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 of the command line
46 utility list has 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() & 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 the `openssl` command without arguments is equivalent to `openssl
253 * Renamed EVP_PKEY_cmp() to EVP_PKEY_eq() and
254 EVP_PKEY_cmp_parameters() to EVP_PKEY_parameters_eq().
255 While the old function names have been retained for backward compatibility
256 they should not be used in new developments
257 because their return values are confusing: Unlike other `_cmp()` functions
258 they do not return 0 in case their arguments are equal.
262 * Deprecated EC_METHOD_get_field_type(). Applications should switch to
263 EC_GROUP_get_field_type().
267 * Deprecated EC_GFp_simple_method(), EC_GFp_mont_method(),
268 EC_GF2m_simple_method(), EC_GFp_nist_method(), EC_GFp_nistp224_method()
269 EC_GFp_nistp256_method(), and EC_GFp_nistp521_method().
270 Applications should rely on the library automatically assigning a suitable
271 EC_METHOD internally upon EC_GROUP construction.
275 * Deprecated EC_GROUP_new(), EC_GROUP_method_of(), and EC_POINT_method_of().
276 EC_METHOD is now an internal-only concept and a suitable EC_METHOD is
277 assigned internally without application intervention.
278 Users of EC_GROUP_new() should switch to a different suitable constructor.
282 * Add CAdES-BES signature verification support, mostly derived
283 from ESSCertIDv2 TS (RFC 5816) contribution by Marek Klein.
285 *Filipe Raimundo da Silva*
287 * Add CAdES-BES signature scheme and attributes support (RFC 5126) to CMS API.
291 * Deprecated EC_POINT_make_affine() and EC_POINTs_make_affine(). These
292 functions are not widely used and now OpenSSL automatically perform this
293 conversion when needed.
297 * Deprecated EC_GROUP_precompute_mult(), EC_GROUP_have_precompute_mult(), and
298 EC_KEY_precompute_mult(). These functions are not widely used and
299 applications should instead switch to named curves which OpenSSL has
300 hardcoded lookup tables for.
304 * Deprecated EC_POINTs_mul(). This function is not widely used and applications
305 should instead use the L<EC_POINT_mul(3)> function.
309 * Removed FIPS_mode() and FIPS_mode_set(). These functions are legacy API's
310 that are not applicable to the new provider model. Applications should
311 instead use EVP_default_properties_is_fips_enabled() and
312 EVP_default_properties_enable_fips().
316 * The SSL option SSL_OP_IGNORE_UNEXPECTED_EOF is introduced. If that option
317 is set, an unexpected EOF is ignored, it pretends a close notify was received
318 instead and so the returned error becomes SSL_ERROR_ZERO_RETURN.
322 * Deprecated EC_POINT_set_Jprojective_coordinates_GFp() and
323 EC_POINT_get_Jprojective_coordinates_GFp(). These functions are not widely
324 used and applications should instead use the
325 L<EC_POINT_set_affine_coordinates(3)> and
326 L<EC_POINT_get_affine_coordinates(3)> functions.
330 * Added OSSL_PARAM_BLD to the public interface. This allows OSSL_PARAM
331 arrays to be more easily constructed via a series of utility functions.
332 Create a parameter builder using OSSL_PARAM_BLD_new(), add parameters using
333 the various push functions and finally convert to a passable OSSL_PARAM
334 array using OSSL_PARAM_BLD_to_param().
338 * The security strength of SHA1 and MD5 based signatures in TLS has been
339 reduced. This results in SSL 3, TLS 1.0, TLS 1.1 and DTLS 1.0 no longer
340 working at the default security level of 1 and instead requires security
341 level 0. The security level can be changed either using the cipher string
342 with @SECLEVEL, or calling SSL_CTX_set_security_level().
346 * EVP_PKEY_get0_RSA(), EVP_PKEY_get0_DSA(), EVP_PKEY_get0_DH(), and
347 EVP_PKEY_get0_EC_KEY() can now handle EVP_PKEYs with provider side
348 internal keys, if they correspond to one of those built in types.
352 * Added EVP_PKEY_set_type_by_keymgmt(), to initialise an EVP_PKEY to
353 contain a provider side internal key.
357 * ASN1_verify(), ASN1_digest() and ASN1_sign() have been deprecated.
358 They are old functions that we don't use, and that you could disable with
359 the macro NO_ASN1_OLD. This goes all the way back to OpenSSL 0.9.7.
363 * Project text documents not yet having a proper file name extension
364 (`HACKING`, `LICENSE`, `NOTES*`, `README*`, `VERSION`) have been renamed to
365 `*.md` as far as reasonable, else `*.txt`, for better use with file managers.
369 * The main project documents (README, NEWS, CHANGES, INSTALL, SUPPORT)
370 have been converted to Markdown with the goal to produce documents
371 which not only look pretty when viewed online in the browser, but
372 remain well readable inside a plain text editor.
374 To achieve this goal, a 'minimalistic' Markdown style has been applied
375 which avoids formatting elements that interfere too much with the
376 reading flow in the text file. For example, it
378 * avoids [ATX headings][] and uses [setext headings][] instead
379 (which works for `<h1>` and `<h2>` headings only).
380 * avoids [inline links][] and uses [reference links][] instead.
381 * avoids [fenced code blocks][] and uses [indented code blocks][] instead.
383 [ATX headings]: https://github.github.com/gfm/#atx-headings
384 [setext headings]: https://github.github.com/gfm/#setext-headings
385 [inline links]: https://github.github.com/gfm/#inline-link
386 [reference links]: https://github.github.com/gfm/#reference-link
387 [fenced code blocks]: https://github.github.com/gfm/#fenced-code-blocks
388 [indented code blocks]: https://github.github.com/gfm/#indented-code-blocks
390 *Matthias St. Pierre*
392 * The test suite is changed to preserve results of each test recipe.
393 A new directory test-runs/ with subdirectories named like the
394 test recipes are created in the build tree for this purpose.
398 * Added an implementation of CMP and CRMF (RFC 4210, RFC 4211 RFC 6712).
399 This adds crypto/cmp/, crpyto/crmf/, apps/cmp.c, and test/cmp_*.
400 See L<openssl-cmp(1)> and L<OSSL_CMP_exec_IR_ses(3)> as starting points.
402 *David von Oheimb, Martin Peylo*
404 * Generalized the HTTP client code from crypto/ocsp/ into crpyto/http/.
405 The legacy OCSP-focused and only partly documented API is retained.
406 See L<OSSL_CMP_MSG_http_perform(3)> etc. for details.
410 * Added `util/check-format.pl`, a tool for checking adherence to the
411 OpenSSL coding style <https://www.openssl.org/policies/codingstyle.html>.
412 The checks performed are incomplete and yield some false positives.
413 Still the tool should be useful for detecting most typical glitches.
417 * BIO_do_connect and BIO_do_handshake have been extended:
418 If domain name resolution yields multiple IP addresses all of them are tried
419 after connect() failures.
423 * All of the low level RSA functions have been deprecated including:
425 RSA_new_method, RSA_size, RSA_security_bits, RSA_get0_pss_params,
426 RSA_get_version, RSA_get0_engine, RSA_generate_key_ex,
427 RSA_generate_multi_prime_key, RSA_X931_derive_ex, RSA_X931_generate_key_ex,
428 RSA_check_key, RSA_check_key_ex, RSA_public_encrypt, RSA_private_encrypt,
429 RSA_public_decrypt, RSA_private_decrypt, RSA_set_default_method,
430 RSA_get_default_method, RSA_null_method, RSA_get_method, RSA_set_method,
431 RSA_PKCS1_OpenSSL, RSA_print_fp, RSA_print, RSA_sign, RSA_verify,
432 RSA_sign_ASN1_OCTET_STRING, RSA_verify_ASN1_OCTET_STRING, RSA_blinding_on,
433 RSA_blinding_off, RSA_setup_blinding, RSA_padding_add_PKCS1_type_1,
434 RSA_padding_check_PKCS1_type_1, RSA_padding_add_PKCS1_type_2,
435 RSA_padding_check_PKCS1_type_2, PKCS1_MGF1, RSA_padding_add_PKCS1_OAEP,
436 RSA_padding_check_PKCS1_OAEP, RSA_padding_add_PKCS1_OAEP_mgf1,
437 RSA_padding_check_PKCS1_OAEP_mgf1, RSA_padding_add_SSLv23,
438 RSA_padding_check_SSLv23, RSA_padding_add_none, RSA_padding_check_none,
439 RSA_padding_add_X931, RSA_padding_check_X931, RSA_X931_hash_id,
440 RSA_verify_PKCS1_PSS, RSA_padding_add_PKCS1_PSS, RSA_verify_PKCS1_PSS_mgf1,
441 RSA_padding_add_PKCS1_PSS_mgf1, RSA_set_ex_data, RSA_get_ex_data,
442 RSA_meth_new, RSA_meth_free, RSA_meth_dup, RSA_meth_get0_name,
443 RSA_meth_set1_name, RSA_meth_get_flags, RSA_meth_set_flags,
444 RSA_meth_get0_app_data, RSA_meth_set0_app_data, RSA_meth_get_pub_enc,
445 RSA_meth_set_pub_enc, RSA_meth_get_pub_dec, RSA_meth_set_pub_dec,
446 RSA_meth_get_priv_enc, RSA_meth_set_priv_enc, RSA_meth_get_priv_dec,
447 RSA_meth_set_priv_dec, RSA_meth_get_mod_exp, RSA_meth_set_mod_exp,
448 RSA_meth_get_bn_mod_exp, RSA_meth_set_bn_mod_exp, RSA_meth_get_init,
449 RSA_meth_set_init, RSA_meth_get_finish, RSA_meth_set_finish,
450 RSA_meth_get_sign, RSA_meth_set_sign, RSA_meth_get_verify,
451 RSA_meth_set_verify, RSA_meth_get_keygen, RSA_meth_set_keygen,
452 RSA_meth_get_multi_prime_keygen and RSA_meth_set_multi_prime_keygen.
454 Use of these low level functions has been informally discouraged for a long
455 time. Instead applications should use L<EVP_PKEY_encrypt_init(3)>,
456 L<EVP_PKEY_encrypt(3)>, L<EVP_PKEY_decrypt_init(3)> and
457 L<EVP_PKEY_decrypt(3)>.
461 * X509 certificates signed using SHA1 are no longer allowed at security
463 In TLS/SSL the default security level is 1. It can be set either
464 using the cipher string with @SECLEVEL, or calling
465 SSL_CTX_set_security_level(). If the leaf certificate is signed with SHA-1,
466 a call to SSL_CTX_use_certificate() will fail if the security level is not
468 Outside TLS/SSL, the default security level is -1 (effectively 0). It can
469 be set using X509_VERIFY_PARAM_set_auth_level() or using the -auth_level
474 * The command line utilities dhparam, dsa, gendsa and dsaparam have been
475 modified to use PKEY APIs. These commands are now in maintenance mode
476 and no new features will be added to them.
480 * The command line utility rsautl has been deprecated.
481 Instead use the pkeyutl program.
485 * The command line utilities genrsa and rsa have been modified to use PKEY
486 APIs. They now write PKCS#8 keys by default. These commands are now in
487 maintenance mode and no new features will be added to them.
491 * All of the low level DH functions have been deprecated including:
493 DH_OpenSSL, DH_set_default_method, DH_get_default_method, DH_set_method,
494 DH_new_method, DH_new, DH_free, DH_up_ref, DH_bits, DH_set0_pqg, DH_size,
495 DH_security_bits, DH_get_ex_new_index, DH_set_ex_data, DH_get_ex_data,
496 DH_generate_parameters_ex, DH_check_params_ex, DH_check_ex, DH_check_pub_key_ex,
497 DH_check, DH_check_pub_key, DH_generate_key, DH_compute_key,
498 DH_compute_key_padded, DHparams_print_fp, DHparams_print, DH_get_nid,
499 DH_KDF_X9_42, DH_get0_engine, DH_meth_new, DH_meth_free, DH_meth_dup,
500 DH_meth_get0_name, DH_meth_set1_name, DH_meth_get_flags, DH_meth_set_flags,
501 DH_meth_get0_app_data, DH_meth_set0_app_data, DH_meth_get_generate_key,
502 DH_meth_set_generate_key, DH_meth_get_compute_key, DH_meth_set_compute_key,
503 DH_meth_get_bn_mod_exp, DH_meth_set_bn_mod_exp, DH_meth_get_init,
504 DH_meth_set_init, DH_meth_get_finish, DH_meth_set_finish,
505 DH_meth_get_generate_params and DH_meth_set_generate_params.
507 Use of these low level functions has been informally discouraged for a long
508 time. Instead applications should use L<EVP_PKEY_derive_init(3)>
509 and L<EVP_PKEY_derive(3)>.
511 Additionally functions that read and write DH objects such as d2i_DHparams,
512 i2d_DHparams, PEM_read_DHparam, PEM_write_DHparams and other similar
513 functions have also been deprecated. Applications should instead use the
514 OSSL_DECODER and OSSL_ENCODER APIs to read and write DH files.
516 Finaly functions that assign or obtain DH objects from an EVP_PKEY such as
517 EVP_PKEY_assign_DH(), EVP_PKEY_get0_DH, EVP_PKEY_get1_DH, EVP_PKEY_set1_DH
518 are also deprecated. Applications should instead either read or write an
519 EVP_PKEY directly using the OSSL_DECODER and OSSL_ENCODER APIs. Or load an
520 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 * Added several checks to X509_verify_cert() according to requirements in
918 RFC 5280 in case `X509_V_FLAG_X509_STRICT` is set
919 (which may be done by using the CLI option `-x509_strict`):
920 * The basicConstraints of CA certificates must be marked critical.
921 * CA certificates must explicitly include the keyUsage extension.
922 * If a pathlenConstraint is given the key usage keyCertSign must be allowed.
923 * The issuer name of any certificate must not be empty.
924 * The subject name of CA certs, certs with keyUsage crlSign,
925 and certs without subjectAlternativeName must not be empty.
926 * If a subjectAlternativeName extension is given it must not be empty.
927 * The signatureAlgorithm field and the cert signature must be consistent.
928 * Any given authorityKeyIdentifier and any given subjectKeyIdentifier
929 must not be marked critical.
930 * The authorityKeyIdentifier must be given for X.509v3 certs
931 unless they are self-signed.
932 * The subjectKeyIdentifier must be given for all X.509v3 CA certs.
936 * Certificate verification using X509_verify_cert() meanwhile rejects EC keys
937 with explicit curve parameters (specifiedCurve) as required by RFC 5480.
941 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
942 used even when parsing explicit parameters, when loading a encoded key
943 or calling `EC_GROUP_new_from_ecpkparameters()`/
944 `EC_GROUP_new_from_ecparameters()`.
945 This prevents bypass of security hardening and performance gains,
946 especially for curves with specialized EC_METHODs.
947 By default, if a key encoded with explicit parameters is loaded and later
948 encoded, the output is still encoded with explicit parameters, even if
949 internally a "named" EC_GROUP is used for computation.
953 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
954 this change, EC_GROUP_set_generator would accept order and/or cofactor as
955 NULL. After this change, only the cofactor parameter can be NULL. It also
956 does some minimal sanity checks on the passed order.
961 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
962 An attack is simple, if the first CMS_recipientInfo is valid but the
963 second CMS_recipientInfo is chosen ciphertext. If the second
964 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
965 encryption key will be replaced by garbage, and the message cannot be
966 decoded, but if the RSA decryption fails, the correct encryption key is
967 used and the recipient will not notice the attack.
968 As a work around for this potential attack the length of the decrypted
969 key must be equal to the cipher default key length, in case the
970 certifiate is not given and all recipientInfo are tried out.
971 The old behaviour can be re-enabled in the CMS code by setting the
972 CMS_DEBUG_DECRYPT flag.
976 * Early start up entropy quality from the DEVRANDOM seed source has been
977 improved for older Linux systems. The RAND subsystem will wait for
978 /dev/random to be producing output before seeding from /dev/urandom.
979 The seeded state is stored for future library initialisations using
980 a system global shared memory segment. The shared memory identifier
981 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
982 the desired value. The default identifier is 114.
986 * Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
987 when primes for RSA keys are computed.
988 Since we previously always generated primes == 2 (mod 3) for RSA keys,
989 the 2-prime and 3-prime RSA modules were easy to distinguish, since
990 `N = p*q = 1 (mod 3)`, but `N = p*q*r = 2 (mod 3)`. Therefore fingerprinting
991 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
992 This avoids possible fingerprinting of newly generated RSA modules.
996 * Correct the extended master secret constant on EBCDIC systems. Without this
997 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
998 negotiate EMS will fail. Unfortunately this also means that TLS connections
999 between EBCDIC systems with this fix, and EBCDIC systems without this
1000 fix will fail if they negotiate EMS.
1004 * Changed the library initialisation so that the config file is now loaded
1005 by default. This was already the case for libssl. It now occurs for both
1006 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
1007 OPENSSL_init_crypto() to suppress automatic loading of a config file.
1011 * Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
1012 where the former acts as a replacement for ERR_put_error(), and the
1013 latter replaces the combination ERR_put_error()+ERR_add_error_data().
1014 ERR_raise_data() adds more flexibility by taking a format string and
1015 an arbitrary number of arguments following it, to be processed with
1020 * Introduced a new function, OSSL_PROVIDER_available(), which can be used
1021 to check if a named provider is loaded and available. When called, it
1022 will also activate all fallback providers if such are still present.
1026 * Enforce a minimum DH modulus size of 512 bits.
1030 * Changed DH parameters to generate the order q subgroup instead of 2q.
1031 Previously generated DH parameters are still accepted by DH_check
1032 but DH_generate_key works around that by clearing bit 0 of the
1033 private key for those. This avoids leaking bit 0 of the private key.
1037 * Significantly reduce secure memory usage by the randomness pools.
1041 * `{CRYPTO,OPENSSL}_mem_debug_{push,pop}` are now no-ops and have been
1046 * A new type, EVP_KEYEXCH, has been introduced to represent key exchange
1047 algorithms. An implementation of a key exchange algorithm can be obtained
1048 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
1049 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
1050 the older EVP_PKEY_derive_init() function. See the man pages for the new
1051 functions for further details.
1055 * The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
1059 * Removed the function names from error messages and deprecated the
1062 * Removed NextStep support and the macro OPENSSL_UNISTD
1066 * Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
1067 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
1068 Also removed "export var as function" capability; we do not export
1069 variables, only functions.
1073 * RC5_32_set_key has been changed to return an int type, with 0 indicating
1074 an error and 1 indicating success. In previous versions of OpenSSL this
1075 was a void type. If a key was set longer than the maximum possible this
1080 * Support SM2 signing and verification schemes with X509 certificate.
1084 * Use SHA256 as the default digest for TS query in the ts app.
1088 * Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
1089 This checks that the salt length is at least 128 bits, the derived key
1090 length is at least 112 bits, and that the iteration count is at least 1000.
1091 For backwards compatibility these checks are disabled by default in the
1092 default provider, but are enabled by default in the fips provider.
1093 To enable or disable these checks use the control
1094 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
1098 * Default cipher lists/suites are now available via a function, the
1099 #defines are deprecated.
1103 * Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
1104 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
1105 for Windows Store apps easier. Also, the "no-uplink" option has been added.
1109 * Join the directories crypto/x509 and crypto/x509v3
1113 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1114 This changes the size when using the genpkey app when no size is given. It
1115 fixes an omission in earlier changes that changed all RSA, DSA and DH
1116 generation apps to use 2048 bits by default.
1120 * Added command 'openssl kdf' that uses the EVP_KDF API.
1124 * Added command 'openssl mac' that uses the EVP_MAC API.
1128 * Added OPENSSL_info() to get diverse built-in OpenSSL data, such
1129 as default directories. Also added the command 'openssl info'
1130 for scripting purposes.
1134 * The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
1135 deprecated. These undocumented functions were never integrated into the EVP
1136 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
1137 Bi-directional IGE mode. These modes were never formally standardised and
1138 usage of these functions is believed to be very small. In particular
1139 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
1140 is ever used. The security implications are believed to be minimal, but
1141 this issue was never fixed for backwards compatibility reasons. New code
1142 should not use these modes.
1146 * Add prediction resistance to the DRBG reseeding process.
1150 * Limit the number of blocks in a data unit for AES-XTS to 2^20 as
1151 mandated by IEEE Std 1619-2018.
1155 * Added newline escaping functionality to a filename when using openssl dgst.
1156 This output format is to replicate the output format found in the `*sum`
1157 checksum programs. This aims to preserve backward compatibility.
1159 *Matt Eaton, Richard Levitte, and Paul Dale*
1161 * Removed the heartbeat message in DTLS feature, as it has very
1162 little usage and doesn't seem to fulfill a valuable purpose.
1163 The configuration option is now deprecated.
1167 * Changed the output of 'openssl {digestname} < file' to display the
1168 digest name in its output.
1172 * Added a new generic trace API which provides support for enabling
1173 instrumentation through trace output. This feature is mainly intended
1174 as an aid for developers and is disabled by default. To utilize it,
1175 OpenSSL needs to be configured with the `enable-trace` option.
1177 If the tracing API is enabled, the application can activate trace output
1178 by registering BIOs as trace channels for a number of tracing and debugging
1181 The 'openssl' application has been expanded to enable any of the types
1182 available via environment variables defined by the user, and serves as
1183 one possible example on how to use this functionality.
1185 *Richard Levitte & Matthias St. Pierre*
1187 * Added build tests for C++. These are generated files that only do one
1188 thing, to include one public OpenSSL head file each. This tests that
1189 the public header files can be usefully included in a C++ application.
1191 This test isn't enabled by default. It can be enabled with the option
1192 'enable-buildtest-c++'.
1196 * Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
1200 * Add KMAC to EVP_MAC.
1204 * Added property based algorithm implementation selection framework to
1209 * Added SCA hardening for modular field inversion in EC_GROUP through
1210 a new dedicated field_inv() pointer in EC_METHOD.
1211 This also addresses a leakage affecting conversions from projective
1212 to affine coordinates.
1214 *Billy Bob Brumley, Nicola Tuveri*
1216 * Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
1217 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
1218 those algorithms that were already supported through the EVP_PKEY API
1219 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
1220 and scrypt are now wrappers that call EVP_KDF.
1224 * Build devcrypto engine as a dynamic engine.
1226 *Eneas U de Queiroz*
1228 * Add keyed BLAKE2 to EVP_MAC.
1232 * Fix a bug in the computation of the endpoint-pair shared secret used
1233 by DTLS over SCTP. This breaks interoperability with older versions
1234 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
1235 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
1236 interoperability with such broken implementations. However, enabling
1237 this switch breaks interoperability with correct implementations.
1239 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
1240 re-used X509_PUBKEY object if the second PUBKEY is malformed.
1244 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
1248 * Change the license to the Apache License v2.0.
1252 * Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
1254 - Major releases (indicated by incrementing the MAJOR release number)
1255 may introduce incompatible API/ABI changes.
1256 - Minor releases (indicated by incrementing the MINOR release number)
1257 may introduce new features but retain API/ABI compatibility.
