert" <deengert@anl.gov>.
12008 * Add support for shared libraries for Unixware-7
12009 (Boyd Lynn Gerber <gerberb@zenez.com>).
12013 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12014 be done prior to destruction. Use this to unload error strings from
12015 ENGINEs that load their own error strings. NB: This adds two new API
12016 functions to "get" and "set" this destroy handler in an ENGINE.
12020 * Alter all existing ENGINE implementations (except "openssl" and
12021 "openbsd") to dynamically instantiate their own error strings. This
12022 makes them more flexible to be built both as statically-linked ENGINEs
12023 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12024 Also, add stub code to each that makes building them as self-contained
12025 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12029 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12030 implementations into applications that are completely implemented in
12031 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12032 commands that can be used to configure what shared-library to load and
12033 to control aspects of the way it is handled. Also, made an update to
12034 the [README-Engine.md](README-Engine.md) file
12035 that brings its information up-to-date and
12036 provides some information and instructions on the "dynamic" ENGINE
12037 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12041 * Make it possible to unload ranges of ERR strings with a new
12042 "ERR_unload_strings" function.
12046 * Add a copy() function to EVP_MD.
12050 * Make EVP_MD routines take a context pointer instead of just the
12051 md_data void pointer.
12055 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12056 that the digest can only process a single chunk of data
12057 (typically because it is provided by a piece of
12058 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12059 is only going to provide a single chunk of data, and hence the
12060 framework needn't accumulate the data for oneshot drivers.
12064 * As with "ERR", make it possible to replace the underlying "ex_data"
12065 functions. This change also alters the storage and management of global
12066 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12067 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12068 index counters. The API functions that use this state have been changed
12069 to take a "class_index" rather than pointers to the class's local STACK
12070 and counter, and there is now an API function to dynamically create new
12071 classes. This centralisation allows us to (a) plug a lot of the
12072 thread-safety problems that existed, and (b) makes it possible to clean
12073 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12074 such data would previously have always leaked in application code and
12075 workarounds were in place to make the memory debugging turn a blind eye
12076 to it. Application code that doesn't use this new function will still
12077 leak as before, but their memory debugging output will announce it now
12078 rather than letting it slide.
12080 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12081 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12082 has a return value to indicate success or failure.
12086 * Make it possible to replace the underlying "ERR" functions such that the
12087 global state (2 LHASH tables and 2 locks) is only used by the "default"
12088 implementation. This change also adds two functions to "get" and "set"
12089 the implementation prior to it being automatically set the first time
12090 any other ERR function takes place. Ie. an application can call "get",
12091 pass the return value to a module it has just loaded, and that module
12092 can call its own "set" function using that value. This means the
12093 module's "ERR" operations will use (and modify) the error state in the
12094 application and not in its own statically linked copy of OpenSSL code.
12098 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12099 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12100 the operation, and provides a more encapsulated way for external code
12101 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12102 to use these functions rather than manually incrementing the counts.
12104 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12108 * Add EVP test program.
12112 * Add symmetric cipher support to ENGINE. Expect the API to change!
12116 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12117 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12118 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12119 These allow a CRL to be built without having to access X509_CRL fields
12120 directly. Modify 'ca' application to use new functions.
12124 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12125 bug workarounds. Rollback attack detection is a security feature.
12126 The problem will only arise on OpenSSL servers when TLSv1 is not
12127 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12128 Software authors not wanting to support TLSv1 will have special reasons
12129 for their choice and can explicitly enable this option.
12131 *Bodo Moeller, Lutz Jaenicke*
12133 * Rationalise EVP so it can be extended: don't include a union of
12134 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12135 (similar to those existing for EVP_CIPHER_CTX).
12140 EVP_MD_CTX_init(&md); /* new function call */
12141 EVP_DigestInit(&md, EVP_sha1());
12142 EVP_DigestUpdate(&md, in, len);
12143 EVP_DigestFinal(&md, out, NULL);
12144 EVP_MD_CTX_cleanup(&md); /* new function call */
12148 * Make DES key schedule conform to the usual scheme, as well as
12149 correcting its structure. This means that calls to DES functions
12150 now have to pass a pointer to a des_key_schedule instead of a
12151 plain des_key_schedule (which was actually always a pointer
12154 des_key_schedule ks;
12156 des_set_key_checked(..., &ks);
12157 des_ncbc_encrypt(..., &ks, ...);
12159 (Note that a later change renames 'des_...' into 'DES_...'.)
12163 * Initial reduction of linker bloat: the use of some functions, such as
12164 PEM causes large amounts of unused functions to be linked in due to
12165 poor organisation. For example pem_all.c contains every PEM function
12166 which has a knock on effect of linking in large amounts of (unused)
12167 ASN1 code. Grouping together similar functions and splitting unrelated
12168 functions prevents this.
12172 * Cleanup of EVP macros.
12176 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12177 correct `_ecb suffix`.
12181 * Add initial OCSP responder support to ocsp application. The
12182 revocation information is handled using the text based index
12183 use by the ca application. The responder can either handle
12184 requests generated internally, supplied in files (for example
12185 via a CGI script) or using an internal minimal server.
12189 * Add configuration choices to get zlib compression for TLS.
12193 * Changes to Kerberos SSL for RFC 2712 compliance:
12194 1. Implemented real KerberosWrapper, instead of just using
12195 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12196 2. Implemented optional authenticator field of KerberosWrapper.
12198 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12199 and authenticator structs; see crypto/krb5/.
12201 Generalized Kerberos calls to support multiple Kerberos libraries.
12202 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12203 via Richard Levitte*
12205 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12206 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12207 values for each of the key sizes rather than having just
12208 parameters (and 'speed' generating keys each time).
12212 * Speed up EVP routines.
12215 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12216 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12217 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12218 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12220 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12221 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12222 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12225 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12227 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12231 * Added the OS2-EMX target.
12233 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12235 * Rewrite commands to use `NCONF` routines instead of the old `CONF`.
12236 New functions to support `NCONF `routines in extension code.
12237 New function `CONF_set_nconf()`
12238 to allow functions which take an `NCONF` to also handle the old `LHASH`
12239 structure: this means that the old `CONF` compatible routines can be
12240 retained (in particular w.rt. extensions) without having to duplicate the
12241 code. New function `X509V3_add_ext_nconf_sk()` to add extensions to a stack.
12245 * Enhance the general user interface with mechanisms for inner control
12246 and with possibilities to have yes/no kind of prompts.
12250 * Change all calls to low level digest routines in the library and
12251 applications to use EVP. Add missing calls to HMAC_cleanup() and
12252 don't assume HMAC_CTX can be copied using memcpy().
12254 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12256 * Add the possibility to control engines through control names but with
12257 arbitrary arguments instead of just a string.
12258 Change the key loaders to take a UI_METHOD instead of a callback
12259 function pointer. NOTE: this breaks binary compatibility with earlier
12260 versions of OpenSSL [engine].
12261 Adapt the nCipher code for these new conditions and add a card insertion
12266 * Enhance the general user interface with mechanisms to better support
12267 dialog box interfaces, application-defined prompts, the possibility
12268 to use defaults (for example default passwords from somewhere else)
12269 and interrupts/cancellations.
12273 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12274 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12278 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12279 tidy up some unnecessarily weird code in 'sk_new()').
12281 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12283 * Change the key loading routines for ENGINEs to use the same kind
12284 callback (pem_password_cb) as all other routines that need this
12289 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12290 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12291 than this minimum value is recommended.
12295 * New random seeder for OpenVMS, using the system process statistics
12296 that are easily reachable.
12300 * Windows apparently can't transparently handle global
12301 variables defined in DLLs. Initialisations such as:
12303 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12305 won't compile. This is used by the any applications that need to
12306 declare their own ASN1 modules. This was fixed by adding the option
12307 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12308 needed for static libraries under Win32.
12312 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12313 setting of purpose and trust fields. New X509_STORE trust and
12314 purpose functions and tidy up setting in other SSL functions.
12318 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12319 structure. These are inherited by X509_STORE_CTX when it is
12320 initialised. This allows various defaults to be set in the
12321 X509_STORE structure (such as flags for CRL checking and custom
12322 purpose or trust settings) for functions which only use X509_STORE_CTX
12323 internally such as S/MIME.
12325 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12326 trust settings if they are not set in X509_STORE. This allows X509_STORE
12327 purposes and trust (in S/MIME for example) to override any set by default.
12329 Add command line options for CRL checking to smime, s_client and s_server
12334 * Initial CRL based revocation checking. If the CRL checking flag(s)
12335 are set then the CRL is looked up in the X509_STORE structure and
12336 its validity and signature checked, then if the certificate is found
12337 in the CRL the verify fails with a revoked error.
12339 Various new CRL related callbacks added to X509_STORE_CTX structure.
12341 Command line options added to 'verify' application to support this.
12343 This needs some additional work, such as being able to handle multiple
12344 CRLs with different times, extension based lookup (rather than just
12345 by subject name) and ultimately more complete V2 CRL extension
12350 * Add a general user interface API (crypto/ui/). This is designed
12351 to replace things like des_read_password and friends (backward
12352 compatibility functions using this new API are provided).
12353 The purpose is to remove prompting functions from the DES code
12354 section as well as provide for prompting through dialog boxes in
12355 a window system and the like.
12359 * Add "ex_data" support to ENGINE so implementations can add state at a
12360 per-structure level rather than having to store it globally.
12364 * Make it possible for ENGINE structures to be copied when retrieved by
12365 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12366 This causes the "original" ENGINE structure to act like a template,
12367 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12368 operational state can be localised to each ENGINE structure, despite the
12369 fact they all share the same "methods". New ENGINE structures returned in
12370 this case have no functional references and the return value is the single
12371 structural reference. This matches the single structural reference returned
12372 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12377 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12378 needs to match any other type at all we need to manually clear the
12383 * Changes to the "openssl engine" utility to include;
12384 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12385 about an ENGINE's available control commands.
12386 - executing control commands from command line arguments using the
12387 '-pre' and '-post' switches. '-post' is only used if '-t' is
12388 specified and the ENGINE is successfully initialised. The syntax for
12389 the individual commands are colon-separated, for example;
12390 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12394 * New dynamic control command support for ENGINEs. ENGINEs can now
12395 declare their own commands (numbers), names (strings), descriptions,
12396 and input types for run-time discovery by calling applications. A
12397 subset of these commands are implicitly classed as "executable"
12398 depending on their input type, and only these can be invoked through
12399 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12400 can be based on user input, config files, etc). The distinction is
12401 that "executable" commands cannot return anything other than a boolean
12402 result and can only support numeric or string input, whereas some
12403 discoverable commands may only be for direct use through
12404 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12405 pointers, or other custom uses. The "executable" commands are to
12406 support parameterisations of ENGINE behaviour that can be
12407 unambiguously defined by ENGINEs and used consistently across any
12408 OpenSSL-based application. Commands have been added to all the
12409 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12410 control over shared-library paths without source code alterations.
12414 * Changed all ENGINE implementations to dynamically allocate their
12415 ENGINEs rather than declaring them statically. Apart from this being
12416 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12417 this also allows the implementations to compile without using the
12418 internal engine_int.h header.
12422 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12423 'const' value. Any code that should be able to modify a RAND_METHOD
12424 should already have non-const pointers to it (ie. they should only
12425 modify their own ones).
12429 * Made a variety of little tweaks to the ENGINE code.
12430 - "atalla" and "ubsec" string definitions were moved from header files
12431 to C code. "nuron" string definitions were placed in variables
12432 rather than hard-coded - allowing parameterisation of these values
12433 later on via ctrl() commands.
12434 - Removed unused "#if 0"'d code.
12435 - Fixed engine list iteration code so it uses ENGINE_free() to release
12436 structural references.
12437 - Constified the RAND_METHOD element of ENGINE structures.
12438 - Constified various get/set functions as appropriate and added
12439 missing functions (including a catch-all ENGINE_cpy that duplicates
12440 all ENGINE values onto a new ENGINE except reference counts/state).
12441 - Removed NULL parameter checks in get/set functions. Setting a method
12442 or function to NULL is a way of cancelling out a previously set
12443 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12444 and doesn't justify the extra error symbols and code.
12445 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12446 flags from engine_int.h to engine.h.
12447 - Changed prototypes for ENGINE handler functions (init(), finish(),
12448 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12452 * Implement binary inversion algorithm for BN_mod_inverse in addition
12453 to the algorithm using long division. The binary algorithm can be
12454 used only if the modulus is odd. On 32-bit systems, it is faster
12455 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12456 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12457 up to 450 bits. In 64-bit environments, the binary algorithm
12458 appears to be advantageous for much longer moduli; here we use it
12459 for moduli up to 2048 bits.
12463 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12464 could not support the combine flag in choice fields.
12468 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12469 extensions from a certificate request to the certificate.
12473 * Allow multiple 'certopt' and 'nameopt' options to be separated
12474 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12475 file: this allows the display of the certificate about to be
12476 signed to be customised, to allow certain fields to be included
12477 or excluded and extension details. The old system didn't display
12478 multicharacter strings properly, omitted fields not in the policy
12479 and couldn't display additional details such as extensions.
12483 * Function EC_POINTs_mul for multiple scalar multiplication
12484 of an arbitrary number of elliptic curve points
12485 \sum scalars[i]*points[i],
12486 optionally including the generator defined for the EC_GROUP:
12487 scalar*generator + \sum scalars[i]*points[i].
12489 EC_POINT_mul is a simple wrapper function for the typical case
12490 that the point list has just one item (besides the optional
12495 * First EC_METHODs for curves over GF(p):
12497 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12498 operations and provides various method functions that can also
12499 operate with faster implementations of modular arithmetic.
12501 EC_GFp_mont_method() reuses most functions that are part of
12502 EC_GFp_simple_method, but uses Montgomery arithmetic.
12504 *Bodo Moeller; point addition and point doubling
12505 implementation directly derived from source code provided by
12506 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12508 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12509 crypto/ec/ec_lib.c):
12511 Curves are EC_GROUP objects (with an optional group generator)
12512 based on EC_METHODs that are built into the library.
12514 Points are EC_POINT objects based on EC_GROUP objects.
12516 Most of the framework would be able to handle curves over arbitrary
12517 finite fields, but as there are no obvious types for fields other
12518 than GF(p), some functions are limited to that for now.
12522 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12523 that the file contains a complete HTTP response.
12527 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12528 change the def and num file printf format specifier from "%-40sXXX"
12529 to "%-39s XXX". The latter will always guarantee a space after the
12530 field while the former will cause them to run together if the field
12531 is 40 of more characters long.
12535 * Constify the cipher and digest 'method' functions and structures
12536 and modify related functions to take constant EVP_MD and EVP_CIPHER
12541 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12542 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12546 * Modify `EVP_Digest*()` routines so they now return values. Although the
12547 internal software routines can never fail additional hardware versions
12552 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12554 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12555 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12558 ERR_R_NESTED_ASN1_ERROR
12560 ERR_R_MISSING_ASN1_EOS
12561 were 4 .. 9, conflicting with
12562 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12564 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12565 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12567 Add new error code 'ERR_R_INTERNAL_ERROR'.
12571 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12576 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12577 sets the subject name for a new request or supersedes the
12578 subject name in a given request. Formats that can be parsed are
12579 'CN=Some Name, OU=myOU, C=IT'
12581 'CN=Some Name/OU=myOU/C=IT'.
12583 Add options '-batch' and '-verbose' to 'openssl req'.
12585 *Massimiliano Pala <madwolf@hackmasters.net>*
12587 * Introduce the possibility to access global variables through
12588 functions on platform were that's the best way to handle exporting
12589 global variables in shared libraries. To enable this functionality,
12590 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12591 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12592 is normally done by Configure or something similar).
12594 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12595 in the source file (foo.c) like this:
12597 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12598 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12600 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12601 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12603 OPENSSL_DECLARE_GLOBAL(int,foo);
12604 #define foo OPENSSL_GLOBAL_REF(foo)
12605 OPENSSL_DECLARE_GLOBAL(double,bar);
12606 #define bar OPENSSL_GLOBAL_REF(bar)
12608 The #defines are very important, and therefore so is including the
12609 header file everywhere where the defined globals are used.
12611 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12612 of ASN.1 items, but that structure is a bit different.
12614 The largest change is in util/mkdef.pl which has been enhanced with
12615 better and easier to understand logic to choose which symbols should
12616 go into the Windows .def files as well as a number of fixes and code
12617 cleanup (among others, algorithm keywords are now sorted
12618 lexicographically to avoid constant rewrites).
12622 * In BN_div() keep a copy of the sign of 'num' before writing the
12623 result to 'rm' because if rm==num the value will be overwritten
12624 and produce the wrong result if 'num' is negative: this caused
12625 problems with BN_mod() and BN_nnmod().
12629 * Function OCSP_request_verify(). This checks the signature on an
12630 OCSP request and verifies the signer certificate. The signer
12631 certificate is just checked for a generic purpose and OCSP request
12636 * Add OCSP_check_validity() function to check the validity of OCSP
12637 responses. OCSP responses are prepared in real time and may only
12638 be a few seconds old. Simply checking that the current time lies
12639 between thisUpdate and nextUpdate max reject otherwise valid responses
12640 caused by either OCSP responder or client clock inaccuracy. Instead
12641 we allow thisUpdate and nextUpdate to fall within a certain period of
12642 the current time. The age of the response can also optionally be
12643 checked. Two new options -validity_period and -status_age added to
12648 * If signature or public key algorithm is unrecognized print out its
12649 OID rather that just UNKNOWN.
12653 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12654 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12655 ID to be generated from the issuer certificate alone which can then be
12656 passed to OCSP_id_issuer_cmp().
12660 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12661 ASN1 modules to export functions returning ASN1_ITEM pointers
12662 instead of the ASN1_ITEM structures themselves. This adds several
12663 new macros which allow the underlying ASN1 function/structure to
12664 be accessed transparently. As a result code should not use ASN1_ITEM
12665 references directly (such as &X509_it) but instead use the relevant
12666 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12667 use of the new ASN1 code on platforms where exporting structures
12668 is problematical (for example in shared libraries) but exporting
12669 functions returning pointers to structures is not.
12673 * Add support for overriding the generation of SSL/TLS session IDs.
12674 These callbacks can be registered either in an SSL_CTX or per SSL.
12675 The purpose of this is to allow applications to control, if they wish,
12676 the arbitrary values chosen for use as session IDs, particularly as it
12677 can be useful for session caching in multiple-server environments. A
12678 command-line switch for testing this (and any client code that wishes
12679 to use such a feature) has been added to "s_server".
12681 *Geoff Thorpe, Lutz Jaenicke*
12683 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12684 of the form `#if defined(...) || defined(...) || ...` and
12685 `#if !defined(...) && !defined(...) && ...`. This also avoids
12686 the growing number of special cases it was previously handling.
12690 * Make all configuration macros available for application by making
12691 sure they are available in opensslconf.h, by giving them names starting
12692 with `OPENSSL_` to avoid conflicts with other packages and by making
12693 sure e_os2.h will cover all platform-specific cases together with
12695 Additionally, it is now possible to define configuration/platform-
12696 specific names (called "system identities"). In the C code, these
12697 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12698 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12699 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12704 * New option -set_serial to 'req' and 'x509' this allows the serial
12705 number to use to be specified on the command line. Previously self
12706 signed certificates were hard coded with serial number 0 and the
12707 CA options of 'x509' had to use a serial number in a file which was
12712 * New options to 'ca' utility to support V2 CRL entry extensions.
12713 Currently CRL reason, invalidity date and hold instruction are
12714 supported. Add new CRL extensions to V3 code and some new objects.
12718 * New function EVP_CIPHER_CTX_set_padding() this is used to
12719 disable standard block padding (aka PKCS#5 padding) in the EVP
12720 API, which was previously mandatory. This means that the data is
12721 not padded in any way and so the total length much be a multiple
12722 of the block size, otherwise an error occurs.
12726 * Initial (incomplete) OCSP SSL support.
12730 * New function OCSP_parse_url(). This splits up a URL into its host,
12731 port and path components: primarily to parse OCSP URLs. New -url
12732 option to ocsp utility.
12736 * New nonce behavior. The return value of OCSP_check_nonce() now
12737 reflects the various checks performed. Applications can decide
12738 whether to tolerate certain situations such as an absent nonce
12739 in a response when one was present in a request: the ocsp application
12740 just prints out a warning. New function OCSP_add1_basic_nonce()
12741 this is to allow responders to include a nonce in a response even if
12742 the request is nonce-less.
12746 * Disable stdin buffering in `load_cert()` (`apps/apps.c`) so that no certs are
12747 skipped when using openssl x509 multiple times on a single input file,
12748 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12752 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12753 set string type: to handle setting ASN1_TIME structures. Fix ca
12754 utility to correctly initialize revocation date of CRLs.
12758 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12759 the clients preferred ciphersuites and rather use its own preferences.
12760 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12761 Internet Explorer by ensuring unchanged hash method during stepup.