1258 - Patch releases (indicated by incrementing the PATCH number)
1259 are intended for bug fixes and other improvements of existing
1260 features only (like improving performance or adding documentation)
1261 and retain API/ABI compatibility.
1265 * Add support for RFC5297 SIV mode (siv128), including AES-SIV.
1269 * Remove the 'dist' target and add a tarball building script. The
1270 'dist' target has fallen out of use, and it shouldn't be
1271 necessary to configure just to create a source distribution.
1275 * Recreate the OS390-Unix config target. It no longer relies on a
1276 special script like it did for OpenSSL pre-1.1.0.
1280 * Instead of having the source directories listed in Configure, add
1281 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
1286 * Add GMAC to EVP_MAC.
1290 * Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
1294 * Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
1295 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
1296 to facilitate the continued use of MACs through raw private keys in
1297 functionality such as `EVP_DigestSign*` and `EVP_DigestVerify*`.
1301 * Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
1302 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
1306 * Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
1307 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
1308 are retained for backwards compatibility.
1312 * AES-XTS mode now enforces that its two keys are different to mitigate
1313 the attacked described in "Efficient Instantiations of Tweakable
1314 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
1315 Details of this attack can be obtained from:
1316 <http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf>
1320 * Rename the object files, i.e. give them other names than in previous
1321 versions. Their names now include the name of the final product, as
1322 well as its type mnemonic (bin, lib, shlib).
1326 * Added new option for 'openssl list', '-objects', which will display the
1327 list of built in objects, i.e. OIDs with names.
1331 * Added the options `-crl_lastupdate` and `-crl_nextupdate` to `openssl ca`,
1332 allowing the `lastUpdate` and `nextUpdate` fields in the generated CRL to
1337 * Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
1338 improves application performance by removing data copies and providing
1339 applications with zero-copy system calls such as sendfile and splice.
1343 * The SSL option SSL_OP_CLEANSE_PLAINTEXT is introduced. If that
1344 option is set, openssl cleanses (zeroize) plaintext bytes from
1345 internal buffers after delivering them to the application. Note,
1346 the application is still responsible for cleansing other copies
1347 (e.g.: data received by SSL_read(3)).
1351 * `PKCS12_parse` now maintains the order of the parsed certificates
1352 when outputting them via `*ca` (rather than reversing it).
1356 * Deprecated pthread fork support methods. These were unused so no
1357 replacement is required.
1359 - OPENSSL_fork_prepare()
1360 - OPENSSL_fork_parent()
1361 - OPENSSL_fork_child()
1368 ### Changes between 1.1.1h and 1.1.1i [xx XXX xxxx]
1370 * Fixed NULL pointer deref in the GENERAL_NAME_cmp function
1371 This function could crash if both GENERAL_NAMEs contain an EDIPARTYNAME.
1372 If an attacker can control both items being compared then this could lead
1373 to a possible denial of service attack. OpenSSL itself uses the
1374 GENERAL_NAME_cmp function for two purposes:
1375 1) Comparing CRL distribution point names between an available CRL and a
1376 CRL distribution point embedded in an X509 certificate
1377 2) When verifying that a timestamp response token signer matches the
1378 timestamp authority name (exposed via the API functions
1379 TS_RESP_verify_response and TS_RESP_verify_token)
1384 ### Changes between 1.1.1g and 1.1.1h [22 Sep 2020]
1386 * Certificates with explicit curve parameters are now disallowed in
1387 verification chains if the X509_V_FLAG_X509_STRICT flag is used.
1391 * The 'MinProtocol' and 'MaxProtocol' configuration commands now silently
1392 ignore TLS protocol version bounds when configuring DTLS-based contexts, and
1393 conversely, silently ignore DTLS protocol version bounds when configuring
1394 TLS-based contexts. The commands can be repeated to set bounds of both
1395 types. The same applies with the corresponding "min_protocol" and
1396 "max_protocol" command-line switches, in case some application uses both TLS
1399 SSL_CTX instances that are created for a fixed protocol version (e.g.
1400 TLSv1_server_method()) also silently ignore version bounds. Previously
1401 attempts to apply bounds to these protocol versions would result in an
1402 error. Now only the "version-flexible" SSL_CTX instances are subject to
1403 limits in configuration files in command-line options.
1407 * Handshake now fails if Extended Master Secret extension is dropped
1412 * The Oracle Developer Studio compiler will start reporting deprecated APIs
1414 ### Changes between 1.1.1f and 1.1.1g [21 Apr 2020]
1416 * Fixed segmentation fault in SSL_check_chain()
1417 Server or client applications that call the SSL_check_chain() function
1418 during or after a TLS 1.3 handshake may crash due to a NULL pointer
1419 dereference as a result of incorrect handling of the
1420 "signature_algorithms_cert" TLS extension. The crash occurs if an invalid
1421 or unrecognised signature algorithm is received from the peer. This could
1422 be exploited by a malicious peer in a Denial of Service attack.
1427 * Added AES consttime code for no-asm configurations
1428 an optional constant time support for AES was added
1429 when building openssl for no-asm.
1430 Enable with: ./config no-asm -DOPENSSL_AES_CONST_TIME
1431 Disable with: ./config no-asm -DOPENSSL_NO_AES_CONST_TIME
1432 At this time this feature is by default disabled.
1433 It will be enabled by default in 3.0.
1437 ### Changes between 1.1.1e and 1.1.1f [31 Mar 2020]
1439 * Revert the change of EOF detection while reading in libssl to avoid
1440 regressions in applications depending on the current way of reporting
1441 the EOF. As the existing method is not fully accurate the change to
1442 reporting the EOF via SSL_ERROR_SSL is kept on the current development
1443 branch and will be present in the 3.0 release.
1447 * Revised BN_generate_prime_ex to not avoid factors 3..17863 in p-1
1448 when primes for RSA keys are computed.
1449 Since we previously always generated primes == 2 (mod 3) for RSA keys,
1450 the 2-prime and 3-prime RSA modules were easy to distinguish, since
1451 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
1452 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
1453 This avoids possible fingerprinting of newly generated RSA modules.
1457 ### Changes between 1.1.1d and 1.1.1e [17 Mar 2020]
1459 * Properly detect EOF while reading in libssl. Previously if we hit an EOF
1460 while reading in libssl then we would report an error back to the
1461 application (SSL_ERROR_SYSCALL) but errno would be 0. We now add
1462 an error to the stack (which means we instead return SSL_ERROR_SSL) and
1463 therefore give a hint as to what went wrong.
1467 * Check that ed25519 and ed448 are allowed by the security level. Previously
1468 signature algorithms not using an MD were not being checked that they were
1469 allowed by the security level.
1473 * Fixed SSL_get_servername() behaviour. The behaviour of SSL_get_servername()
1474 was not quite right. The behaviour was not consistent between resumption
1475 and normal handshakes, and also not quite consistent with historical
1476 behaviour. The behaviour in various scenarios has been clarified and
1477 it has been updated to make it match historical behaviour as closely as
1482 * *[VMS only]* The header files that the VMS compilers include automatically,
1483 `__DECC_INCLUDE_PROLOGUE.H` and `__DECC_INCLUDE_EPILOGUE.H`, use pragmas
1484 that the C++ compiler doesn't understand. This is a shortcoming in the
1485 compiler, but can be worked around with `__cplusplus` guards.
1487 C++ applications that use OpenSSL libraries must be compiled using the
1488 qualifier `/NAMES=(AS_IS,SHORTENED)` to be able to use all the OpenSSL
1489 functions. Otherwise, only functions with symbols of less than 31
1490 characters can be used, as the linker will not be able to successfully
1491 resolve symbols with longer names.
1495 * Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
1496 The presence of this system service is determined at run-time.
1500 * Added newline escaping functionality to a filename when using openssl dgst.
1501 This output format is to replicate the output format found in the `*sum`
1502 checksum programs. This aims to preserve backward compatibility.
1504 *Matt Eaton, Richard Levitte, and Paul Dale*
1506 * Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
1511 ### Changes between 1.1.1c and 1.1.1d [10 Sep 2019]
1513 * Fixed a fork protection issue. OpenSSL 1.1.1 introduced a rewritten random
1514 number generator (RNG). This was intended to include protection in the
1515 event of a fork() system call in order to ensure that the parent and child
1516 processes did not share the same RNG state. However this protection was not
1517 being used in the default case.
1519 A partial mitigation for this issue is that the output from a high
1520 precision timer is mixed into the RNG state so the likelihood of a parent
1521 and child process sharing state is significantly reduced.
1523 If an application already calls OPENSSL_init_crypto() explicitly using
1524 OPENSSL_INIT_ATFORK then this problem does not occur at all.
1527 *Matthias St. Pierre*
1529 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
1530 used even when parsing explicit parameters, when loading a encoded key
1531 or calling `EC_GROUP_new_from_ecpkparameters()`/
1532 `EC_GROUP_new_from_ecparameters()`.
1533 This prevents bypass of security hardening and performance gains,
1534 especially for curves with specialized EC_METHODs.
1535 By default, if a key encoded with explicit parameters is loaded and later
1536 encoded, the output is still encoded with explicit parameters, even if
1537 internally a "named" EC_GROUP is used for computation.
1541 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
1542 this change, EC_GROUP_set_generator would accept order and/or cofactor as
1543 NULL. After this change, only the cofactor parameter can be NULL. It also
1544 does some minimal sanity checks on the passed order.
1549 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
1550 An attack is simple, if the first CMS_recipientInfo is valid but the
1551 second CMS_recipientInfo is chosen ciphertext. If the second
1552 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
1553 encryption key will be replaced by garbage, and the message cannot be
1554 decoded, but if the RSA decryption fails, the correct encryption key is
1555 used and the recipient will not notice the attack.
1556 As a work around for this potential attack the length of the decrypted
1557 key must be equal to the cipher default key length, in case the
1558 certifiate is not given and all recipientInfo are tried out.
1559 The old behaviour can be re-enabled in the CMS code by setting the
1560 CMS_DEBUG_DECRYPT flag.
1565 * Early start up entropy quality from the DEVRANDOM seed source has been
1566 improved for older Linux systems. The RAND subsystem will wait for
1567 /dev/random to be producing output before seeding from /dev/urandom.
1568 The seeded state is stored for future library initialisations using
1569 a system global shared memory segment. The shared memory identifier
1570 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
1571 the desired value. The default identifier is 114.
1575 * Correct the extended master secret constant on EBCDIC systems. Without this
1576 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
1577 negotiate EMS will fail. Unfortunately this also means that TLS connections
1578 between EBCDIC systems with this fix, and EBCDIC systems without this
1579 fix will fail if they negotiate EMS.
1583 * Use Windows installation paths in the mingw builds
1585 Mingw isn't a POSIX environment per se, which means that Windows
1586 paths should be used for installation.
1591 * Changed DH_check to accept parameters with order q and 2q subgroups.
1592 With order 2q subgroups the bit 0 of the private key is not secret
1593 but DH_generate_key works around that by clearing bit 0 of the
1594 private key for those. This avoids leaking bit 0 of the private key.
1598 * Significantly reduce secure memory usage by the randomness pools.
1602 * Revert the DEVRANDOM_WAIT feature for Linux systems
1604 The DEVRANDOM_WAIT feature added a select() call to wait for the
1605 /dev/random device to become readable before reading from the
1606 /dev/urandom device.
1608 It turned out that this change had negative side effects on
1609 performance which were not acceptable. After some discussion it
1610 was decided to revert this feature and leave it up to the OS
1611 resp. the platform maintainer to ensure a proper initialization
1612 during early boot time.
1614 *Matthias St. Pierre*
1616 ### Changes between 1.1.1b and 1.1.1c [28 May 2019]
1618 * Add build tests for C++. These are generated files that only do one
1619 thing, to include one public OpenSSL head file each. This tests that
1620 the public header files can be usefully included in a C++ application.
1622 This test isn't enabled by default. It can be enabled with the option
1623 'enable-buildtest-c++'.
1627 * Enable SHA3 pre-hashing for ECDSA and DSA.
1631 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
1632 This changes the size when using the genpkey app when no size is given. It
1633 fixes an omission in earlier changes that changed all RSA, DSA and DH
1634 generation apps to use 2048 bits by default.
1638 * Reorganize the manual pages to consistently have RETURN VALUES,
1639 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
1640 util/fix-doc-nits accordingly.
1642 *Paul Yang, Joshua Lock*
1644 * Add the missing accessor EVP_PKEY_get0_engine()
1648 * Have apps like 's_client' and 's_server' output the signature scheme
1649 along with other cipher suite parameters when debugging.
1653 * Make OPENSSL_config() error agnostic again.
1657 * Do the error handling in RSA decryption constant time.
1661 * Prevent over long nonces in ChaCha20-Poly1305.
1663 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
1664 for every encryption operation. RFC 7539 specifies that the nonce value
1665 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
1666 and front pads the nonce with 0 bytes if it is less than 12
1667 bytes. However it also incorrectly allows a nonce to be set of up to 16
1668 bytes. In this case only the last 12 bytes are significant and any
1669 additional leading bytes are ignored.
1671 It is a requirement of using this cipher that nonce values are
1672 unique. Messages encrypted using a reused nonce value are susceptible to
1673 serious confidentiality and integrity attacks. If an application changes
1674 the default nonce length to be longer than 12 bytes and then makes a
1675 change to the leading bytes of the nonce expecting the new value to be a
1676 new unique nonce then such an application could inadvertently encrypt
1677 messages with a reused nonce.
1679 Additionally the ignored bytes in a long nonce are not covered by the
1680 integrity guarantee of this cipher. Any application that relies on the
1681 integrity of these ignored leading bytes of a long nonce may be further
1682 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
1683 is safe because no such use sets such a long nonce value. However user
1684 applications that use this cipher directly and set a non-default nonce
1685 length to be longer than 12 bytes may be vulnerable.
1687 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
1693 * Add DEVRANDOM_WAIT feature for Linux systems
1695 On older Linux systems where the getrandom() system call is not available,
1696 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
1697 Contrary to getrandom(), the /dev/urandom device will not block during
1698 early boot when the kernel CSPRNG has not been seeded yet.
1700 To mitigate this known weakness, use select() to wait for /dev/random to
1701 become readable before reading from /dev/urandom.
1703 * Ensure that SM2 only uses SM3 as digest algorithm
1707 ### Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
1709 * Change the info callback signals for the start and end of a post-handshake
1710 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
1711 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
1712 confused by this and assume that a TLSv1.2 renegotiation has started. This
1713 can break KeyUpdate handling. Instead we no longer signal the start and end
1714 of a post handshake message exchange (although the messages themselves are
1715 still signalled). This could break some applications that were expecting
1716 the old signals. However without this KeyUpdate is not usable for many
1721 ### Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
1723 * Timing vulnerability in DSA signature generation
1725 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
1726 timing side channel attack. An attacker could use variations in the signing
1727 algorithm to recover the private key.
1729 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
1734 * Timing vulnerability in ECDSA signature generation
1736 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
1737 timing side channel attack. An attacker could use variations in the signing
1738 algorithm to recover the private key.
1740 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
1745 * Fixed the issue that RAND_add()/RAND_seed() silently discards random input
1746 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
1747 of two gigabytes and the error handling improved.
1749 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
1750 categorized as a normal bug, not a security issue, because the DRBG reseeds
1751 automatically and is fully functional even without additional randomness
1752 provided by the application.
1754 ### Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
1756 * Add a new ClientHello callback. Provides a callback interface that gives
1757 the application the ability to adjust the nascent SSL object at the
1758 earliest stage of ClientHello processing, immediately after extensions have
1759 been collected but before they have been processed. In particular, this
1760 callback can adjust the supported TLS versions in response to the contents
1765 * Add SM2 base algorithm support.
1769 * s390x assembly pack: add (improved) hardware-support for the following
1770 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
1771 aes-cfb/cfb8, aes-ecb.
1775 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
1776 parameter is no longer accepted, as it leads to a corrupt table. NULL
1777 pem_str is reserved for alias entries only.
1781 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1782 step for prime curves. The new implementation is based on formulae from
1783 differential addition-and-doubling in homogeneous projective coordinates
1784 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
1785 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
1786 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
1787 to work in projective coordinates.
1789 *Billy Bob Brumley, Nicola Tuveri*
1791 * Change generating and checking of primes so that the error rate of not
1792 being prime depends on the intended use based on the size of the input.
1793 For larger primes this will result in more rounds of Miller-Rabin.
1794 The maximal error rate for primes with more than 1080 bits is lowered
1797 *Kurt Roeckx, Annie Yousar*
1799 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
1803 * The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
1804 moving between systems, and to avoid confusion when a Windows build is
1805 done with mingw vs with MSVC. For POSIX installs, there's still a
1806 symlink or copy named 'tsget' to avoid that confusion as well.
1810 * Revert blinding in ECDSA sign and instead make problematic addition
1811 length-invariant. Switch even to fixed-length Montgomery multiplication.
1815 * Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
1816 step for binary curves. The new implementation is based on formulae from
1817 differential addition-and-doubling in mixed Lopez-Dahab projective
1818 coordinates, modified to independently blind the operands.
1820 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1822 * Add a scaffold to optionally enhance the Montgomery ladder implementation
1823 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
1824 EC_METHODs to implement their own specialized "ladder step", to take
1825 advantage of more favorable coordinate systems or more efficient
1826 differential addition-and-doubling algorithms.
1828 *Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri*
1830 * Modified the random device based seed sources to keep the relevant
1831 file descriptors open rather than reopening them on each access.
1832 This allows such sources to operate in a chroot() jail without
1833 the associated device nodes being available. This behaviour can be
1834 controlled using RAND_keep_random_devices_open().
1838 * Numerous side-channel attack mitigations have been applied. This may have
1839 performance impacts for some algorithms for the benefit of improved
1840 security. Specific changes are noted in this change log by their respective
1845 * AIX shared library support overhaul. Switch to AIX "natural" way of
1846 handling shared libraries, which means collecting shared objects of
1847 different versions and bitnesses in one common archive. This allows to
1848 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
1849 doesn't affect the way 3rd party applications are linked, only how
1850 multi-version installation is managed.
1854 * Make ec_group_do_inverse_ord() more robust and available to other
1855 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
1856 mitigations are applied to the fallback BN_mod_inverse().
1857 When using this function rather than BN_mod_inverse() directly, new
1858 EC cryptosystem implementations are then safer-by-default.
1862 * Add coordinate blinding for EC_POINT and implement projective
1863 coordinate blinding for generic prime curves as a countermeasure to
1864 chosen point SCA attacks.
1866 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
1868 * Add blinding to ECDSA and DSA signatures to protect against side channel
1869 attacks discovered by Keegan Ryan (NCC Group).
1873 * Enforce checking in the pkeyutl command line app to ensure that the input
1874 length does not exceed the maximum supported digest length when performing
1875 a sign, verify or verifyrecover operation.
1879 * SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
1880 I/O in combination with something like select() or poll() will hang. This
1881 can be turned off again using SSL_CTX_clear_mode().
1882 Many applications do not properly handle non-application data records, and
1883 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
1884 around the problems in those applications, but can also break some.
1885 It's recommended to read the manpages about SSL_read(), SSL_write(),
1886 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
1887 SSL_CTX_set_read_ahead() again.
1891 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
1892 now allow empty (zero character) pass phrases.
1896 * Apply blinding to binary field modular inversion and remove patent
1897 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
1901 * Deprecate ec2_mult.c and unify scalar multiplication code paths for
1902 binary and prime elliptic curves.
1906 * Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
1907 constant time fixed point multiplication.
1911 * Revise elliptic curve scalar multiplication with timing attack
1912 defenses: ec_wNAF_mul redirects to a constant time implementation
1913 when computing fixed point and variable point multiplication (which
1914 in OpenSSL are mostly used with secret scalars in keygen, sign,
1915 ECDH derive operations).
1916 *Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
1919 * Updated CONTRIBUTING
1923 * Updated DRBG / RAND to request nonce and additional low entropy
1924 randomness from the system.
1926 *Matthias St. Pierre*
1928 * Updated 'openssl rehash' to use OpenSSL consistent default.
1932 * Moved the load of the ssl_conf module to libcrypto, which helps
1933 loading engines that libssl uses before libssl is initialised.
1937 * Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
1941 * Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
1943 *Ingo Schwarze, Rich Salz*
1945 * Added output of accepting IP address and port for 'openssl s_server'
1949 * Added a new API for TLSv1.3 ciphersuites:
1950 SSL_CTX_set_ciphersuites()
1951 SSL_set_ciphersuites()
1955 * Memory allocation failures consistently add an error to the error
1960 * Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
1961 in libcrypto when run as setuid/setgid.
1965 * Load any config file by default when libssl is used.
1969 * Added new public header file <openssl/rand_drbg.h> and documentation
1970 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
1972 *Matthias St. Pierre*
1974 * QNX support removed (cannot find contributors to get their approval
1975 for the license change).
1979 * TLSv1.3 replay protection for early data has been implemented. See the
1980 SSL_read_early_data() man page for further details.
1984 * Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
1985 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
1986 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
1987 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
1988 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
1989 configuration has been separated out. See the ciphers man page or the
1990 SSL_CTX_set_ciphersuites() man page for more information.
1994 * On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
1995 in responder mode now supports the new "-multi" option, which
1996 spawns the specified number of child processes to handle OCSP
1997 requests. The "-timeout" option now also limits the OCSP
1998 responder's patience to wait to receive the full client request
1999 on a newly accepted connection. Child processes are respawned
2000 as needed, and the CA index file is automatically reloaded
2001 when changed. This makes it possible to run the "ocsp" responder
2002 as a long-running service, making the OpenSSL CA somewhat more
2003 feature-complete. In this mode, most diagnostic messages logged
2004 after entering the event loop are logged via syslog(3) rather than
2009 * Added support for X448 and Ed448. Heavily based on original work by
2014 * Extend OSSL_STORE with capabilities to search and to narrow the set of
2015 objects loaded. This adds the functions OSSL_STORE_expect() and
2016 OSSL_STORE_find() as well as needed tools to construct searches and
2017 get the search data out of them.
2021 * Support for TLSv1.3 added. Note that users upgrading from an earlier
2022 version of OpenSSL should review their configuration settings to ensure
2023 that they are still appropriate for TLSv1.3. For further information see:
2024 <https://wiki.openssl.org/index.php/TLS1.3>
2028 * Grand redesign of the OpenSSL random generator
2030 The default RAND method now utilizes an AES-CTR DRBG according to
2031 NIST standard SP 800-90Ar1. The new random generator is essentially
2032 a port of the default random generator from the OpenSSL FIPS 2.0
2033 object module. It is a hybrid deterministic random bit generator
2034 using an AES-CTR bit stream and which seeds and reseeds itself
2035 automatically using trusted system entropy sources.
2037 Some of its new features are:
2038 - Support for multiple DRBG instances with seed chaining.
2039 - The default RAND method makes use of a DRBG.
2040 - There is a public and private DRBG instance.
2041 - The DRBG instances are fork-safe.
2042 - Keep all global DRBG instances on the secure heap if it is enabled.
2043 - The public and private DRBG instance are per thread for lock free
2046 *Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre*
2048 * Changed Configure so it only says what it does and doesn't dump
2049 so much data. Instead, ./configdata.pm should be used as a script
2050 to display all sorts of configuration data.
2054 * Added processing of "make variables" to Configure.