12762 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12766 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12767 to aes and add a new 'exist' option to print out symbols that don't
12772 * Additional options to ocsp utility to allow flags to be set and
12773 additional certificates supplied.
12777 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12778 OCSP client a number of certificate to only verify the response
12783 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12784 handle the new API. Currently only ECB, CBC modes supported. Add new
12787 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12788 Encryption Standard (AES) Ciphersuites for Transport Layer
12789 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12790 not enabled by default and were not part of the "ALL" ciphersuite
12791 alias because they were not yet official; they could be
12792 explicitly requested by specifying the "AESdraft" ciphersuite
12793 group alias. In the final release of OpenSSL 0.9.7, the group
12794 alias is called "AES" and is part of "ALL".)
12796 *Ben Laurie, Steve Henson, Bodo Moeller*
12798 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12799 request to response.
12803 * Functions for OCSP responders. OCSP_request_onereq_count(),
12804 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12805 extract information from a certificate request. OCSP_response_create()
12806 creates a response and optionally adds a basic response structure.
12807 OCSP_basic_add1_status() adds a complete single response to a basic
12808 response and returns the OCSP_SINGLERESP structure just added (to allow
12809 extensions to be included for example). OCSP_basic_add1_cert() adds a
12810 certificate to a basic response and OCSP_basic_sign() signs a basic
12811 response with various flags. New helper functions ASN1_TIME_check()
12812 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12813 (converts ASN1_TIME to GeneralizedTime).
12817 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12818 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12819 structure from a certificate. X509_pubkey_digest() digests the public_key
12820 contents: this is used in various key identifiers.
12824 * Make sk_sort() tolerate a NULL argument.
12826 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12828 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12829 passed by the function are trusted implicitly. If any of them signed the
12830 response then it is assumed to be valid and is not verified.
12834 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12835 to data. This was previously part of the PKCS7 ASN1 code. This
12836 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12837 *Steve Henson, reported by Kenneth R. Robinette
12838 <support@securenetterm.com>*
12840 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12841 routines: without these tracing memory leaks is very painful.
12842 Fix leaks in PKCS12 and PKCS7 routines.
12846 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12847 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12848 effectively meant GeneralizedTime would never be used. Now it
12849 is initialised to -1 but X509_time_adj() now has to check the value
12850 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12851 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12852 *Steve Henson, reported by Kenneth R. Robinette
12853 <support@securenetterm.com>*
12855 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12856 result in a zero length in the ASN1_INTEGER structure which was
12857 not consistent with the structure when d2i_ASN1_INTEGER() was used
12858 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12859 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12860 where it did not print out a minus for negative ASN1_INTEGER.
12864 * Add summary printout to ocsp utility. The various functions which
12865 convert status values to strings have been renamed to:
12866 OCSP_response_status_str(), OCSP_cert_status_str() and
12867 OCSP_crl_reason_str() and are no longer static. New options
12868 to verify nonce values and to disable verification. OCSP response
12869 printout format cleaned up.
12873 * Add additional OCSP certificate checks. These are those specified
12874 in RFC2560. This consists of two separate checks: the CA of the
12875 certificate being checked must either be the OCSP signer certificate
12876 or the issuer of the OCSP signer certificate. In the latter case the
12877 OCSP signer certificate must contain the OCSP signing extended key
12878 usage. This check is performed by attempting to match the OCSP
12879 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12880 in the OCSP_CERTID structures of the response.
12884 * Initial OCSP certificate verification added to OCSP_basic_verify()
12885 and related routines. This uses the standard OpenSSL certificate
12886 verify routines to perform initial checks (just CA validity) and
12887 to obtain the certificate chain. Then additional checks will be
12888 performed on the chain. Currently the root CA is checked to see
12889 if it is explicitly trusted for OCSP signing. This is used to set
12890 a root CA as a global signing root: that is any certificate that
12891 chains to that CA is an acceptable OCSP signing certificate.
12895 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12896 extensions from a separate configuration file.
12897 As when reading extensions from the main configuration file,
12898 the '-extensions ...' option may be used for specifying the
12901 *Massimiliano Pala <madwolf@comune.modena.it>*
12903 * New OCSP utility. Allows OCSP requests to be generated or
12904 read. The request can be sent to a responder and the output
12905 parsed, outputted or printed in text form. Not complete yet:
12906 still needs to check the OCSP response validity.
12910 * New subcommands for 'openssl ca':
12911 `openssl ca -status <serial>` prints the status of the cert with
12912 the given serial number (according to the index file).
12913 `openssl ca -updatedb` updates the expiry status of certificates
12916 *Massimiliano Pala <madwolf@comune.modena.it>*
12918 * New '-newreq-nodes' command option to CA.pl. This is like
12919 '-newreq', but calls 'openssl req' with the '-nodes' option
12920 so that the resulting key is not encrypted.
12922 *Damien Miller <djm@mindrot.org>*
12924 * New configuration for the GNU Hurd.
12926 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12928 * Initial code to implement OCSP basic response verify. This
12929 is currently incomplete. Currently just finds the signer's
12930 certificate and verifies the signature on the response.
12934 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12935 value of OPENSSLDIR. This is available via the new '-d' option
12936 to 'openssl version', and is also included in 'openssl version -a'.
12940 * Allowing defining memory allocation callbacks that will be given
12941 file name and line number information in additional arguments
12942 (a `const char*` and an int). The basic functionality remains, as
12943 well as the original possibility to just replace malloc(),
12944 realloc() and free() by functions that do not know about these
12945 additional arguments. To register and find out the current
12946 settings for extended allocation functions, the following
12947 functions are provided:
12949 CRYPTO_set_mem_ex_functions
12950 CRYPTO_set_locked_mem_ex_functions
12951 CRYPTO_get_mem_ex_functions
12952 CRYPTO_get_locked_mem_ex_functions
12954 These work the same way as CRYPTO_set_mem_functions and friends.
12955 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12956 extended allocation function is enabled.
12957 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12958 a conventional allocation function is enabled.
12960 *Richard Levitte, Bodo Moeller*
12962 * Finish off removing the remaining LHASH function pointer casts.
12963 There should no longer be any prototype-casting required when using
12964 the LHASH abstraction, and any casts that remain are "bugs". See
12965 the callback types and macros at the head of lhash.h for details
12966 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12970 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12971 If /dev/[u]random devices are not available or do not return enough
12972 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12974 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12975 /etc/entropy will be queried once each in this sequence, querying stops
12976 when enough entropy was collected without querying more sockets.
12980 * Change the Unix RAND_poll() variant to be able to poll several
12981 random devices, as specified by DEVRANDOM, until a sufficient amount
12982 of data has been collected. We spend at most 10 ms on each file
12983 (select timeout) and read in non-blocking mode. DEVRANDOM now
12984 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12985 (previously it was just the string "/dev/urandom"), so on typical
12986 platforms the 10 ms delay will never occur.
12987 Also separate out the Unix variant to its own file, rand_unix.c.
12988 For VMS, there's a currently-empty rand_vms.c.
12992 * Move OCSP client related routines to ocsp_cl.c. These
12993 provide utility functions which an application needing
12994 to issue a request to an OCSP responder and analyse the
12995 response will typically need: as opposed to those which an
12996 OCSP responder itself would need which will be added later.
12998 OCSP_request_sign() signs an OCSP request with an API similar
12999 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
13000 response. OCSP_response_get1_basic() extracts basic response
13001 from response. OCSP_resp_find_status(): finds and extracts status
13002 information from an OCSP_CERTID structure (which will be created
13003 when the request structure is built). These are built from lower
13004 level functions which work on OCSP_SINGLERESP structures but
13005 won't normally be used unless the application wishes to examine
13006 extensions in the OCSP response for example.
13008 Replace nonce routines with a pair of functions.
13009 OCSP_request_add1_nonce() adds a nonce value and optionally
13010 generates a random value. OCSP_check_nonce() checks the
13011 validity of the nonce in an OCSP response.
13015 * Change function OCSP_request_add() to OCSP_request_add0_id().
13016 This doesn't copy the supplied OCSP_CERTID and avoids the
13017 need to free up the newly created id. Change return type
13018 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13019 This can then be used to add extensions to the request.
13020 Deleted OCSP_request_new(), since most of its functionality
13021 is now in OCSP_REQUEST_new() (and the case insensitive name
13022 clash) apart from the ability to set the request name which
13023 will be added elsewhere.
13027 * Update OCSP API. Remove obsolete extensions argument from
13028 various functions. Extensions are now handled using the new
13029 OCSP extension code. New simple OCSP HTTP function which
13030 can be used to send requests and parse the response.
13034 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13035 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13036 uses the special reorder version of SET OF to sort the attributes
13037 and reorder them to match the encoded order. This resolves a long
13038 standing problem: a verify on a PKCS7 structure just after signing
13039 it used to fail because the attribute order did not match the
13040 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13041 it uses the received order. This is necessary to tolerate some broken
13042 software that does not order SET OF. This is handled by encoding
13043 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13044 to produce the required SET OF.
13048 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13049 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13050 files to get correct declarations of the ASN.1 item variables.
13054 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13055 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13056 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13057 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13058 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13059 ASN1_ITEM and no wrapper functions.
13063 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13064 replace the old function pointer based I/O routines. Change most of
13065 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13069 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13070 lines, recognize more "algorithms" that can be deselected, and make
13071 it complain about algorithm deselection that isn't recognised.
13075 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13076 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13077 to use new functions. Add NO_ASN1_OLD which can be set to remove
13078 some old style ASN1 functions: this can be used to determine if old
13079 code will still work when these eventually go away.
13083 * New extension functions for OCSP structures, these follow the
13084 same conventions as certificates and CRLs.
13088 * New function X509V3_add1_i2d(). This automatically encodes and
13089 adds an extension. Its behaviour can be customised with various
13090 flags to append, replace or delete. Various wrappers added for
13091 certificates and CRLs.
13095 * Fix to avoid calling the underlying ASN1 print routine when
13096 an extension cannot be parsed. Correct a typo in the
13097 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13101 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13102 entries for variables.
13106 * Add functionality to `apps/openssl.c` for detecting locking
13107 problems: As the program is single-threaded, all we have
13108 to do is register a locking callback using an array for
13109 storing which locks are currently held by the program.
13113 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13114 SSL_get_ex_data_X509_STORE_idx(), which is used in
13115 ssl_verify_cert_chain() and thus can be called at any time
13116 during TLS/SSL handshakes so that thread-safety is essential.
13117 Unfortunately, the ex_data design is not at all suited
13118 for multi-threaded use, so it probably should be abolished.
13122 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13124 *Broadcom, tweaked and integrated by Geoff Thorpe*
13126 * Move common extension printing code to new function
13127 X509V3_print_extensions(). Reorganise OCSP print routines and
13128 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13132 * New function X509_signature_print() to remove duplication in some
13137 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13138 set (this was treated exactly the same as SET OF previously). This
13139 is used to reorder the STACK representing the structure to match the
13140 encoding. This will be used to get round a problem where a PKCS7
13141 structure which was signed could not be verified because the STACK
13142 order did not reflect the encoded order.
13146 * Reimplement the OCSP ASN1 module using the new code.
13150 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13151 for its ASN1 operations. The old style function pointers still exist
13152 for now but they will eventually go away.
13156 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13157 completely replaces the old ASN1 functionality with a table driven
13158 encoder and decoder which interprets an ASN1_ITEM structure describing
13159 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13160 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13161 has also been converted to the new form.
13165 * Change BN_mod_exp_recp so that negative moduli are tolerated
13166 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13167 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13168 for negative moduli.
13172 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13173 of not touching the result's sign bit.
13177 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13182 * Changed the LHASH code to use prototypes for callbacks, and created
13183 macros to declare and implement thin (optionally static) functions
13184 that provide type-safety and avoid function pointer casting for the
13185 type-specific callbacks.
13189 * Added Kerberos Cipher Suites to be used with TLS, as written in
13191 *Veers Staats <staatsvr@asc.hpc.mil>,
13192 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13194 * Reformat the FAQ so the different questions and answers can be divided
13195 in sections depending on the subject.
13199 * Have the zlib compression code load ZLIB.DLL dynamically under
13204 * New function BN_mod_sqrt for computing square roots modulo a prime
13205 (using the probabilistic Tonelli-Shanks algorithm unless
13206 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13207 be handled deterministically).
13209 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13211 * Make BN_mod_inverse faster by explicitly handling small quotients
13212 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13213 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13217 * New function BN_kronecker.
13221 * Fix BN_gcd so that it works on negative inputs; the result is
13222 positive unless both parameters are zero.
13223 Previously something reasonably close to an infinite loop was
13224 possible because numbers could be growing instead of shrinking
13225 in the implementation of Euclid's algorithm.
13229 * Fix BN_is_word() and BN_is_one() macros to take into account the
13230 sign of the number in question.
13232 Fix BN_is_word(a,w) to work correctly for w == 0.
13234 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13235 because its test if the absolute value of 'a' equals 'w'.
13236 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13237 it exists mostly for use in the implementations of BN_is_zero(),
13238 BN_is_one(), and BN_is_word().
13242 * New function BN_swap.
13246 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13247 the exponentiation functions are more likely to produce reasonable
13248 results on negative inputs.
13252 * Change BN_mod_mul so that the result is always non-negative.
13253 Previously, it could be negative if one of the factors was negative;
13254 I don't think anyone really wanted that behaviour.
13258 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13259 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13260 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13261 and add new functions:
13270 BN_mod_lshift1_quick
13272 BN_mod_lshift_quick
13274 These functions always generate non-negative results.
13276 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13277 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13279 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13280 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13281 be reduced modulo `m`.
13283 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13286 The following entry accidentally appeared in the CHANGES file
13287 distributed with OpenSSL 0.9.7. The modifications described in
13288 it do *not* apply to OpenSSL 0.9.7.
13290 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13291 was actually never needed) and in BN_mul(). The removal in BN_mul()
13292 required a small change in bn_mul_part_recursive() and the addition
13293 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13294 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13295 bn_sub_words() and bn_add_words() except they take arrays with
13301 * In 'openssl passwd', verify passwords read from the terminal
13302 unless the '-salt' option is used (which usually means that
13303 verification would just waste user's time since the resulting
13304 hash is going to be compared with some given password hash)
13305 or the new '-noverify' option is used.
13307 This is an incompatible change, but it does not affect
13308 non-interactive use of 'openssl passwd' (passwords on the command
13309 line, '-stdin' option, '-in ...' option) and thus should not
13310 cause any problems.
13314 * Remove all references to RSAref, since there's no more need for it.
13318 * Make DSO load along a path given through an environment variable
13319 (SHLIB_PATH) with shl_load().
13323 * Constify the ENGINE code as a result of BIGNUM constification.
13324 Also constify the RSA code and most things related to it. In a
13325 few places, most notable in the depth of the ASN.1 code, ugly
13326 casts back to non-const were required (to be solved at a later
13331 * Make it so the openssl application has all engines loaded by default.
13335 * Constify the BIGNUM routines a little more.
13339 * Add the following functions:
13341 ENGINE_load_cswift()
13343 ENGINE_load_atalla()
13344 ENGINE_load_nuron()
13345 ENGINE_load_builtin_engines()
13347 That way, an application can itself choose if external engines that
13348 are built-in in OpenSSL shall ever be used or not. The benefit is
13349 that applications won't have to be linked with libdl or other dso
13350 libraries unless it's really needed.
13352 Changed 'openssl engine' to load all engines on demand.
13353 Changed the engine header files to avoid the duplication of some
13354 declarations (they differed!).
13358 * 'openssl engine' can now list capabilities.
13362 * Better error reporting in 'openssl engine'.
13366 * Never call load_dh_param(NULL) in s_server.
13370 * Add engine application. It can currently list engines by name and
13371 identity, and test if they are actually available.
13375 * Improve RPM specification file by forcing symbolic linking and making
13376 sure the installed documentation is also owned by root.root.
13378 *Damien Miller <djm@mindrot.org>*
13380 * Give the OpenSSL applications more possibilities to make use of
13381 keys (public as well as private) handled by engines.
13385 * Add OCSP code that comes from CertCo.
13389 * Add VMS support for the Rijndael code.
13393 * Added untested support for Nuron crypto accelerator.
13397 * Add support for external cryptographic devices. This code was
13398 previously distributed separately as the "engine" branch.
13400 *Geoff Thorpe, Richard Levitte*
13402 * Rework the filename-translation in the DSO code. It is now possible to
13403 have far greater control over how a "name" is turned into a filename
13404 depending on the operating environment and any oddities about the
13405 different shared library filenames on each system.
13409 * Support threads on FreeBSD-elf in Configure.
13413 * Fix for SHA1 assembly problem with MASM: it produces
13414 warnings about corrupt line number information when assembling
13415 with debugging information. This is caused by the overlapping
13418 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13421 NCONF_get_number() has no error checking at all. As a replacement,
13422 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13423 promoted strongly. The old NCONF_get_number is kept around for
13424 binary backward compatibility.
13425 Make it possible for methods to load from something other than a BIO,
13426 by providing a function pointer that is given a name instead of a BIO.
13427 For example, this could be used to load configuration data from an
13432 * Fix for non blocking accept BIOs. Added new I/O special reason
13433 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13434 with non blocking I/O was not possible because no retry code was
13435 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13440 * Added the beginnings of Rijndael support.
13444 * Fix for bug in DirectoryString mask setting. Add support for
13445 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13446 to allow certificate printing to more controllable, additional
13447 'certopt' option to 'x509' to allow new printing options to be
13452 * Clean old EAY MD5 hack from e_os.h.
13456 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13458 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13459 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13461 *Joe Orton, Steve Henson*
13463 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13465 * Fix additional bug revealed by the NISCC test suite:
13467 Stop bug triggering large recursion when presented with
13468 certain ASN.1 tags ([CVE-2003-0851])
13472 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13474 * Fix various bugs revealed by running the NISCC test suite:
13476 Stop out of bounds reads in the ASN1 code when presented with
13477 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13479 If verify callback ignores invalid public key errors don't try to check
13480 certificate signature with the NULL public key.
13484 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13485 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13490 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13491 extra data after the compression methods not only for TLS 1.0
13492 but also for SSL 3.0 (as required by the specification).
13494 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13496 * Change X509_certificate_type() to mark the key as exported/exportable
13497 when it's 512 *bits* long, not 512 bytes.
13501 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13503 * Countermeasure against the Klima-Pokorny-Rosa extension of
13504 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13505 a protocol version number mismatch like a decryption error
13506 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13510 * Turn on RSA blinding by default in the default implementation
13511 to avoid a timing attack. Applications that don't want it can call
13512 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13513 They would be ill-advised to do so in most cases.
13515 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13517 * Change RSA blinding code so that it works when the PRNG is not
13518 seeded (in this case, the secret RSA exponent is abused as
13519 an unpredictable seed -- if it is not unpredictable, there
13520 is no point in blinding anyway). Make RSA blinding thread-safe
13521 by remembering the creator's thread ID in rsa->blinding and
13522 having all other threads use local one-time blinding factors
13523 (this requires more computation than sharing rsa->blinding, but
13524 avoids excessive locking; and if an RSA object is not shared
13525 between threads, blinding will still be very fast).
13529 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13531 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13532 via timing by performing a MAC computation even if incorrect
13533 block cipher padding has been found. This is a countermeasure
13534 against active attacks where the attacker has to distinguish
13535 between bad padding and a MAC verification error. ([CVE-2003-0078])
13537 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13538 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13539 Martin Vuagnoux (EPFL, Ilion)*
13541 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13543 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13544 memory from its contents. This is done with a counter that will
13545 place alternating values in each byte. This can be used to solve
13546 two issues: 1) the removal of calls to memset() by highly optimizing
13547 compilers, and 2) cleansing with other values than 0, since those can
13548 be read through on certain media, for example a swap space on disk.
13552 * Bugfix: client side session caching did not work with external caching,
13553 because the session->cipher setting was not restored when reloading
13554 from the external cache. This problem was masked, when
13555 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13556 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13560 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13561 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13563 *Zeev Lieber <zeev-l@yahoo.com>*
13565 * Undo an undocumented change introduced in 0.9.6e which caused
13566 repeated calls to OpenSSL_add_all_ciphers() and
13567 OpenSSL_add_all_digests() to be ignored, even after calling
13572 * Change the default configuration reader to deal with last line not
13573 being properly terminated.
13577 * Change X509_NAME_cmp() so it applies the special rules on handling
13578 DN values that are of type PrintableString, as well as RDNs of type
13579 emailAddress where the value has the type ia5String.
13581 *stefank@valicert.com via Richard Levitte*
13583 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13584 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13585 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13586 the bitwise-OR of the two for use by the majority of applications
13587 wanting this behaviour, and update the docs. The documented
13588 behaviour and actual behaviour were inconsistent and had been
13589 changing anyway, so this is more a bug-fix than a behavioural
13592 *Geoff Thorpe, diagnosed by Nadav Har'El*
13594 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13595 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13599 * Fix initialization code race conditions in
13600 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13601 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13602 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13603 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13604 ssl2_get_cipher_by_char(),
13605 ssl3_get_cipher_by_char().
13607 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13609 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13610 the cached sessions are flushed, as the remove_cb() might use ex_data
13611 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13612 (see [openssl.org #212]).
13614 *Geoff Thorpe, Lutz Jaenicke*
13616 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13617 length, instead of the encoding length to d2i_ASN1_OBJECT.
13621 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13623 * [In 0.9.6g-engine release:]
13624 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13626 *Lynn Gazis <lgazis@rainbow.com>*
13628 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13630 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13631 and get fix the header length calculation.
13632 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13633 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13635 * Use proper error handling instead of 'assertions' in buffer
13636 overflow checks added in 0.9.6e. This prevents DoS (the
13637 assertions could call abort()).
13639 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13641 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13643 * Add various sanity checks to asn1_get_length() to reject
13644 the ASN1 length bytes if they exceed sizeof(long), will appear
13645 negative or the content length exceeds the length of the
13648 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13650 * Fix cipher selection routines: ciphers without encryption had no flags
13651 for the cipher strength set and where therefore not handled correctly
13652 by the selection routines (PR #130).
13656 * Fix EVP_dsa_sha macro.
13661 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13662 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13663 that was added in OpenSSL 0.9.6d.
13665 As the countermeasure turned out to be incompatible with some
13666 broken SSL implementations, the new option is part of SSL_OP_ALL.
13667 SSL_OP_ALL is usually employed when compatibility with weird SSL
13668 implementations is desired (e.g. '-bugs' option to 's_client' and
13669 's_server'), so the new option is automatically set in many
13674 * Changes in security patch:
13676 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13677 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13678 Air Force Materiel Command, USAF, under agreement number
13681 * Add various sanity checks to asn1_get_length() to reject
13682 the ASN1 length bytes if they exceed sizeof(long), will appear
13683 negative or the content length exceeds the length of the
13684 supplied buffer. ([CVE-2002-0659])
13686 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13688 * Assertions for various potential buffer overflows, not known to
13689 happen in practice.
13691 *Ben Laurie (CHATS)*
13693 * Various temporary buffers to hold ASCII versions of integers were
13694 too small for 64 bit platforms. ([CVE-2002-0655])
13695 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13697 * Remote buffer overflow in SSL3 protocol - an attacker could
13698 supply an oversized session ID to a client. ([CVE-2002-0656])
13700 *Ben Laurie (CHATS)*
13702 * Remote buffer overflow in SSL2 protocol - an attacker could
13703 supply an oversized client master key. ([CVE-2002-0656])
13705 *Ben Laurie (CHATS)*
13707 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13709 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13710 encoded as NULL) with id-dsa-with-sha1.
13712 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13714 * Check various `X509_...()` return values in `apps/req.c`.
13716 *Nils Larsch <nla@trustcenter.de>*
13718 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13719 an end-of-file condition would erroneously be flagged, when the CRLF
13720 was just at the end of a processed block. The bug was discovered when
13721 processing data through a buffering memory BIO handing the data to a
13722 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13723 <ptsekov@syntrex.com> and Nedelcho Stanev.
13727 * Implement a countermeasure against a vulnerability recently found
13728 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13729 before application data chunks to avoid the use of known IVs
13730 with data potentially chosen by the attacker.
13734 * Fix length checks in ssl3_get_client_hello().
13738 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13739 to prevent ssl3_read_internal() from incorrectly assuming that
13740 ssl3_read_bytes() found application data while handshake
13741 processing was enabled when in fact s->s3->in_read_app_data was
13742 merely automatically cleared during the initial handshake.
13744 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13746 * Fix object definitions for Private and Enterprise: they were not
13747 recognized in their shortname (=lowercase) representation. Extend
13748 obj_dat.pl to issue an error when using undefined keywords instead
13749 of silently ignoring the problem (Svenning Sorensen
13750 <sss@sss.dnsalias.net>).
13754 * Fix DH_generate_parameters() so that it works for 'non-standard'
13755 generators, i.e. generators other than 2 and 5. (Previously, the
13756 code did not properly initialise the 'add' and 'rem' values to
13757 BN_generate_prime().)
13759 In the new general case, we do not insist that 'generator' is
13760 actually a primitive root: This requirement is rather pointless;
13761 a generator of the order-q subgroup is just as good, if not
13766 * Map new X509 verification errors to alerts. Discovered and submitted by
13767 Tom Wu <tom@arcot.com>.
13771 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13772 returning non-zero before the data has been completely received
13773 when using non-blocking I/O.
13775 *Bodo Moeller; problem pointed out by John Hughes*
13777 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13779 *Ben Laurie, Lutz Jaenicke*
13781 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13782 Yoram Zahavi <YoramZ@gilian.com>).
13786 * Add information about CygWin 1.3 and on, and preserve proper
13787 configuration for the versions before that.
13789 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13791 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13792 check whether we deal with a copy of a session and do not delete from
13793 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13794 <izhar@checkpoint.com>.
13798 * Do not store session data into the internal session cache, if it
13799 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13800 flag is set). Proposed by Aslam <aslam@funk.com>.
13804 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13809 * [In 0.9.6d-engine release:]
13810 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13812 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13814 * Add the configuration target linux-s390x.
13816 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13818 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13819 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13820 variable as an indication that a ClientHello message has been
13821 received. As the flag value will be lost between multiple
13822 invocations of ssl3_accept when using non-blocking I/O, the
13823 function may not be aware that a handshake has actually taken
13824 place, thus preventing a new session from being added to the
13827 To avoid this problem, we now set s->new_session to 2 instead of
13828 using a local variable.
13830 *Lutz Jaenicke, Bodo Moeller*
13832 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13833 if the SSL_R_LENGTH_MISMATCH error is detected.
13835 *Geoff Thorpe, Bodo Moeller*
13837 * New 'shared_ldflag' column in Configure platform table.
13841 * Fix EVP_CIPHER_mode macro.
13843 *"Dan S. Camper" <dan@bti.net>*
13845 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13846 type, we must throw them away by setting rr->length to 0.
13848 *D P Chang <dpc@qualys.com>*
13850 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13852 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13853 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13854 worked incorrectly for those cases where range = `10..._2` and
13855 `3*range` is two bits longer than range.)
13859 * Only add signing time to PKCS7 structures if it is not already
13864 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13865 OBJ_ld_ce should be OBJ_id_ce.
13866 Also some ip-pda OIDs in crypto/objects/objects.txt were
13867 incorrect (cf. RFC 3039).
13869 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13871 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13872 returns early because it has nothing to do.
13874 *Andy Schneider <andy.schneider@bjss.co.uk>*
13876 * [In 0.9.6c-engine release:]
13877 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13879 *Andy Schneider <andy.schneider@bjss.co.uk>*
13881 * [In 0.9.6c-engine release:]
13882 Add support for Cryptographic Appliance's keyserver technology.
13883 (Use engine 'keyclient')
13885 *Cryptographic Appliances and Geoff Thorpe*
13887 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13888 is called via tools/c89.sh because arguments have to be
13889 rearranged (all '-L' options must appear before the first object
13892 *Richard Shapiro <rshapiro@abinitio.com>*
13894 * [In 0.9.6c-engine release:]
13895 Add support for Broadcom crypto accelerator cards, backported
13898 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13900 * [In 0.9.6c-engine release:]
13901 Add support for SureWare crypto accelerator cards from
13902 Baltimore Technologies. (Use engine 'sureware')
13904 *Baltimore Technologies and Mark Cox*
13906 * [In 0.9.6c-engine release:]
13907 Add support for crypto accelerator cards from Accelerated
13908 Encryption Processing, www.aep.ie. (Use engine 'aep')
13910 *AEP Inc. and Mark Cox*
13912 * Add a configuration entry for gcc on UnixWare.
13914 *Gary Benson <gbenson@redhat.com>*
13916 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13917 messages are stored in a single piece (fixed-length part and
13918 variable-length part combined) and fix various bugs found on the way.
13922 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13923 instead. BIO_gethostbyname() does not know what timeouts are
13924 appropriate, so entries would stay in cache even when they have
13926 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13928 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13929 faced with a pathologically small ClientHello fragment that does
13930 not contain client_version: Instead of aborting with an error,
13931 simply choose the highest available protocol version (i.e.,
13932 TLS 1.0 unless it is disabled). In practice, ClientHello
13933 messages are never sent like this, but this change gives us
13934 strictly correct behaviour at least for TLS.
13938 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13939 never resets s->method to s->ctx->method when called from within
13940 one of the SSL handshake functions.
13942 *Bodo Moeller; problem pointed out by Niko Baric*
13944 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13945 (sent using the client's version number) if client_version is
13946 smaller than the protocol version in use. Also change
13947 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13948 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13949 the client will at least see that alert.
13953 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13958 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13959 client receives HelloRequest while in a handshake.
13961 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13963 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13964 should end in 'break', not 'goto end' which circumvents various
13965 cleanups done in state SSL_ST_OK. But session related stuff
13966 must be disabled for SSL_ST_OK in the case that we just sent a
13969 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13970 before just sending a HelloRequest.
13972 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13974 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13975 reveal whether illegal block cipher padding was found or a MAC
13976 verification error occurred. (Neither SSLerr() codes nor alerts
13977 are directly visible to potential attackers, but the information
13978 may leak via logfiles.)
13980 Similar changes are not required for the SSL 2.0 implementation
13981 because the number of padding bytes is sent in clear for SSL 2.0,
13982 and the extra bytes are just ignored. However ssl/s2_pkt.c
13983 failed to verify that the purported number of padding bytes is in
13988 * Add OpenUNIX-8 support including shared libraries
13989 (Boyd Lynn Gerber <gerberb@zenez.com>).
13993 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13994 'wristwatch attack' using huge encoding parameters (cf.
13995 James H. Manger's CRYPTO 2001 paper). Note that the
13996 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13997 encoding parameters and hence was not vulnerable.
14001 * BN_sqr() bug fix.
14003 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14005 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14006 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14007 followed by modular reduction.
14009 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14011 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14012 equivalent based on BN_pseudo_rand() instead of BN_rand().
14016 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14017 This function was broken, as the check for a new client hello message
14018 to handle SGC did not allow these large messages.
14019 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14023 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14027 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14028 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14032 * Rework the configuration and shared library support for Tru64 Unix.
14033 The configuration part makes use of modern compiler features and
14034 still retains old compiler behavior for those that run older versions
14035 of the OS. The shared library support part includes a variant that
14036 uses the RPATH feature, and is available through the special
14037 configuration target "alpha-cc-rpath", which will never be selected
14040 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14042 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14043 with the same message size as in ssl3_get_certificate_request().
14044 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14045 messages might inadvertently be reject as too long.
14047 *Petr Lampa <lampa@fee.vutbr.cz>*
14049 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14053 * Modified SSL library such that the verify_callback that has been set
14054 specifically for an SSL object with SSL_set_verify() is actually being
14055 used. Before the change, a verify_callback set with this function was
14056 ignored and the verify_callback() set in the SSL_CTX at the time of
14057 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14058 to allow the necessary settings.
14062 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14063 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14064 done automatically (in contradiction to the requirements of the C
14065 standard). This made problems when used from OpenSSH.
14069 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14070 dh->length and always used
14072 BN_rand_range(priv_key, dh->p).
14074 BN_rand_range() is not necessary for Diffie-Hellman, and this
14075 specific range makes Diffie-Hellman unnecessarily inefficient if
14076 dh->length (recommended exponent length) is much smaller than the
14077 length of dh->p. We could use BN_rand_range() if the order of
14078 the subgroup was stored in the DH structure, but we only have
14083 BN_rand(priv_key, l, ...)
14085 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14092 RSA_eay_public_encrypt
14093 RSA_eay_private_decrypt
14094 RSA_eay_private_encrypt (signing)
14095 RSA_eay_public_decrypt (signature verification)
14097 (default implementations for RSA_public_encrypt,
14098 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14099 always reject numbers >= n.
14103 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14104 to synchronize access to 'locking_thread'. This is necessary on
14105 systems where access to 'locking_thread' (an 'unsigned long'
14106 variable) is not atomic.
14110 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14111 *before* setting the 'crypto_lock_rand' flag. The previous code had
14112 a race condition if 0 is a valid thread ID.
14114 *Travis Vitek <vitek@roguewave.com>*
14116 * Add support for shared libraries under Irix.
14118 *Albert Chin-A-Young <china@thewrittenword.com>*
14120 * Add configuration option to build on Linux on both big-endian and
14121 little-endian MIPS.
14123 *Ralf Baechle <ralf@uni-koblenz.de>*
14125 * Add the possibility to create shared libraries on HP-UX.
14129 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14131 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14132 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14133 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14134 PRNG state recovery was possible based on the output of
14135 one PRNG request appropriately sized to gain knowledge on
14136 'md' followed by enough consecutive 1-byte PRNG requests
14137 to traverse all of 'state'.
14139 1. When updating 'md_local' (the current thread's copy of 'md')
14140 during PRNG output generation, hash all of the previous
14141 'md_local' value, not just the half used for PRNG output.
14143 2. Make the number of bytes from 'state' included into the hash
14144 independent from the number of PRNG bytes requested.
14146 The first measure alone would be sufficient to avoid
14147 Markku-Juhani's attack. (Actually it had never occurred
14148 to me that the half of 'md_local' used for chaining was the
14149 half from which PRNG output bytes were taken -- I had always
14150 assumed that the secret half would be used.) The second
14151 measure makes sure that additional data from 'state' is never
14152 mixed into 'md_local' in small portions; this heuristically
14153 further strengthens the PRNG.
14157 * Fix crypto/bn/asm/mips3.s.
14161 * When only the key is given to "enc", the IV is undefined. Print out
14162 an error message in this case.
14166 * Handle special case when X509_NAME is empty in X509 printing routines.
14170 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14171 positive and less than q.
14175 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14176 used: it isn't thread safe and the add_lock_callback should handle
14179 *Paul Rose <Paul.Rose@bridge.com>*
14181 * Verify that incoming data obeys the block size in
14182 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14188 *Ulf Möller, Bodo Möller*
14190 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14191 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14192 when fixing the server behaviour for backwards-compatible 'client
14193 hello' messages. (Note that the attack is impractical against
14194 SSL 3.0 and TLS 1.0 anyway because length and version checking
14195 means that the probability of guessing a valid ciphertext is
14196 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14199 Before 0.9.5, the countermeasure (hide the error by generating a
14200 random 'decryption result') did not work properly because
14201 ERR_clear_error() was missing, meaning that SSL_get_error() would
14202 detect the supposedly ignored error.
14204 Both problems are now fixed.
14208 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14209 (previously it was 1024).
14213 * Fix for compatibility mode trust settings: ignore trust settings
14214 unless some valid trust or reject settings are present.
14218 * Fix for blowfish EVP: its a variable length cipher.
14222 * Fix various bugs related to DSA S/MIME verification. Handle missing
14223 parameters in DSA public key structures and return an error in the
14224 DSA routines if parameters are absent.
14228 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14229 in the current directory if neither $RANDFILE nor $HOME was set.
14230 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14231 caused some confusion to Windows users who haven't defined $HOME.
14232 Thus RAND_file_name() is changed again: e_os.h can define a
14233 DEFAULT_HOME, which will be used if $HOME is not set.
14234 For Windows, we use "C:"; on other platforms, we still require
14235 environment variables.
14237 * Move 'if (!initialized) RAND_poll()' into regions protected by
14238 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14239 having multiple threads call RAND_poll() concurrently.
14243 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14244 combination of a flag and a thread ID variable.
14245 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14246 flag), *other* threads can enter ssleay_add_bytes without obeying
14247 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14248 that they do not hold after the first thread unsets add_do_not_lock).
14252 * Change bctest again: '-x' expressions are not available in all
14253 versions of 'test'.
14257 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14259 * Fix a couple of memory leaks in PKCS7_dataDecode()
14261 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14263 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14264 the default extension for executables, if any. Also, make the perl
14265 scripts that use symlink() to test if it really exists and use "cp"
14266 if it doesn't. All this made OpenSSL compilable and installable in
14271 * Fix for asn1_GetSequence() for indefinite length constructed data.
14272 If SEQUENCE is length is indefinite just set c->slen to the total
14273 amount of data available.
14275 *Steve Henson, reported by shige@FreeBSD.org*
14277 *This change does not apply to 0.9.7.*
14279 * Change bctest to avoid here-documents inside command substitution
14280 (workaround for FreeBSD /bin/sh bug).
14281 For compatibility with Ultrix, avoid shell functions (introduced
14282 in the bctest version that searches along $PATH).
14286 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14287 with des_encrypt() defined on some operating systems, like Solaris
14292 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14293 On the Importance of Eliminating Errors in Cryptographic
14294 Computations, J. Cryptology 14 (2001) 2, 101-119,
14295 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14299 * MIPS assembler BIGNUM division bug fix.
14303 * Disabled incorrect Alpha assembler code.
14307 * Fix PKCS#7 decode routines so they correctly update the length
14308 after reading an EOC for the EXPLICIT tag.
14312 *This change does not apply to 0.9.7.*
14314 * Fix bug in PKCS#12 key generation routines. This was triggered
14315 if a 3DES key was generated with a 0 initial byte. Include
14316 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14317 (but broken) behaviour.
14321 * Enhance bctest to search for a working bc along $PATH and print
14324 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14326 * Fix memory leaks in err.c: free err_data string if necessary;
14327 don't write to the wrong index in ERR_set_error_data.
14331 * Implement ssl23_peek (analogous to ssl23_read), which previously
14336 * Replace rdtsc with `_emit` statements for VC++ version 5.
14338 *Jeremy Cooper <jeremy@baymoo.org>*
14340 * Make it possible to reuse SSLv2 sessions.
14344 * In copy_email() check for >= 0 as a return value for
14345 X509_NAME_get_index_by_NID() since 0 is a valid index.
14347 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14349 * Avoid coredump with unsupported or invalid public keys by checking if
14350 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14351 PKCS7_verify() fails with non detached data.
14355 * Don't use getenv in library functions when run as setuid/setgid.
14356 New function OPENSSL_issetugid().
14360 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14361 due to incorrect handling of multi-threading:
14363 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14365 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14367 3. Count how many times MemCheck_off() has been called so that
14368 nested use can be treated correctly. This also avoids
14369 inband-signalling in the previous code (which relied on the
14370 assumption that thread ID 0 is impossible).
14374 * Add "-rand" option also to s_client and s_server.
14378 * Fix CPU detection on Irix 6.x.
14379 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14380 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14382 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14387 *This change does not apply to 0.9.7.*
14389 * Use the cached encoding of an X509_NAME structure rather than
14390 copying it. This is apparently the reason for the libsafe "errors"
14391 but the code is actually correct.
14395 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14396 Bleichenbacher's DSA attack.
14397 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14398 to be set and top=0 forces the highest bit to be set; top=-1 is new
14399 and leaves the highest bit random.
14401 *Ulf Moeller, Bodo Moeller*
14403 * In the `NCONF_...`-based implementations for `CONF_...` queries
14404 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14405 a temporary CONF structure with the data component set to NULL
14406 (which gives segmentation faults in lh_retrieve).
14407 Instead, use NULL for the CONF pointer in CONF_get_string and
14408 CONF_get_number (which may use environment variables) and directly
14409 return NULL from CONF_get_section.
14413 * Fix potential buffer overrun for EBCDIC.
14417 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14418 keyUsage if basicConstraints absent for a CA.
14422 * Make SMIME_write_PKCS7() write mail header values with a format that
14423 is more generally accepted (no spaces before the semicolon), since
14424 some programs can't parse those values properly otherwise. Also make
14425 sure BIO's that break lines after each write do not create invalid
14430 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14431 macros previously used would not encode an empty SEQUENCE OF
14432 and break the signature.
14436 *This change does not apply to 0.9.7.*
14438 * Zero the premaster secret after deriving the master secret in
14443 * Add some EVP_add_digest_alias registrations (as found in
14444 OpenSSL_add_all_digests()) to SSL_library_init()
14445 aka OpenSSL_add_ssl_algorithms(). This provides improved
14446 compatibility with peers using X.509 certificates
14447 with unconventional AlgorithmIdentifier OIDs.
14451 * Fix for Irix with NO_ASM.
14453 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14455 * ./config script fixes.
14457 *Ulf Moeller, Richard Levitte*
14459 * Fix 'openssl passwd -1'.
14463 * Change PKCS12_key_gen_asc() so it can cope with non null
14464 terminated strings whose length is passed in the passlen
14465 parameter, for example from PEM callbacks. This was done
14466 by adding an extra length parameter to asc2uni().
14468 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14470 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14471 call failed, free the DSA structure.
14475 * Fix to uni2asc() to cope with zero length Unicode strings.
14476 These are present in some PKCS#12 files.
14480 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14481 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14482 when writing a 32767 byte record.
14484 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14486 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14487 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14489 (RSA objects have a reference count access to which is protected
14490 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14491 so they are meant to be shared between threads.)