2058 * Added SHA512/224 and SHA512/256 algorithm support.
2062 * The last traces of Netware support, first removed in 1.1.0, have
2067 * Get rid of Makefile.shared, and in the process, make the processing
2068 of certain files (rc.obj, or the .def/.map/.opt files produced from
2069 the ordinal files) more visible and hopefully easier to trace and
2070 debug (or make silent).
2074 * Make it possible to have environment variable assignments as
2075 arguments to config / Configure.
2079 * Add multi-prime RSA (RFC 8017) support.
2083 * Add SM3 implemented according to GB/T 32905-2016
2084 *Jack Lloyd <jack.lloyd@ribose.com>,*
2085 *Ronald Tse <ronald.tse@ribose.com>,*
2086 *Erick Borsboom <erick.borsboom@ribose.com>*
2088 * Add 'Maximum Fragment Length' TLS extension negotiation and support
2089 as documented in RFC6066.
2090 Based on a patch from Tomasz Moń
2092 *Filipe Raimundo da Silva*
2094 * Add SM4 implemented according to GB/T 32907-2016.
2095 *Jack Lloyd <jack.lloyd@ribose.com>,*
2096 *Ronald Tse <ronald.tse@ribose.com>,*
2097 *Erick Borsboom <erick.borsboom@ribose.com>*
2099 * Reimplement -newreq-nodes and ERR_error_string_n; the
2100 original author does not agree with the license change.
2104 * Add ARIA AEAD TLS support.
2108 * Some macro definitions to support VS6 have been removed. Visual
2109 Studio 6 has not worked since 1.1.0
2113 * Add ERR_clear_last_mark(), to allow callers to clear the last mark
2114 without clearing the errors.
2118 * Add "atfork" functions. If building on a system that without
2119 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
2120 requirements. The RAND facility now uses/requires this.
2128 * The UI API becomes a permanent and integral part of libcrypto, i.e.
2129 not possible to disable entirely. However, it's still possible to
2130 disable the console reading UI method, UI_OpenSSL() (use UI_null()
2133 To disable, configure with 'no-ui-console'. 'no-ui' is still
2134 possible to use as an alias. Check at compile time with the
2135 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
2136 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
2140 * Add a STORE module, which implements a uniform and URI based reader of
2141 stores that can contain keys, certificates, CRLs and numerous other
2142 objects. The main API is loosely based on a few stdio functions,
2143 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
2144 OSSL_STORE_error and OSSL_STORE_close.
2145 The implementation uses backends called "loaders" to implement arbitrary
2146 URI schemes. There is one built in "loader" for the 'file' scheme.
2150 * Add devcrypto engine. This has been implemented against cryptodev-linux,
2151 then adjusted to work on FreeBSD 8.4 as well.
2152 Enable by configuring with 'enable-devcryptoeng'. This is done by default
2153 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
2157 * Module names can prefixed with OSSL_ or OPENSSL_. This affects
2158 util/mkerr.pl, which is adapted to allow those prefixes, leading to
2159 error code calls like this:
2161 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
2163 With this change, we claim the namespaces OSSL and OPENSSL in a manner
2164 that can be encoded in C. For the foreseeable future, this will only
2167 *Richard Levitte and Tim Hudson*
2169 * Removed BSD cryptodev engine.
2173 * Add a build target 'build_all_generated', to build all generated files
2174 and only that. This can be used to prepare everything that requires
2175 things like perl for a system that lacks perl and then move everything
2176 to that system and do the rest of the build there.
2180 * In the UI interface, make it possible to duplicate the user data. This
2181 can be used by engines that need to retain the data for a longer time
2182 than just the call where this user data is passed.
2186 * Ignore the '-named_curve auto' value for compatibility of applications
2189 *Tomas Mraz <tmraz@fedoraproject.org>*
2191 * Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
2192 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
2193 alerts across multiple records (some of which could be empty). In practice
2194 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
2195 prohibits this altogether and other libraries (BoringSSL, NSS) do not
2196 support this at all. Supporting it adds significant complexity to the
2197 record layer, and its removal is unlikely to cause interoperability
2202 * Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
2203 with Z. These are meant to replace LONG and ZLONG and to be size safe.
2204 The use of LONG and ZLONG is discouraged and scheduled for deprecation
2209 * Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
2210 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
2212 *Richard Levitte, Andy Polyakov*
2214 * Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
2219 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2220 platform rather than 'mingw'.
2224 * The functions X509_STORE_add_cert and X509_STORE_add_crl return
2225 success if they are asked to add an object which already exists
2226 in the store. This change cascades to other functions which load
2227 certificates and CRLs.
2231 * x86_64 assembly pack: annotate code with DWARF CFI directives to
2232 facilitate stack unwinding even from assembly subroutines.
2236 * Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
2237 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
2241 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2242 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2243 which is the minimum version we support.
2247 * Certificate time validation (X509_cmp_time) enforces stricter
2248 compliance with RFC 5280. Fractional seconds and timezone offsets
2249 are no longer allowed.
2253 * Add support for ARIA
2257 * s_client will now send the Server Name Indication (SNI) extension by
2258 default unless the new "-noservername" option is used. The server name is
2259 based on the host provided to the "-connect" option unless overridden by
2260 using "-servername".
2264 * Add support for SipHash
2268 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
2269 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
2270 prevent issues where no progress is being made and the peer continually
2271 sends unrecognised record types, using up resources processing them.
2275 * 'openssl passwd' can now produce SHA256 and SHA512 based output,
2276 using the algorithm defined in
2277 <https://www.akkadia.org/drepper/SHA-crypt.txt>
2281 * Heartbeat support has been removed; the ABI is changed for now.
2283 *Richard Levitte, Rich Salz*
2285 * Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
2289 * The RSA "null" method, which was partially supported to avoid patent
2290 issues, has been replaced to always returns NULL.
2297 ### Changes between 1.1.0k and 1.1.0l [10 Sep 2019]
2299 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
2300 used even when parsing explicit parameters, when loading a encoded key
2301 or calling `EC_GROUP_new_from_ecpkparameters()`/
2302 `EC_GROUP_new_from_ecparameters()`.
2303 This prevents bypass of security hardening and performance gains,
2304 especially for curves with specialized EC_METHODs.
2305 By default, if a key encoded with explicit parameters is loaded and later
2306 encoded, the output is still encoded with explicit parameters, even if
2307 internally a "named" EC_GROUP is used for computation.
2311 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
2312 this change, EC_GROUP_set_generator would accept order and/or cofactor as
2313 NULL. After this change, only the cofactor parameter can be NULL. It also
2314 does some minimal sanity checks on the passed order.
2319 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
2320 An attack is simple, if the first CMS_recipientInfo is valid but the
2321 second CMS_recipientInfo is chosen ciphertext. If the second
2322 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
2323 encryption key will be replaced by garbage, and the message cannot be
2324 decoded, but if the RSA decryption fails, the correct encryption key is
2325 used and the recipient will not notice the attack.
2326 As a work around for this potential attack the length of the decrypted
2327 key must be equal to the cipher default key length, in case the
2328 certifiate is not given and all recipientInfo are tried out.
2329 The old behaviour can be re-enabled in the CMS code by setting the
2330 CMS_DEBUG_DECRYPT flag.
2335 * Use Windows installation paths in the mingw builds
2337 Mingw isn't a POSIX environment per se, which means that Windows
2338 paths should be used for installation.
2343 ### Changes between 1.1.0j and 1.1.0k [28 May 2019]
2345 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
2346 This changes the size when using the genpkey app when no size is given. It
2347 fixes an omission in earlier changes that changed all RSA, DSA and DH
2348 generation apps to use 2048 bits by default.
2352 * Prevent over long nonces in ChaCha20-Poly1305.
2354 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
2355 for every encryption operation. RFC 7539 specifies that the nonce value
2356 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
2357 and front pads the nonce with 0 bytes if it is less than 12
2358 bytes. However it also incorrectly allows a nonce to be set of up to 16
2359 bytes. In this case only the last 12 bytes are significant and any
2360 additional leading bytes are ignored.
2362 It is a requirement of using this cipher that nonce values are
2363 unique. Messages encrypted using a reused nonce value are susceptible to
2364 serious confidentiality and integrity attacks. If an application changes
2365 the default nonce length to be longer than 12 bytes and then makes a
2366 change to the leading bytes of the nonce expecting the new value to be a
2367 new unique nonce then such an application could inadvertently encrypt
2368 messages with a reused nonce.
2370 Additionally the ignored bytes in a long nonce are not covered by the
2371 integrity guarantee of this cipher. Any application that relies on the
2372 integrity of these ignored leading bytes of a long nonce may be further
2373 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
2374 is safe because no such use sets such a long nonce value. However user
2375 applications that use this cipher directly and set a non-default nonce
2376 length to be longer than 12 bytes may be vulnerable.
2378 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
2384 * Added SCA hardening for modular field inversion in EC_GROUP through
2385 a new dedicated field_inv() pointer in EC_METHOD.
2386 This also addresses a leakage affecting conversions from projective
2387 to affine coordinates.
2389 *Billy Bob Brumley, Nicola Tuveri*
2391 * Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
2392 re-used X509_PUBKEY object if the second PUBKEY is malformed.
2396 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
2400 * Remove the 'dist' target and add a tarball building script. The
2401 'dist' target has fallen out of use, and it shouldn't be
2402 necessary to configure just to create a source distribution.
2406 ### Changes between 1.1.0i and 1.1.0j [20 Nov 2018]
2408 * Timing vulnerability in DSA signature generation
2410 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
2411 timing side channel attack. An attacker could use variations in the signing
2412 algorithm to recover the private key.
2414 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
2419 * Timing vulnerability in ECDSA signature generation
2421 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
2422 timing side channel attack. An attacker could use variations in the signing
2423 algorithm to recover the private key.
2425 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
2430 * Add coordinate blinding for EC_POINT and implement projective
2431 coordinate blinding for generic prime curves as a countermeasure to
2432 chosen point SCA attacks.
2434 *Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley*
2436 ### Changes between 1.1.0h and 1.1.0i [14 Aug 2018]
2438 * Client DoS due to large DH parameter
2440 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
2441 malicious server can send a very large prime value to the client. This will
2442 cause the client to spend an unreasonably long period of time generating a
2443 key for this prime resulting in a hang until the client has finished. This
2444 could be exploited in a Denial Of Service attack.
2446 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
2451 * Cache timing vulnerability in RSA Key Generation
2453 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
2454 a cache timing side channel attack. An attacker with sufficient access to
2455 mount cache timing attacks during the RSA key generation process could
2456 recover the private key.
2458 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
2459 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
2464 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
2465 parameter is no longer accepted, as it leads to a corrupt table. NULL
2466 pem_str is reserved for alias entries only.
2470 * Revert blinding in ECDSA sign and instead make problematic addition
2471 length-invariant. Switch even to fixed-length Montgomery multiplication.
2475 * Change generating and checking of primes so that the error rate of not
2476 being prime depends on the intended use based on the size of the input.
2477 For larger primes this will result in more rounds of Miller-Rabin.
2478 The maximal error rate for primes with more than 1080 bits is lowered
2481 *Kurt Roeckx, Annie Yousar*
2483 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
2487 * Add blinding to ECDSA and DSA signatures to protect against side channel
2488 attacks discovered by Keegan Ryan (NCC Group).
2492 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
2493 now allow empty (zero character) pass phrases.
2497 * Certificate time validation (X509_cmp_time) enforces stricter
2498 compliance with RFC 5280. Fractional seconds and timezone offsets
2499 are no longer allowed.
2503 * Fixed a text canonicalisation bug in CMS
2505 Where a CMS detached signature is used with text content the text goes
2506 through a canonicalisation process first prior to signing or verifying a
2507 signature. This process strips trailing space at the end of lines, converts
2508 line terminators to CRLF and removes additional trailing line terminators
2509 at the end of a file. A bug in the canonicalisation process meant that
2510 some characters, such as form-feed, were incorrectly treated as whitespace
2511 and removed. This is contrary to the specification (RFC5485). This fix
2512 could mean that detached text data signed with an earlier version of
2513 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
2514 signed with a fixed OpenSSL may fail to verify with an earlier version of
2515 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
2516 and use the "-binary" flag (for the "cms" command line application) or set
2517 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
2521 ### Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
2523 * Constructed ASN.1 types with a recursive definition could exceed the stack
2525 Constructed ASN.1 types with a recursive definition (such as can be found
2526 in PKCS7) could eventually exceed the stack given malicious input with
2527 excessive recursion. This could result in a Denial Of Service attack. There
2528 are no such structures used within SSL/TLS that come from untrusted sources
2529 so this is considered safe.
2531 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
2537 * Incorrect CRYPTO_memcmp on HP-UX PA-RISC
2539 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
2540 effectively reduced to only comparing the least significant bit of each
2541 byte. This allows an attacker to forge messages that would be considered as
2542 authenticated in an amount of tries lower than that guaranteed by the
2543 security claims of the scheme. The module can only be compiled by the
2544 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
2546 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
2552 * Add a build target 'build_all_generated', to build all generated files
2553 and only that. This can be used to prepare everything that requires
2554 things like perl for a system that lacks perl and then move everything
2555 to that system and do the rest of the build there.
2559 * Backport SSL_OP_NO_RENGOTIATION
2561 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
2562 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
2563 changes this is no longer possible in 1.1.0. Therefore the new
2564 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
2565 1.1.0 to provide equivalent functionality.
2567 Note that if an application built against 1.1.0h headers (or above) is run
2568 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
2569 accepted but nothing will happen, i.e. renegotiation will not be prevented.
2573 * Removed the OS390-Unix config target. It relied on a script that doesn't
2578 * rsaz_1024_mul_avx2 overflow bug on x86_64
2580 There is an overflow bug in the AVX2 Montgomery multiplication procedure
2581 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
2582 Analysis suggests that attacks against RSA and DSA as a result of this
2583 defect would be very difficult to perform and are not believed likely.
2584 Attacks against DH1024 are considered just feasible, because most of the
2585 work necessary to deduce information about a private key may be performed
2586 offline. The amount of resources required for such an attack would be
2587 significant. However, for an attack on TLS to be meaningful, the server
2588 would have to share the DH1024 private key among multiple clients, which is
2589 no longer an option since CVE-2016-0701.
2591 This only affects processors that support the AVX2 but not ADX extensions
2592 like Intel Haswell (4th generation).
2594 This issue was reported to OpenSSL by David Benjamin (Google). The issue
2595 was originally found via the OSS-Fuzz project.
2600 ### Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
2602 * bn_sqrx8x_internal carry bug on x86_64
2604 There is a carry propagating bug in the x86_64 Montgomery squaring
2605 procedure. No EC algorithms are affected. Analysis suggests that attacks
2606 against RSA and DSA as a result of this defect would be very difficult to
2607 perform and are not believed likely. Attacks against DH are considered just
2608 feasible (although very difficult) because most of the work necessary to
2609 deduce information about a private key may be performed offline. The amount
2610 of resources required for such an attack would be very significant and
2611 likely only accessible to a limited number of attackers. An attacker would
2612 additionally need online access to an unpatched system using the target
2613 private key in a scenario with persistent DH parameters and a private
2614 key that is shared between multiple clients.
2616 This only affects processors that support the BMI1, BMI2 and ADX extensions
2617 like Intel Broadwell (5th generation) and later or AMD Ryzen.
2619 This issue was reported to OpenSSL by the OSS-Fuzz project.
2624 * Malformed X.509 IPAddressFamily could cause OOB read
2626 If an X.509 certificate has a malformed IPAddressFamily extension,
2627 OpenSSL could do a one-byte buffer overread. The most likely result
2628 would be an erroneous display of the certificate in text format.
2630 This issue was reported to OpenSSL by the OSS-Fuzz project.
2635 ### Changes between 1.1.0e and 1.1.0f [25 May 2017]
2637 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
2638 platform rather than 'mingw'.
2642 * Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
2643 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
2644 which is the minimum version we support.
2648 ### Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
2650 * Encrypt-Then-Mac renegotiation crash
2652 During a renegotiation handshake if the Encrypt-Then-Mac extension is
2653 negotiated where it was not in the original handshake (or vice-versa) then
2654 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
2655 and servers are affected.
2657 This issue was reported to OpenSSL by Joe Orton (Red Hat).
2662 ### Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
2664 * Truncated packet could crash via OOB read
2666 If one side of an SSL/TLS path is running on a 32-bit host and a specific
2667 cipher is being used, then a truncated packet can cause that host to
2668 perform an out-of-bounds read, usually resulting in a crash.
2670 This issue was reported to OpenSSL by Robert Święcki of Google.
2675 * Bad (EC)DHE parameters cause a client crash
2677 If a malicious server supplies bad parameters for a DHE or ECDHE key
2678 exchange then this can result in the client attempting to dereference a
2679 NULL pointer leading to a client crash. This could be exploited in a Denial
2682 This issue was reported to OpenSSL by Guido Vranken.
2687 * BN_mod_exp may produce incorrect results on x86_64
2689 There is a carry propagating bug in the x86_64 Montgomery squaring
2690 procedure. No EC algorithms are affected. Analysis suggests that attacks
2691 against RSA and DSA as a result of this defect would be very difficult to
2692 perform and are not believed likely. Attacks against DH are considered just
2693 feasible (although very difficult) because most of the work necessary to
2694 deduce information about a private key may be performed offline. The amount
2695 of resources required for such an attack would be very significant and
2696 likely only accessible to a limited number of attackers. An attacker would
2697 additionally need online access to an unpatched system using the target
2698 private key in a scenario with persistent DH parameters and a private
2699 key that is shared between multiple clients. For example this can occur by
2700 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
2701 similar to CVE-2015-3193 but must be treated as a separate problem.
2703 This issue was reported to OpenSSL by the OSS-Fuzz project.
2708 ### Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
2710 * ChaCha20/Poly1305 heap-buffer-overflow
2712 TLS connections using `*-CHACHA20-POLY1305` ciphersuites are susceptible to
2713 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
2714 crash. This issue is not considered to be exploitable beyond a DoS.
2716 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
2721 * CMS Null dereference
2723 Applications parsing invalid CMS structures can crash with a NULL pointer
2724 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
2725 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
2726 structure callback if an attempt is made to free certain invalid encodings.
2727 Only CHOICE structures using a callback which do not handle NULL value are
2730 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
2735 * Montgomery multiplication may produce incorrect results
2737 There is a carry propagating bug in the Broadwell-specific Montgomery
2738 multiplication procedure that handles input lengths divisible by, but
2739 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
2740 and DH private keys are impossible. This is because the subroutine in
2741 question is not used in operations with the private key itself and an input
2742 of the attacker's direct choice. Otherwise the bug can manifest itself as
2743 transient authentication and key negotiation failures or reproducible
2744 erroneous outcome of public-key operations with specially crafted input.
2745 Among EC algorithms only Brainpool P-512 curves are affected and one
2746 presumably can attack ECDH key negotiation. Impact was not analyzed in
2747 detail, because pre-requisites for attack are considered unlikely. Namely
2748 multiple clients have to choose the curve in question and the server has to
2749 share the private key among them, neither of which is default behaviour.
2750 Even then only clients that chose the curve will be affected.
2752 This issue was publicly reported as transient failures and was not
2753 initially recognized as a security issue. Thanks to Richard Morgan for
2754 providing reproducible case.
2759 * Removed automatic addition of RPATH in shared libraries and executables,
2760 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
2764 ### Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
2766 * Fix Use After Free for large message sizes
2768 The patch applied to address CVE-2016-6307 resulted in an issue where if a
2769 message larger than approx 16k is received then the underlying buffer to
2770 store the incoming message is reallocated and moved. Unfortunately a
2771 dangling pointer to the old location is left which results in an attempt to
2772 write to the previously freed location. This is likely to result in a
2773 crash, however it could potentially lead to execution of arbitrary code.
2775 This issue only affects OpenSSL 1.1.0a.
2777 This issue was reported to OpenSSL by Robert Święcki.
2782 ### Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
2784 * OCSP Status Request extension unbounded memory growth
2786 A malicious client can send an excessively large OCSP Status Request
2787 extension. If that client continually requests renegotiation, sending a
2788 large OCSP Status Request extension each time, then there will be unbounded
2789 memory growth on the server. This will eventually lead to a Denial Of
2790 Service attack through memory exhaustion. Servers with a default
2791 configuration are vulnerable even if they do not support OCSP. Builds using
2792 the "no-ocsp" build time option are not affected.
2794 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2799 * SSL_peek() hang on empty record
2801 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
2802 sends an empty record. This could be exploited by a malicious peer in a
2803 Denial Of Service attack.
2805 This issue was reported to OpenSSL by Alex Gaynor.
2810 * Excessive allocation of memory in tls_get_message_header() and
2811 dtls1_preprocess_fragment()
2813 A (D)TLS message includes 3 bytes for its length in the header for the
2814 message. This would allow for messages up to 16Mb in length. Messages of
2815 this length are excessive and OpenSSL includes a check to ensure that a
2816 peer is sending reasonably sized messages in order to avoid too much memory
2817 being consumed to service a connection. A flaw in the logic of version
2818 1.1.0 means that memory for the message is allocated too early, prior to
2819 the excessive message length check. Due to way memory is allocated in
2820 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
2821 to service a connection. This could lead to a Denial of Service through
2822 memory exhaustion. However, the excessive message length check still takes
2823 place, and this would cause the connection to immediately fail. Assuming
2824 that the application calls SSL_free() on the failed connection in a timely
2825 manner then the 21Mb of allocated memory will then be immediately freed
2826 again. Therefore the excessive memory allocation will be transitory in
2827 nature. This then means that there is only a security impact if:
2829 1) The application does not call SSL_free() in a timely manner in the event
2830 that the connection fails
2832 2) The application is working in a constrained environment where there is
2833 very little free memory
2835 3) The attacker initiates multiple connection attempts such that there are
2836 multiple connections in a state where memory has been allocated for the
2837 connection; SSL_free() has not yet been called; and there is insufficient
2838 memory to service the multiple requests.
2840 Except in the instance of (1) above any Denial Of Service is likely to be
2841 transitory because as soon as the connection fails the memory is
2842 subsequently freed again in the SSL_free() call. However there is an
2843 increased risk during this period of application crashes due to the lack of
2844 memory - which would then mean a more serious Denial of Service.
2846 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
2847 (CVE-2016-6307 and CVE-2016-6308)
2851 * solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
2852 had to be removed. Primary reason is that vendor assembler can't
2853 assemble our modules with -KPIC flag. As result it, assembly
2854 support, was not even available as option. But its lack means
2855 lack of side-channel resistant code, which is incompatible with
2856 security by todays standards. Fortunately gcc is readily available
2857 prepackaged option, which we firmly point at...
2861 ### Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
2863 * Windows command-line tool supports UTF-8 opt-in option for arguments
2864 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
2865 (to any value) allows Windows user to access PKCS#12 file generated
2866 with Windows CryptoAPI and protected with non-ASCII password, as well
2867 as files generated under UTF-8 locale on Linux also protected with
2872 * To mitigate the SWEET32 attack ([CVE-2016-2183]), 3DES cipher suites
2873 have been disabled by default and removed from DEFAULT, just like RC4.