14492 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14493 "Reddie, Steven" <Steven.Reddie@ca.com>*
14495 * Fix a deadlock in CRYPTO_mem_leaks().
14499 * Use better test patterns in bntest.
14503 * rand_win.c fix for Borland C.
14507 * BN_rshift bugfix for n == 0.
14511 * Add a 'bctest' script that checks for some known 'bc' bugs
14512 so that 'make test' does not abort just because 'bc' is broken.
14516 * Store verify_result within SSL_SESSION also for client side to
14517 avoid potential security hole. (Re-used sessions on the client side
14518 always resulted in verify_result==X509_V_OK, not using the original
14519 result of the server certificate verification.)
14523 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14524 SSL3_RT_APPLICATION_DATA, return 0.
14525 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14530 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14531 releases, have been re-implemented by renaming the previous
14532 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14533 and ssl3_read_internal, respectively, and adding 'peek' parameters
14534 to them. The new ssl[23]_{read,peek} functions are calls to
14535 ssl[23]_read_internal with the 'peek' flag set appropriately.
14536 A 'peek' parameter has also been added to ssl3_read_bytes, which
14537 does the actual work for ssl3_read_internal.
14541 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14542 the method-specific "init()" handler. Also clean up ex_data after
14543 calling the method-specific "finish()" handler. Previously, this was
14544 happening the other way round.
14548 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14549 The previous value, 12, was not always sufficient for BN_mod_exp().
14553 * Make sure that shared libraries get the internal name engine with
14554 the full version number and not just 0. This should mark the
14555 shared libraries as not backward compatible. Of course, this should
14556 be changed again when we can guarantee backward binary compatibility.
14560 * Fix typo in get_cert_by_subject() in by_dir.c
14562 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14564 * Rework the system to generate shared libraries:
14566 - Make note of the expected extension for the shared libraries and
14567 if there is a need for symbolic links from for example libcrypto.so.0
14568 to libcrypto.so.0.9.7. There is extended info in Configure for
14571 - Make as few rebuilds of the shared libraries as possible.
14573 - Still avoid linking the OpenSSL programs with the shared libraries.
14575 - When installing, install the shared libraries separately from the
14580 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14582 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14583 and not in SSL_clear because the latter is also used by the
14584 accept/connect functions; previously, the settings made by
14585 SSL_set_read_ahead would be lost during the handshake.
14587 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14589 * Correct util/mkdef.pl to be selective about disabled algorithms.
14590 Previously, it would create entries for disabled algorithms no
14595 * Added several new manual pages for SSL_* function.
14599 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14601 * In ssl23_get_client_hello, generate an error message when faced
14602 with an initial SSL 3.0/TLS record that is too small to contain the
14603 first two bytes of the ClientHello message, i.e. client_version.
14604 (Note that this is a pathologic case that probably has never happened
14605 in real life.) The previous approach was to use the version number
14606 from the record header as a substitute; but our protocol choice
14607 should not depend on that one because it is not authenticated
14608 by the Finished messages.
14612 * More robust randomness gathering functions for Windows.
14614 *Jeffrey Altman <jaltman@columbia.edu>*
14616 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14617 not set then we don't setup the error code for issuer check errors
14618 to avoid possibly overwriting other errors which the callback does
14619 handle. If an application does set the flag then we assume it knows
14620 what it is doing and can handle the new informational codes
14625 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14626 a general "ANY" type, as such it should be able to decode anything
14627 including tagged types. However it didn't check the class so it would
14628 wrongly interpret tagged types in the same way as their universal
14629 counterpart and unknown types were just rejected. Changed so that the
14630 tagged and unknown types are handled in the same way as a SEQUENCE:
14631 that is the encoding is stored intact. There is also a new type
14632 "V_ASN1_OTHER" which is used when the class is not universal, in this
14633 case we have no idea what the actual type is so we just lump them all
14638 * On VMS, stdout may very well lead to a file that is written to
14639 in a record-oriented fashion. That means that every write() will
14640 write a separate record, which will be read separately by the
14641 programs trying to read from it. This can be very confusing.
14643 The solution is to put a BIO filter in the way that will buffer
14644 text until a linefeed is reached, and then write everything a
14645 line at a time, so every record written will be an actual line,
14646 not chunks of lines and not (usually doesn't happen, but I've
14647 seen it once) several lines in one record. BIO_f_linebuffer() is
14650 Currently, it's a VMS-only method, because that's where it has
14651 been tested well enough.
14655 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14656 it can return incorrect results.
14657 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14658 but it was in 0.9.6-beta[12].)
14662 * Disable the check for content being present when verifying detached
14663 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14664 include zero length content when signing messages.
14668 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14669 BIO_ctrl (for BIO pairs).
14673 * Add DSO method for VMS.
14677 * Bug fix: Montgomery multiplication could produce results with the
14682 * Add RPM specification openssl.spec and modify it to build three
14683 packages. The default package contains applications, application
14684 documentation and run-time libraries. The devel package contains
14685 include files, static libraries and function documentation. The
14686 doc package contains the contents of the doc directory. The original
14687 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14691 * Add a large number of documentation files for many SSL routines.
14693 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14695 * Add a configuration entry for Sony News 4.
14697 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14699 * Don't set the two most significant bits to one when generating a
14700 random number < q in the DSA library.
14704 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14705 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14706 the underlying transport is blocking) if a handshake took place.
14707 (The default behaviour is needed by applications such as s_client
14708 and s_server that use select() to determine when to use SSL_read;
14709 but for applications that know in advance when to expect data, it
14710 just makes things more complicated.)
14714 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14719 * Add a few more EBCDIC conditionals that make `req` and `x509`
14720 work better on such systems.
14722 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14724 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14725 Update PKCS12_parse() so it copies the friendlyName and the
14726 keyid to the certificates aux info.
14730 * Fix bug in PKCS7_verify() which caused an infinite loop
14731 if there was more than one signature.
14733 *Sven Uszpelkat <su@celocom.de>*
14735 * Major change in util/mkdef.pl to include extra information
14736 about each symbol, as well as presenting variables as well
14737 as functions. This change means that there's n more need
14738 to rebuild the .num files when some algorithms are excluded.
14742 * Allow the verify time to be set by an application,
14743 rather than always using the current time.
14747 * Phase 2 verify code reorganisation. The certificate
14748 verify code now looks up an issuer certificate by a
14749 number of criteria: subject name, authority key id
14750 and key usage. It also verifies self signed certificates
14751 by the same criteria. The main comparison function is
14752 X509_check_issued() which performs these checks.
14754 Lot of changes were necessary in order to support this
14755 without completely rewriting the lookup code.
14757 Authority and subject key identifier are now cached.
14759 The LHASH 'certs' is X509_STORE has now been replaced
14760 by a STACK_OF(X509_OBJECT). This is mainly because an
14761 LHASH can't store or retrieve multiple objects with
14762 the same hash value.
14764 As a result various functions (which were all internal
14765 use only) have changed to handle the new X509_STORE
14766 structure. This will break anything that messed round
14767 with X509_STORE internally.
14769 The functions X509_STORE_add_cert() now checks for an
14770 exact match, rather than just subject name.
14772 The X509_STORE API doesn't directly support the retrieval
14773 of multiple certificates matching a given criteria, however
14774 this can be worked round by performing a lookup first
14775 (which will fill the cache with candidate certificates)
14776 and then examining the cache for matches. This is probably
14777 the best we can do without throwing out X509_LOOKUP
14778 entirely (maybe later...).
14780 The X509_VERIFY_CTX structure has been enhanced considerably.
14782 All certificate lookup operations now go via a get_issuer()
14783 callback. Although this currently uses an X509_STORE it
14784 can be replaced by custom lookups. This is a simple way
14785 to bypass the X509_STORE hackery necessary to make this
14786 work and makes it possible to use more efficient techniques
14787 in future. A very simple version which uses a simple
14788 STACK for its trusted certificate store is also provided
14789 using X509_STORE_CTX_trusted_stack().
14791 The verify_cb() and verify() callbacks now have equivalents
14792 in the X509_STORE_CTX structure.
14794 X509_STORE_CTX also has a 'flags' field which can be used
14795 to customise the verify behaviour.
14799 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14800 excludes S/MIME capabilities.
14804 * When a certificate request is read in keep a copy of the
14805 original encoding of the signed data and use it when outputting
14806 again. Signatures then use the original encoding rather than
14807 a decoded, encoded version which may cause problems if the
14808 request is improperly encoded.
14812 * For consistency with other BIO_puts implementations, call
14813 buffer_write(b, ...) directly in buffer_puts instead of calling
14816 In BIO_puts, increment b->num_write as in BIO_write.
14818 *Peter.Sylvester@EdelWeb.fr*
14820 * Fix BN_mul_word for the case where the word is 0. (We have to use
14821 BN_zero, we may not return a BIGNUM with an array consisting of
14822 words set to zero.)
14826 * Avoid calling abort() from within the library when problems are
14827 detected, except if preprocessor symbols have been defined
14828 (such as REF_CHECK, BN_DEBUG etc.).
14832 * New openssl application 'rsautl'. This utility can be
14833 used for low level RSA operations. DER public key
14834 BIO/fp routines also added.
14838 * New Configure entry and patches for compiling on QNX 4.
14840 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14842 * A demo state-machine implementation was sponsored by
14843 Nuron (<http://www.nuron.com/>) and is now available in
14844 demos/state_machine.
14848 * New options added to the 'dgst' utility for signature
14849 generation and verification.
14853 * Unrecognized PKCS#7 content types are now handled via a
14854 catch all ASN1_TYPE structure. This allows unsupported
14855 types to be stored as a "blob" and an application can
14856 encode and decode it manually.
14860 * Fix various signed/unsigned issues to make a_strex.c
14861 compile under VC++.
14863 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14865 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14866 length if passed a buffer. ASN1_INTEGER_to_BN failed
14867 if passed a NULL BN and its argument was negative.
14869 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14871 * Modification to PKCS#7 encoding routines to output definite
14872 length encoding. Since currently the whole structures are in
14873 memory there's not real point in using indefinite length
14874 constructed encoding. However if OpenSSL is compiled with
14875 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14879 * Added BIO_vprintf() and BIO_vsnprintf().
14883 * Added more prefixes to parse for in the strings written
14884 through a logging bio, to cover all the levels that are available
14885 through syslog. The prefixes are now:
14887 PANIC, EMERG, EMR => LOG_EMERG
14888 ALERT, ALR => LOG_ALERT
14889 CRIT, CRI => LOG_CRIT
14890 ERROR, ERR => LOG_ERR
14891 WARNING, WARN, WAR => LOG_WARNING
14892 NOTICE, NOTE, NOT => LOG_NOTICE
14893 INFO, INF => LOG_INFO
14894 DEBUG, DBG => LOG_DEBUG
14896 and as before, if none of those prefixes are present at the
14897 beginning of the string, LOG_ERR is chosen.
14899 On Win32, the `LOG_*` levels are mapped according to this:
14901 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14902 LOG_WARNING => EVENTLOG_WARNING_TYPE
14903 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14907 * Made it possible to reconfigure with just the configuration
14908 argument "reconf" or "reconfigure". The command line arguments
14909 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14910 and are retrieved from there when reconfiguring.
14916 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14918 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14922 * The obj_dat.pl script was messing up the sorting of object
14923 names. The reason was that it compared the quoted version
14924 of strings as a result "OCSP" > "OCSP Signing" because
14925 " > SPACE. Changed script to store unquoted versions of
14926 names and add quotes on output. It was also omitting some
14927 names from the lookup table if they were given a default
14928 value (that is if SN is missing it is given the same
14929 value as LN and vice versa), these are now added on the
14930 grounds that if an object has a name we should be able to
14931 look it up. Finally added warning output when duplicate
14932 short or long names are found.
14936 * Changes needed for Tandem NSK.
14938 *Scott Uroff <scott@xypro.com>*
14940 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14941 RSA_padding_check_SSLv23(), special padding was never detected
14942 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14943 version rollback attacks was not effective.
14945 In s23_clnt.c, don't use special rollback-attack detection padding
14946 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14947 client; similarly, in s23_srvr.c, don't do the rollback check if
14948 SSL 2.0 is the only protocol enabled in the server.
14952 * Make it possible to get hexdumps of unprintable data with 'openssl
14953 asn1parse'. By implication, the functions ASN1_parse_dump() and
14954 BIO_dump_indent() are added.
14958 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14959 these print out strings and name structures based on various
14960 flags including RFC2253 support and proper handling of
14961 multibyte characters. Added options to the 'x509' utility
14962 to allow the various flags to be set.
14966 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14967 Also change the functions X509_cmp_current_time() and
14968 X509_gmtime_adj() work with an ASN1_TIME structure,
14969 this will enable certificates using GeneralizedTime in validity
14970 dates to be checked.
14974 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14975 negative public key encodings) on by default,
14976 NO_NEG_PUBKEY_BUG can be set to disable it.
14980 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14981 content octets. An i2c_ASN1_OBJECT is unnecessary because
14982 the encoding can be trivially obtained from the structure.
14986 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14987 not read locks (`CRYPTO_r_[un]lock`).
14991 * A first attempt at creating official support for shared
14992 libraries through configuration. I've kept it so the
14993 default is static libraries only, and the OpenSSL programs
14994 are always statically linked for now, but there are
14995 preparations for dynamic linking in place.
14996 This has been tested on Linux and Tru64.
15000 * Randomness polling function for Win9x, as described in:
15001 Peter Gutmann, Software Generation of Practically Strong
15006 * Fix so PRNG is seeded in req if using an already existing
15011 * New options to smime application. -inform and -outform
15012 allow alternative formats for the S/MIME message including
15013 PEM and DER. The -content option allows the content to be
15014 specified separately. This should allow things like Netscape
15015 form signing output easier to verify.
15019 * Fix the ASN1 encoding of tags using the 'long form'.
15023 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15024 STRING types. These convert content octets to and from the
15025 underlying type. The actual tag and length octets are
15026 already assumed to have been read in and checked. These
15027 are needed because all other string types have virtually
15028 identical handling apart from the tag. By having versions
15029 of the ASN1 functions that just operate on content octets
15030 IMPLICIT tagging can be handled properly. It also allows
15031 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15032 and ASN1_INTEGER are identical apart from the tag.
15036 * Change the handling of OID objects as follows:
15038 - New object identifiers are inserted in objects.txt, following
15039 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15040 - objects.pl is used to process obj_mac.num and create a new
15042 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15045 This is currently kind of a hack, and the perl code in objects.pl
15046 isn't very elegant, but it works as I intended. The simplest way
15047 to check that it worked correctly is to look in obj_dat.h and
15048 check the array nid_objs and make sure the objects haven't moved
15049 around (this is important!). Additions are OK, as well as
15050 consistent name changes.
15054 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15058 * Addition of the command line parameter '-rand file' to 'openssl req'.
15059 The given file adds to whatever has already been seeded into the
15060 random pool through the RANDFILE configuration file option or
15061 environment variable, or the default random state file.
15065 * mkstack.pl now sorts each macro group into lexical order.
15066 Previously the output order depended on the order the files
15067 appeared in the directory, resulting in needless rewriting
15072 * Patches to make OpenSSL compile under Win32 again. Mostly
15073 work arounds for the VC++ problem that it treats func() as
15074 func(void). Also stripped out the parts of mkdef.pl that
15075 added extra typesafe functions: these no longer exist.
15079 * Reorganisation of the stack code. The macros are now all
15080 collected in safestack.h . Each macro is defined in terms of
15081 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15082 DEBUG_SAFESTACK is now handled in terms of function casts,
15083 this has the advantage of retaining type safety without the
15084 use of additional functions. If DEBUG_SAFESTACK is not defined
15085 then the non typesafe macros are used instead. Also modified the
15086 mkstack.pl script to handle the new form. Needs testing to see
15087 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15088 the default if no major problems. Similar behaviour for ASN1_SET_OF
15089 and PKCS12_STACK_OF.
15093 * When some versions of IIS use the 'NET' form of private key the
15094 key derivation algorithm is different. Normally MD5(password) is
15095 used as a 128 bit RC4 key. In the modified case
15096 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15097 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15098 as the old Netscape_RSA functions except they have an additional
15099 'sgckey' parameter which uses the modified algorithm. Also added
15100 an -sgckey command line option to the rsa utility. Thanks to
15101 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15102 algorithm to openssl-dev.
15106 * The evp_local.h macros were using 'c.##kname' which resulted in
15107 invalid expansion on some systems (SCO 5.0.5 for example).
15108 Corrected to 'c.kname'.
15110 *Phillip Porch <root@theporch.com>*
15112 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15113 a STACK of email addresses from a certificate or request, these look
15114 in the subject name and the subject alternative name extensions and
15115 omit any duplicate addresses.
15119 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15120 This makes DSA verification about 2 % faster.
15124 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15125 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15126 plus overhead for 1024 bit moduli).
15127 This makes exponentiations about 0.5 % faster for 1024 bit
15128 exponents (as measured by "openssl speed rsa2048").
15132 * Rename memory handling macros to avoid conflicts with other
15134 Malloc => OPENSSL_malloc
15135 Malloc_locked => OPENSSL_malloc_locked
15136 Realloc => OPENSSL_realloc
15137 Free => OPENSSL_free
15141 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15142 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15146 * CygWin32 support.
15148 *John Jarvie <jjarvie@newsguy.com>*
15150 * The type-safe stack code has been rejigged. It is now only compiled
15151 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15152 by default all type-specific stack functions are "#define"d back to
15153 standard stack functions. This results in more streamlined output
15154 but retains the type-safety checking possibilities of the original
15159 * The STACK code has been cleaned up, and certain type declarations
15160 that didn't make a lot of sense have been brought in line. This has
15161 also involved a cleanup of sorts in safestack.h to more correctly
15162 map type-safe stack functions onto their plain stack counterparts.
15163 This work has also resulted in a variety of "const"ifications of
15164 lots of the code, especially `_cmp` operations which should normally
15165 be prototyped with "const" parameters anyway.
15169 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15170 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15171 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15172 where all of 'md' is used each time the PRNG is used, but 'state'
15173 is used only indexed by a cyclic counter. As entropy may not be
15174 well distributed from the beginning, 'md' is important as a
15175 chaining variable. However, the output function chains only half
15176 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15177 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15178 in all of 'state' being rewritten, with the new values depending
15179 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15183 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15184 the handshake is continued after ssl_verify_cert_chain();
15185 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15186 can lead to 'unexplainable' connection aborts later.
15188 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15190 * Major EVP API cipher revision.
15191 Add hooks for extra EVP features. This allows various cipher
15192 parameters to be set in the EVP interface. Support added for variable
15193 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15194 setting of RC2 and RC5 parameters.
15196 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15199 Remove lots of duplicated code from the EVP library. For example *every*
15200 cipher init() function handles the 'iv' in the same way according to the
15201 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15202 for CFB and OFB modes they zero ctx->num.
15204 New functionality allows removal of S/MIME code RC2 hack.
15206 Most of the routines have the same form and so can be declared in terms
15209 By shifting this to the top level EVP_CipherInit() it can be removed from
15210 all individual ciphers. If the cipher wants to handle IVs or keys
15211 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15214 Change lots of functions like EVP_EncryptUpdate() to now return a
15215 value: although software versions of the algorithms cannot fail
15216 any installed hardware versions can.
15220 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15221 this option is set, tolerate broken clients that send the negotiated
15222 protocol version number instead of the requested protocol version
15227 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15228 i.e. non-zero for export ciphersuites, zero otherwise.
15229 Previous versions had this flag inverted, inconsistent with
15230 rsa_tmp_cb (..._TMP_RSA_CB).
15232 *Bodo Moeller; problem reported by Amit Chopra*
15234 * Add missing DSA library text string. Work around for some IIS
15235 key files with invalid SEQUENCE encoding.
15239 * Add a document (doc/standards.txt) that list all kinds of standards
15240 and so on that are implemented in OpenSSL.
15244 * Enhance c_rehash script. Old version would mishandle certificates
15245 with the same subject name hash and wouldn't handle CRLs at all.
15246 Added -fingerprint option to crl utility, to support new c_rehash
15251 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15255 * Fix for SSL server purpose checking. Server checking was
15256 rejecting certificates which had extended key usage present
15257 but no ssl client purpose.
15259 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15261 * Make PKCS#12 code work with no password. The PKCS#12 spec
15262 is a little unclear about how a blank password is handled.
15263 Since the password in encoded as a BMPString with terminating
15264 double NULL a zero length password would end up as just the
15265 double NULL. However no password at all is different and is
15266 handled differently in the PKCS#12 key generation code. NS
15267 treats a blank password as zero length. MSIE treats it as no
15268 password on export: but it will try both on import. We now do
15269 the same: PKCS12_parse() tries zero length and no password if
15270 the password is set to "" or NULL (NULL is now a valid password:
15271 it wasn't before) as does the pkcs12 application.
15275 * Bugfixes in `apps/x509.c`: Avoid a memory leak; and don't use
15276 perror when PEM_read_bio_X509_REQ fails, the error message must
15277 be obtained from the error queue.
15281 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15282 it in ERR_remove_state if appropriate, and change ERR_get_state
15283 accordingly to avoid race conditions (this is necessary because
15284 thread_hash is no longer constant once set).
15288 * Bugfix for linux-elf makefile.one.
15292 * RSA_get_default_method() will now cause a default
15293 RSA_METHOD to be chosen if one doesn't exist already.
15294 Previously this was only set during a call to RSA_new()
15295 or RSA_new_method(NULL) meaning it was possible for
15296 RSA_get_default_method() to return NULL.
15300 * Added native name translation to the existing DSO code
15301 that will convert (if the flag to do so is set) filenames
15302 that are sufficiently small and have no path information
15303 into a canonical native form. Eg. "blah" converted to
15304 "libblah.so" or "blah.dll" etc.
15308 * New function ERR_error_string_n(e, buf, len) which is like
15309 ERR_error_string(e, buf), but writes at most 'len' bytes
15310 including the 0 terminator. For ERR_error_string_n, 'buf'
15313 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15315 * CONF library reworked to become more general. A new CONF
15316 configuration file reader "class" is implemented as well as a
15317 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15318 old `CONF_*` functions are still there, but are reimplemented to
15319 work in terms of the new functions. Also, a set of functions
15320 to handle the internal storage of the configuration data is
15321 provided to make it easier to write new configuration file
15322 reader "classes" (I can definitely see something reading a
15323 configuration file in XML format, for example), called `_CONF_*`,
15324 or "the configuration storage API"...
15326 The new configuration file reading functions are:
15328 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15329 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15331 NCONF_default, NCONF_WIN32
15333 NCONF_dump_fp, NCONF_dump_bio
15335 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15336 NCONF_new creates a new CONF object. This works in the same way
15337 as other interfaces in OpenSSL, like the BIO interface.
15338 `NCONF_dump_*` dump the internal storage of the configuration file,
15339 which is useful for debugging. All other functions take the same
15340 arguments as the old `CONF_*` functions with the exception of the
15341 first that must be a `CONF *` instead of a `LHASH *`.
15343 To make it easier to use the new classes with the old `CONF_*` functions,
15344 the function CONF_set_default_method is provided.
15348 * Add '-tls1' option to 'openssl ciphers', which was already
15349 mentioned in the documentation but had not been implemented.
15350 (This option is not yet really useful because even the additional
15351 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15355 * Initial DSO code added into libcrypto for letting OpenSSL (and
15356 OpenSSL-based applications) load shared libraries and bind to
15357 them in a portable way.
15359 *Geoff Thorpe, with contributions from Richard Levitte*
15361 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15363 * Make sure _lrotl and _lrotr are only used with MSVC.
15365 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15366 (the default implementation of RAND_status).
15368 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15369 to '-clrext' (= clear extensions), as intended and documented.
15370 *Bodo Moeller; inconsistency pointed out by Michael Attili
15371 <attili@amaxo.com>*
15373 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15374 was larger than the MD block size.
15376 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15378 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15379 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15380 using the passed key: if the passed key was a private key the result
15381 of X509_print(), for example, would be to print out all the private key
15386 * des_quad_cksum() byte order bug fix.
15387 *Ulf Möller, using the problem description in krb4-0.9.7, where
15388 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15390 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15393 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15395 * For easily testing in shell scripts whether some command
15396 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15397 returns with exit code 0 iff no command of the given name is available.
15398 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15399 the output goes to stdout and nothing is printed to stderr.
15400 Additional arguments are always ignored.
15402 Since for each cipher there is a command of the same name,
15403 the 'no-cipher' compilation switches can be tested this way.
15405 ('openssl no-XXX' is not able to detect pseudo-commands such
15406 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15410 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15414 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15415 is set; it will be thrown away anyway because each handshake creates
15417 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15418 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15419 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15420 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15424 * New s_client option -ign_eof: EOF at stdin is ignored, and
15425 'Q' and 'R' lose their special meanings (quit/renegotiate).
15426 This is part of what -quiet does; unlike -quiet, -ign_eof
15427 does not suppress any output.
15431 * Add compatibility options to the purpose and trust code. The
15432 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15433 accepts a certificate or CA, this was the previous behaviour,
15434 with all the associated security issues.
15436 X509_TRUST_COMPAT is the old trust behaviour: only and
15437 automatically trust self signed roots in certificate store. A
15438 new trust setting X509_TRUST_DEFAULT is used to specify that
15439 a purpose has no associated trust setting and it should instead
15440 use the value in the default purpose.
15444 * Fix the PKCS#8 DSA private key code so it decodes keys again
15445 and fix a memory leak.
15449 * In util/mkerr.pl (which implements 'make errors'), preserve
15450 reason strings from the previous version of the .c file, as
15451 the default to have only downcase letters (and digits) in
15452 automatically generated reasons codes is not always appropriate.
15456 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15457 using strerror. Previously, ERR_reason_error_string() returned
15458 library names as reason strings for SYSerr; but SYSerr is a special
15459 case where small numbers are errno values, not library numbers.
15463 * Add '-dsaparam' option to 'openssl dhparam' application. This
15464 converts DSA parameters into DH parameters. (When creating parameters,
15465 DSA_generate_parameters is used.)
15469 * Include 'length' (recommended exponent length) in C code generated
15470 by 'openssl dhparam -C'.
15474 * The second argument to set_label in perlasm was already being used
15475 so couldn't be used as a "file scope" flag. Moved to third argument
15480 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15481 instead of RAND_bytes for encryption IVs and salts.
15485 * Include RAND_status() into RAND_METHOD instead of implementing
15486 it only for md_rand.c Otherwise replacing the PRNG by calling
15487 RAND_set_rand_method would be impossible.
15491 * Don't let DSA_generate_key() enter an infinite loop if the random
15492 number generation fails.
15496 * New 'rand' application for creating pseudo-random output.
15500 * Added configuration support for Linux/IA64
15502 *Rolf Haberrecker <rolf@suse.de>*
15504 * Assembler module support for Mingw32.
15508 * Shared library support for HPUX (in shlib/).
15510 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15512 * Shared library support for Solaris gcc.
15514 *Lutz Behnke <behnke@trustcenter.de>*
15516 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15518 * PKCS7_encrypt() was adding text MIME headers twice because they
15519 were added manually and by SMIME_crlf_copy().
15523 * In bntest.c don't call BN_rand with zero bits argument.
15525 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15527 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15528 case was implemented. This caused BN_div_recp() to fail occasionally.
15532 * Add an optional second argument to the set_label() in the perl
15533 assembly language builder. If this argument exists and is set
15534 to 1 it signals that the assembler should use a symbol whose
15535 scope is the entire file, not just the current function. This
15536 is needed with MASM which uses the format label:: for this scope.
15538 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15540 * Change the ASN1 types so they are typedefs by default. Before
15541 almost all types were #define'd to ASN1_STRING which was causing
15542 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15547 * Change names of new functions to the new get1/get0 naming
15548 convention: After 'get1', the caller owns a reference count
15549 and has to call `..._free`; 'get0' returns a pointer to some
15550 data structure without incrementing reference counters.
15551 (Some of the existing 'get' functions increment a reference
15552 counter, some don't.)
15553 Similarly, 'set1' and 'add1' functions increase reference
15554 counters or duplicate objects.
15558 * Allow for the possibility of temp RSA key generation failure:
15559 the code used to assume it always worked and crashed on failure.
15563 * Fix potential buffer overrun problem in BIO_printf().
15564 *Ulf Möller, using public domain code by Patrick Powell; problem
15565 pointed out by David Sacerdote <das33@cornell.edu>*
15567 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15568 RAND_egd() and RAND_status(). In the command line application,
15569 the EGD socket can be specified like a seed file using RANDFILE
15574 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15575 Some CAs (e.g. Verisign) distribute certificates in this form.
15579 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15580 list to exclude them. This means that no special compilation option
15581 is needed to use anonymous DH: it just needs to be included in the
15586 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15587 EVP_MD_type. The old functionality is available in a new macro called
15588 EVP_MD_md(). Change code that uses it and update docs.
15592 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15593 where the `void *` argument is replaced by a function pointer argument.
15594 Previously `void *` was abused to point to functions, which works on
15595 many platforms, but is not correct. As these functions are usually
15596 called by macros defined in OpenSSL header files, most source code
15597 should work without changes.
15601 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15602 sections with information on -D... compiler switches used for
15603 compiling the library so that applications can see them. To enable
15604 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15605 must be defined. E.g.,
15606 #define OPENSSL_ALGORITHM_DEFINES
15607 #include <openssl/opensslconf.h>
15608 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15610 *Richard Levitte, Ulf and Bodo Möller*
15612 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15617 * Change the 'other' type in certificate aux info to a STACK_OF
15618 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15619 the required ASN1 format: arbitrary types determined by an OID.
15623 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15624 argument to 'req'. This is not because the function is newer or
15625 better than others it just uses the work 'NEW' in the certificate
15626 request header lines. Some software needs this.
15630 * Reorganise password command line arguments: now passwords can be
15631 obtained from various sources. Delete the PEM_cb function and make
15632 it the default behaviour: i.e. if the callback is NULL and the
15633 usrdata argument is not NULL interpret it as a null terminated pass
15634 phrase. If usrdata and the callback are NULL then the pass phrase
15635 is prompted for as usual.
15639 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15640 the support is automatically enabled. The resulting binaries will
15641 autodetect the card and use it if present.
15643 *Ben Laurie and Compaq Inc.*
15645 * Work around for Netscape hang bug. This sends certificate request
15646 and server done in one record. Since this is perfectly legal in the
15647 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15648 the bugs/SSLv3 entry for more info.
15652 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15656 * Add -rand argument to smime and pkcs12 applications and read/write
15661 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15665 * Add command line password options to the remaining applications.
15669 * Bug fix for BN_div_recp() for numerators with an even number of
15674 * More tests in bntest.c, and changed test_bn output.
15678 * ./config recognizes MacOS X now.
15682 * Bug fix for BN_div() when the first words of num and divisor are
15683 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15687 * Add support for various broken PKCS#8 formats, and command line
15688 options to produce them.
15692 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15693 get temporary BIGNUMs from a BN_CTX.
15697 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15702 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15703 include a #define from the old name to the new. The original intent
15704 was that statically linked binaries could for example just call
15705 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15706 link with digests. This never worked because SSLeay_add_all_digests()
15707 and SSLeay_add_all_ciphers() were in the same source file so calling
15708 one would link with the other. They are now in separate source files.
15712 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15716 * Use a less unusual form of the Miller-Rabin primality test (it used
15717 a binary algorithm for exponentiation integrated into the Miller-Rabin
15718 loop, our standard modexp algorithms are faster).
15722 * Support for the EBCDIC character set completed.
15724 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15726 * Source code cleanups: use const where appropriate, eliminate casts,
15727 use `void *` instead of `char *` in lhash.
15731 * Bugfix: ssl3_send_server_key_exchange was not restartable
15732 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15733 this the server could overwrite ephemeral keys that the client
15738 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15739 using 50 iterations of the Rabin-Miller test.
15741 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15742 iterations of the Rabin-Miller test as required by the appendix
15743 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15744 As BN_is_prime_fasttest includes trial division, DSA parameter
15745 generation becomes much faster.
15747 This implies a change for the callback functions in DSA_is_prime
15748 and DSA_generate_parameters: The callback function is called once
15749 for each positive witness in the Rabin-Miller test, not just
15750 occasionally in the inner loop; and the parameters to the
15751 callback function now provide an iteration count for the outer
15752 loop rather than for the current invocation of the inner loop.
15753 DSA_generate_parameters additionally can call the callback
15754 function with an 'iteration count' of -1, meaning that a
15755 candidate has passed the trial division test (when q is generated
15756 from an application-provided seed, trial division is skipped).
15760 * New function BN_is_prime_fasttest that optionally does trial
15761 division before starting the Rabin-Miller test and has
15762 an additional BN_CTX * argument (whereas BN_is_prime always
15763 has to allocate at least one BN_CTX).
15764 'callback(1, -1, cb_arg)' is called when a number has passed the
15765 trial division stage.
15769 * Fix for bug in CRL encoding. The validity dates weren't being handled
15774 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15778 * New function BN_pseudo_rand().
15782 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15783 bignum version of BN_from_montgomery() with the working code from
15784 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15789 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15790 made it impossible to use the same SSL_SESSION data structure in
15791 SSL2 clients in multiple threads.
15795 * The return value of RAND_load_file() no longer counts bytes obtained
15796 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15797 to seed the PRNG (previously an explicit byte count was required).
15799 *Ulf Möller, Bodo Möller*
15801 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15802 used `char *` instead of `void *` and had casts all over the place.
15806 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15810 * Retain source code compatibility for BN_prime_checks macro:
15811 BN_is_prime(..., BN_prime_checks, ...) now uses
15812 BN_prime_checks_for_size to determine the appropriate number of
15813 Rabin-Miller iterations.
15817 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15818 DH_CHECK_P_NOT_SAFE_PRIME.
15819 (Check if this is true? OpenPGP calls them "strong".)
15823 * Merge the functionality of "dh" and "gendh" programs into a new program
15824 "dhparam". The old programs are retained for now but will handle DH keys
15825 (instead of parameters) in future.
15829 * Make the ciphers, s_server and s_client programs check the return values
15830 when a new cipher list is set.
15834 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15835 ciphers. Before when the 56bit ciphers were enabled the sorting was
15838 The syntax for the cipher sorting has been extended to support sorting by
15839 cipher-strength (using the strength_bits hard coded in the tables).
15840 The new command is `@STRENGTH` (see also `doc/apps/ciphers.pod`).
15842 Fix a bug in the cipher-command parser: when supplying a cipher command
15843 string with an "undefined" symbol (neither command nor alphanumeric
15844 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15845 an error is flagged.
15847 Due to the strength-sorting extension, the code of the
15848 ssl_create_cipher_list() function was completely rearranged. I hope that
15849 the readability was also increased :-)
15851 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15853 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15854 for the first serial number and places 2 in the serial number file. This
15855 avoids problems when the root CA is created with serial number zero and
15856 the first user certificate has the same issuer name and serial number
15861 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15862 the new code. Add documentation for this stuff.
15866 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15867 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15868 structures and behave in an analogous way to the X509v3 functions:
15869 they shouldn't be called directly but wrapper functions should be used
15872 So we also now have some wrapper functions that call the X509at functions
15873 when passed certificate requests. (TO DO: similar things can be done with
15874 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15875 things. Some of these need some d2i or i2d and print functionality
15876 because they handle more complex structures.)
15880 * Add missing #ifndefs that caused missing symbols when building libssl
15881 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15882 NO_RSA in `ssl/s2*.c`.
15884 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15886 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15887 has a return value which indicates the quality of the random data
15888 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15889 error queue. New function RAND_pseudo_bytes() generates output that is
15890 guaranteed to be unique but not unpredictable. RAND_add is like
15891 RAND_seed, but takes an extra argument for an entropy estimate
15892 (RAND_seed always assumes full entropy).
15896 * Do more iterations of Rabin-Miller probable prime test (specifically,
15897 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15898 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15899 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15900 false-positive rate of at most 2^-80 for random input.
15904 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15908 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15909 in the 0.9.5 release), this returns the chain
15910 from an X509_CTX structure with a dup of the stack and all
15911 the X509 reference counts upped: so the stack will exist
15912 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15915 Also make SSL_SESSION_print() print out the verify return
15920 * Add manpage for the pkcs12 command. Also change the default
15921 behaviour so MAC iteration counts are used unless the new
15922 -nomaciter option is used. This improves file security and
15923 only older versions of MSIE (4.0 for example) need it.
15927 * Honor the no-xxx Configure options when creating .DEF files.
15931 * Add PKCS#10 attributes to field table: challengePassword,
15932 unstructuredName and unstructuredAddress. These are taken from
15933 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15934 international characters are used.
15936 More changes to X509_ATTRIBUTE code: allow the setting of types
15937 based on strings. Remove the 'loc' parameter when adding
15938 attributes because these will be a SET OF encoding which is sorted
15943 * Initial changes to the 'req' utility to allow request generation
15944 automation. This will allow an application to just generate a template
15945 file containing all the field values and have req construct the
15948 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15949 used all over the place including certificate requests and PKCS#7
15950 structures. They are currently handled manually where necessary with
15951 some primitive wrappers for PKCS#7. The new functions behave in a
15952 manner analogous to the X509 extension functions: they allow
15953 attributes to be looked up by NID and added.
15955 Later something similar to the X509V3 code would be desirable to
15956 automatically handle the encoding, decoding and printing of the
15957 more complex types. The string types like challengePassword can
15958 be handled by the string table functions.
15960 Also modified the multi byte string table handling. Now there is
15961 a 'global mask' which masks out certain types. The table itself
15962 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15963 is useful when for example there is only one permissible type
15964 (as in countryName) and using the mask might result in no valid
15969 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15970 SSL_get_peer_finished to allow applications to obtain the latest
15971 Finished messages sent to the peer or expected from the peer,
15972 respectively. (SSL_get_peer_finished is usually the Finished message
15973 actually received from the peer, otherwise the protocol will be aborted.)
15975 As the Finished message are message digests of the complete handshake
15976 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15977 be used for external authentication procedures when the authentication
15978 provided by SSL/TLS is not desired or is not enough.
15982 * Enhanced support for Alpha Linux is added. Now ./config checks if
15983 the host supports BWX extension and if Compaq C is present on the
15984 $PATH. Just exploiting of the BWX extension results in 20-30%
15985 performance kick for some algorithms, e.g. DES and RC4 to mention
15986 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15991 * Add support for MS "fast SGC". This is arguably a violation of the
15992 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15993 weak crypto and after checking the certificate is SGC a second one
15994 with strong crypto. MS SGC stops the first handshake after receiving
15995 the server certificate message and sends a second client hello. Since
15996 a server will typically do all the time consuming operations before
15997 expecting any further messages from the client (server key exchange
15998 is the most expensive) there is little difference between the two.
16000 To get OpenSSL to support MS SGC we have to permit a second client
16001 hello message after we have sent server done. In addition we have to
16002 reset the MAC if we do get this second client hello.
16006 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16007 if a DER encoded private key is RSA or DSA traditional format. Changed
16008 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16009 format DER encoded private key. Newer code should use PKCS#8 format which
16010 has the key type encoded in the ASN1 structure. Added DER private key
16011 support to pkcs8 application.
16015 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16016 ciphersuites has been selected (as required by the SSL 3/TLS 1
16017 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16018 is set, we interpret this as a request to violate the specification
16019 (the worst that can happen is a handshake failure, and 'correct'
16020 behaviour would result in a handshake failure anyway).
16024 * In SSL_CTX_add_session, take into account that there might be multiple
16025 SSL_SESSION structures with the same session ID (e.g. when two threads
16026 concurrently obtain them from an external cache).
16027 The internal cache can handle only one SSL_SESSION with a given ID,
16028 so if there's a conflict, we now throw out the old one to achieve
16033 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16034 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16035 some routines that use cipher OIDs: some ciphers do not have OIDs
16036 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16041 * Simplify the trust setting structure and code. Now we just have
16042 two sequences of OIDs for trusted and rejected settings. These will
16043 typically have values the same as the extended key usage extension
16044 and any application specific purposes.
16046 The trust checking code now has a default behaviour: it will just
16047 check for an object with the same NID as the passed id. Functions can
16048 be provided to override either the default behaviour or the behaviour
16049 for a given id. SSL client, server and email already have functions
16050 in place for compatibility: they check the NID and also return "trusted"
16051 if the certificate is self signed.
16055 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16056 traditional format into an EVP_PKEY structure.
16060 * Add a password callback function PEM_cb() which either prompts for
16061 a password if usr_data is NULL or otherwise assumes it is a null
16062 terminated password. Allow passwords to be passed on command line
16063 environment or config files in a few more utilities.
16067 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16068 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16069 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16070 Update documentation.
16074 * Support for ASN1 "NULL" type. This could be handled before by using
16075 ASN1_TYPE but there wasn't any function that would try to read a NULL
16076 and produce an error if it couldn't. For compatibility we also have
16077 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16078 don't allocate anything because they don't need to.
16082 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16085 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16087 * Rebuild of the memory allocation routines used by OpenSSL code and
16088 possibly others as well. The purpose is to make an interface that
16089 provide hooks so anyone can build a separate set of allocation and
16090 deallocation routines to be used by OpenSSL, for example memory
16091 pool implementations, or something else, which was previously hard
16092 since Malloc(), Realloc() and Free() were defined as macros having
16093 the values malloc, realloc and free, respectively (except for Win32
16094 compilations). The same is provided for memory debugging code.
16095 OpenSSL already comes with functionality to find memory leaks, but
16096 this gives people a chance to debug other memory problems.