2874 See the RC4 item below to re-enable both.
2878 * The method for finding the storage location for the Windows RAND seed file
2879 has changed. First we check %RANDFILE%. If that is not set then we check
2880 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
2881 all else fails we fall back to C:\.
2885 * The EVP_EncryptUpdate() function has had its return type changed from void
2886 to int. A return of 0 indicates and error while a return of 1 indicates
2891 * The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
2892 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
2893 off the constant time implementation for RSA, DSA and DH have been made
2894 no-ops and deprecated.
2898 * Windows RAND implementation was simplified to only get entropy by
2899 calling CryptGenRandom(). Various other RAND-related tickets
2902 *Joseph Wylie Yandle, Rich Salz*
2904 * The stack and lhash API's were renamed to start with `OPENSSL_SK_`
2905 and `OPENSSL_LH_`, respectively. The old names are available
2906 with API compatibility. They new names are now completely documented.
2910 * Unify TYPE_up_ref(obj) methods signature.
2911 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
2912 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
2913 int (instead of void) like all others TYPE_up_ref() methods.
2914 So now these methods also check the return value of CRYPTO_atomic_add(),
2915 and the validity of object reference counter.
2917 *fdasilvayy@gmail.com*
2919 * With Windows Visual Studio builds, the .pdb files are installed
2920 alongside the installed libraries and executables. For a static
2921 library installation, ossl_static.pdb is the associate compiler
2922 generated .pdb file to be used when linking programs.
2926 * Remove openssl.spec. Packaging files belong with the packagers.
2930 * Automatic Darwin/OSX configuration has had a refresh, it will now
2931 recognise x86_64 architectures automatically. You can still decide
2932 to build for a different bitness with the environment variable
2933 KERNEL_BITS (can be 32 or 64), for example:
2935 KERNEL_BITS=32 ./config
2939 * Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
2940 256 bit AES and HMAC with SHA256.
2944 * Remove support for MIPS o32 ABI on IRIX (and IRIX only).
2948 * Triple-DES ciphers have been moved from HIGH to MEDIUM.
2952 * To enable users to have their own config files and build file templates,
2953 Configure looks in the directory indicated by the environment variable
2954 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
2955 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
2956 name and is used as is.
2960 * The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
2961 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
2962 X509_CERT_FILE_CTX was removed.
2966 * "shared" builds are now the default. To create only static libraries use
2967 the "no-shared" Configure option.
2971 * Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
2972 All of these option have not worked for some while and are fundamental
2977 * Make various cleanup routines no-ops and mark them as deprecated. Most
2978 global cleanup functions are no longer required because they are handled
2979 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
2980 Explicitly de-initing can cause problems (e.g. where a library that uses
2981 OpenSSL de-inits, but an application is still using it). The affected
2982 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
2983 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
2984 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
2985 COMP_zlib_cleanup().
2989 * --strict-warnings no longer enables runtime debugging options
2990 such as REF_DEBUG. Instead, debug options are automatically
2991 enabled with '--debug' builds.
2993 *Andy Polyakov, Emilia Käsper*
2995 * Made DH and DH_METHOD opaque. The structures for managing DH objects
2996 have been moved out of the public header files. New functions for managing
2997 these have been added.
3001 * Made RSA and RSA_METHOD opaque. The structures for managing RSA
3002 objects have been moved out of the public header files. New
3003 functions for managing these have been added.
3007 * Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
3008 have been moved out of the public header files. New functions for managing
3009 these have been added.
3013 * Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
3014 moved out of the public header files. New functions for managing these
3019 * Removed no-rijndael as a config option. Rijndael is an old name for AES.
3023 * Removed the mk1mf build scripts.
3027 * Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
3028 it is always safe to #include a header now.
3032 * Removed the aged BC-32 config and all its supporting scripts
3036 * Removed support for Ultrix, Netware, and OS/2.
3040 * Add support for HKDF.
3042 *Alessandro Ghedini*
3044 * Add support for blake2b and blake2s
3048 * Added support for "pipelining". Ciphers that have the
3049 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
3050 encryptions/decryptions simultaneously. There are currently no built-in
3051 ciphers with this property but the expectation is that engines will be able
3052 to offer it to significantly improve throughput. Support has been extended
3053 into libssl so that multiple records for a single connection can be
3054 processed in one go (for >=TLS 1.1).
3058 * Added the AFALG engine. This is an async capable engine which is able to
3059 offload work to the Linux kernel. In this initial version it only supports
3060 AES128-CBC. The kernel must be version 4.1.0 or greater.
3064 * OpenSSL now uses a new threading API. It is no longer necessary to
3065 set locking callbacks to use OpenSSL in a multi-threaded environment. There
3066 are two supported threading models: pthreads and windows threads. It is
3067 also possible to configure OpenSSL at compile time for "no-threads". The
3068 old threading API should no longer be used. The functions have been
3069 replaced with "no-op" compatibility macros.
3071 *Alessandro Ghedini, Matt Caswell*
3073 * Modify behavior of ALPN to invoke callback after SNI/servername
3074 callback, such that updates to the SSL_CTX affect ALPN.
3078 * Add SSL_CIPHER queries for authentication and key-exchange.
3082 * Changes to the DEFAULT cipherlist:
3083 - Prefer (EC)DHE handshakes over plain RSA.
3084 - Prefer AEAD ciphers over legacy ciphers.
3085 - Prefer ECDSA over RSA when both certificates are available.
3086 - Prefer TLSv1.2 ciphers/PRF.
3087 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
3092 * Change the ECC default curve list to be this, in order: x25519,
3093 secp256r1, secp521r1, secp384r1.
3097 * RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
3098 disabled by default. They can be re-enabled using the
3099 enable-weak-ssl-ciphers option to Configure.
3103 * If the server has ALPN configured, but supports no protocols that the
3104 client advertises, send a fatal "no_application_protocol" alert.
3105 This behaviour is SHALL in RFC 7301, though it isn't universally
3106 implemented by other servers.
3110 * Add X25519 support.
3111 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
3112 for public and private key encoding using the format documented in
3113 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
3114 key generation and key derivation.
3116 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
3121 * Deprecate SRP_VBASE_get_by_user.
3122 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
3123 In order to fix an unavoidable memory leak ([CVE-2016-0798]),
3124 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
3125 seed, even if the seed is configured.
3127 Users should use SRP_VBASE_get1_by_user instead. Note that in
3128 SRP_VBASE_get1_by_user, caller must free the returned value. Note
3129 also that even though configuring the SRP seed attempts to hide
3130 invalid usernames by continuing the handshake with fake
3131 credentials, this behaviour is not constant time and no strong
3132 guarantees are made that the handshake is indistinguishable from
3133 that of a valid user.
3137 * Configuration change; it's now possible to build dynamic engines
3138 without having to build shared libraries and vice versa. This
3139 only applies to the engines in engines/, those in crypto/engine/
3140 will always be built into libcrypto (i.e. "static").
3142 Building dynamic engines is enabled by default; to disable, use
3143 the configuration option "disable-dynamic-engine".
3145 The only requirements for building dynamic engines are the
3146 presence of the DSO module and building with position independent
3147 code, so they will also automatically be disabled if configuring
3148 with "disable-dso" or "disable-pic".
3150 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
3151 are also taken away from openssl/opensslconf.h, as they are
3156 * Configuration change; if there is a known flag to compile
3157 position independent code, it will always be applied on the
3158 libcrypto and libssl object files, and never on the application
3159 object files. This means other libraries that use routines from
3160 libcrypto / libssl can be made into shared libraries regardless
3161 of how OpenSSL was configured.
3163 If this isn't desirable, the configuration options "disable-pic"
3164 or "no-pic" can be used to disable the use of PIC. This will
3165 also disable building shared libraries and dynamic engines.
3169 * Removed JPAKE code. It was experimental and has no wide use.
3173 * The INSTALL_PREFIX Makefile variable has been renamed to
3174 DESTDIR. That makes for less confusion on what this variable
3175 is for. Also, the configuration option --install_prefix is
3180 * Heartbeat for TLS has been removed and is disabled by default
3181 for DTLS; configure with enable-heartbeats. Code that uses the
3182 old #define's might need to be updated.
3184 *Emilia Käsper, Rich Salz*
3186 * Rename REF_CHECK to REF_DEBUG.
3190 * New "unified" build system
3192 The "unified" build system is aimed to be a common system for all
3193 platforms we support. With it comes new support for VMS.
3195 This system builds supports building in a different directory tree
3196 than the source tree. It produces one Makefile (for unix family
3197 or lookalikes), or one descrip.mms (for VMS).
3199 The source of information to make the Makefile / descrip.mms is
3200 small files called 'build.info', holding the necessary
3201 information for each directory with source to compile, and a
3202 template in Configurations, like unix-Makefile.tmpl or
3205 With this change, the library names were also renamed on Windows
3206 and on VMS. They now have names that are closer to the standard
3207 on Unix, and include the major version number, and in certain
3208 cases, the architecture they are built for. See "Notes on shared
3209 libraries" in INSTALL.
3211 We rely heavily on the perl module Text::Template.
3215 * Added support for auto-initialisation and de-initialisation of the library.
3216 OpenSSL no longer requires explicit init or deinit routines to be called,
3217 except in certain circumstances. See the OPENSSL_init_crypto() and
3218 OPENSSL_init_ssl() man pages for further information.
3222 * The arguments to the DTLSv1_listen function have changed. Specifically the
3223 "peer" argument is now expected to be a BIO_ADDR object.
3225 * Rewrite of BIO networking library. The BIO library lacked consistent
3226 support of IPv6, and adding it required some more extensive
3227 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
3228 which hold all types of addresses and chains of address information.
3229 It also introduces a new API, with functions like BIO_socket,
3230 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
3231 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
3232 have been adapted accordingly.
3236 * RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
3241 * CRIME protection: disable compression by default, even if OpenSSL is
3242 compiled with zlib enabled. Applications can still enable compression
3243 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
3244 using the SSL_CONF library to configure compression.
3248 * The signature of the session callback configured with
3249 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
3250 was explicitly marked as `const unsigned char*` instead of
3255 * Always DPURIFY. Remove the use of uninitialized memory in the
3256 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
3260 * Removed many obsolete configuration items, including
3261 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
3262 MD2_CHAR, MD2_INT, MD2_LONG
3264 IDEA_SHORT, IDEA_LONG
3265 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
3267 *Rich Salz, with advice from Andy Polyakov*
3269 * Many BN internals have been moved to an internal header file.
3271 *Rich Salz with help from Andy Polyakov*
3273 * Configuration and writing out the results from it has changed.
3274 Files such as Makefile include/openssl/opensslconf.h and are now
3275 produced through general templates, such as Makefile.in and
3276 crypto/opensslconf.h.in and some help from the perl module
3279 Also, the center of configuration information is no longer
3280 Makefile. Instead, Configure produces a perl module in
3281 configdata.pm which holds most of the config data (in the hash
3282 table %config), the target data that comes from the target
3283 configuration in one of the `Configurations/*.conf` files (in
3288 * To clarify their intended purposes, the Configure options
3289 --prefix and --openssldir change their semantics, and become more
3290 straightforward and less interdependent.
3292 --prefix shall be used exclusively to give the location INSTALLTOP
3293 where programs, scripts, libraries, include files and manuals are
3294 going to be installed. The default is now /usr/local.
3296 --openssldir shall be used exclusively to give the default
3297 location OPENSSLDIR where certificates, private keys, CRLs are
3298 managed. This is also where the default openssl.cnf gets
3300 If the directory given with this option is a relative path, the
3301 values of both the --prefix value and the --openssldir value will
3302 be combined to become OPENSSLDIR.
3303 The default for --openssldir is INSTALLTOP/ssl.
3305 Anyone who uses --openssldir to specify where OpenSSL is to be
3306 installed MUST change to use --prefix instead.
3310 * The GOST engine was out of date and therefore it has been removed. An up
3311 to date GOST engine is now being maintained in an external repository.
3312 See: <https://wiki.openssl.org/index.php/Binaries>. Libssl still retains
3313 support for GOST ciphersuites (these are only activated if a GOST engine
3318 * EGD is no longer supported by default; use enable-egd when
3321 *Ben Kaduk and Rich Salz*
3323 * The distribution now has Makefile.in files, which are used to
3324 create Makefile's when Configure is run. *Configure must be run
3325 before trying to build now.*
3329 * The return value for SSL_CIPHER_description() for error conditions
3334 * Support for RFC6698/RFC7671 DANE TLSA peer authentication.
3336 Obtaining and performing DNSSEC validation of TLSA records is
3337 the application's responsibility. The application provides
3338 the TLSA records of its choice to OpenSSL, and these are then
3339 used to authenticate the peer.
3341 The TLSA records need not even come from DNS. They can, for
3342 example, be used to implement local end-entity certificate or
3343 trust-anchor "pinning", where the "pin" data takes the form
3344 of TLSA records, which can augment or replace verification
3345 based on the usual WebPKI public certification authorities.
3349 * Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
3350 continues to support deprecated interfaces in default builds.
3351 However, applications are strongly advised to compile their
3352 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
3353 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
3354 or the 1.1.0 releases.
3356 In environments in which all applications have been ported to
3357 not use any deprecated interfaces OpenSSL's Configure script
3358 should be used with the --api=1.1.0 option to entirely remove
3359 support for the deprecated features from the library and
3360 unconditionally disable them in the installed headers.
3361 Essentially the same effect can be achieved with the "no-deprecated"
3362 argument to Configure, except that this will always restrict
3363 the build to just the latest API, rather than a fixed API
3366 As applications are ported to future revisions of the API,
3367 they should update their compile-time OPENSSL_API_COMPAT define
3368 accordingly, but in most cases should be able to continue to
3369 compile with later releases.
3371 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
3372 0x10000000L and 0x00908000L, respectively. However those
3373 versions did not support the OPENSSL_API_COMPAT feature, and
3374 so applications are not typically tested for explicit support
3375 of just the undeprecated features of either release.
3379 * Add support for setting the minimum and maximum supported protocol.
3380 It can bet set via the SSL_set_min_proto_version() and
3381 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
3382 MaxProtocol. It's recommended to use the new APIs to disable
3383 protocols instead of disabling individual protocols using
3384 SSL_set_options() or SSL_CONF's Protocol. This change also
3385 removes support for disabling TLS 1.2 in the OpenSSL TLS
3386 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
3390 * Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
3394 * New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
3395 and integrates ECDSA and ECDH functionality into EC. Implementations can
3396 now redirect key generation and no longer need to convert to or from
3399 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
3400 include the ec.h header file instead.
3404 * Remove support for all 40 and 56 bit ciphers. This includes all the export
3405 ciphers who are no longer supported and drops support the ephemeral RSA key
3406 exchange. The LOW ciphers currently doesn't have any ciphers in it.
3410 * Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
3411 opaque. For HMAC_CTX, the following constructors and destructors
3414 HMAC_CTX *HMAC_CTX_new(void);
3415 void HMAC_CTX_free(HMAC_CTX *ctx);
3417 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
3418 destroy such methods has been added. See EVP_MD_meth_new(3) and
3419 EVP_CIPHER_meth_new(3) for documentation.
3422 1) `EVP_MD_CTX_cleanup()`, `EVP_CIPHER_CTX_cleanup()` and
3423 `HMAC_CTX_cleanup()` were removed. `HMAC_CTX_reset()` and
3424 `EVP_MD_CTX_reset()` should be called instead to reinitialise
3425 an already created structure.
3426 2) For consistency with the majority of our object creators and
3427 destructors, `EVP_MD_CTX_(create|destroy)` were renamed to
3428 `EVP_MD_CTX_(new|free)`. The old names are retained as macros
3429 for deprecated builds.
3433 * Added ASYNC support. Libcrypto now includes the async sub-library to enable
3434 cryptographic operations to be performed asynchronously as long as an
3435 asynchronous capable engine is used. See the ASYNC_start_job() man page for
3436 further details. Libssl has also had this capability integrated with the
3437 introduction of the new mode SSL_MODE_ASYNC and associated error
3438 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
3439 pages. This work was developed in partnership with Intel Corp.
3443 * SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
3444 always enabled now. If you want to disable the support you should
3445 exclude it using the list of supported ciphers. This also means that the
3446 "-no_ecdhe" option has been removed from s_server.
3450 * SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
3451 SSL_{CTX_}set1_curves() which can set a list.
3455 * Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
3456 curve you want to support using SSL_{CTX_}set1_curves().
3460 * State machine rewrite. The state machine code has been significantly
3461 refactored in order to remove much duplication of code and solve issues
3462 with the old code (see [ssl/statem/README.md](ssl/statem/README.md) for
3463 further details). This change does have some associated API changes.
3464 Notably the SSL_state() function has been removed and replaced by
3465 SSL_get_state which now returns an "OSSL_HANDSHAKE_STATE" instead of an int.
3466 SSL_set_state() has been removed altogether. The previous handshake states
3467 defined in ssl.h and ssl3.h have also been removed.
3471 * All instances of the string "ssleay" in the public API were replaced
3472 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
3473 Some error codes related to internal RSA_eay API's were renamed.
3477 * The demo files in crypto/threads were moved to demo/threads.
3481 * Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
3484 *Matt Caswell, Rich Salz*
3486 * New ASN.1 embed macro.
3488 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
3489 structure is not allocated: it is part of the parent. That is instead of
3497 This reduces memory fragmentation and make it impossible to accidentally
3498 set a mandatory field to NULL.
3500 This currently only works for some fields specifically a SEQUENCE, CHOICE,
3501 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
3502 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
3507 * Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
3511 * Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
3512 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
3513 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
3514 DES and RC4 ciphersuites.
3518 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
3519 This changes the decoding behaviour for some invalid messages,
3520 though the change is mostly in the more lenient direction, and
3521 legacy behaviour is preserved as much as possible.
3525 * Fix no-stdio build.
3526 *David Woodhouse <David.Woodhouse@intel.com> and also*
3527 *Ivan Nestlerode <ivan.nestlerode@sonos.com>*
3529 * New testing framework
3530 The testing framework has been largely rewritten and is now using
3531 perl and the perl modules Test::Harness and an extended variant of
3532 Test::More called OpenSSL::Test to do its work. All test scripts in
3533 test/ have been rewritten into test recipes, and all direct calls to
3534 executables in test/Makefile have become individual recipes using the
3535 simplified testing OpenSSL::Test::Simple.
3537 For documentation on our testing modules, do:
3539 perldoc test/testlib/OpenSSL/Test/Simple.pm
3540 perldoc test/testlib/OpenSSL/Test.pm
3544 * Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
3545 are used; the latter aborts on memory leaks (usually checked on exit).
3546 Some undocumented "set malloc, etc., hooks" functions were removed
3547 and others were changed. All are now documented.
3551 * In DSA_generate_parameters_ex, if the provided seed is too short,
3554 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
3556 * Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
3557 from RFC4279, RFC4785, RFC5487, RFC5489.
3559 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
3560 original RSA_PSK patch.
3564 * Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
3565 era flag was never set throughout the codebase (only read). Also removed
3566 SSL3_FLAGS_POP_BUFFER which was only used if
3567 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
3571 * Changed the default name options in the "ca", "crl", "req" and "x509"
3572 to be "oneline" instead of "compat".
3576 * Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
3577 not aware of clients that still exhibit this bug, and the workaround
3578 hasn't been working properly for a while.
3582 * The return type of BIO_number_read() and BIO_number_written() as well as
3583 the corresponding num_read and num_write members in the BIO structure has
3584 changed from unsigned long to uint64_t. On platforms where an unsigned
3585 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
3590 * Given the pervasive nature of TLS extensions it is inadvisable to run
3591 OpenSSL without support for them. It also means that maintaining
3592 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
3593 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
3597 * Removed support for the two export grade static DH ciphersuites
3598 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
3599 were newly added (along with a number of other static DH ciphersuites) to
3600 1.0.2. However the two export ones have *never* worked since they were
3601 introduced. It seems strange in any case to be adding new export
3602 ciphersuites, and given "logjam" it also does not seem correct to fix them.
3606 * Version negotiation has been rewritten. In particular SSLv23_method(),
3607 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
3608 and turned into macros which simply call the new preferred function names
3609 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
3610 should use the new names instead. Also as part of this change the ssl23.h
3611 header file has been removed.
3615 * Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
3616 code and the associated standard is no longer considered fit-for-purpose.
3620 * RT2547 was closed. When generating a private key, try to make the
3621 output file readable only by the owner. This behavior change might
3622 be noticeable when interacting with other software.
3624 * Documented all exdata functions. Added CRYPTO_free_ex_index.
3629 * Added HTTP GET support to the ocsp command.
3633 * Changed default digest for the dgst and enc commands from MD5 to
3638 * RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
3642 * Added support for TLS extended master secret from
3643 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
3644 initial patch which was a great help during development.
3648 * All libssl internal structures have been removed from the public header
3649 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
3650 now redundant). Users should not attempt to access internal structures
3651 directly. Instead they should use the provided API functions.
3655 * config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
3656 Access to deprecated functions can be re-enabled by running config with
3657 "enable-deprecated". In addition applications wishing to use deprecated
3658 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
3659 will, by default, disable some transitive includes that previously existed
3660 in the header files (e.g. ec.h will no longer, by default, include bn.h)
3664 * Added support for OCB mode. OpenSSL has been granted a patent license
3665 compatible with the OpenSSL license for use of OCB. Details are available
3666 at <https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf>. Support
3667 for OCB can be removed by calling config with no-ocb.
3671 * SSLv2 support has been removed. It still supports receiving a SSLv2
3672 compatible client hello.
3676 * Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
3677 done while fixing the error code for the key-too-small case.
3679 *Annie Yousar <a.yousar@informatik.hu-berlin.de>*
3681 * CA.sh has been removed; use CA.pl instead.
3685 * Removed old DES API.
3689 * Remove various unsupported platforms:
3695 Sinix/ReliantUNIX RM400
3700 16-bit platforms such as WIN16
3704 * Clean up OPENSSL_NO_xxx #define's
3705 - Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
3706 - Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
3707 - OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
3708 - OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
3709 - OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
3710 - Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
3711 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
3712 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
3713 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
3714 - Remove MS_STATIC; it's a relic from platforms <32 bits.
3718 * Cleaned up dead code
3719 Remove all but one '#ifdef undef' which is to be looked at.
3723 * Clean up calling of xxx_free routines.
3724 Just like free(), fix most of the xxx_free routines to accept
3725 NULL. Remove the non-null checks from callers. Save much code.
3729 * Add secure heap for storage of private keys (when possible).
3730 Add BIO_s_secmem(), CBIGNUM, etc.
3731 Contributed by Akamai Technologies under our Corporate CLA.
3735 * Experimental support for a new, fast, unbiased prime candidate generator,
3736 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
3738 *Felix Laurie von Massenbach <felix@erbridge.co.uk>*
3740 * New output format NSS in the sess_id command line tool. This allows
3741 exporting the session id and the master key in NSS keylog format.
3743 *Martin Kaiser <martin@kaiser.cx>*
3745 * Harmonize version and its documentation. -f flag is used to display
3748 *mancha <mancha1@zoho.com>*
3750 * Fix eckey_priv_encode so it immediately returns an error upon a failure
3751 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
3753 *mancha <mancha1@zoho.com>*
3755 * Fix some double frees. These are not thought to be exploitable.