16098 With these changes, a new set of functions and macros have appeared:
16100 CRYPTO_set_mem_debug_functions() [F]
16101 CRYPTO_get_mem_debug_functions() [F]
16102 CRYPTO_dbg_set_options() [F]
16103 CRYPTO_dbg_get_options() [F]
16104 CRYPTO_malloc_debug_init() [M]
16106 The memory debug functions are NULL by default, unless the library
16107 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16108 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16109 gives the standard debugging functions that come with OpenSSL) or
16110 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16111 provided by the library user) must be used. When the standard
16112 debugging functions are used, CRYPTO_dbg_set_options can be used to
16113 request additional information:
16114 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16115 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16117 Also, things like CRYPTO_set_mem_functions will always give the
16118 expected result (the new set of functions is used for allocation
16119 and deallocation) at all times, regardless of platform and compiler
16122 To finish it up, some functions that were never use in any other
16123 way than through macros have a new API and new semantic:
16125 CRYPTO_dbg_malloc()
16126 CRYPTO_dbg_realloc()
16129 All macros of value have retained their old syntax.
16131 *Richard Levitte and Bodo Moeller*
16133 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16134 ordering of SMIMECapabilities wasn't in "strength order" and there
16135 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16140 * Some ASN1 types with illegal zero length encoding (INTEGER,
16141 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16143 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16145 * Merge in my S/MIME library for OpenSSL. This provides a simple
16146 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16147 functionality to handle multipart/signed properly) and a utility
16148 called 'smime' to call all this stuff. This is based on code I
16149 originally wrote for Celo who have kindly allowed it to be
16150 included in OpenSSL.
16154 * Add variants des_set_key_checked and des_set_key_unchecked of
16155 des_set_key (aka des_key_sched). Global variable des_check_key
16156 decides which of these is called by des_set_key; this way
16157 des_check_key behaves as it always did, but applications and
16158 the library itself, which was buggy for des_check_key == 1,
16159 have a cleaner way to pick the version they need.
16163 * New function PKCS12_newpass() which changes the password of a
16168 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16169 dynamic mix. In both cases the ids can be used as an index into the
16170 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16171 functions so they accept a list of the field values and the
16172 application doesn't need to directly manipulate the X509_TRUST
16177 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16182 * Modify the way the V3 extension code looks up extensions. This now
16183 works in a similar way to the object code: we have some "standard"
16184 extensions in a static table which is searched with OBJ_bsearch()
16185 and the application can add dynamic ones if needed. The file
16186 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16187 updated whenever a new extension is added to the core code and kept
16188 in ext_nid order. There is a simple program 'tabtest.c' which checks
16189 this. New extensions are not added too often so this file can readily
16190 be maintained manually.
16192 There are two big advantages in doing things this way. The extensions
16193 can be looked up immediately and no longer need to be "added" using
16194 X509V3_add_standard_extensions(): this function now does nothing.
16195 Side note: I get *lots* of email saying the extension code doesn't
16196 work because people forget to call this function.
16197 Also no dynamic allocation is done unless new extensions are added:
16198 so if we don't add custom extensions there is no need to call
16199 X509V3_EXT_cleanup().
16203 * Modify enc utility's salting as follows: make salting the default. Add a
16204 magic header, so unsalted files fail gracefully instead of just decrypting
16205 to garbage. This is because not salting is a big security hole, so people
16206 should be discouraged from doing it.
16210 * Fixes and enhancements to the 'x509' utility. It allowed a message
16211 digest to be passed on the command line but it only used this
16212 parameter when signing a certificate. Modified so all relevant
16213 operations are affected by the digest parameter including the
16214 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16215 DSA key was used because it didn't fix the digest.
16219 * Initial certificate chain verify code. Currently tests the untrusted
16220 certificates for consistency with the verify purpose (which is set
16221 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16223 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16224 this is because it will reject chains with invalid extensions whereas
16225 every previous version of OpenSSL and SSLeay made no checks at all.
16227 Trust code: checks the root CA for the relevant trust settings. Trust
16228 settings have an initial value consistent with the verify purpose: e.g.
16229 if the verify purpose is for SSL client use it expects the CA to be
16230 trusted for SSL client use. However the default value can be changed to
16231 permit custom trust settings: one example of this would be to only trust
16232 certificates from a specific "secure" set of CAs.
16234 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16235 which should be used for version portability: especially since the
16236 verify structure is likely to change more often now.
16238 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16239 to set them. If not set then assume SSL clients will verify SSL servers
16242 Two new options to the verify program: -untrusted allows a set of
16243 untrusted certificates to be passed in and -purpose which sets the
16244 intended purpose of the certificate. If a purpose is set then the
16245 new chain verify code is used to check extension consistency.
16249 * Support for the authority information access extension.
16253 * Modify RSA and DSA PEM read routines to transparently handle
16254 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16255 public keys in a format compatible with certificate
16256 SubjectPublicKeyInfo structures. Unfortunately there were already
16257 functions called *_PublicKey_* which used various odd formats so
16258 these are retained for compatibility: however the DSA variants were
16259 never in a public release so they have been deleted. Changed dsa/rsa
16260 utilities to handle the new format: note no releases ever handled public
16261 keys so we should be OK.
16263 The primary motivation for this change is to avoid the same fiasco
16264 that dogs private keys: there are several incompatible private key
16265 formats some of which are standard and some OpenSSL specific and
16266 require various evil hacks to allow partial transparent handling and
16267 even then it doesn't work with DER formats. Given the option anything
16268 other than PKCS#8 should be dumped: but the other formats have to
16269 stay in the name of compatibility.
16271 With public keys and the benefit of hindsight one standard format
16272 is used which works with EVP_PKEY, RSA or DSA structures: though
16273 it clearly returns an error if you try to read the wrong kind of key.
16275 Added a -pubkey option to the 'x509' utility to output the public key.
16276 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16277 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16278 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16279 that do the same as the `EVP_PKEY_assign_*()` except they up the
16280 reference count of the added key (they don't "swallow" the
16285 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16286 CRLs would fail if the file contained no certificates or no CRLs:
16287 added a new function to read in both types and return the number
16288 read: this means that if none are read it will be an error. The
16289 DER versions of the certificate and CRL reader would always fail
16290 because it isn't possible to mix certificates and CRLs in DER format
16291 without choking one or the other routine. Changed this to just read
16292 a certificate: this is the best we can do. Also modified the code
16293 in `apps/verify.c` to take notice of return codes: it was previously
16294 attempting to read in certificates from NULL pointers and ignoring
16295 any errors: this is one reason why the cert and CRL reader seemed
16296 to work. It doesn't check return codes from the default certificate
16297 routines: these may well fail if the certificates aren't installed.
16301 * Code to support otherName option in GeneralName.
16305 * First update to verify code. Change the verify utility
16306 so it warns if it is passed a self signed certificate:
16307 for consistency with the normal behaviour. X509_verify
16308 has been modified to it will now verify a self signed
16309 certificate if *exactly* the same certificate appears
16310 in the store: it was previously impossible to trust a
16311 single self signed certificate. This means that:
16312 openssl verify ss.pem
16313 now gives a warning about a self signed certificate but
16314 openssl verify -CAfile ss.pem ss.pem
16319 * For servers, store verify_result in SSL_SESSION data structure
16320 (and add it to external session representation).
16321 This is needed when client certificate verifications fails,
16322 but an application-provided verification callback (set by
16323 SSL_CTX_set_cert_verify_callback) allows accepting the session
16324 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16325 but returns 1): When the session is reused, we have to set
16326 ssl->verify_result to the appropriate error code to avoid
16329 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16331 * Fix a bug in the new PKCS#7 code: it didn't consider the
16332 case in PKCS7_dataInit() where the signed PKCS7 structure
16333 didn't contain any existing data because it was being created.
16335 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16337 * Add a salt to the key derivation routines in enc.c. This
16338 forms the first 8 bytes of the encrypted file. Also add a
16339 -S option to allow a salt to be input on the command line.
16343 * New function X509_cmp(). Oddly enough there wasn't a function
16344 to compare two certificates. We do this by working out the SHA1
16345 hash and comparing that. X509_cmp() will be needed by the trust
16350 * SSL_get1_session() is like SSL_get_session(), but increments
16351 the reference count in the SSL_SESSION returned.
16353 *Geoff Thorpe <geoff@eu.c2.net>*
16355 * Fix for 'req': it was adding a null to request attributes.
16356 Also change the X509_LOOKUP and X509_INFO code to handle
16357 certificate auxiliary information.
16361 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16366 * Add the possibility to add extra information to the memory leak
16367 detecting output, to form tracebacks, showing from where each
16368 allocation was originated: CRYPTO_push_info("constant string") adds
16369 the string plus current file name and line number to a per-thread
16370 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16371 is like calling CYRPTO_pop_info() until the stack is empty.
16372 Also updated memory leak detection code to be multi-thread-safe.
16376 * Add options -text and -noout to pkcs7 utility and delete the
16377 encryption options which never did anything. Update docs.
16381 * Add options to some of the utilities to allow the pass phrase
16382 to be included on either the command line (not recommended on
16383 OSes like Unix) or read from the environment. Update the
16384 manpages and fix a few bugs.
16388 * Add a few manpages for some of the openssl commands.
16392 * Fix the -revoke option in ca. It was freeing up memory twice,
16393 leaking and not finding already revoked certificates.
16397 * Extensive changes to support certificate auxiliary information.
16398 This involves the use of X509_CERT_AUX structure and X509_AUX
16399 functions. An X509_AUX function such as PEM_read_X509_AUX()
16400 can still read in a certificate file in the usual way but it
16401 will also read in any additional "auxiliary information". By
16402 doing things this way a fair degree of compatibility can be
16403 retained: existing certificates can have this information added
16404 using the new 'x509' options.
16406 Current auxiliary information includes an "alias" and some trust
16407 settings. The trust settings will ultimately be used in enhanced
16408 certificate chain verification routines: currently a certificate
16409 can only be trusted if it is self signed and then it is trusted
16414 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16415 The problem was that one of the replacement routines had not been working
16416 since SSLeay releases. For now the offending routine has been replaced
16417 with non-optimised assembler. Even so, this now gives around 95%
16418 performance improvement for 1024 bit RSA signs.
16422 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16423 handling. Most clients have the effective key size in bits equal to
16424 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16425 A few however don't do this and instead use the size of the decrypted key
16426 to determine the RC2 key length and the AlgorithmIdentifier to determine
16427 the effective key length. In this case the effective key length can still
16428 be 40 bits but the key length can be 168 bits for example. This is fixed
16429 by manually forcing an RC2 key into the EVP_PKEY structure because the
16430 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16431 the key length and effective key length are equal.
16435 * Add a bunch of functions that should simplify the creation of
16436 X509_NAME structures. Now you should be able to do:
16437 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16438 and have it automatically work out the correct field type and fill in
16439 the structures. The more adventurous can try:
16440 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16441 and it will (hopefully) work out the correct multibyte encoding.
16445 * Change the 'req' utility to use the new field handling and multibyte
16446 copy routines. Before the DN field creation was handled in an ad hoc
16447 way in req, ca, and x509 which was rather broken and didn't support
16448 BMPStrings or UTF8Strings. Since some software doesn't implement
16449 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16450 using the dirstring_type option. See the new comment in the default
16451 openssl.cnf for more info.
16455 * Make crypto/rand/md_rand.c more robust:
16456 - Assure unique random numbers after fork().
16457 - Make sure that concurrent threads access the global counter and
16458 md serializably so that we never lose entropy in them
16459 or use exactly the same state in multiple threads.
16460 Access to the large state is not always serializable because
16461 the additional locking could be a performance killer, and
16462 md should be large enough anyway.
16466 * New file `apps/app_rand.c` with commonly needed functionality
16467 for handling the random seed file.
16469 Use the random seed file in some applications that previously did not:
16471 dsaparam -genkey (which also ignored its '-rand' option),
16474 x509 (when signing).
16475 Except on systems with /dev/urandom, it is crucial to have a random
16476 seed file at least for key creation, DSA signing, and for DH exchanges;
16477 for RSA signatures we could do without one.
16479 gendh and gendsa (unlike genrsa) used to read only the first byte
16480 of each file listed in the '-rand' option. The function as previously
16481 found in genrsa is now in app_rand.c and is used by all programs
16482 that support '-rand'.
16486 * In RAND_write_file, use mode 0600 for creating files;
16487 don't just chmod when it may be too late.
16491 * Report an error from X509_STORE_load_locations
16492 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16496 * New function ASN1_mbstring_copy() this copies a string in either
16497 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16498 into an ASN1_STRING type. A mask of permissible types is passed
16499 and it chooses the "minimal" type to use or an error if not type
16504 * Add function equivalents to the various macros in asn1.h. The old
16505 macros are retained with an `M_` prefix. Code inside the library can
16506 use the `M_` macros. External code (including the openssl utility)
16507 should *NOT* in order to be "shared library friendly".
16511 * Add various functions that can check a certificate's extensions
16512 to see if it usable for various purposes such as SSL client,
16513 server or S/MIME and CAs of these types. This is currently
16514 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16515 verification. Also added a -purpose flag to x509 utility to
16516 print out all the purposes.
16520 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16525 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16526 for, obtain and decode and extension and obtain its critical flag.
16527 This allows all the necessary extension code to be handled in a
16528 single function call.
16532 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16533 platforms. See crypto/rc4/rc4_enc.c for further details.
16537 * New -noout option to asn1parse. This causes no output to be produced
16538 its main use is when combined with -strparse and -out to extract data
16539 from a file (which may not be in ASN.1 format).
16543 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16544 when producing the local key id.
16546 *Richard Levitte <levitte@stacken.kth.se>*
16548 * New option -dhparam in s_server. This allows a DH parameter file to be
16549 stated explicitly. If it is not stated then it tries the first server
16550 certificate file. The previous behaviour hard coded the filename
16555 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16556 a public key to be input or output. For example:
16557 openssl rsa -in key.pem -pubout -out pubkey.pem
16558 Also added necessary DSA public key functions to handle this.
16562 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16563 in the message. This was handled by allowing
16564 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16566 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16568 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16569 to the end of the strings whereas this didn't. This would cause problems
16570 if strings read with d2i_ASN1_bytes() were later modified.
16572 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16574 * Fix for base64 decode bug. When a base64 bio reads only one line of
16575 data and it contains EOF it will end up returning an error. This is
16576 caused by input 46 bytes long. The cause is due to the way base64
16577 BIOs find the start of base64 encoded data. They do this by trying a
16578 trial decode on each line until they find one that works. When they
16579 do a flag is set and it starts again knowing it can pass all the
16580 data directly through the decoder. Unfortunately it doesn't reset
16581 the context it uses. This means that if EOF is reached an attempt
16582 is made to pass two EOFs through the context and this causes the
16583 resulting error. This can also cause other problems as well. As is
16584 usual with these problems it takes *ages* to find and the fix is
16585 trivial: move one line.
16587 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16589 * Ugly workaround to get s_client and s_server working under Windows. The
16590 old code wouldn't work because it needed to select() on sockets and the
16591 tty (for keypresses and to see if data could be written). Win32 only
16592 supports select() on sockets so we select() with a 1s timeout on the
16593 sockets and then see if any characters are waiting to be read, if none
16594 are present then we retry, we also assume we can always write data to
16595 the tty. This isn't nice because the code then blocks until we've
16596 received a complete line of data and it is effectively polling the
16597 keyboard at 1s intervals: however it's quite a bit better than not
16598 working at all :-) A dedicated Windows application might handle this
16599 with an event loop for example.
16603 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16604 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16605 will be called when RSA_sign() and RSA_verify() are used. This is useful
16606 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16607 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16608 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16609 This necessitated the support of an extra signature type NID_md5_sha1
16610 for SSL signatures and modifications to the SSL library to use it instead
16611 of calling RSA_public_decrypt() and RSA_private_encrypt().
16615 * Add new -verify -CAfile and -CApath options to the crl program, these
16616 will lookup a CRL issuers certificate and verify the signature in a
16617 similar way to the verify program. Tidy up the crl program so it
16618 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16619 less strict. It will now permit CRL extensions even if it is not
16620 a V2 CRL: this will allow it to tolerate some broken CRLs.
16624 * Initialize all non-automatic variables each time one of the openssl
16625 sub-programs is started (this is necessary as they may be started
16626 multiple times from the "OpenSSL>" prompt).
16628 *Lennart Bang, Bodo Moeller*
16630 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16631 removing all other RSA functionality (this is what NO_RSA does). This
16632 is so (for example) those in the US can disable those operations covered
16633 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16638 * Non-copying interface to BIO pairs.
16639 (still largely untested)
16643 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16644 ASCII string. This was handled independently in various places before.
16648 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16649 UTF8 strings a character at a time.
16653 * Use client_version from client hello to select the protocol
16654 (s23_srvr.c) and for RSA client key exchange verification
16655 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16659 * Add various utility functions to handle SPKACs, these were previously
16660 handled by poking round in the structure internals. Added new function
16661 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16662 print, verify and generate SPKACs. Based on an original idea from
16663 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16667 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16671 * Allow the config file extension section to be overwritten on the
16672 command line. Based on an original idea from Massimiliano Pala
16673 <madwolf@comune.modena.it>. The new option is called -extensions
16674 and can be applied to ca, req and x509. Also -reqexts to override
16675 the request extensions in req and -crlexts to override the crl extensions
16680 * Add new feature to the SPKAC handling in ca. Now you can include
16681 the same field multiple times by preceding it by "XXXX." for example:
16684 this is the same syntax as used in the req config file.
16688 * Allow certificate extensions to be added to certificate requests. These
16689 are specified in a 'req_extensions' option of the req section of the
16690 config file. They can be printed out with the -text option to req but
16691 are otherwise ignored at present.
16695 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16696 data read consists of only the final block it would not decrypted because
16697 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16698 A misplaced 'break' also meant the decrypted final block might not be
16699 copied until the next read.
16703 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16704 a few extra parameters to the DH structure: these will be useful if
16705 for example we want the value of 'q' or implement X9.42 DH.
16709 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16710 provides hooks that allow the default DSA functions or functions on a
16711 "per key" basis to be replaced. This allows hardware acceleration and
16712 hardware key storage to be handled without major modification to the
16713 library. Also added low level modexp hooks and CRYPTO_EX structure and
16714 associated functions.
16718 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16719 as "read only": it can't be written to and the buffer it points to will
16720 not be freed. Reading from a read only BIO is much more efficient than
16721 a normal memory BIO. This was added because there are several times when
16722 an area of memory needs to be read from a BIO. The previous method was
16723 to create a memory BIO and write the data to it, this results in two
16724 copies of the data and an O(n^2) reading algorithm. There is a new
16725 function BIO_new_mem_buf() which creates a read only memory BIO from
16726 an area of memory. Also modified the PKCS#7 routines to use read only
16731 * Bugfix: ssl23_get_client_hello did not work properly when called in
16732 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16733 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16734 but a retry condition occurred while trying to read the rest.
16738 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16739 NID_pkcs7_encrypted by default: this was wrong since this should almost
16740 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16741 the encrypted data type: this is a more sensible place to put it and it
16742 allows the PKCS#12 code to be tidied up that duplicated this
16747 * Changed obj_dat.pl script so it takes its input and output files on
16748 the command line. This should avoid shell escape redirection problems
16753 * Initial support for certificate extension requests, these are included
16754 in things like Xenroll certificate requests. Included functions to allow
16755 extensions to be obtained and added.
16759 * -crlf option to s_client and s_server for sending newlines as
16760 CRLF (as required by many protocols).
16764 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16766 * Install libRSAglue.a when OpenSSL is built with RSAref.
16768 *Ralf S. Engelschall*
16770 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16772 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16774 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16779 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16780 DH parameters/keys (q is lost during that conversion, but the resulting
16781 DH parameters contain its length).
16783 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16784 much faster than DH_generate_parameters (which creates parameters
16785 where `p = 2*q + 1`), and also the smaller q makes DH computations
16786 much more efficient (160-bit exponentiation instead of 1024-bit
16787 exponentiation); so this provides a convenient way to support DHE
16788 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16789 utter importance to use
16790 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16792 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16793 when such DH parameters are used, because otherwise small subgroup
16794 attacks may become possible!
16798 * Avoid memory leak in i2d_DHparams.
16802 * Allow the -k option to be used more than once in the enc program:
16803 this allows the same encrypted message to be read by multiple recipients.
16807 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16808 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16809 it will always use the numerical form of the OID, even if it has a short
16814 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16815 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16816 otherwise bn_mod_exp was called. In the case of hardware keys for example
16817 no private key components need be present and it might store extra data
16818 in the RSA structure, which cannot be accessed from bn_mod_exp.
16819 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16820 private key operations.
16824 * Added support for SPARC Linux.
16828 * pem_password_cb function type incompatibly changed from
16829 typedef int pem_password_cb(char *buf, int size, int rwflag);
16831 ....(char *buf, int size, int rwflag, void *userdata);
16832 so that applications can pass data to their callbacks:
16833 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16834 additional void * argument, which is just handed through whenever
16835 the password callback is called.
16837 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16839 New function SSL_CTX_set_default_passwd_cb_userdata.
16841 Compatibility note: As many C implementations push function arguments
16842 onto the stack in reverse order, the new library version is likely to
16843 interoperate with programs that have been compiled with the old
16844 pem_password_cb definition (PEM_whatever takes some data that
16845 happens to be on the stack as its last argument, and the callback
16846 just ignores this garbage); but there is no guarantee whatsoever that
16849 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16850 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16851 problems not only on Windows, but also on some Unix platforms.
16852 To avoid problematic command lines, these definitions are now in an
16853 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16854 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16858 * MIPS III/IV assembler module is reimplemented.
16862 * More DES library cleanups: remove references to srand/rand and
16863 delete an unused file.
16867 * Add support for the free Netwide assembler (NASM) under Win32,
16868 since not many people have MASM (ml) and it can be hard to obtain.
16869 This is currently experimental but it seems to work OK and pass all
16870 the tests. Check out INSTALL.W32 for info.
16874 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16875 without temporary keys kept an extra copy of the server key,
16876 and connections with temporary keys did not free everything in case
16881 * New function RSA_check_key and new openssl rsa option -check
16882 for verifying the consistency of RSA keys.
16884 *Ulf Moeller, Bodo Moeller*
16886 * Various changes to make Win32 compile work:
16887 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16888 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16889 comparison" warnings.
16890 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16894 * Add a debugging option to PKCS#5 v2 key generation function: when
16895 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16896 derived keys are printed to stderr.
16900 * Copy the flags in ASN1_STRING_dup().
16902 *Roman E. Pavlov <pre@mo.msk.ru>*
16904 * The x509 application mishandled signing requests containing DSA
16905 keys when the signing key was also DSA and the parameters didn't match.
16907 It was supposed to omit the parameters when they matched the signing key:
16908 the verifying software was then supposed to automatically use the CA's
16909 parameters if they were absent from the end user certificate.
16911 Omitting parameters is no longer recommended. The test was also
16912 the wrong way round! This was probably due to unusual behaviour in
16913 EVP_cmp_parameters() which returns 1 if the parameters match.
16914 This meant that parameters were omitted when they *didn't* match and
16915 the certificate was useless. Certificates signed with 'ca' didn't have
16918 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16920 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16921 The interface is as follows:
16922 Applications can use
16923 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16924 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16925 "off" is now the default.
16926 The library internally uses
16927 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16928 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16929 to disable memory-checking temporarily.
16931 Some inconsistent states that previously were possible (and were
16932 even the default) are now avoided.
16934 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16935 with each memory chunk allocated; this is occasionally more helpful
16936 than just having a counter.
16938 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16940 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16945 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16946 which largely parallels "options", but is for changing API behaviour,
16947 whereas "options" are about protocol behaviour.
16948 Initial "mode" flags are:
16950 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16951 a single record has been written.
16952 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16953 retries use the same buffer location.
16954 (But all of the contents must be
16959 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16962 * Fix problems with no-hmac etc.
16964 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16966 * New functions RSA_get_default_method(), RSA_set_method() and
16967 RSA_get_method(). These allows replacement of RSA_METHODs without having
16968 to mess around with the internals of an RSA structure.
16972 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16973 Also really enable memory leak checks in openssl.c and in some
16976 *Chad C. Mulligan, Bodo Moeller*
16978 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16979 up the length of negative integers. This has now been simplified to just
16980 store the length when it is first determined and use it later, rather
16981 than trying to keep track of where data is copied and updating it to
16983 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16985 * Add a new function PKCS7_signatureVerify. This allows the verification
16986 of a PKCS#7 signature but with the signing certificate passed to the
16987 function itself. This contrasts with PKCS7_dataVerify which assumes the
16988 certificate is present in the PKCS#7 structure. This isn't always the
16989 case: certificates can be omitted from a PKCS#7 structure and be
16990 distributed by "out of band" means (such as a certificate database).
16994 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16995 function prototypes in pem.h, also change util/mkdef.pl to add the
16996 necessary function names.
17000 * mk1mf.pl (used by Windows builds) did not properly read the
17001 options set by Configure in the top level Makefile, and Configure
17002 was not even able to write more than one option correctly.
17003 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17007 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17008 file to be loaded from a BIO or FILE pointer. The BIO version will
17009 for example allow memory BIOs to contain config info.
17013 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17014 Whoever hopes to achieve shared-library compatibility across versions
17015 must use this, not the compile-time macro.
17016 (Exercise 0.9.4: Which is the minimum library version required by
17018 Note: All this applies only to multi-threaded programs, others don't
17023 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17024 through a BIO pair triggered the default case, i.e.
17025 SSLerr(...,SSL_R_UNKNOWN_STATE).
17029 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17030 can use the SSL library even if none of the specific BIOs is
17035 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17036 for the encoded length.
17038 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17040 * Add initial documentation of the X509V3 functions.
17044 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17045 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17046 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17047 secure PKCS#8 private key format with a high iteration count.
17051 * Fix determination of Perl interpreter: A perl or perl5
17052 *directory* in $PATH was also accepted as the interpreter.
17054 *Ralf S. Engelschall*
17056 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17057 wrong with it but it was very old and did things like calling
17058 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17059 unusual formatting.
17063 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17064 to use the new extension code.
17068 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17069 with macros. This should make it easier to change their form, add extra
17070 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17075 * Add to configuration table a new entry that can specify an alternative
17076 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17077 according to Mark Crispin <MRC@Panda.COM>.
17081 * DES CBC did not update the IV. Weird.
17085 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17086 Changing the behaviour of the former might break existing programs --
17087 where IV updating is needed, des_ncbc_encrypt can be used.
17090 * When bntest is run from "make test" it drives bc to check its
17091 calculations, as well as internally checking them. If an internal check
17092 fails, it needs to cause bc to give a non-zero result or make test carries
17093 on without noticing the failure. Fixed.
17097 * DES library cleanups.
17101 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17102 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17103 ciphers. NOTE: although the key derivation function has been verified
17104 against some published test vectors it has not been extensively tested
17105 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17110 * Instead of "mkdir -p", which is not fully portable, use new
17111 Perl script "util/mkdir-p.pl".
17115 * Rewrite the way password based encryption (PBE) is handled. It used to
17116 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17117 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17118 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17119 the 'parameter' field of the AlgorithmIdentifier is passed to the
17120 underlying key generation function so it must do its own ASN1 parsing.
17121 This has also changed the EVP_PBE_CipherInit() function which now has a
17122 'parameter' argument instead of literal salt and iteration count values
17123 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17127 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17128 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17129 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17130 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17131 value was just used as a "magic string" and not used directly its
17132 value doesn't matter.
17136 * Introduce some semblance of const correctness to BN. Shame C doesn't
17141 * "linux-sparc64" configuration (ultrapenguin).
17143 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17144 "linux-sparc" configuration.
17146 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17148 * config now generates no-xxx options for missing ciphers.
17152 * Support the EBCDIC character set (work in progress).
17153 File ebcdic.c not yet included because it has a different license.
17155 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17157 * Support BS2000/OSD-POSIX.
17159 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17161 * Make callbacks for key generation use `void *` instead of `char *`.
17165 * Make S/MIME samples compile (not yet tested).
17169 * Additional typesafe stacks.
17173 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17177 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17179 * New configuration variant "sco5-gcc".
17181 * Updated some demos.
17183 *Sean O Riordain, Wade Scholine*
17185 * Add missing BIO_free at exit of pkcs12 application.
17189 * Fix memory leak in conf.c.
17193 * Updates for Win32 to assembler version of MD5.
17197 * Set #! path to perl in `apps/der_chop` to where we found it
17198 instead of using a fixed path.
17202 * SHA library changes for irix64-mips4-cc.
17206 * Improvements for VMS support.
17210 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17212 * Bignum library bug fix. IRIX 6 passes "make test" now!
17213 This also avoids the problems with SC4.2 and unpatched SC5.
17215 *Andy Polyakov <appro@fy.chalmers.se>*
17217 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17218 These are required because of the typesafe stack would otherwise break
17219 existing code. If old code used a structure member which used to be STACK
17220 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17221 sk_num or sk_value it would produce an error because the num, data members
17222 are not present in STACK_OF. Now it just produces a warning. sk_set
17223 replaces the old method of assigning a value to sk_value
17224 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17225 that does this will no longer work (and should use sk_set instead) but
17226 this could be regarded as a "questionable" behaviour anyway.
17230 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17231 correctly handle encrypted S/MIME data.
17235 * Change type of various DES function arguments from des_cblock
17236 (which means, in function argument declarations, pointer to char)
17237 to des_cblock * (meaning pointer to array with 8 char elements),
17238 which allows the compiler to do more typechecking; it was like
17239 that back in SSLeay, but with lots of ugly casts.
17241 Introduce new type const_des_cblock.
17245 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17246 problems: find RecipientInfo structure that matches recipient certificate
17247 and initialise the ASN1 structures properly based on passed cipher.
17251 * Belatedly make the BN tests actually check the results.
17255 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17256 to and from BNs: it was completely broken. New compilation option
17257 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17258 key elements as negative integers.
17262 * Reorganize and speed up MD5.
17264 *Andy Polyakov <appro@fy.chalmers.se>*
17268 *Richard Levitte <richard@levitte.org>*
17270 * New option -out to asn1parse to allow the parsed structure to be
17271 output to a file. This is most useful when combined with the -strparse
17272 option to examine the output of things like OCTET STRINGS.
17276 * Make SSL library a little more fool-proof by not requiring any longer
17277 that `SSL_set_{accept,connect}_state` be called before
17278 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17279 in many applications because usually everything *appeared* to work as
17280 intended anyway -- now it really works as intended).
17284 * Move openssl.cnf out of lib/.
17288 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17289 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17290 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17292 *Ralf S. Engelschall*
17294 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17295 handle PKCS#7 enveloped data properly.
17297 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17299 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17300 copying pointers. The cert_st handling is changed by this in
17301 various ways (and thus what used to be known as ctx->default_cert
17302 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17303 any longer when s->cert does not give us what we need).
17304 ssl_cert_instantiate becomes obsolete by this change.
17305 As soon as we've got the new code right (possibly it already is?),
17306 we have solved a couple of bugs of the earlier code where s->cert
17307 was used as if it could not have been shared with other SSL structures.
17309 Note that using the SSL API in certain dirty ways now will result
17310 in different behaviour than observed with earlier library versions:
17311 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17312 does not influence s as it used to.
17314 In order to clean up things more thoroughly, inside SSL_SESSION
17315 we don't use CERT any longer, but a new structure SESS_CERT
17316 that holds per-session data (if available); currently, this is
17317 the peer's certificate chain and, for clients, the server's certificate
17318 and temporary key. CERT holds only those values that can have
17319 meaningful defaults in an SSL_CTX.
17323 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17324 from the internal representation. Various PKCS#7 fixes: remove some
17325 evil casts and set the enc_dig_alg field properly based on the signing
17330 * Allow PKCS#12 password to be set from the command line or the
17331 environment. Let 'ca' get its config file name from the environment
17332 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17337 * Allow certificate policies extension to use an IA5STRING for the
17338 organization field. This is contrary to the PKIX definition but
17339 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17344 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17345 without disallowing inline assembler and the like for non-pedantic builds.
17349 * Support Borland C++ builder.
17351 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17357 * SHA-1 cleanups and performance enhancements.
17359 *Andy Polyakov <appro@fy.chalmers.se>*
17361 * Sparc v8plus assembler for the bignum library.
17363 *Andy Polyakov <appro@fy.chalmers.se>*
17365 * Accept any -xxx and +xxx compiler options in Configure.
17369 * Update HPUX configuration.
17373 * Add missing `sk_<type>_unshift()` function to safestack.h
17375 *Ralf S. Engelschall*
17377 * New function SSL_CTX_use_certificate_chain_file that sets the
17378 "extra_cert"s in addition to the certificate. (This makes sense
17379 only for "PEM" format files, as chains as a whole are not
17384 * Support verify_depth from the SSL API.
17385 x509_vfy.c had what can be considered an off-by-one-error:
17386 Its depth (which was not part of the external interface)
17387 was actually counting the number of certificates in a chain;
17388 now it really counts the depth.
17392 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17393 instead of X509err, which often resulted in confusing error
17394 messages since the error codes are not globally unique
17395 (e.g. an alleged error in ssl3_accept when a certificate
17396 didn't match the private key).
17398 * New function SSL_CTX_set_session_id_context that allows to set a default
17399 value (so that you don't need SSL_set_session_id_context for each
17400 connection using the SSL_CTX).
17404 * OAEP decoding bug fix.
17408 * Support INSTALL_PREFIX for package builders, as proposed by
17413 * New Configure options "threads" and "no-threads". For systems
17414 where the proper compiler options are known (currently Solaris
17415 and Linux), "threads" is the default.
17419 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17423 * Install various scripts to $(OPENSSLDIR)/misc, not to
17424 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17425 such as /usr/local/bin.
17429 * "make linux-shared" to build shared libraries.
17431 *Niels Poppe <niels@netbox.org>*
17433 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17437 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17438 extension adding in x509 utility.
17442 * Remove NOPROTO sections and error code comments.
17446 * Partial rewrite of the DEF file generator to now parse the ANSI
17451 * New Configure options --prefix=DIR and --openssldir=DIR.
17455 * Complete rewrite of the error code script(s). It is all now handled
17456 by one script at the top level which handles error code gathering,
17457 header rewriting and C source file generation. It should be much better
17458 than the old method: it now uses a modified version of Ulf's parser to
17459 read the ANSI prototypes in all header files (thus the old K&R definitions
17460 aren't needed for error creation any more) and do a better job of
17461 translating function codes into names. The old 'ASN1 error code embedded
17462 in a comment' is no longer necessary and it doesn't use .err files which
17463 have now been deleted. Also the error code call doesn't have to appear all
17464 on one line (which resulted in some large lines...).
17468 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17472 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17473 0 (which usually indicates a closed connection), but continue reading.
17477 * Fix some race conditions.
17481 * Add support for CRL distribution points extension. Add Certificate
17482 Policies and CRL distribution points documentation.
17486 * Move the autogenerated header file parts to crypto/opensslconf.h.
17490 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17491 8 of keying material. Merlin has also confirmed interop with this fix
17492 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17494 *Merlin Hughes <merlin@baltimore.ie>*
17496 * Fix lots of warnings.
17498 *Richard Levitte <levitte@stacken.kth.se>*
17500 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17501 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17503 *Richard Levitte <levitte@stacken.kth.se>*
17505 * Fix problems with sizeof(long) == 8.
17507 *Andy Polyakov <appro@fy.chalmers.se>*
17509 * Change functions to ANSI C.
17513 * Fix typos in error codes.
17515 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17517 * Remove defunct assembler files from Configure.
17521 * SPARC v8 assembler BIGNUM implementation.
17523 *Andy Polyakov <appro@fy.chalmers.se>*
17525 * Support for Certificate Policies extension: both print and set.
17526 Various additions to support the r2i method this uses.
17530 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17531 return a const string when you are expecting an allocated buffer.
17535 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17536 types DirectoryString and DisplayText.
17540 * Add code to allow r2i extensions to access the configuration database,
17541 add an LHASH database driver and add several ctx helper functions.
17545 * Fix an evil bug in bn_expand2() which caused various BN functions to
17546 fail when they extended the size of a BIGNUM.
17550 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17551 support typesafe stack.
17555 * Fix typo in SSL_[gs]et_options().
17557 *Nils Frostberg <nils@medcom.se>*
17559 * Delete various functions and files that belonged to the (now obsolete)
17560 old X509V3 handling code.
17564 * New Configure option "rsaref".
17568 * Don't auto-generate pem.h.
17572 * Introduce type-safe ASN.1 SETs.
17576 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17578 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17580 * Introduce type-safe STACKs. This will almost certainly break lots of code
17581 that links with OpenSSL (well at least cause lots of warnings), but fear
17582 not: the conversion is trivial, and it eliminates loads of evil casts. A
17583 few STACKed things have been converted already. Feel free to convert more.
17584 In the fullness of time, I'll do away with the STACK type altogether.
17588 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17589 specified in `<certfile>` by updating the entry in the index.txt file.
17590 This way one no longer has to edit the index.txt file manually for
17591 revoking a certificate. The -revoke option does the gory details now.
17593 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17595 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17596 `-text` option at all and this way the `-noout -text` combination was
17597 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17599 *Ralf S. Engelschall*
17601 * Make sure a corresponding plain text error message exists for the
17602 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17603 verify callback function determined that a certificate was revoked.
17605 *Ralf S. Engelschall*
17607 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17608 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17609 all available ciphers including rc5, which was forgotten until now.
17610 In order to let the testing shell script know which algorithms
17611 are available, a new (up to now undocumented) command
17612 `openssl list-cipher-commands` is used.
17616 * Bugfix: s_client occasionally would sleep in select() when
17617 it should have checked SSL_pending() first.
17621 * New functions DSA_do_sign and DSA_do_verify to provide access to
17622 the raw DSA values prior to ASN.1 encoding.
17626 * Tweaks to Configure
17628 *Niels Poppe <niels@netbox.org>*
17630 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17635 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17639 * New config option to avoid instructions that are illegal on the 80386.
17640 The default code is faster, but requires at least a 486.
17644 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17645 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17646 same as SSL2_VERSION anyway.
17650 * New "-showcerts" option for s_client.
17654 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17655 application. Various cleanups and fixes.
17659 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17660 modify error routines to work internally. Add error codes and PBE init
17661 to library startup routines.
17665 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17666 packing functions to asn1 and evp. Changed function names and error
17667 codes along the way.
17671 * PKCS12 integration: and so it begins... First of several patches to
17672 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17673 objects to objects.h
17677 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17678 and display support for Thawte strong extranet extension.
17682 * Add LinuxPPC support.
17684 *Jeff Dubrule <igor@pobox.org>*
17686 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17687 bn_div_words in alpha.s.
17689 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17691 * Make sure the RSA OAEP test is skipped under -DRSAref because
17692 OAEP isn't supported when OpenSSL is built with RSAref.
17694 *Ulf Moeller <ulf@fitug.de>*
17696 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17697 so they no longer are missing under -DNOPROTO.
17699 *Soren S. Jorvang <soren@t.dk>*
17701 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17703 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17704 doesn't work when the session is reused. Coming soon!
17708 * Fix a security hole, that allows sessions to be reused in the wrong
17709 context thus bypassing client cert protection! All software that uses
17710 client certs and session caches in multiple contexts NEEDS PATCHING to
17711 allow session reuse! A fuller solution is in the works.
17713 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17715 * Some more source tree cleanups (removed obsolete files
17716 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17717 permission on "config" script to be executable) and a fix for the INSTALL
17720 *Ulf Moeller <ulf@fitug.de>*
17722 * Remove some legacy and erroneous uses of malloc, free instead of
17725 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17727 * Make rsa_oaep_test return non-zero on error.
17729 *Ulf Moeller <ulf@fitug.de>*
17731 * Add support for native Solaris shared libraries. Configure
17732 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17733 if someone would make that last step automatic.
17735 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17737 * ctx_size was not built with the right compiler during "make links". Fixed.
17741 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17742 except NULL ciphers". This means the default cipher list will no longer
17743 enable NULL ciphers. They need to be specifically enabled e.g. with
17744 the string "DEFAULT:eNULL".
17748 * Fix to RSA private encryption routines: if p < q then it would
17749 occasionally produce an invalid result. This will only happen with
17750 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17754 * Be less restrictive and allow also `perl util/perlpath.pl
17755 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17756 because this way one can also use an interpreter named `perl5` (which is
17757 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17758 installed as `perl`).
17760 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17762 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17764 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17766 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17767 advapi32.lib to Win32 build and change the pem test comparison
17768 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17769 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17770 and crypto/des/ede_cbcm_enc.c.
17774 * DES quad checksum was broken on big-endian architectures. Fixed.
17778 * Comment out two functions in bio.h that aren't implemented. Fix up the
17779 Win32 test batch file so it (might) work again. The Win32 test batch file
17780 is horrible: I feel ill....
17784 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17785 in e_os.h. Audit of header files to check ANSI and non ANSI
17786 sections: 10 functions were absent from non ANSI section and not exported
17787 from Windows DLLs. Fixed up libeay.num for new functions.
17791 * Make `openssl version` output lines consistent.
17793 *Ralf S. Engelschall*
17795 * Fix Win32 symbol export lists for BIO functions: Added
17796 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17797 to ms/libeay{16,32}.def.
17799 *Ralf S. Engelschall*
17801 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17802 fine under Unix and passes some trivial tests I've now added. But the
17803 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17804 added to make sure no one expects that this stuff really works in the
17805 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17806 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17809 *Ralf S. Engelschall*
17811 * Fix the generation of two part addresses in perl.
17813 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17815 * Add config entry for Linux on MIPS.
17817 *John Tobey <jtobey@channel1.com>*
17819 * Make links whenever Configure is run, unless we are on Windoze.