3757 *mancha <mancha1@zoho.com>*
3759 * A missing bounds check in the handling of the TLS heartbeat extension
3760 can be used to reveal up to 64k of memory to a connected client or
3763 Thanks for Neel Mehta of Google Security for discovering this bug and to
3764 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3765 preparing the fix ([CVE-2014-0160])
3767 *Adam Langley, Bodo Moeller*
3769 * Fix for the attack described in the paper "Recovering OpenSSL
3770 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3771 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3772 <http://eprint.iacr.org/2014/140>
3774 Thanks to Yuval Yarom and Naomi Benger for discovering this
3775 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
3777 *Yuval Yarom and Naomi Benger*
3779 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3780 this fixes a limitation in previous versions of OpenSSL.
3784 * Experimental encrypt-then-mac support.
3786 Experimental support for encrypt then mac from
3787 draft-gutmann-tls-encrypt-then-mac-02.txt
3789 To enable it set the appropriate extension number (0x42 for the test
3790 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
3792 For non-compliant peers (i.e. just about everything) this should have no
3795 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
3799 * Add EVP support for key wrapping algorithms, to avoid problems with
3800 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3801 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3802 algorithms and include tests cases.
3806 * Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
3811 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3812 MGF1 digest and OAEP label.
3816 * Make openssl verify return errors.
3818 *Chris Palmer <palmer@google.com> and Ben Laurie*
3820 * New function ASN1_TIME_diff to calculate the difference between two
3821 ASN1_TIME structures or one structure and the current time.
3825 * Update fips_test_suite to support multiple command line options. New
3826 test to induce all self test errors in sequence and check expected
3831 * Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
3832 sign or verify all in one operation.
3836 * Add fips_algvs: a multicall fips utility incorporating all the algorithm
3837 test programs and fips_test_suite. Includes functionality to parse
3838 the minimal script output of fipsalgest.pl directly.
3842 * Add authorisation parameter to FIPS_module_mode_set().
3846 * Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
3850 * Use separate DRBG fields for internal and external flags. New function
3851 FIPS_drbg_health_check() to perform on demand health checking. Add
3852 generation tests to fips_test_suite with reduced health check interval to
3853 demonstrate periodic health checking. Add "nodh" option to
3854 fips_test_suite to skip very slow DH test.
3858 * New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
3863 * More extensive health check for DRBG checking many more failure modes.
3864 New function FIPS_selftest_drbg_all() to handle every possible DRBG
3865 combination: call this in fips_test_suite.
3869 * Add support for canonical generation of DSA parameter 'g'. See
3872 * Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
3873 POST to handle HMAC cases.
3877 * Add functions FIPS_module_version() and FIPS_module_version_text()
3878 to return numerical and string versions of the FIPS module number.
3882 * Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
3883 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
3884 outside the validated module in the FIPS capable OpenSSL.
3888 * Minor change to DRBG entropy callback semantics. In some cases
3889 there is no multiple of the block length between min_len and
3890 max_len. Allow the callback to return more than max_len bytes
3891 of entropy but discard any extra: it is the callback's responsibility
3892 to ensure that the extra data discarded does not impact the
3893 requested amount of entropy.
3897 * Add PRNG security strength checks to RSA, DSA and ECDSA using
3898 information in FIPS186-3, SP800-57 and SP800-131A.
3902 * CCM support via EVP. Interface is very similar to GCM case except we
3903 must supply all data in one chunk (i.e. no update, final) and the
3904 message length must be supplied if AAD is used. Add algorithm test
3909 * Initial version of POST overhaul. Add POST callback to allow the status
3910 of POST to be monitored and/or failures induced. Modify fips_test_suite
3911 to use callback. Always run all selftests even if one fails.
3915 * XTS support including algorithm test driver in the fips_gcmtest program.
3916 Note: this does increase the maximum key length from 32 to 64 bytes but
3917 there should be no binary compatibility issues as existing applications
3918 will never use XTS mode.
3922 * Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
3923 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
3924 performs algorithm blocking for unapproved PRNG types. Also do not
3925 set PRNG type in FIPS_mode_set(): leave this to the application.
3926 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
3927 the standard OpenSSL PRNG: set additional data to a date time vector.
3931 * Rename old X9.31 PRNG functions of the form `FIPS_rand*` to `FIPS_x931*`.
3932 This shouldn't present any incompatibility problems because applications
3933 shouldn't be using these directly and any that are will need to rethink
3934 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
3938 * Extensive self tests and health checking required by SP800-90 DRBG.
3939 Remove strength parameter from FIPS_drbg_instantiate and always
3940 instantiate at maximum supported strength.
3944 * Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
3948 * New algorithm test program fips_dhvs to handle DH primitives only testing.
3952 * New function DH_compute_key_padded() to compute a DH key and pad with
3953 leading zeroes if needed: this complies with SP800-56A et al.
3957 * Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
3958 anything, incomplete, subject to change and largely untested at present.
3962 * Modify fipscanisteronly build option to only build the necessary object
3963 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
3967 * Add experimental option FIPSSYMS to give all symbols in
3968 fipscanister.o and FIPS or fips prefix. This will avoid
3969 conflicts with future versions of OpenSSL. Add perl script
3970 util/fipsas.pl to preprocess assembly language source files
3971 and rename any affected symbols.
3975 * Add selftest checks and algorithm block of non-fips algorithms in
3976 FIPS mode. Remove DES2 from selftests.
3980 * Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
3981 return internal method without any ENGINE dependencies. Add new
3982 tiny fips sign and verify functions.
3986 * New build option no-ec2m to disable characteristic 2 code.
3990 * New build option "fipscanisteronly". This only builds fipscanister.o
3991 and (currently) associated fips utilities. Uses the file Makefile.fips
3992 instead of Makefile.org as the prototype.
3996 * Add some FIPS mode restrictions to GCM. Add internal IV generator.
3997 Update fips_gcmtest to use IV generator.
4001 * Initial, experimental EVP support for AES-GCM. AAD can be input by
4002 setting output buffer to NULL. The `*Final` function must be
4003 called although it will not retrieve any additional data. The tag
4004 can be set or retrieved with a ctrl. The IV length is by default 12
4005 bytes (96 bits) but can be set to an alternative value. If the IV
4006 length exceeds the maximum IV length (currently 16 bytes) it cannot be
4011 * New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
4012 underlying do_cipher function handles all cipher semantics itself
4013 including padding and finalisation. This is useful if (for example)
4014 an ENGINE cipher handles block padding itself. The behaviour of
4015 do_cipher is subtly changed if this flag is set: the return value
4016 is the number of characters written to the output buffer (zero is
4017 no longer an error code) or a negative error code. Also if the
4018 input buffer is NULL and length 0 finalisation should be performed.
4022 * If a candidate issuer certificate is already part of the constructed
4023 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
4027 * Improve forward-security support: add functions
4029 void SSL_CTX_set_not_resumable_session_callback(
4030 SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
4031 void SSL_set_not_resumable_session_callback(
4032 SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
4034 for use by SSL/TLS servers; the callback function will be called whenever a
4035 new session is created, and gets to decide whether the session may be
4036 cached to make it resumable (return 0) or not (return 1). (As by the
4037 SSL/TLS protocol specifications, the session_id sent by the server will be
4038 empty to indicate that the session is not resumable; also, the server will
4039 not generate RFC 4507 (RFC 5077) session tickets.)
4041 A simple reasonable callback implementation is to return is_forward_secure.
4042 This parameter will be set to 1 or 0 depending on the ciphersuite selected
4043 by the SSL/TLS server library, indicating whether it can provide forward
4046 *Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)*
4048 * New -verify_name option in command line utilities to set verification
4053 * Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
4054 Add CMAC pkey methods.
4058 * Experimental renegotiation in s_server -www mode. If the client
4059 browses /reneg connection is renegotiated. If /renegcert it is
4060 renegotiated requesting a certificate.
4064 * Add an "external" session cache for debugging purposes to s_server. This
4065 should help trace issues which normally are only apparent in deployed
4066 multi-process servers.
4070 * Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
4071 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
4072 BIO_set_cipher() and some obscure PEM functions were changed so they
4073 can now return an error. The RAND changes required a change to the
4074 RAND_METHOD structure.
4078 * New macro `__owur` for "OpenSSL Warn Unused Result". This makes use of
4079 a gcc attribute to warn if the result of a function is ignored. This
4080 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
4081 whose return value is often ignored.
4085 * New -noct, -requestct, -requirect and -ctlogfile options for s_client.
4086 These allow SCTs (signed certificate timestamps) to be requested and
4087 validated when establishing a connection.
4089 *Rob Percival <robpercival@google.com>*
4094 ### Changes between 1.0.2s and 1.0.2t [10 Sep 2019]
4096 * For built-in EC curves, ensure an EC_GROUP built from the curve name is
4097 used even when parsing explicit parameters, when loading a encoded key
4098 or calling `EC_GROUP_new_from_ecpkparameters()`/
4099 `EC_GROUP_new_from_ecparameters()`.
4100 This prevents bypass of security hardening and performance gains,
4101 especially for curves with specialized EC_METHODs.
4102 By default, if a key encoded with explicit parameters is loaded and later
4103 encoded, the output is still encoded with explicit parameters, even if
4104 internally a "named" EC_GROUP is used for computation.
4108 * Compute ECC cofactors if not provided during EC_GROUP construction. Before
4109 this change, EC_GROUP_set_generator would accept order and/or cofactor as
4110 NULL. After this change, only the cofactor parameter can be NULL. It also
4111 does some minimal sanity checks on the passed order.
4116 * Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
4117 An attack is simple, if the first CMS_recipientInfo is valid but the
4118 second CMS_recipientInfo is chosen ciphertext. If the second
4119 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
4120 encryption key will be replaced by garbage, and the message cannot be
4121 decoded, but if the RSA decryption fails, the correct encryption key is
4122 used and the recipient will not notice the attack.
4123 As a work around for this potential attack the length of the decrypted
4124 key must be equal to the cipher default key length, in case the
4125 certifiate is not given and all recipientInfo are tried out.
4126 The old behaviour can be re-enabled in the CMS code by setting the
4127 CMS_DEBUG_DECRYPT flag.
4132 * Document issue with installation paths in diverse Windows builds
4134 '/usr/local/ssl' is an unsafe prefix for location to install OpenSSL
4135 binaries and run-time config file.
4140 ### Changes between 1.0.2r and 1.0.2s [28 May 2019]
4142 * Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
4143 This changes the size when using the genpkey app when no size is given. It
4144 fixes an omission in earlier changes that changed all RSA, DSA and DH
4145 generation apps to use 2048 bits by default.
4149 * Add FIPS support for Android Arm 64-bit
4151 Support for Android Arm 64-bit was added to the OpenSSL FIPS Object
4152 Module in Version 2.0.10. For some reason, the corresponding target
4153 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be
4154 built with FIPS support on Android Arm 64-bit. This omission has been
4157 *Matthias St. Pierre*
4159 ### Changes between 1.0.2q and 1.0.2r [26 Feb 2019]
4161 * 0-byte record padding oracle
4163 If an application encounters a fatal protocol error and then calls
4164 SSL_shutdown() twice (once to send a close_notify, and once to receive one)
4165 then OpenSSL can respond differently to the calling application if a 0 byte
4166 record is received with invalid padding compared to if a 0 byte record is
4167 received with an invalid MAC. If the application then behaves differently
4168 based on that in a way that is detectable to the remote peer, then this
4169 amounts to a padding oracle that could be used to decrypt data.
4171 In order for this to be exploitable "non-stitched" ciphersuites must be in
4172 use. Stitched ciphersuites are optimised implementations of certain
4173 commonly used ciphersuites. Also the application must call SSL_shutdown()
4174 twice even if a protocol error has occurred (applications should not do
4175 this but some do anyway).
4177 This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod
4178 Aviram, with additional investigation by Steven Collison and Andrew
4179 Hourselt. It was reported to OpenSSL on 10th December 2018.
4184 * Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
4188 ### Changes between 1.0.2p and 1.0.2q [20 Nov 2018]
4190 * Microarchitecture timing vulnerability in ECC scalar multiplication
4192 OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been
4193 shown to be vulnerable to a microarchitecture timing side channel attack.
4194 An attacker with sufficient access to mount local timing attacks during
4195 ECDSA signature generation could recover the private key.
4197 This issue was reported to OpenSSL on 26th October 2018 by Alejandro
4198 Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and
4204 * Timing vulnerability in DSA signature generation
4206 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
4207 timing side channel attack. An attacker could use variations in the signing
4208 algorithm to recover the private key.
4210 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
4215 * Resolve a compatibility issue in EC_GROUP handling with the FIPS Object
4216 Module, accidentally introduced while backporting security fixes from the
4217 development branch and hindering the use of ECC in FIPS mode.
4221 ### Changes between 1.0.2o and 1.0.2p [14 Aug 2018]
4223 * Client DoS due to large DH parameter
4225 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
4226 malicious server can send a very large prime value to the client. This will
4227 cause the client to spend an unreasonably long period of time generating a
4228 key for this prime resulting in a hang until the client has finished. This
4229 could be exploited in a Denial Of Service attack.
4231 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
4236 * Cache timing vulnerability in RSA Key Generation
4238 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
4239 a cache timing side channel attack. An attacker with sufficient access to
4240 mount cache timing attacks during the RSA key generation process could
4241 recover the private key.
4243 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
4244 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
4249 * Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
4250 parameter is no longer accepted, as it leads to a corrupt table. NULL
4251 pem_str is reserved for alias entries only.
4255 * Revert blinding in ECDSA sign and instead make problematic addition
4256 length-invariant. Switch even to fixed-length Montgomery multiplication.
4260 * Change generating and checking of primes so that the error rate of not
4261 being prime depends on the intended use based on the size of the input.
4262 For larger primes this will result in more rounds of Miller-Rabin.
4263 The maximal error rate for primes with more than 1080 bits is lowered
4266 *Kurt Roeckx, Annie Yousar*
4268 * Increase the number of Miller-Rabin rounds for DSA key generating to 64.
4272 * Add blinding to ECDSA and DSA signatures to protect against side channel
4273 attacks discovered by Keegan Ryan (NCC Group).
4277 * When unlocking a pass phrase protected PEM file or PKCS#8 container, we
4278 now allow empty (zero character) pass phrases.
4282 * Certificate time validation (X509_cmp_time) enforces stricter
4283 compliance with RFC 5280. Fractional seconds and timezone offsets
4284 are no longer allowed.
4288 ### Changes between 1.0.2n and 1.0.2o [27 Mar 2018]
4290 * Constructed ASN.1 types with a recursive definition could exceed the stack
4292 Constructed ASN.1 types with a recursive definition (such as can be found
4293 in PKCS7) could eventually exceed the stack given malicious input with
4294 excessive recursion. This could result in a Denial Of Service attack. There
4295 are no such structures used within SSL/TLS that come from untrusted sources
4296 so this is considered safe.
4298 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
4304 ### Changes between 1.0.2m and 1.0.2n [7 Dec 2017]
4306 * Read/write after SSL object in error state
4308 OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state"
4309 mechanism. The intent was that if a fatal error occurred during a handshake
4310 then OpenSSL would move into the error state and would immediately fail if
4311 you attempted to continue the handshake. This works as designed for the
4312 explicit handshake functions (SSL_do_handshake(), SSL_accept() and
4313 SSL_connect()), however due to a bug it does not work correctly if
4314 SSL_read() or SSL_write() is called directly. In that scenario, if the
4315 handshake fails then a fatal error will be returned in the initial function
4316 call. If SSL_read()/SSL_write() is subsequently called by the application
4317 for the same SSL object then it will succeed and the data is passed without
4318 being decrypted/encrypted directly from the SSL/TLS record layer.
4320 In order to exploit this issue an application bug would have to be present
4321 that resulted in a call to SSL_read()/SSL_write() being issued after having
4322 already received a fatal error.
4324 This issue was reported to OpenSSL by David Benjamin (Google).
4329 * rsaz_1024_mul_avx2 overflow bug on x86_64
4331 There is an overflow bug in the AVX2 Montgomery multiplication procedure
4332 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
4333 Analysis suggests that attacks against RSA and DSA as a result of this
4334 defect would be very difficult to perform and are not believed likely.
4335 Attacks against DH1024 are considered just feasible, because most of the
4336 work necessary to deduce information about a private key may be performed
4337 offline. The amount of resources required for such an attack would be
4338 significant. However, for an attack on TLS to be meaningful, the server
4339 would have to share the DH1024 private key among multiple clients, which is
4340 no longer an option since CVE-2016-0701.
4342 This only affects processors that support the AVX2 but not ADX extensions
4343 like Intel Haswell (4th generation).
4345 This issue was reported to OpenSSL by David Benjamin (Google). The issue
4346 was originally found via the OSS-Fuzz project.
4351 ### Changes between 1.0.2l and 1.0.2m [2 Nov 2017]
4353 * bn_sqrx8x_internal carry bug on x86_64
4355 There is a carry propagating bug in the x86_64 Montgomery squaring
4356 procedure. No EC algorithms are affected. Analysis suggests that attacks
4357 against RSA and DSA as a result of this defect would be very difficult to
4358 perform and are not believed likely. Attacks against DH are considered just
4359 feasible (although very difficult) because most of the work necessary to
4360 deduce information about a private key may be performed offline. The amount
4361 of resources required for such an attack would be very significant and
4362 likely only accessible to a limited number of attackers. An attacker would
4363 additionally need online access to an unpatched system using the target
4364 private key in a scenario with persistent DH parameters and a private
4365 key that is shared between multiple clients.
4367 This only affects processors that support the BMI1, BMI2 and ADX extensions
4368 like Intel Broadwell (5th generation) and later or AMD Ryzen.
4370 This issue was reported to OpenSSL by the OSS-Fuzz project.
4375 * Malformed X.509 IPAddressFamily could cause OOB read
4377 If an X.509 certificate has a malformed IPAddressFamily extension,
4378 OpenSSL could do a one-byte buffer overread. The most likely result
4379 would be an erroneous display of the certificate in text format.
4381 This issue was reported to OpenSSL by the OSS-Fuzz project.
4386 ### Changes between 1.0.2k and 1.0.2l [25 May 2017]
4388 * Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
4389 platform rather than 'mingw'.
4393 ### Changes between 1.0.2j and 1.0.2k [26 Jan 2017]
4395 * Truncated packet could crash via OOB read
4397 If one side of an SSL/TLS path is running on a 32-bit host and a specific
4398 cipher is being used, then a truncated packet can cause that host to
4399 perform an out-of-bounds read, usually resulting in a crash.
4401 This issue was reported to OpenSSL by Robert Święcki of Google.
4406 * BN_mod_exp may produce incorrect results on x86_64
4408 There is a carry propagating bug in the x86_64 Montgomery squaring
4409 procedure. No EC algorithms are affected. Analysis suggests that attacks
4410 against RSA and DSA as a result of this defect would be very difficult to
4411 perform and are not believed likely. Attacks against DH are considered just
4412 feasible (although very difficult) because most of the work necessary to
4413 deduce information about a private key may be performed offline. The amount
4414 of resources required for such an attack would be very significant and
4415 likely only accessible to a limited number of attackers. An attacker would
4416 additionally need online access to an unpatched system using the target
4417 private key in a scenario with persistent DH parameters and a private
4418 key that is shared between multiple clients. For example this can occur by
4419 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
4420 similar to CVE-2015-3193 but must be treated as a separate problem.
4422 This issue was reported to OpenSSL by the OSS-Fuzz project.
4427 * Montgomery multiplication may produce incorrect results
4429 There is a carry propagating bug in the Broadwell-specific Montgomery
4430 multiplication procedure that handles input lengths divisible by, but
4431 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
4432 and DH private keys are impossible. This is because the subroutine in
4433 question is not used in operations with the private key itself and an input
4434 of the attacker's direct choice. Otherwise the bug can manifest itself as
4435 transient authentication and key negotiation failures or reproducible
4436 erroneous outcome of public-key operations with specially crafted input.
4437 Among EC algorithms only Brainpool P-512 curves are affected and one
4438 presumably can attack ECDH key negotiation. Impact was not analyzed in
4439 detail, because pre-requisites for attack are considered unlikely. Namely
4440 multiple clients have to choose the curve in question and the server has to
4441 share the private key among them, neither of which is default behaviour.
4442 Even then only clients that chose the curve will be affected.
4444 This issue was publicly reported as transient failures and was not
4445 initially recognized as a security issue. Thanks to Richard Morgan for
4446 providing reproducible case.
4451 * OpenSSL now fails if it receives an unrecognised record type in TLS1.0
4452 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
4453 prevent issues where no progress is being made and the peer continually
4454 sends unrecognised record types, using up resources processing them.
4458 ### Changes between 1.0.2i and 1.0.2j [26 Sep 2016]
4460 * Missing CRL sanity check
4462 A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0
4463 but was omitted from OpenSSL 1.0.2i. As a result any attempt to use
4464 CRLs in OpenSSL 1.0.2i will crash with a null pointer exception.
4466 This issue only affects the OpenSSL 1.0.2i
4471 ### Changes between 1.0.2h and 1.0.2i [22 Sep 2016]
4473 * OCSP Status Request extension unbounded memory growth
4475 A malicious client can send an excessively large OCSP Status Request
4476 extension. If that client continually requests renegotiation, sending a
4477 large OCSP Status Request extension each time, then there will be unbounded
4478 memory growth on the server. This will eventually lead to a Denial Of
4479 Service attack through memory exhaustion. Servers with a default
4480 configuration are vulnerable even if they do not support OCSP. Builds using
4481 the "no-ocsp" build time option are not affected.
4483 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4488 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
4491 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
4497 * OOB write in MDC2_Update()
4499 An overflow can occur in MDC2_Update() either if called directly or
4500 through the EVP_DigestUpdate() function using MDC2. If an attacker
4501 is able to supply very large amounts of input data after a previous
4502 call to EVP_EncryptUpdate() with a partial block then a length check
4503 can overflow resulting in a heap corruption.
4505 The amount of data needed is comparable to SIZE_MAX which is impractical
4508 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4513 * Malformed SHA512 ticket DoS
4515 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
4516 DoS attack where a malformed ticket will result in an OOB read which will
4519 The use of SHA512 in TLS session tickets is comparatively rare as it requires
4520 a custom server callback and ticket lookup mechanism.
4522 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4527 * OOB write in BN_bn2dec()
4529 The function BN_bn2dec() does not check the return value of BN_div_word().
4530 This can cause an OOB write if an application uses this function with an
4531 overly large BIGNUM. This could be a problem if an overly large certificate
4532 or CRL is printed out from an untrusted source. TLS is not affected because
4533 record limits will reject an oversized certificate before it is parsed.
4535 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4540 * OOB read in TS_OBJ_print_bio()
4542 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
4543 the total length the OID text representation would use and not the amount
4544 of data written. This will result in OOB reads when large OIDs are
4547 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4552 * Pointer arithmetic undefined behaviour
4554 Avoid some undefined pointer arithmetic
4556 A common idiom in the codebase is to check limits in the following manner:
4559 Where "p" points to some malloc'd data of SIZE bytes and
4562 "len" here could be from some externally supplied data (e.g. from a TLS
4565 The rules of C pointer arithmetic are such that "p + len" is only well
4566 defined where len <= SIZE. Therefore the above idiom is actually
4567 undefined behaviour.
4569 For example this could cause problems if some malloc implementation
4570 provides an address for "p" such that "p + len" actually overflows for
4571 values of len that are too big and therefore p + len < limit.
4573 This issue was reported to OpenSSL by Guido Vranken
4578 * Constant time flag not preserved in DSA signing
4580 Operations in the DSA signing algorithm should run in constant time in
4581 order to avoid side channel attacks. A flaw in the OpenSSL DSA
4582 implementation means that a non-constant time codepath is followed for
4583 certain operations. This has been demonstrated through a cache-timing
4584 attack to be sufficient for an attacker to recover the private DSA key.
4586 This issue was reported by César Pereida (Aalto University), Billy Brumley
4587 (Tampere University of Technology), and Yuval Yarom (The University of
4588 Adelaide and NICTA).
4593 * DTLS buffered message DoS
4595 In a DTLS connection where handshake messages are delivered out-of-order
4596 those messages that OpenSSL is not yet ready to process will be buffered
4597 for later use. Under certain circumstances, a flaw in the logic means that
4598 those messages do not get removed from the buffer even though the handshake
4599 has been completed. An attacker could force up to approx. 15 messages to
4600 remain in the buffer when they are no longer required. These messages will
4601 be cleared when the DTLS connection is closed. The default maximum size for
4602 a message is 100k. Therefore the attacker could force an additional 1500k
4603 to be consumed per connection. By opening many simulataneous connections an
4604 attacker could cause a DoS attack through memory exhaustion.
4606 This issue was reported to OpenSSL by Quan Luo.
4611 * DTLS replay protection DoS
4613 A flaw in the DTLS replay attack protection mechanism means that records
4614 that arrive for future epochs update the replay protection "window" before
4615 the MAC for the record has been validated. This could be exploited by an
4616 attacker by sending a record for the next epoch (which does not have to
4617 decrypt or have a valid MAC), with a very large sequence number. This means
4618 that all subsequent legitimate packets are dropped causing a denial of
4619 service for a specific DTLS connection.
4621 This issue was reported to OpenSSL by the OCAP audit team.
4626 * Certificate message OOB reads
4628 In OpenSSL 1.0.2 and earlier some missing message length checks can result
4629 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
4630 theoretical DoS risk but this has not been observed in practice on common
4633 The messages affected are client certificate, client certificate request
4634 and server certificate. As a result the attack can only be performed
4635 against a client or a server which enables client authentication.
4637 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
4642 ### Changes between 1.0.2g and 1.0.2h [3 May 2016]
4644 * Prevent padding oracle in AES-NI CBC MAC check
4646 A MITM attacker can use a padding oracle attack to decrypt traffic
4647 when the connection uses an AES CBC cipher and the server support
4650 This issue was introduced as part of the fix for Lucky 13 padding
4651 attack ([CVE-2013-0169]). The padding check was rewritten to be in
4652 constant time by making sure that always the same bytes are read and
4653 compared against either the MAC or padding bytes. But it no longer
4654 checked that there was enough data to have both the MAC and padding
4657 This issue was reported by Juraj Somorovsky using TLS-Attacker.
4662 * Fix EVP_EncodeUpdate overflow
4664 An overflow can occur in the EVP_EncodeUpdate() function which is used for
4665 Base64 encoding of binary data. If an attacker is able to supply very large
4666 amounts of input data then a length check can overflow resulting in a heap
4669 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
4670 the `PEM_write_bio*` family of functions. These are mainly used within the
4671 OpenSSL command line applications, so any application which processes data
4672 from an untrusted source and outputs it as a PEM file should be considered
4673 vulnerable to this issue. User applications that call these APIs directly
4674 with large amounts of untrusted data may also be vulnerable.
4676 This issue was reported by Guido Vranken.
4681 * Fix EVP_EncryptUpdate overflow
4683 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
4684 is able to supply very large amounts of input data after a previous call to
4685 EVP_EncryptUpdate() with a partial block then a length check can overflow
4686 resulting in a heap corruption. Following an analysis of all OpenSSL
4687 internal usage of the EVP_EncryptUpdate() function all usage is one of two
4688 forms. The first form is where the EVP_EncryptUpdate() call is known to be
4689 the first called function after an EVP_EncryptInit(), and therefore that
4690 specific call must be safe. The second form is where the length passed to
4691 EVP_EncryptUpdate() can be seen from the code to be some small value and
4692 therefore there is no possibility of an overflow. Since all instances are
4693 one of these two forms, it is believed that there can be no overflows in
4694 internal code due to this problem. It should be noted that
4695 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
4696 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
4697 of these calls have also been analysed too and it is believed there are no
4698 instances in internal usage where an overflow could occur.
4700 This issue was reported by Guido Vranken.
4705 * Prevent ASN.1 BIO excessive memory allocation
4707 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
4708 a short invalid encoding can cause allocation of large amounts of memory
4709 potentially consuming excessive resources or exhausting memory.
4711 Any application parsing untrusted data through d2i BIO functions is
4712 affected. The memory based functions such as d2i_X509() are *not* affected.
4713 Since the memory based functions are used by the TLS library, TLS
4714 applications are not affected.
4716 This issue was reported by Brian Carpenter.
4723 ASN1 Strings that are over 1024 bytes can cause an overread in applications
4724 using the X509_NAME_oneline() function on EBCDIC systems. This could result
4725 in arbitrary stack data being returned in the buffer.
4727 This issue was reported by Guido Vranken.
4732 * Modify behavior of ALPN to invoke callback after SNI/servername
4733 callback, such that updates to the SSL_CTX affect ALPN.
4737 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
4742 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
4743 methods are enabled and ssl2 is disabled the methods return NULL.
4747 ### Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
4749 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
4750 Builds that are not configured with "enable-weak-ssl-ciphers" will not
4751 provide any "EXPORT" or "LOW" strength ciphers.
4755 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
4756 is by default disabled at build-time. Builds that are not configured with
4757 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
4758 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
4759 will need to explicitly call either of:
4761 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
4763 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
4765 as appropriate. Even if either of those is used, or the application
4766 explicitly uses the version-specific SSLv2_method() or its client and
4767 server variants, SSLv2 ciphers vulnerable to exhaustive search key
4768 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
4769 ciphers, and SSLv2 56-bit DES are no longer available.
4774 * Fix a double-free in DSA code
4776 A double free bug was discovered when OpenSSL parses malformed DSA private
4777 keys and could lead to a DoS attack or memory corruption for applications
4778 that receive DSA private keys from untrusted sources. This scenario is
4781 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
4787 * Disable SRP fake user seed to address a server memory leak.
4789 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
4791 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
4792 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
4793 was changed to ignore the "fake user" SRP seed, even if the seed
4796 Users should use SRP_VBASE_get1_by_user instead. Note that in
4797 SRP_VBASE_get1_by_user, caller must free the returned value. Note
4798 also that even though configuring the SRP seed attempts to hide
4799 invalid usernames by continuing the handshake with fake
4800 credentials, this behaviour is not constant time and no strong
4801 guarantees are made that the handshake is indistinguishable from
4802 that of a valid user.
4807 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
4809 In the BN_hex2bn function the number of hex digits is calculated using an
4810 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
4811 large values of `i` this can result in `bn_expand` not allocating any
4812 memory because `i * 4` is negative. This can leave the internal BIGNUM data
4813 field as NULL leading to a subsequent NULL ptr deref. For very large values
4814 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
4815 In this case memory is allocated to the internal BIGNUM data field, but it
4816 is insufficiently sized leading to heap corruption. A similar issue exists
4817 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
4818 is ever called by user applications with very large untrusted hex/dec data.
4819 This is anticipated to be a rare occurrence.
4821 All OpenSSL internal usage of these functions use data that is not expected
4822 to be untrusted, e.g. config file data or application command line
4823 arguments. If user developed applications generate config file data based
4824 on untrusted data then it is possible that this could also lead to security
4825 consequences. This is also anticipated to be rare.
4827 This issue was reported to OpenSSL by Guido Vranken.
4832 * Fix memory issues in `BIO_*printf` functions
4834 The internal `fmtstr` function used in processing a "%s" format string in
4835 the `BIO_*printf` functions could overflow while calculating the length of a
4836 string and cause an OOB read when printing very long strings.
4838 Additionally the internal `doapr_outch` function can attempt to write to an
4839 OOB memory location (at an offset from the NULL pointer) in the event of a
4840 memory allocation failure. In 1.0.2 and below this could be caused where
4841 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
4842 could be in processing a very long "%s" format string. Memory leaks can
4845 The first issue may mask the second issue dependent on compiler behaviour.
4846 These problems could enable attacks where large amounts of untrusted data
4847 is passed to the `BIO_*printf` functions. If applications use these functions
4848 in this way then they could be vulnerable. OpenSSL itself uses these
4849 functions when printing out human-readable dumps of ASN.1 data. Therefore
4850 applications that print this data could be vulnerable if the data is from
4851 untrusted sources. OpenSSL command line applications could also be
4852 vulnerable where they print out ASN.1 data, or if untrusted data is passed
4853 as command line arguments.
4855 Libssl is not considered directly vulnerable. Additionally certificates etc
4856 received via remote connections via libssl are also unlikely to be able to
4857 trigger these issues because of message size limits enforced within libssl.
4859 This issue was reported to OpenSSL Guido Vranken.
4864 * Side channel attack on modular exponentiation
4866 A side-channel attack was found which makes use of cache-bank conflicts on
4867 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
4868 of RSA keys. The ability to exploit this issue is limited as it relies on
4869 an attacker who has control of code in a thread running on the same
4870 hyper-threaded core as the victim thread which is performing decryptions.
4872 This issue was reported to OpenSSL by Yuval Yarom, The University of
4873 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
4874 Nadia Heninger, University of Pennsylvania with more information at
4875 <http://cachebleed.info>.
4880 * Change the req app to generate a 2048-bit RSA/DSA key by default,
4881 if no keysize is specified with default_bits. This fixes an
4882 omission in an earlier change that changed all RSA/DSA key generation
4883 apps to use 2048 bits by default.
4887 ### Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
4889 * DH small subgroups
4891 Historically OpenSSL only ever generated DH parameters based on "safe"
4892 primes. More recently (in version 1.0.2) support was provided for
4893 generating X9.42 style parameter files such as those required for RFC 5114
4894 support. The primes used in such files may not be "safe". Where an
4895 application is using DH configured with parameters based on primes that are
4896 not "safe" then an attacker could use this fact to find a peer's private
4897 DH exponent. This attack requires that the attacker complete multiple
4898 handshakes in which the peer uses the same private DH exponent. For example
4899 this could be used to discover a TLS server's private DH exponent if it's
4900 reusing the private DH exponent or it's using a static DH ciphersuite.
4902 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
4903 TLS. It is not on by default. If the option is not set then the server
4904 reuses the same private DH exponent for the life of the server process and
4905 would be vulnerable to this attack. It is believed that many popular
4906 applications do set this option and would therefore not be at risk.
4908 The fix for this issue adds an additional check where a "q" parameter is
4909 available (as is the case in X9.42 based parameters). This detects the
4910 only known attack, and is the only possible defense for static DH
4911 ciphersuites. This could have some performance impact.
4913 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
4914 default and cannot be disabled. This could have some performance impact.
4916 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
4921 * SSLv2 doesn't block disabled ciphers
4923 A malicious client can negotiate SSLv2 ciphers that have been disabled on
4924 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
4925 been disabled, provided that the SSLv2 protocol was not also disabled via
4928 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
4929 and Sebastian Schinzel.
4934 ### Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
4936 * BN_mod_exp may produce incorrect results on x86_64
4938 There is a carry propagating bug in the x86_64 Montgomery squaring
4939 procedure. No EC algorithms are affected. Analysis suggests that attacks
4940 against RSA and DSA as a result of this defect would be very difficult to
4941 perform and are not believed likely. Attacks against DH are considered just
4942 feasible (although very difficult) because most of the work necessary to
4943 deduce information about a private key may be performed offline. The amount
4944 of resources required for such an attack would be very significant and
4945 likely only accessible to a limited number of attackers. An attacker would
4946 additionally need online access to an unpatched system using the target
4947 private key in a scenario with persistent DH parameters and a private
4948 key that is shared between multiple clients. For example this can occur by
4949 default in OpenSSL DHE based SSL/TLS ciphersuites.
4951 This issue was reported to OpenSSL by Hanno Böck.
4956 * Certificate verify crash with missing PSS parameter
4958 The signature verification routines will crash with a NULL pointer
4959 dereference if presented with an ASN.1 signature using the RSA PSS
4960 algorithm and absent mask generation function parameter. Since these
4961 routines are used to verify certificate signature algorithms this can be
4962 used to crash any certificate verification operation and exploited in a
4963 DoS attack. Any application which performs certificate verification is
4964 vulnerable including OpenSSL clients and servers which enable client
4967 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
4972 * X509_ATTRIBUTE memory leak
4974 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
4975 memory. This structure is used by the PKCS#7 and CMS routines so any
4976 application which reads PKCS#7 or CMS data from untrusted sources is
4977 affected. SSL/TLS is not affected.
4979 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
4985 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
4986 This changes the decoding behaviour for some invalid messages,
4987 though the change is mostly in the more lenient direction, and
4988 legacy behaviour is preserved as much as possible.
4992 * In DSA_generate_parameters_ex, if the provided seed is too short,
4995 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
4997 ### Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
4999 * Alternate chains certificate forgery
5001 During certificate verification, OpenSSL will attempt to find an
5002 alternative certificate chain if the first attempt to build such a chain
5003 fails. An error in the implementation of this logic can mean that an
5004 attacker could cause certain checks on untrusted certificates to be
5005 bypassed, such as the CA flag, enabling them to use a valid leaf
5006 certificate to act as a CA and "issue" an invalid certificate.
5008 This issue was reported to OpenSSL by Adam Langley/David Benjamin
5013 ### Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
5015 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
5016 incompatibility in the handling of HMAC. The previous ABI has now been
5021 ### Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
5023 * Malformed ECParameters causes infinite loop
5025 When processing an ECParameters structure OpenSSL enters an infinite loop
5026 if the curve specified is over a specially malformed binary polynomial
5029 This can be used to perform denial of service against any
5030 system which processes public keys, certificate requests or
5031 certificates. This includes TLS clients and TLS servers with
5032 client authentication enabled.
5034 This issue was reported to OpenSSL by Joseph Barr-Pixton.
5039 * Exploitable out-of-bounds read in X509_cmp_time
5041 X509_cmp_time does not properly check the length of the ASN1_TIME
5042 string and can read a few bytes out of bounds. In addition,
5043 X509_cmp_time accepts an arbitrary number of fractional seconds in the
5046 An attacker can use this to craft malformed certificates and CRLs of
5047 various sizes and potentially cause a segmentation fault, resulting in
5048 a DoS on applications that verify certificates or CRLs. TLS clients
5049 that verify CRLs are affected. TLS clients and servers with client
5050 authentication enabled may be affected if they use custom verification
5053 This issue was reported to OpenSSL by Robert Swiecki (Google), and
5054 independently by Hanno Böck.
5059 * PKCS7 crash with missing EnvelopedContent
5061 The PKCS#7 parsing code does not handle missing inner EncryptedContent
5062 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
5063 with missing content and trigger a NULL pointer dereference on parsing.
5065 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
5066 structures from untrusted sources are affected. OpenSSL clients and
5067 servers are not affected.
5069 This issue was reported to OpenSSL by Michal Zalewski (Google).
5074 * CMS verify infinite loop with unknown hash function
5076 When verifying a signedData message the CMS code can enter an infinite loop
5077 if presented with an unknown hash function OID. This can be used to perform
5078 denial of service against any system which verifies signedData messages using
5080 This issue was reported to OpenSSL by Johannes Bauer.
5085 * Race condition handling NewSessionTicket
5087 If a NewSessionTicket is received by a multi-threaded client when attempting to
5088 reuse a previous ticket then a race condition can occur potentially leading to
5089 a double free of the ticket data.
5094 * Only support 256-bit or stronger elliptic curves with the
5095 'ecdh_auto' setting (server) or by default (client). Of supported
5096 curves, prefer P-256 (both).
5100 ### Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
5102 * ClientHello sigalgs DoS fix
5104 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
5105 invalid signature algorithms extension a NULL pointer dereference will
5106 occur. This can be exploited in a DoS attack against the server.
5108 This issue was was reported to OpenSSL by David Ramos of Stanford
5112 *Stephen Henson and Matt Caswell*
5114 * Multiblock corrupted pointer fix
5116 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
5117 feature only applies on 64 bit x86 architecture platforms that support AES
5118 NI instructions. A defect in the implementation of "multiblock" can cause
5119 OpenSSL's internal write buffer to become incorrectly set to NULL when
5120 using non-blocking IO. Typically, when the user application is using a
5121 socket BIO for writing, this will only result in a failed connection.
5122 However if some other BIO is used then it is likely that a segmentation
5123 fault will be triggered, thus enabling a potential DoS attack.
5125 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
5130 * Segmentation fault in DTLSv1_listen fix
5132 The DTLSv1_listen function is intended to be stateless and processes the
5133 initial ClientHello from many peers. It is common for user code to loop
5134 over the call to DTLSv1_listen until a valid ClientHello is received with
5135 an associated cookie. A defect in the implementation of DTLSv1_listen means
5136 that state is preserved in the SSL object from one invocation to the next
5137 that can lead to a segmentation fault. Errors processing the initial
5138 ClientHello can trigger this scenario. An example of such an error could be
5139 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
5142 This issue was reported to OpenSSL by Per Allansson.
5147 * Segmentation fault in ASN1_TYPE_cmp fix
5149 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
5150 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
5151 certificate signature algorithm consistency this can be used to crash any
5152 certificate verification operation and exploited in a DoS attack. Any
5153 application which performs certificate verification is vulnerable including
5154 OpenSSL clients and servers which enable client authentication.
5159 * Segmentation fault for invalid PSS parameters fix
5161 The signature verification routines will crash with a NULL pointer
5162 dereference if presented with an ASN.1 signature using the RSA PSS
5163 algorithm and invalid parameters. Since these routines are used to verify
5164 certificate signature algorithms this can be used to crash any
5165 certificate verification operation and exploited in a DoS attack. Any
5166 application which performs certificate verification is vulnerable including
5167 OpenSSL clients and servers which enable client authentication.
5169 This issue was was reported to OpenSSL by Brian Carpenter.
5174 * ASN.1 structure reuse memory corruption fix
5176 Reusing a structure in ASN.1 parsing may allow an attacker to cause
5177 memory corruption via an invalid write. Such reuse is and has been
5178 strongly discouraged and is believed to be rare.
5180 Applications that parse structures containing CHOICE or ANY DEFINED BY
5181 components may be affected. Certificate parsing (d2i_X509 and related
5182 functions) are however not affected. OpenSSL clients and servers are
5188 * PKCS7 NULL pointer dereferences fix
5190 The PKCS#7 parsing code does not handle missing outer ContentInfo
5191 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
5192 missing content and trigger a NULL pointer dereference on parsing.
5194 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
5195 otherwise parse PKCS#7 structures from untrusted sources are
5196 affected. OpenSSL clients and servers are not affected.
5198 This issue was reported to OpenSSL by Michal Zalewski (Google).
5203 * DoS via reachable assert in SSLv2 servers fix
5205 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
5206 servers that both support SSLv2 and enable export cipher suites by sending
5207 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
5209 This issue was discovered by Sean Burford (Google) and Emilia Käsper
5210 (OpenSSL development team).
5215 * Empty CKE with client auth and DHE fix
5217 If client auth is used then a server can seg fault in the event of a DHE
5218 ciphersuite being selected and a zero length ClientKeyExchange message
5219 being sent by the client. This could be exploited in a DoS attack.
5224 * Handshake with unseeded PRNG fix
5226 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
5227 with an unseeded PRNG. The conditions are:
5228 - The client is on a platform where the PRNG has not been seeded
5229 automatically, and the user has not seeded manually
5230 - A protocol specific client method version has been used (i.e. not
5231 SSL_client_methodv23)
5232 - A ciphersuite is used that does not require additional random data from
5233 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
5235 If the handshake succeeds then the client random that has been used will
5236 have been generated from a PRNG with insufficient entropy and therefore the
5237 output may be predictable.
5239 For example using the following command with an unseeded openssl will
5240 succeed on an unpatched platform:
5242 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
5247 * Use After Free following d2i_ECPrivatekey error fix
5249 A malformed EC private key file consumed via the d2i_ECPrivateKey function
5250 could cause a use after free condition. This, in turn, could cause a double
5251 free in several private key parsing functions (such as d2i_PrivateKey
5252 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
5253 for applications that receive EC private keys from untrusted
5254 sources. This scenario is considered rare.
5256 This issue was discovered by the BoringSSL project and fixed in their
5262 * X509_to_X509_REQ NULL pointer deref fix
5264 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
5265 the certificate key is invalid. This function is rarely used in practice.
5267 This issue was discovered by Brian Carpenter.
5272 * Removed the export ciphers from the DEFAULT ciphers
5276 ### Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
5278 * Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
5279 ARMv5 through ARMv8, as opposite to "locking" it to single one.
5280 So far those who have to target multiple platforms would compromise
5281 and argue that binary targeting say ARMv5 would still execute on
5282 ARMv8. "Universal" build resolves this compromise by providing
5283 near-optimal performance even on newer platforms.
5287 * Accelerated NIST P-256 elliptic curve implementation for x86_64
5288 (other platforms pending).
5290 *Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov*
5292 * Add support for the SignedCertificateTimestampList certificate and
5293 OCSP response extensions from RFC6962.
5297 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
5298 for corner cases. (Certain input points at infinity could lead to
5299 bogus results, with non-infinity inputs mapped to infinity too.)
5303 * Initial support for PowerISA 2.0.7, first implemented in POWER8.
5304 This covers AES, SHA256/512 and GHASH. "Initial" means that most
5305 common cases are optimized and there still is room for further
5306 improvements. Vector Permutation AES for Altivec is also added.
5310 * Add support for little-endian ppc64 Linux target.
5312 *Marcelo Cerri (IBM)*
5314 * Initial support for AMRv8 ISA crypto extensions. This covers AES,
5315 SHA1, SHA256 and GHASH. "Initial" means that most common cases
5316 are optimized and there still is room for further improvements.
5317 Both 32- and 64-bit modes are supported.
5319 *Andy Polyakov, Ard Biesheuvel (Linaro)*
5321 * Improved ARMv7 NEON support.
5325 * Support for SPARC Architecture 2011 crypto extensions, first
5326 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
5327 SHA256/512, MD5, GHASH and modular exponentiation.
5329 *Andy Polyakov, David Miller*
5331 * Accelerated modular exponentiation for Intel processors, a.k.a.
5334 *Shay Gueron & Vlad Krasnov (Intel Corp)*
5336 * Support for new and upcoming Intel processors, including AVX2,
5337 BMI and SHA ISA extensions. This includes additional "stitched"
5338 implementations, AESNI-SHA256 and GCM, and multi-buffer support
5341 This work was sponsored by Intel Corp.
5345 * Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
5346 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
5347 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
5351 * Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
5352 this fixes a limitation in previous versions of OpenSSL.