17823 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17824 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17829 * Add a useful kludge to allow package maintainers to specify compiler and
17830 other platforms details on the command line without having to patch the
17831 Configure script every time: One now can use
17832 `perl Configure <id>:<details>`,
17833 i.e. platform ids are allowed to have details appended
17834 to them (separated by colons). This is treated as there would be a static
17835 pre-configured entry in Configure's %table under key `<id>` with value
17836 `<details>` and `perl Configure <id>` is called. So, when you want to
17837 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17838 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17839 now, which overrides the FreeBSD-elf entry on-the-fly.
17841 *Ralf S. Engelschall*
17843 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17847 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17848 on the `perl Configure ...` command line. This way one can compile
17849 OpenSSL libraries with Position Independent Code (PIC) which is needed
17850 for linking it into DSOs.
17852 *Ralf S. Engelschall*
17854 * Remarkably, export ciphers were totally broken and no-one had noticed!
17859 * Cleaned up the LICENSE document: The official contact for any license
17860 questions now is the OpenSSL core team under openssl-core@openssl.org.
17861 And add a paragraph about the dual-license situation to make sure people
17862 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17863 to the OpenSSL toolkit.
17865 *Ralf S. Engelschall*
17867 * General source tree makefile cleanups: Made `making xxx in yyy...`
17868 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17869 Additionally cleaned up the `make links` target: Remove unnecessary
17870 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17871 to speed processing and no longer clutter the display with confusing
17872 stuff. Instead only the actually done links are displayed.
17874 *Ralf S. Engelschall*
17876 * Permit null encryption ciphersuites, used for authentication only. It used
17877 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17878 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17883 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17884 signed attributes when verifying signatures (this would break them),
17885 the detached data encoding was wrong and public keys obtained using
17886 X509_get_pubkey() weren't freed.
17890 * Add text documentation for the BUFFER functions. Also added a work around
17891 to a Win95 console bug. This was triggered by the password read stuff: the
17892 last character typed gets carried over to the next fread(). If you were
17893 generating a new cert request using 'req' for example then the last
17894 character of the passphrase would be CR which would then enter the first
17899 * Added the new 'Includes OpenSSL Cryptography Software' button as
17900 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17901 button and can be used by applications based on OpenSSL to show the
17902 relationship to the OpenSSL project.
17904 *Ralf S. Engelschall*
17906 * Remove confusing variables in function signatures in files
17907 ssl/ssl_lib.c and ssl/ssl.h.
17909 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17911 * Don't install bss_file.c under PREFIX/include/
17913 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17915 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17916 functions that return function pointers and has support for NT specific
17917 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17918 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17919 unsigned to signed types: this was killing the Win32 compile.
17923 * Add new certificate file to stack functions,
17924 SSL_add_dir_cert_subjects_to_stack() and
17925 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17926 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17927 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17928 This means that Apache-SSL and similar packages don't have to mess around
17929 to add as many CAs as they want to the preferred list.
17933 * Experiment with doxygen documentation. Currently only partially applied to
17935 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17936 openssl.doxy as the configuration file.
17940 * Get rid of remaining C++-style comments which strict C compilers hate.
17942 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17944 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17945 compiled in by default: it has problems with large keys.
17949 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17950 DH private keys and/or callback functions which directly correspond to
17951 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17952 is needed for applications which have to configure certificates on a
17953 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17955 For the RSA certificate situation is makes no difference, but
17956 for the DSA certificate situation this fixes the "no shared cipher"
17957 problem where the OpenSSL cipher selection procedure failed because the
17958 temporary keys were not overtaken from the context and the API provided
17959 no way to reconfigure them.
17960 The new functions now let applications reconfigure the stuff and they
17961 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17962 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17963 non-public-API function ssl_cert_instantiate() is used as a helper
17964 function and also to reduce code redundancy inside ssl_rsa.c.
17966 *Ralf S. Engelschall*
17968 * Move s_server -dcert and -dkey options out of the undocumented feature
17969 area because they are useful for the DSA situation and should be
17970 recognized by the users.
17972 *Ralf S. Engelschall*
17974 * Fix the cipher decision scheme for export ciphers: the export bits are
17975 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17976 SSL_EXP_MASK. So, the original variable has to be used instead of the
17977 already masked variable.
17979 *Richard Levitte <levitte@stacken.kth.se>*
17981 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17983 *Richard Levitte <levitte@stacken.kth.se>*
17985 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17986 from `int` to `unsigned int` because it is a length and initialized by
17987 EVP_DigestFinal() which expects an `unsigned int *`.
17989 *Richard Levitte <levitte@stacken.kth.se>*
17991 * Don't hard-code path to Perl interpreter on shebang line of Configure
17992 script. Instead use the usual Shell->Perl transition trick.
17994 *Ralf S. Engelschall*
17996 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17997 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17998 -noout -modulus` as it's already the case for `openssl rsa -noout
17999 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
18000 currently the public key is printed (a decision which was already done by
18001 `openssl dsa -modulus` in the past) which serves a similar purpose.
18002 Additionally the NO_RSA no longer completely removes the whole -modulus
18003 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18006 *Ralf S. Engelschall*
18008 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18009 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18011 *Arne Ansper <arne@ats.cyber.ee>*
18013 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18014 to be added. Now both 'req' and 'ca' can use new objects defined in the
18019 * Add cool BIO that does syslog (or event log on NT).
18021 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18023 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18024 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18025 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18026 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18030 * Add preliminary config info for new extension code.
18034 * Make RSA_NO_PADDING really use no padding.
18036 *Ulf Moeller <ulf@fitug.de>*
18038 * Generate errors when private/public key check is done.
18042 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18043 for some CRL extensions and new objects added.
18047 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18048 key usage extension and fuller support for authority key id.
18052 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18053 padding method for RSA, which is recommended for new applications in PKCS
18054 #1 v2.0 (RFC 2437, October 1998).
18055 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18056 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18057 against Bleichbacher's attack on RSA.
18058 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18061 * Updates to the new SSL compression code
18063 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18065 * Fix so that the version number in the master secret, when passed
18066 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18067 (because the server will not accept higher), that the version number
18068 is 0x03,0x01, not 0x03,0x00
18070 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18072 * Run extensive memory leak checks on SSL commands. Fixed *lots* of memory
18073 leaks in `ssl/` relating to new `X509_get_pubkey()` behaviour. Also fixes
18074 in `apps/` and an unrelated leak in `crypto/dsa/dsa_vrf.c`.
18078 * Support for RAW extensions where an arbitrary extension can be
18079 created by including its DER encoding. See `apps/openssl.cnf` for
18084 * Make sure latest Perl versions don't interpret some generated C array
18085 code as Perl array code in the crypto/err/err_genc.pl script.
18087 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18089 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18090 not many people have the assembler. Various Win32 compilation fixes and
18091 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18092 build instructions.
18096 * Modify configure script 'Configure' to automatically create crypto/date.h
18097 file under Win32 and also build pem.h from pem.org. New script
18098 util/mkfiles.pl to create the MINFO file on environments that can't do a
18099 'make files': perl util/mkfiles.pl >MINFO should work.
18103 * Major rework of DES function declarations, in the pursuit of correctness
18104 and purity. As a result, many evil casts evaporated, and some weirdness,
18105 too. You may find this causes warnings in your code. Zapping your evil
18106 casts will probably fix them. Mostly.
18110 * Fix for a typo in asn1.h. Bug fix to object creation script
18111 obj_dat.pl. It considered a zero in an object definition to mean
18112 "end of object": none of the objects in objects.h have any zeros
18113 so it wasn't spotted.
18115 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18117 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18118 Masking (CBCM). In the absence of test vectors, the best I have been able
18119 to do is check that the decrypt undoes the encrypt, so far. Send me test
18120 vectors if you have them.
18124 * Correct calculation of key length for export ciphers (too much space was
18125 allocated for null ciphers). This has not been tested!
18129 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18130 message is now correct (it understands "crypto" and "ssl" on its
18131 command line). There is also now an "update" option. This will update
18132 the util/ssleay.num and util/libeay.num files with any new functions.
18134 perl util/mkdef.pl crypto ssl update
18135 it will update them.
18139 * Overhauled the Perl interface:
18140 - ported BN stuff to OpenSSL's different BN library
18141 - made the perl/ source tree CVS-aware
18142 - renamed the package from SSLeay to OpenSSL (the files still contain
18143 their history because I've copied them in the repository)
18144 - removed obsolete files (the test scripts will be replaced
18145 by better Test::Harness variants in the future)
18147 *Ralf S. Engelschall*
18149 * First cut for a very conservative source tree cleanup:
18150 1. merge various obsolete readme texts into doc/ssleay.txt
18151 where we collect the old documents and readme texts.
18152 2. remove the first part of files where I'm already sure that we no
18153 longer need them because of three reasons: either they are just temporary
18154 files which were left by Eric or they are preserved original files where
18155 I've verified that the diff is also available in the CVS via "cvs diff
18156 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18157 the crypto/md/ stuff).
18159 *Ralf S. Engelschall*
18161 * More extension code. Incomplete support for subject and issuer alt
18162 name, issuer and authority key id. Change the i2v function parameters
18163 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18164 what that's for :-) Fix to ASN1 macro which messed up
18165 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18169 * Preliminary support for ENUMERATED type. This is largely copied from the
18174 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18176 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18178 * Make sure `make rehash` target really finds the `openssl` program.
18180 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18182 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18183 like to hear about it if this slows down other processors.
18187 * Add CygWin32 platform information to Configure script.
18189 *Alan Batie <batie@aahz.jf.intel.com>*
18191 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18193 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18195 * New program nseq to manipulate netscape certificate sequences
18199 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18204 * Fixes to BN code. Previously the default was to define BN_RECURSION
18205 but the BN code had some problems that would cause failures when
18206 doing certificate verification and some other functions.
18208 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18210 * Add ASN1 and PEM code to support netscape certificate sequences.
18214 * Add ASN1 and PEM code to support netscape certificate sequences.
18218 * Add several PKIX and private extended key usage OIDs.
18222 * Modify the 'ca' program to handle the new extension code. Modify
18223 openssl.cnf for new extension format, add comments.
18227 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18228 and add a sample to openssl.cnf so req -x509 now adds appropriate
18233 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18234 error code, add initial support to X509_print() and x509 application.
18238 * Takes a deep breath and start adding X509 V3 extension support code. Add
18239 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18240 stuff is currently isolated and isn't even compiled yet.
18244 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18245 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18246 Removed the versions check from X509 routines when loading extensions:
18247 this allows certain broken certificates that don't set the version
18248 properly to be processed.
18252 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18253 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18254 can still be regenerated with "make depend".
18258 * Spelling mistake in C version of CAST-128.
18260 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18262 * Changes to the error generation code. The perl script err-code.pl
18263 now reads in the old error codes and retains the old numbers, only
18264 adding new ones if necessary. It also only changes the .err files if new
18265 codes are added. The makefiles have been modified to only insert errors
18266 when needed (to avoid needlessly modifying header files). This is done
18267 by only inserting errors if the .err file is newer than the auto generated
18268 C file. To rebuild all the error codes from scratch (the old behaviour)
18269 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18270 or delete all the .err files.
18274 * CAST-128 was incorrectly implemented for short keys. The C version has
18275 been fixed, but is untested. The assembler versions are also fixed, but
18276 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18277 to regenerate it if needed.
18278 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18279 Hagino <itojun@kame.net>*
18281 * File was opened incorrectly in randfile.c.
18283 *Ulf Möller <ulf@fitug.de>*
18285 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18286 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18287 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18288 al: it's just almost always a UTCTime. Note this patch adds new error
18289 codes so do a "make errors" if there are problems.
18293 * Correct Linux 1 recognition in config.
18295 *Ulf Möller <ulf@fitug.de>*
18297 * Remove pointless MD5 hash when using DSA keys in ca.
18299 *Anonymous <nobody@replay.com>*
18301 * Generate an error if given an empty string as a cert directory. Also
18302 generate an error if handed NULL (previously returned 0 to indicate an
18303 error, but didn't set one).
18305 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18307 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18311 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18312 parameters. This was causing a warning which killed off the Win32 compile.
18316 * Remove C++ style comments from crypto/bn/bn_local.h.
18318 *Neil Costigan <neil.costigan@celocom.com>*
18320 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18321 based on a text string, looking up short and long names and finally
18322 "dot" format. The "dot" format stuff didn't work. Added new function
18323 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18324 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18325 OID is not part of the table.
18329 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18330 X509_LOOKUP_by_alias().
18334 * Sort openssl functions by name.
18338 * Get the `gendsa` command working and add it to the `list` command. Remove
18339 encryption from sample DSA keys (in case anyone is interested the password
18344 * Make *all* `*_free` functions accept a NULL pointer.
18346 *Frans Heymans <fheymans@isaserver.be>*
18348 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18351 *Anonymous <nobody@replay.com>*
18353 * s_server should send the CAfile as acceptable CAs, not its own cert.
18355 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18357 * Don't blow it for numeric `-newkey` arguments to `apps/req`.
18359 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18361 * Temp key "for export" tests were wrong in s3_srvr.c.
18363 *Anonymous <nobody@replay.com>*
18365 * Add prototype for temp key callback functions
18366 SSL_CTX_set_tmp_{rsa,dh}_callback().
18370 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18371 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18375 * X509_name_add_entry() freed the wrong thing after an error.
18377 *Arne Ansper <arne@ats.cyber.ee>*
18379 * rsa_eay.c would attempt to free a NULL context.
18381 *Arne Ansper <arne@ats.cyber.ee>*
18383 * BIO_s_socket() had a broken should_retry() on Windoze.
18385 *Arne Ansper <arne@ats.cyber.ee>*
18387 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18389 *Arne Ansper <arne@ats.cyber.ee>*
18391 * Make sure the already existing X509_STORE->depth variable is initialized
18392 in X509_STORE_new(), but document the fact that this variable is still
18393 unused in the certificate verification process.
18395 *Ralf S. Engelschall*
18397 * Fix the various library and `apps/` files to free up pkeys obtained from
18398 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18402 * Fix reference counting in X509_PUBKEY_get(). This makes
18403 demos/maurice/example2.c work, amongst others, probably.
18405 *Steve Henson and Ben Laurie*
18407 * First cut of a cleanup for `apps/`. First the `ssleay` program is now named
18408 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18409 are no longer created. This way we have a single and consistent command
18410 line interface `openssl <command>`, similar to `cvs <command>`.
18412 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18414 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18415 BIT STRING wrapper always have zero unused bits.
18419 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18423 * Make the top-level INSTALL documentation easier to understand.
18427 * Makefiles updated to exit if an error occurs in a sub-directory
18428 make (including if user presses ^C) [Paul Sutton]
18430 * Make Montgomery context stuff explicit in RSA data structure.
18434 * Fix build order of pem and err to allow for generated pem.h.
18438 * Fix renumbering bug in X509_NAME_delete_entry().
18442 * Enhanced the err-ins.pl script so it makes the error library number
18443 global and can add a library name. This is needed for external ASN1 and
18444 other error libraries.
18448 * Fixed sk_insert which never worked properly.
18452 * Fix ASN1 macros so they can handle indefinite length constructed
18453 EXPLICIT tags. Some non standard certificates use these: they can now
18458 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18459 into a single doc/ssleay.txt bundle. This way the information is still
18460 preserved but no longer messes up this directory. Now it's new room for
18461 the new set of documentation files.
18463 *Ralf S. Engelschall*
18465 * SETs were incorrectly DER encoded. This was a major pain, because they
18466 shared code with SEQUENCEs, which aren't coded the same. This means that
18467 almost everything to do with SETs or SEQUENCEs has either changed name or
18468 number of arguments.
18470 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18472 * Fix test data to work with the above.
18476 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18477 was already fixed by Eric for 0.9.1 it seems.
18479 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18481 * Autodetect FreeBSD3.
18485 * Fix various bugs in Configure. This affects the following platforms:
18489 unixware-2.0-pentium
18494 * Eliminate generated files from CVS. Reorder tests to regenerate files
18495 before they are needed.
18499 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18503 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18505 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18506 changed SSLeay to OpenSSL in version strings.
18508 *Ralf S. Engelschall*
18510 * Some fixups to the top-level documents.
18514 * Fixed the nasty bug where rsaref.h was not found under compile-time
18515 because the symlink to include/ was missing.
18517 *Ralf S. Engelschall*
18519 * Incorporated the popular no-RSA/DSA-only patches
18520 which allow to compile a RSA-free SSLeay.
18522 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18524 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18525 when "ssleay" is still not found.
18527 *Ralf S. Engelschall*
18529 * Added more platforms to Configure: Cray T3E, HPUX 11,
18531 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18533 * Updated the README file.
18535 *Ralf S. Engelschall*
18537 * Added various .cvsignore files in the CVS repository subdirs
18538 to make a "cvs update" really silent.
18540 *Ralf S. Engelschall*
18542 * Recompiled the error-definition header files and added
18543 missing symbols to the Win32 linker tables.
18545 *Ralf S. Engelschall*
18547 * Cleaned up the top-level documents;
18548 o new files: CHANGES and LICENSE
18549 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18550 o merged COPYRIGHT into LICENSE
18551 o removed obsolete TODO file
18552 o renamed MICROSOFT to INSTALL.W32
18554 *Ralf S. Engelschall*
18556 * Removed dummy files from the 0.9.1b source tree:
18558 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18559 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18560 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18561 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18562 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18565 *Ralf S. Engelschall*
18567 * Added various platform portability fixes.
18571 * The Genesis of the OpenSSL rpject:
18572 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18573 Young and Tim J. Hudson created while they were working for C2Net until
18576 *The OpenSSL Project*
18578 ### Changes between 0.9.0b and 0.9.1b [not released]
18580 * Updated a few CA certificates under certs/
18584 * Changed some BIGNUM api stuff.
18588 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18589 DGUX x86, Linux Alpha, etc.
18593 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18594 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18599 * Add -strparse option to asn1pars program which parses nested
18602 *Dr Stephen Henson <shenson@bigfoot.com>*
18604 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18608 * DSA fix for "ca" program.
18612 * Added "-genkey" option to "dsaparam" program.
18616 * Added RIPE MD160 (rmd160) message digest.
18620 * Added -a (all) option to "ssleay version" command.
18624 * Added PLATFORM define which is the id given to Configure.
18628 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18632 * Extended the ASN.1 parser routines.
18636 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18640 * Added a BN_CTX to the BN library.
18644 * Fixed the weak key values in DES library
18648 * Changed API in EVP library for cipher aliases.
18652 * Added support for RC2/64bit cipher.
18656 * Converted the lhash library to the crypto/mem.c functions.
18660 * Added more recognized ASN.1 object ids.
18664 * Added more RSA padding checks for SSL/TLS.
18668 * Added BIO proxy/filter functionality.
18672 * Added extra_certs to SSL_CTX which can be used
18673 send extra CA certificates to the client in the CA cert chain sending
18674 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18678 * Now Fortezza is denied in the authentication phase because
18679 this is key exchange mechanism is not supported by SSLeay at all.
18683 * Additional PKCS1 checks.
18687 * Support the string "TLSv1" for all TLS v1 ciphers.
18691 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18692 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18696 * Fixed a few memory leaks.
18700 * Fixed various code and comment typos.
18704 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18705 bytes sent in the client random.
18707 *Edward Bishop <ebishop@spyglass.com>*
18711 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18712 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18713 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18714 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18715 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18716 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18717 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18718 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18719 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18720 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18721 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18722 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18723 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18724 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18725 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18726 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18727 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18728 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18729 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18730 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18731 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18732 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18733 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18734 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18735 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18736 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18737 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18738 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18739 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18740 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18741 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18742 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18743 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18744 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18745 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18746 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18747 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18748 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18749 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18750 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18751 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18752 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18753 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18754 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18755 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18756 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18757 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18758 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18759 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18760 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18761 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18762 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18763 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18764 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18765 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18766 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18767 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18768 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18769 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18770 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18771 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18772 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18773 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18774 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18775 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18776 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18777 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18778 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18779 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18780 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18781 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18782 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18783 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18784 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18785 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18786 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18787 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18788 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18789 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18790 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18791 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18792 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18793 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18794 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18795 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18796 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18797 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18798 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18799 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18800 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18801 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18802 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18803 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18804 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18805 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18806 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18807 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18808 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18809 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18810 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18811 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18812 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18813 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18814 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18815 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18816 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18817 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18818 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18819 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18820 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18821 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18822 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18823 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18824 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18825 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18826 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18827 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18828 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18829 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18830 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18831 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18832 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18833 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18834 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18835 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18836 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18837 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18838 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18839 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18840 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18841 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18842 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18843 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18844 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18845 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18846 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18847 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18848 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18849 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18850 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18851 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18852 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18853 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18854 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18855 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18856 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18857 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18858 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18859 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18860 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18861 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18862 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18863 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18864 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18865 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18866 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18867 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18868 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18869 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18870 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18871 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18872 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655