5356 * Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
5357 MGF1 digest and OAEP label.
5361 * Add EVP support for key wrapping algorithms, to avoid problems with
5362 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
5363 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
5364 algorithms and include tests cases.
5368 * Add functions to allocate and set the fields of an ECDSA_METHOD
5371 *Douglas E. Engert, Steve Henson*
5373 * New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
5374 difference in days and seconds between two tm or ASN1_TIME structures.
5378 * Add -rev test option to s_server to just reverse order of characters
5379 received by client and send back to server. Also prints an abbreviated
5380 summary of the connection parameters.
5384 * New option -brief for s_client and s_server to print out a brief summary
5385 of connection parameters.
5389 * Add callbacks for arbitrary TLS extensions.
5391 *Trevor Perrin <trevp@trevp.net> and Ben Laurie*
5393 * New option -crl_download in several openssl utilities to download CRLs
5394 from CRLDP extension in certificates.
5398 * New options -CRL and -CRLform for s_client and s_server for CRLs.
5402 * New function X509_CRL_diff to generate a delta CRL from the difference
5403 of two full CRLs. Add support to "crl" utility.
5407 * New functions to set lookup_crls function and to retrieve
5408 X509_STORE from X509_STORE_CTX.
5412 * Print out deprecated issuer and subject unique ID fields in
5417 * Extend OCSP I/O functions so they can be used for simple general purpose
5418 HTTP as well as OCSP. New wrapper function which can be used to download
5419 CRLs using the OCSP API.
5423 * Delegate command line handling in s_client/s_server to SSL_CONF APIs.
5427 * `SSL_CONF*` functions. These provide a common framework for application
5428 configuration using configuration files or command lines.
5432 * SSL/TLS tracing code. This parses out SSL/TLS records using the
5433 message callback and prints the results. Needs compile time option
5434 "enable-ssl-trace". New options to s_client and s_server to enable
5439 * New ctrl and macro to retrieve supported points extensions.
5440 Print out extension in s_server and s_client.
5444 * New functions to retrieve certificate signature and signature
5449 * Add functions to retrieve and manipulate the raw cipherlist sent by a
5454 * New Suite B modes for TLS code. These use and enforce the requirements
5455 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
5456 only use Suite B curves. The Suite B modes can be set by using the
5457 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
5461 * New chain verification flags for Suite B levels of security. Check
5462 algorithms are acceptable when flags are set in X509_verify_cert.
5466 * Make tls1_check_chain return a set of flags indicating checks passed
5467 by a certificate chain. Add additional tests to handle client
5468 certificates: checks for matching certificate type and issuer name
5473 * If an attempt is made to use a signature algorithm not in the peer
5474 preference list abort the handshake. If client has no suitable
5475 signature algorithms in response to a certificate request do not
5476 use the certificate.
5480 * If server EC tmp key is not in client preference list abort handshake.
5484 * Add support for certificate stores in CERT structure. This makes it
5485 possible to have different stores per SSL structure or one store in
5486 the parent SSL_CTX. Include distinct stores for certificate chain
5487 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
5488 to build and store a certificate chain in CERT structure: returning
5489 an error if the chain cannot be built: this will allow applications
5490 to test if a chain is correctly configured.
5492 Note: if the CERT based stores are not set then the parent SSL_CTX
5493 store is used to retain compatibility with existing behaviour.
5497 * New function ssl_set_client_disabled to set a ciphersuite disabled
5498 mask based on the current session, check mask when sending client
5499 hello and checking the requested ciphersuite.
5503 * New ctrls to retrieve and set certificate types in a certificate
5504 request message. Print out received values in s_client. If certificate
5505 types is not set with custom values set sensible values based on
5506 supported signature algorithms.
5510 * Support for distinct client and server supported signature algorithms.
5514 * Add certificate callback. If set this is called whenever a certificate
5515 is required by client or server. An application can decide which
5516 certificate chain to present based on arbitrary criteria: for example
5517 supported signature algorithms. Add very simple example to s_server.
5518 This fixes many of the problems and restrictions of the existing client
5519 certificate callback: for example you can now clear an existing
5520 certificate and specify the whole chain.
5524 * Add new "valid_flags" field to CERT_PKEY structure which determines what
5525 the certificate can be used for (if anything). Set valid_flags field
5526 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
5527 to have similar checks in it.
5529 Add new "cert_flags" field to CERT structure and include a "strict mode".
5530 This enforces some TLS certificate requirements (such as only permitting
5531 certificate signature algorithms contained in the supported algorithms
5532 extension) which some implementations ignore: this option should be used
5533 with caution as it could cause interoperability issues.
5537 * Update and tidy signature algorithm extension processing. Work out
5538 shared signature algorithms based on preferences and peer algorithms
5539 and print them out in s_client and s_server. Abort handshake if no
5540 shared signature algorithms.
5544 * Add new functions to allow customised supported signature algorithms
5545 for SSL and SSL_CTX structures. Add options to s_client and s_server
5550 * New function SSL_certs_clear() to delete all references to certificates
5551 from an SSL structure. Before this once a certificate had been added
5552 it couldn't be removed.
5556 * Integrate hostname, email address and IP address checking with certificate
5557 verification. New verify options supporting checking in openssl utility.
5561 * Fixes and wildcard matching support to hostname and email checking
5562 functions. Add manual page.
5564 *Florian Weimer (Red Hat Product Security Team)*
5566 * New functions to check a hostname email or IP address against a
5567 certificate. Add options x509 utility to print results of checks against
5572 * Fix OCSP checking.
5574 *Rob Stradling <rob.stradling@comodo.com> and Ben Laurie*
5576 * Initial experimental support for explicitly trusted non-root CAs.
5577 OpenSSL still tries to build a complete chain to a root but if an
5578 intermediate CA has a trust setting included that is used. The first
5579 setting is used: whether to trust (e.g., -addtrust option to the x509
5584 * Add -trusted_first option which attempts to find certificates in the
5585 trusted store even if an untrusted chain is also supplied.
5589 * MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
5590 platform support for Linux and Android.
5594 * Support for linux-x32, ILP32 environment in x86_64 framework.
5598 * Experimental multi-implementation support for FIPS capable OpenSSL.
5599 When in FIPS mode the approved implementations are used as normal,
5600 when not in FIPS mode the internal unapproved versions are used instead.
5601 This means that the FIPS capable OpenSSL isn't forced to use the
5602 (often lower performance) FIPS implementations outside FIPS mode.
5606 * Transparently support X9.42 DH parameters when calling
5607 PEM_read_bio_DHparameters. This means existing applications can handle
5608 the new parameter format automatically.
5612 * Initial experimental support for X9.42 DH parameter format: mainly
5613 to support use of 'q' parameter for RFC5114 parameters.
5617 * Add DH parameters from RFC5114 including test data to dhtest.
5621 * Support for automatic EC temporary key parameter selection. If enabled
5622 the most preferred EC parameters are automatically used instead of
5623 hardcoded fixed parameters. Now a server just has to call:
5624 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
5625 support ECDH and use the most appropriate parameters.
5629 * Enhance and tidy EC curve and point format TLS extension code. Use
5630 static structures instead of allocation if default values are used.
5631 New ctrls to set curves we wish to support and to retrieve shared curves.
5632 Print out shared curves in s_server. New options to s_server and s_client
5633 to set list of supported curves.
5637 * New ctrls to retrieve supported signature algorithms and
5638 supported curve values as an array of NIDs. Extend openssl utility
5639 to print out received values.
5643 * Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
5644 between NIDs and the more common NIST names such as "P-256". Enhance
5645 ecparam utility and ECC method to recognise the NIST names for curves.
5649 * Enhance SSL/TLS certificate chain handling to support different
5650 chains for each certificate instead of one chain in the parent SSL_CTX.
5654 * Support for fixed DH ciphersuite client authentication: where both
5655 server and client use DH certificates with common parameters.
5659 * Support for fixed DH ciphersuites: those requiring DH server
5664 * New function i2d_re_X509_tbs for re-encoding the TBS portion of
5666 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
5667 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
5668 X509_CINF_get_signature were reverted post internal team review.
5673 ### Changes between 1.0.1t and 1.0.1u [22 Sep 2016]
5675 * OCSP Status Request extension unbounded memory growth
5677 A malicious client can send an excessively large OCSP Status Request
5678 extension. If that client continually requests renegotiation, sending a
5679 large OCSP Status Request extension each time, then there will be unbounded
5680 memory growth on the server. This will eventually lead to a Denial Of
5681 Service attack through memory exhaustion. Servers with a default
5682 configuration are vulnerable even if they do not support OCSP. Builds using
5683 the "no-ocsp" build time option are not affected.
5685 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5690 * In order to mitigate the SWEET32 attack, the DES ciphers were moved from
5693 This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan
5699 * OOB write in MDC2_Update()
5701 An overflow can occur in MDC2_Update() either if called directly or
5702 through the EVP_DigestUpdate() function using MDC2. If an attacker
5703 is able to supply very large amounts of input data after a previous
5704 call to EVP_EncryptUpdate() with a partial block then a length check
5705 can overflow resulting in a heap corruption.
5707 The amount of data needed is comparable to SIZE_MAX which is impractical
5710 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5715 * Malformed SHA512 ticket DoS
5717 If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a
5718 DoS attack where a malformed ticket will result in an OOB read which will
5721 The use of SHA512 in TLS session tickets is comparatively rare as it requires
5722 a custom server callback and ticket lookup mechanism.
5724 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5729 * OOB write in BN_bn2dec()
5731 The function BN_bn2dec() does not check the return value of BN_div_word().
5732 This can cause an OOB write if an application uses this function with an
5733 overly large BIGNUM. This could be a problem if an overly large certificate
5734 or CRL is printed out from an untrusted source. TLS is not affected because
5735 record limits will reject an oversized certificate before it is parsed.
5737 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5742 * OOB read in TS_OBJ_print_bio()
5744 The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is
5745 the total length the OID text representation would use and not the amount
5746 of data written. This will result in OOB reads when large OIDs are
5749 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5754 * Pointer arithmetic undefined behaviour
5756 Avoid some undefined pointer arithmetic
5758 A common idiom in the codebase is to check limits in the following manner:
5761 Where "p" points to some malloc'd data of SIZE bytes and
5764 "len" here could be from some externally supplied data (e.g. from a TLS
5767 The rules of C pointer arithmetic are such that "p + len" is only well
5768 defined where len <= SIZE. Therefore the above idiom is actually
5769 undefined behaviour.
5771 For example this could cause problems if some malloc implementation
5772 provides an address for "p" such that "p + len" actually overflows for
5773 values of len that are too big and therefore p + len < limit.
5775 This issue was reported to OpenSSL by Guido Vranken
5780 * Constant time flag not preserved in DSA signing
5782 Operations in the DSA signing algorithm should run in constant time in
5783 order to avoid side channel attacks. A flaw in the OpenSSL DSA
5784 implementation means that a non-constant time codepath is followed for
5785 certain operations. This has been demonstrated through a cache-timing
5786 attack to be sufficient for an attacker to recover the private DSA key.
5788 This issue was reported by César Pereida (Aalto University), Billy Brumley
5789 (Tampere University of Technology), and Yuval Yarom (The University of
5790 Adelaide and NICTA).
5795 * DTLS buffered message DoS
5797 In a DTLS connection where handshake messages are delivered out-of-order
5798 those messages that OpenSSL is not yet ready to process will be buffered
5799 for later use. Under certain circumstances, a flaw in the logic means that
5800 those messages do not get removed from the buffer even though the handshake
5801 has been completed. An attacker could force up to approx. 15 messages to
5802 remain in the buffer when they are no longer required. These messages will
5803 be cleared when the DTLS connection is closed. The default maximum size for
5804 a message is 100k. Therefore the attacker could force an additional 1500k
5805 to be consumed per connection. By opening many simulataneous connections an
5806 attacker could cause a DoS attack through memory exhaustion.
5808 This issue was reported to OpenSSL by Quan Luo.
5813 * DTLS replay protection DoS
5815 A flaw in the DTLS replay attack protection mechanism means that records
5816 that arrive for future epochs update the replay protection "window" before
5817 the MAC for the record has been validated. This could be exploited by an
5818 attacker by sending a record for the next epoch (which does not have to
5819 decrypt or have a valid MAC), with a very large sequence number. This means
5820 that all subsequent legitimate packets are dropped causing a denial of
5821 service for a specific DTLS connection.
5823 This issue was reported to OpenSSL by the OCAP audit team.
5828 * Certificate message OOB reads
5830 In OpenSSL 1.0.2 and earlier some missing message length checks can result
5831 in OOB reads of up to 2 bytes beyond an allocated buffer. There is a
5832 theoretical DoS risk but this has not been observed in practice on common
5835 The messages affected are client certificate, client certificate request
5836 and server certificate. As a result the attack can only be performed
5837 against a client or a server which enables client authentication.
5839 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
5844 ### Changes between 1.0.1s and 1.0.1t [3 May 2016]
5846 * Prevent padding oracle in AES-NI CBC MAC check
5848 A MITM attacker can use a padding oracle attack to decrypt traffic
5849 when the connection uses an AES CBC cipher and the server support
5852 This issue was introduced as part of the fix for Lucky 13 padding
5853 attack ([CVE-2013-0169]). The padding check was rewritten to be in
5854 constant time by making sure that always the same bytes are read and
5855 compared against either the MAC or padding bytes. But it no longer
5856 checked that there was enough data to have both the MAC and padding
5859 This issue was reported by Juraj Somorovsky using TLS-Attacker.
5864 * Fix EVP_EncodeUpdate overflow
5866 An overflow can occur in the EVP_EncodeUpdate() function which is used for
5867 Base64 encoding of binary data. If an attacker is able to supply very large
5868 amounts of input data then a length check can overflow resulting in a heap
5871 Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by
5872 the `PEM_write_bio*` family of functions. These are mainly used within the
5873 OpenSSL command line applications, so any application which processes data
5874 from an untrusted source and outputs it as a PEM file should be considered
5875 vulnerable to this issue. User applications that call these APIs directly
5876 with large amounts of untrusted data may also be vulnerable.
5878 This issue was reported by Guido Vranken.
5883 * Fix EVP_EncryptUpdate overflow
5885 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
5886 is able to supply very large amounts of input data after a previous call to
5887 EVP_EncryptUpdate() with a partial block then a length check can overflow
5888 resulting in a heap corruption. Following an analysis of all OpenSSL
5889 internal usage of the EVP_EncryptUpdate() function all usage is one of two
5890 forms. The first form is where the EVP_EncryptUpdate() call is known to be
5891 the first called function after an EVP_EncryptInit(), and therefore that
5892 specific call must be safe. The second form is where the length passed to
5893 EVP_EncryptUpdate() can be seen from the code to be some small value and
5894 therefore there is no possibility of an overflow. Since all instances are
5895 one of these two forms, it is believed that there can be no overflows in
5896 internal code due to this problem. It should be noted that
5897 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
5898 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
5899 of these calls have also been analysed too and it is believed there are no
5900 instances in internal usage where an overflow could occur.
5902 This issue was reported by Guido Vranken.
5907 * Prevent ASN.1 BIO excessive memory allocation
5909 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
5910 a short invalid encoding can casuse allocation of large amounts of memory
5911 potentially consuming excessive resources or exhausting memory.
5913 Any application parsing untrusted data through d2i BIO functions is
5914 affected. The memory based functions such as d2i_X509() are *not* affected.
5915 Since the memory based functions are used by the TLS library, TLS
5916 applications are not affected.
5918 This issue was reported by Brian Carpenter.
5925 ASN1 Strings that are over 1024 bytes can cause an overread in applications
5926 using the X509_NAME_oneline() function on EBCDIC systems. This could result
5927 in arbitrary stack data being returned in the buffer.
5929 This issue was reported by Guido Vranken.
5934 * Modify behavior of ALPN to invoke callback after SNI/servername
5935 callback, such that updates to the SSL_CTX affect ALPN.
5939 * Remove LOW from the DEFAULT cipher list. This removes singles DES from the
5944 * Only remove the SSLv2 methods with the no-ssl2-method option. When the
5945 methods are enabled and ssl2 is disabled the methods return NULL.
5949 ### Changes between 1.0.1r and 1.0.1s [1 Mar 2016]
5951 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
5952 Builds that are not configured with "enable-weak-ssl-ciphers" will not
5953 provide any "EXPORT" or "LOW" strength ciphers.
5957 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
5958 is by default disabled at build-time. Builds that are not configured with
5959 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
5960 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
5961 will need to explicitly call either of:
5963 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
5965 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
5967 as appropriate. Even if either of those is used, or the application
5968 explicitly uses the version-specific SSLv2_method() or its client and
5969 server variants, SSLv2 ciphers vulnerable to exhaustive search key
5970 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
5971 ciphers, and SSLv2 56-bit DES are no longer available.
5976 * Fix a double-free in DSA code
5978 A double free bug was discovered when OpenSSL parses malformed DSA private
5979 keys and could lead to a DoS attack or memory corruption for applications
5980 that receive DSA private keys from untrusted sources. This scenario is
5983 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
5989 * Disable SRP fake user seed to address a server memory leak.
5991 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
5993 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
5994 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
5995 was changed to ignore the "fake user" SRP seed, even if the seed
5998 Users should use SRP_VBASE_get1_by_user instead. Note that in
5999 SRP_VBASE_get1_by_user, caller must free the returned value. Note
6000 also that even though configuring the SRP seed attempts to hide
6001 invalid usernames by continuing the handshake with fake
6002 credentials, this behaviour is not constant time and no strong
6003 guarantees are made that the handshake is indistinguishable from
6004 that of a valid user.
6009 * Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
6011 In the BN_hex2bn function the number of hex digits is calculated using an
6012 int value `i`. Later `bn_expand` is called with a value of `i * 4`. For
6013 large values of `i` this can result in `bn_expand` not allocating any
6014 memory because `i * 4` is negative. This can leave the internal BIGNUM data
6015 field as NULL leading to a subsequent NULL ptr deref. For very large values
6016 of `i`, the calculation `i * 4` could be a positive value smaller than `i`.
6017 In this case memory is allocated to the internal BIGNUM data field, but it
6018 is insufficiently sized leading to heap corruption. A similar issue exists
6019 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
6020 is ever called by user applications with very large untrusted hex/dec data.
6021 This is anticipated to be a rare occurrence.
6023 All OpenSSL internal usage of these functions use data that is not expected
6024 to be untrusted, e.g. config file data or application command line
6025 arguments. If user developed applications generate config file data based
6026 on untrusted data then it is possible that this could also lead to security
6027 consequences. This is also anticipated to be rare.
6029 This issue was reported to OpenSSL by Guido Vranken.
6034 * Fix memory issues in `BIO_*printf` functions
6036 The internal `fmtstr` function used in processing a "%s" format string in
6037 the `BIO_*printf` functions could overflow while calculating the length of a
6038 string and cause an OOB read when printing very long strings.
6040 Additionally the internal `doapr_outch` function can attempt to write to an
6041 OOB memory location (at an offset from the NULL pointer) in the event of a
6042 memory allocation failure. In 1.0.2 and below this could be caused where
6043 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
6044 could be in processing a very long "%s" format string. Memory leaks can
6047 The first issue may mask the second issue dependent on compiler behaviour.
6048 These problems could enable attacks where large amounts of untrusted data
6049 is passed to the `BIO_*printf` functions. If applications use these functions
6050 in this way then they could be vulnerable. OpenSSL itself uses these
6051 functions when printing out human-readable dumps of ASN.1 data. Therefore
6052 applications that print this data could be vulnerable if the data is from
6053 untrusted sources. OpenSSL command line applications could also be
6054 vulnerable where they print out ASN.1 data, or if untrusted data is passed
6055 as command line arguments.
6057 Libssl is not considered directly vulnerable. Additionally certificates etc
6058 received via remote connections via libssl are also unlikely to be able to
6059 trigger these issues because of message size limits enforced within libssl.
6061 This issue was reported to OpenSSL Guido Vranken.
6066 * Side channel attack on modular exponentiation
6068 A side-channel attack was found which makes use of cache-bank conflicts on
6069 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
6070 of RSA keys. The ability to exploit this issue is limited as it relies on
6071 an attacker who has control of code in a thread running on the same
6072 hyper-threaded core as the victim thread which is performing decryptions.
6074 This issue was reported to OpenSSL by Yuval Yarom, The University of
6075 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
6076 Nadia Heninger, University of Pennsylvania with more information at
6077 <http://cachebleed.info>.
6082 * Change the req app to generate a 2048-bit RSA/DSA key by default,
6083 if no keysize is specified with default_bits. This fixes an
6084 omission in an earlier change that changed all RSA/DSA key generation
6085 apps to use 2048 bits by default.
6089 ### Changes between 1.0.1q and 1.0.1r [28 Jan 2016]
6091 * Protection for DH small subgroup attacks
6093 As a precautionary measure the SSL_OP_SINGLE_DH_USE option has been
6094 switched on by default and cannot be disabled. This could have some
6099 * SSLv2 doesn't block disabled ciphers
6101 A malicious client can negotiate SSLv2 ciphers that have been disabled on
6102 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
6103 been disabled, provided that the SSLv2 protocol was not also disabled via
6106 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
6107 and Sebastian Schinzel.
6112 * Reject DH handshakes with parameters shorter than 1024 bits.
6116 ### Changes between 1.0.1p and 1.0.1q [3 Dec 2015]
6118 * Certificate verify crash with missing PSS parameter
6120 The signature verification routines will crash with a NULL pointer
6121 dereference if presented with an ASN.1 signature using the RSA PSS
6122 algorithm and absent mask generation function parameter. Since these
6123 routines are used to verify certificate signature algorithms this can be
6124 used to crash any certificate verification operation and exploited in a
6125 DoS attack. Any application which performs certificate verification is
6126 vulnerable including OpenSSL clients and servers which enable client
6129 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
6134 * X509_ATTRIBUTE memory leak
6136 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
6137 memory. This structure is used by the PKCS#7 and CMS routines so any
6138 application which reads PKCS#7 or CMS data from untrusted sources is
6139 affected. SSL/TLS is not affected.
6141 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
6147 * Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
6148 This changes the decoding behaviour for some invalid messages,
6149 though the change is mostly in the more lenient direction, and
6150 legacy behaviour is preserved as much as possible.
6154 * In DSA_generate_parameters_ex, if the provided seed is too short,
6155 use a random seed, as already documented.
6157 *Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>*
6159 ### Changes between 1.0.1o and 1.0.1p [9 Jul 2015]
6161 * Alternate chains certificate forgery
6163 During certificate verfification, OpenSSL will attempt to find an
6164 alternative certificate chain if the first attempt to build such a chain
6165 fails. An error in the implementation of this logic can mean that an
6166 attacker could cause certain checks on untrusted certificates to be
6167 bypassed, such as the CA flag, enabling them to use a valid leaf
6168 certificate to act as a CA and "issue" an invalid certificate.
6170 This issue was reported to OpenSSL by Adam Langley/David Benjamin
6176 * Race condition handling PSK identify hint
6178 If PSK identity hints are received by a multi-threaded client then
6179 the values are wrongly updated in the parent SSL_CTX structure. This can
6180 result in a race condition potentially leading to a double free of the
6186 ### Changes between 1.0.1n and 1.0.1o [12 Jun 2015]
6188 * Fix HMAC ABI incompatibility. The previous version introduced an ABI
6189 incompatibility in the handling of HMAC. The previous ABI has now been
6192 ### Changes between 1.0.1m and 1.0.1n [11 Jun 2015]
6194 * Malformed ECParameters causes infinite loop
6196 When processing an ECParameters structure OpenSSL enters an infinite loop
6197 if the curve specified is over a specially malformed binary polynomial
6200 This can be used to perform denial of service against any
6201 system which processes public keys, certificate requests or
6202 certificates. This includes TLS clients and TLS servers with
6203 client authentication enabled.
6205 This issue was reported to OpenSSL by Joseph Barr-Pixton.
6210 * Exploitable out-of-bounds read in X509_cmp_time
6212 X509_cmp_time does not properly check the length of the ASN1_TIME
6213 string and can read a few bytes out of bounds. In addition,
6214 X509_cmp_time accepts an arbitrary number of fractional seconds in the
6217 An attacker can use this to craft malformed certificates and CRLs of
6218 various sizes and potentially cause a segmentation fault, resulting in
6219 a DoS on applications that verify certificates or CRLs. TLS clients
6220 that verify CRLs are affected. TLS clients and servers with client
6221 authentication enabled may be affected if they use custom verification
6224 This issue was reported to OpenSSL by Robert Swiecki (Google), and
6225 independently by Hanno Böck.
6230 * PKCS7 crash with missing EnvelopedContent
6232 The PKCS#7 parsing code does not handle missing inner EncryptedContent
6233 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
6234 with missing content and trigger a NULL pointer dereference on parsing.
6236 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
6237 structures from untrusted sources are affected. OpenSSL clients and
6238 servers are not affected.
6240 This issue was reported to OpenSSL by Michal Zalewski (Google).
6245 * CMS verify infinite loop with unknown hash function
6247 When verifying a signedData message the CMS code can enter an infinite loop
6248 if presented with an unknown hash function OID. This can be used to perform
6249 denial of service against any system which verifies signedData messages using
6251 This issue was reported to OpenSSL by Johannes Bauer.
6256 * Race condition handling NewSessionTicket
6258 If a NewSessionTicket is received by a multi-threaded client when attempting to
6259 reuse a previous ticket then a race condition can occur potentially leading to
6260 a double free of the ticket data.
6265 * Reject DH handshakes with parameters shorter than 768 bits.
6267 *Kurt Roeckx and Emilia Kasper*
6269 * dhparam: generate 2048-bit parameters by default.
6271 *Kurt Roeckx and Emilia Kasper*
6273 ### Changes between 1.0.1l and 1.0.1m [19 Mar 2015]
6275 * Segmentation fault in ASN1_TYPE_cmp fix
6277 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
6278 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
6279 certificate signature algorithm consistency this can be used to crash any
6280 certificate verification operation and exploited in a DoS attack. Any
6281 application which performs certificate verification is vulnerable including
6282 OpenSSL clients and servers which enable client authentication.
6287 * ASN.1 structure reuse memory corruption fix
6289 Reusing a structure in ASN.1 parsing may allow an attacker to cause
6290 memory corruption via an invalid write. Such reuse is and has been
6291 strongly discouraged and is believed to be rare.
6293 Applications that parse structures containing CHOICE or ANY DEFINED BY
6294 components may be affected. Certificate parsing (d2i_X509 and related
6295 functions) are however not affected. OpenSSL clients and servers are
6301 * PKCS7 NULL pointer dereferences fix
6303 The PKCS#7 parsing code does not handle missing outer ContentInfo
6304 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
6305 missing content and trigger a NULL pointer dereference on parsing.
6307 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
6308 otherwise parse PKCS#7 structures from untrusted sources are
6309 affected. OpenSSL clients and servers are not affected.
6311 This issue was reported to OpenSSL by Michal Zalewski (Google).
6316 * DoS via reachable assert in SSLv2 servers fix
6318 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
6319 servers that both support SSLv2 and enable export cipher suites by sending
6320 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
6322 This issue was discovered by Sean Burford (Google) and Emilia Käsper
6323 (OpenSSL development team).
6328 * Use After Free following d2i_ECPrivatekey error fix
6330 A malformed EC private key file consumed via the d2i_ECPrivateKey function
6331 could cause a use after free condition. This, in turn, could cause a double
6332 free in several private key parsing functions (such as d2i_PrivateKey
6333 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
6334 for applications that receive EC private keys from untrusted
6335 sources. This scenario is considered rare.
6337 This issue was discovered by the BoringSSL project and fixed in their
6343 * X509_to_X509_REQ NULL pointer deref fix
6345 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
6346 the certificate key is invalid. This function is rarely used in practice.
6348 This issue was discovered by Brian Carpenter.
6353 * Removed the export ciphers from the DEFAULT ciphers
6357 ### Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
6359 * Build fixes for the Windows and OpenVMS platforms
6361 *Matt Caswell and Richard Levitte*
6363 ### Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
6365 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
6366 message can cause a segmentation fault in OpenSSL due to a NULL pointer
6367 dereference. This could lead to a Denial Of Service attack. Thanks to
6368 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
6373 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
6374 dtls1_buffer_record function under certain conditions. In particular this
6375 could occur if an attacker sent repeated DTLS records with the same
6376 sequence number but for the next epoch. The memory leak could be exploited
6377 by an attacker in a Denial of Service attack through memory exhaustion.
6378 Thanks to Chris Mueller for reporting this issue.
6383 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
6384 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
6385 method would be set to NULL which could later result in a NULL pointer
6386 dereference. Thanks to Frank Schmirler for reporting this issue.
6391 * Abort handshake if server key exchange message is omitted for ephemeral
6394 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
6395 reporting this issue.
6400 * Remove non-export ephemeral RSA code on client and server. This code
6401 violated the TLS standard by allowing the use of temporary RSA keys in
6402 non-export ciphersuites and could be used by a server to effectively
6403 downgrade the RSA key length used to a value smaller than the server
6404 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
6405 INRIA or reporting this issue.
6410 * Fixed issue where DH client certificates are accepted without verification.
6411 An OpenSSL server will accept a DH certificate for client authentication
6412 without the certificate verify message. This effectively allows a client to
6413 authenticate without the use of a private key. This only affects servers
6414 which trust a client certificate authority which issues certificates
6415 containing DH keys: these are extremely rare and hardly ever encountered.
6416 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
6422 * Ensure that the session ID context of an SSL is updated when its
6423 SSL_CTX is updated via SSL_set_SSL_CTX.
6425 The session ID context is typically set from the parent SSL_CTX,
6426 and can vary with the CTX.
6430 * Fix various certificate fingerprint issues.
6432 By using non-DER or invalid encodings outside the signed portion of a
6433 certificate the fingerprint can be changed without breaking the signature.
6434 Although no details of the signed portion of the certificate can be changed
6435 this can cause problems with some applications: e.g. those using the
6436 certificate fingerprint for blacklists.
6438 1. Reject signatures with non zero unused bits.
6440 If the BIT STRING containing the signature has non zero unused bits reject
6441 the signature. All current signature algorithms require zero unused bits.
6443 2. Check certificate algorithm consistency.
6445 Check the AlgorithmIdentifier inside TBS matches the one in the
6446 certificate signature. NB: this will result in signature failure
6447 errors for some broken certificates.
6449 Thanks to Konrad Kraszewski from Google for reporting this issue.
6451 3. Check DSA/ECDSA signatures use DER.
6453 Re-encode DSA/ECDSA signatures and compare with the original received
6454 signature. Return an error if there is a mismatch.
6456 This will reject various cases including garbage after signature
6457 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
6458 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
6459 (negative or with leading zeroes).
6461 Further analysis was conducted and fixes were developed by Stephen Henson
6462 of the OpenSSL core team.
6468 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
6469 results on some platforms, including x86_64. This bug occurs at random
6470 with a very low probability, and is not known to be exploitable in any
6471 way, though its exact impact is difficult to determine. Thanks to Pieter
6472 Wuille (Blockstream) who reported this issue and also suggested an initial
6473 fix. Further analysis was conducted by the OpenSSL development team and
6474 Adam Langley of Google. The final fix was developed by Andy Polyakov of
6475 the OpenSSL core team.
6480 * Do not resume sessions on the server if the negotiated protocol
6481 version does not match the session's version. Resuming with a different
6482 version, while not strictly forbidden by the RFC, is of questionable
6483 sanity and breaks all known clients.
6485 *David Benjamin, Emilia Käsper*
6487 * Tighten handling of the ChangeCipherSpec (CCS) message: reject
6488 early CCS messages during renegotiation. (Note that because
6489 renegotiation is encrypted, this early CCS was not exploitable.)
6493 * Tighten client-side session ticket handling during renegotiation:
6494 ensure that the client only accepts a session ticket if the server sends
6495 the extension anew in the ServerHello. Previously, a TLS client would
6496 reuse the old extension state and thus accept a session ticket if one was
6497 announced in the initial ServerHello.
6499 Similarly, ensure that the client requires a session ticket if one
6500 was advertised in the ServerHello. Previously, a TLS client would
6501 ignore a missing NewSessionTicket message.
6505 ### Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
6509 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
6510 sends a carefully crafted handshake message, to cause OpenSSL to fail
6511 to free up to 64k of memory causing a memory leak. This could be
6512 exploited in a Denial Of Service attack. This issue affects OpenSSL
6513 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
6514 whether SRTP is used or configured. Implementations of OpenSSL that
6515 have been compiled with OPENSSL_NO_SRTP defined are not affected.
6517 The fix was developed by the OpenSSL team.
6522 * Session Ticket Memory Leak.
6524 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
6525 integrity of that ticket is first verified. In the event of a session
6526 ticket integrity check failing, OpenSSL will fail to free memory
6527 causing a memory leak. By sending a large number of invalid session
6528 tickets an attacker could exploit this issue in a Denial Of Service
6534 * Build option no-ssl3 is incomplete.
6536 When OpenSSL is configured with "no-ssl3" as a build option, servers
6537 could accept and complete a SSL 3.0 handshake, and clients could be
6538 configured to send them.
6541 *Akamai and the OpenSSL team*
6543 * Add support for TLS_FALLBACK_SCSV.
6544 Client applications doing fallback retries should call
6545 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
6548 *Adam Langley, Bodo Moeller*
6550 * Add additional DigestInfo checks.
6552 Re-encode DigestInto in DER and check against the original when
6553 verifying RSA signature: this will reject any improperly encoded
6554 DigestInfo structures.
6556 Note: this is a precautionary measure and no attacks are currently known.
6560 ### Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
6562 * Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
6563 SRP code can be overrun an internal buffer. Add sanity check that
6564 g, A, B < N to SRP code.
6566 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
6567 Group for discovering this issue.
6572 * A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
6573 TLS 1.0 instead of higher protocol versions when the ClientHello message
6574 is badly fragmented. This allows a man-in-the-middle attacker to force a
6575 downgrade to TLS 1.0 even if both the server and the client support a
6576 higher protocol version, by modifying the client's TLS records.
6578 Thanks to David Benjamin and Adam Langley (Google) for discovering and
6579 researching this issue.
6584 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
6585 to a denial of service attack. A malicious server can crash the client
6586 with a null pointer dereference (read) by specifying an anonymous (EC)DH
6587 ciphersuite and sending carefully crafted handshake messages.
6589 Thanks to Felix Gröbert (Google) for discovering and researching this
6595 * By sending carefully crafted DTLS packets an attacker could cause openssl
6596 to leak memory. This can be exploited through a Denial of Service attack.
6597 Thanks to Adam Langley for discovering and researching this issue.
6602 * An attacker can force openssl to consume large amounts of memory whilst
6603 processing DTLS handshake messages. This can be exploited through a
6604 Denial of Service attack.
6605 Thanks to Adam Langley for discovering and researching this issue.
6610 * An attacker can force an error condition which causes openssl to crash
6611 whilst processing DTLS packets due to memory being freed twice. This
6612 can be exploited through a Denial of Service attack.
6613 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
6619 * If a multithreaded client connects to a malicious server using a resumed
6620 session and the server sends an ec point format extension it could write
6621 up to 255 bytes to freed memory.
6623 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
6629 * A malicious server can crash an OpenSSL client with a null pointer
6630 dereference (read) by specifying an SRP ciphersuite even though it was not
6631 properly negotiated with the client. This can be exploited through a
6632 Denial of Service attack.
6634 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
6635 discovering and researching this issue.
6640 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
6641 X509_name_oneline, X509_name_print_ex et al. to leak some information
6642 from the stack. Applications may be affected if they echo pretty printing
6643 output to the attacker.
6645 Thanks to Ivan Fratric (Google) for discovering this issue.
6648 *Emilia Käsper, and Steve Henson*
6650 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
6651 for corner cases. (Certain input points at infinity could lead to
6652 bogus results, with non-infinity inputs mapped to infinity too.)
6656 ### Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
6658 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
6659 handshake can force the use of weak keying material in OpenSSL
6660 SSL/TLS clients and servers.
6662 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
6663 researching this issue. ([CVE-2014-0224])
6665 *KIKUCHI Masashi, Steve Henson*
6667 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
6668 OpenSSL DTLS client the code can be made to recurse eventually crashing
6671 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
6674 *Imre Rad, Steve Henson*
6676 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
6677 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
6678 client or server. This is potentially exploitable to run arbitrary
6679 code on a vulnerable client or server.
6681 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
6683 *Jüri Aedla, Steve Henson*
6685 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
6686 are subject to a denial of service attack.
6688 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
6689 this issue. ([CVE-2014-3470])
6691 *Felix Gröbert, Ivan Fratric, Steve Henson*
6693 * Harmonize version and its documentation. -f flag is used to display
6696 *mancha <mancha1@zoho.com>*
6698 * Fix eckey_priv_encode so it immediately returns an error upon a failure
6699 in i2d_ECPrivateKey.
6701 *mancha <mancha1@zoho.com>*
6703 * Fix some double frees. These are not thought to be exploitable.
6705 *mancha <mancha1@zoho.com>*
6707 ### Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
6709 * A missing bounds check in the handling of the TLS heartbeat extension
6710 can be used to reveal up to 64k of memory to a connected client or
6713 Thanks for Neel Mehta of Google Security for discovering this bug and to
6714 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
6715 preparing the fix ([CVE-2014-0160])
6717 *Adam Langley, Bodo Moeller*
6719 * Fix for the attack described in the paper "Recovering OpenSSL
6720 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
6721 by Yuval Yarom and Naomi Benger. Details can be obtained from:
6722 <http://eprint.iacr.org/2014/140>
6724 Thanks to Yuval Yarom and Naomi Benger for discovering this
6725 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
6727 *Yuval Yarom and Naomi Benger*
6729 * TLS pad extension: draft-agl-tls-padding-03
6731 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
6732 TLS client Hello record length value would otherwise be > 255 and
6733 less that 512 pad with a dummy extension containing zeroes so it
6734 is at least 512 bytes long.
6736 *Adam Langley, Steve Henson*
6738 ### Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
6740 * Fix for TLS record tampering bug. A carefully crafted invalid
6741 handshake could crash OpenSSL with a NULL pointer exception.
6742 Thanks to Anton Johansson for reporting this issues.
6745 * Keep original DTLS digest and encryption contexts in retransmission
6746 structures so we can use the previous session parameters if they need
6747 to be resent. ([CVE-2013-6450])
6751 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
6752 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
6753 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
6754 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
6755 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
6756 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
6758 *Rob Stradling, Adam Langley*
6760 ### Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
6762 * Correct fix for CVE-2013-0169. The original didn't work on AES-NI
6763 supporting platforms or when small records were transferred.
6765 *Andy Polyakov, Steve Henson*
6767 ### Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
6769 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
6771 This addresses the flaw in CBC record processing discovered by
6772 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
6773 at: <http://www.isg.rhul.ac.uk/tls/>
6775 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
6776 Security Group at Royal Holloway, University of London
6777 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
6778 Emilia Käsper for the initial patch.
6781 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
6783 * Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
6784 ciphersuites which can be exploited in a denial of service attack.
6785 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
6786 and detecting this bug and to Wolfgang Ettlinger
6787 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
6792 * Return an error when checking OCSP signatures when key is NULL.
6793 This fixes a DoS attack. ([CVE-2013-0166])
6797 * Make openssl verify return errors.
6799 *Chris Palmer <palmer@google.com> and Ben Laurie*
6801 * Call OCSP Stapling callback after ciphersuite has been chosen, so
6802 the right response is stapled. Also change SSL_get_certificate()
6803 so it returns the certificate actually sent.
6804 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
6806 *Rob Stradling <rob.stradling@comodo.com>*
6808 * Fix possible deadlock when decoding public keys.
6812 * Don't use TLS 1.0 record version number in initial client hello
6817 ### Changes between 1.0.1b and 1.0.1c [10 May 2012]
6819 * Sanity check record length before skipping explicit IV in TLS
6820 1.2, 1.1 and DTLS to fix DoS attack.
6822 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
6823 fuzzing as a service testing platform.
6828 * Initialise tkeylen properly when encrypting CMS messages.
6829 Thanks to Solar Designer of Openwall for reporting this issue.
6833 * In FIPS mode don't try to use composite ciphers as they are not
6838 ### Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
6840 * OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
6841 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
6842 mean any application compiled against OpenSSL 1.0.0 headers setting
6843 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
6844 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
6845 0x10000000L Any application which was previously compiled against
6846 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
6847 will need to be recompiled as a result. Letting be results in
6848 inability to disable specifically TLS 1.1 and in client context,
6849 in unlike event, limit maximum offered version to TLS 1.0 [see below].
6853 * In order to ensure interoperability SSL_OP_NO_protocolX does not
6854 disable just protocol X, but all protocols above X *if* there are
6855 protocols *below* X still enabled. In more practical terms it means
6856 that if application wants to disable TLS1.0 in favor of TLS1.1 and
6857 above, it's not sufficient to pass `SSL_OP_NO_TLSv1`, one has to pass
6858 `SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2`. This applies to
6863 ### Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
6865 * Check for potentially exploitable overflows in asn1_d2i_read_bio
6866 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
6867 in CRYPTO_realloc_clean.
6869 Thanks to Tavis Ormandy, Google Security Team, for discovering this
6870 issue and to Adam Langley <agl@chromium.org> for fixing it.
6873 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
6875 * Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
6879 * Workarounds for some broken servers that "hang" if a client hello
6880 record length exceeds 255 bytes.
6882 1. Do not use record version number > TLS 1.0 in initial client
6883 hello: some (but not all) hanging servers will now work.
6884 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
6885 the number of ciphers sent in the client hello. This should be
6886 set to an even number, such as 50, for example by passing:
6887 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
6888 Most broken servers should now work.
6889 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
6890 TLS 1.2 client support entirely.
6894 * Fix SEGV in Vector Permutation AES module observed in OpenSSH.
6898 ### Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
6900 * Add compatibility with old MDC2 signatures which use an ASN1 OCTET
6901 STRING form instead of a DigestInfo.
6905 * The format used for MDC2 RSA signatures is inconsistent between EVP
6906 and the RSA_sign/RSA_verify functions. This was made more apparent when
6907 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
6908 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
6909 the correct format in RSA_verify so both forms transparently work.
6913 * Some servers which support TLS 1.0 can choke if we initially indicate
6914 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
6915 encrypted premaster secret. As a workaround use the maximum permitted
6916 client version in client hello, this should keep such servers happy
6917 and still work with previous versions of OpenSSL.
6921 * Add support for TLS/DTLS heartbeats.
6923 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6925 * Add support for SCTP.
6927 *Robin Seggelmann <seggelmann@fh-muenster.de>*
6929 * Improved PRNG seeding for VOS.
6931 *Paul Green <Paul.Green@stratus.com>*
6933 * Extensive assembler packs updates, most notably:
6935 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
6936 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
6937 - x86_64: bit-sliced AES implementation;
6938 - ARM: NEON support, contemporary platforms optimizations;
6939 - s390x: z196 support;
6940 - `*`: GHASH and GF(2^m) multiplication implementations;
6944 * Make TLS-SRP code conformant with RFC 5054 API cleanup
6945 (removal of unnecessary code)
6947 *Peter Sylvester <peter.sylvester@edelweb.fr>*
6949 * Add TLS key material exporter from RFC 5705.
6953 * Add DTLS-SRTP negotiation from RFC 5764.
6957 * Add Next Protocol Negotiation,
6958 <http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00>. Can be
6959 disabled with a no-npn flag to config or Configure. Code donated
6962 *Adam Langley <agl@google.com> and Ben Laurie*
6964 * Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
6965 NIST-P256, NIST-P521, with constant-time single point multiplication on
6966 typical inputs. Compiler support for the nonstandard type `__uint128_t` is
6967 required to use this (present in gcc 4.4 and later, for 64-bit builds).
6968 Code made available under Apache License version 2.0.
6970 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
6971 line to include this in your build of OpenSSL, and run "make depend" (or
6972 "make update"). This enables the following EC_METHODs:
6974 EC_GFp_nistp224_method()
6975 EC_GFp_nistp256_method()
6976 EC_GFp_nistp521_method()
6978 EC_GROUP_new_by_curve_name() will automatically use these (while
6979 EC_GROUP_new_curve_GFp() currently prefers the more flexible
6982 *Emilia Käsper, Adam Langley, Bodo Moeller (Google)*
6984 * Use type ossl_ssize_t instead of ssize_t which isn't available on
6985 all platforms. Move ssize_t definition from e_os.h to the public
6986 header file e_os2.h as it now appears in public header file cms.h
6990 * New -sigopt option to the ca, req and x509 utilities. Additional
6991 signature parameters can be passed using this option and in
6996 * Add RSA PSS signing function. This will generate and set the
6997 appropriate AlgorithmIdentifiers for PSS based on those in the
6998 corresponding EVP_MD_CTX structure. No application support yet.
7002 * Support for companion algorithm specific ASN1 signing routines.
7003 New function ASN1_item_sign_ctx() signs a pre-initialised
7004 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
7005 the appropriate parameters.
7009 * Add new algorithm specific ASN1 verification initialisation function
7010 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
7011 handling will be the same no matter what EVP_PKEY_METHOD is used.
7012 Add a PSS handler to support verification of PSS signatures: checked
7013 against a number of sample certificates.
7017 * Add signature printing for PSS. Add PSS OIDs.
7019 *Steve Henson, Martin Kaiser <lists@kaiser.cx>*
7021 * Add algorithm specific signature printing. An individual ASN1 method
7022 can now print out signatures instead of the standard hex dump.
7024 More complex signatures (e.g. PSS) can print out more meaningful
7025 information. Include DSA version that prints out the signature
7030 * Password based recipient info support for CMS library: implementing
7035 * Split password based encryption into PBES2 and PBKDF2 functions. This
7036 neatly separates the code into cipher and PBE sections and is required
7037 for some algorithms that split PBES2 into separate pieces (such as
7038 password based CMS).
7042 * Session-handling fixes:
7043 - Fix handling of connections that are resuming with a session ID,
7044 but also support Session Tickets.
7045 - Fix a bug that suppressed issuing of a new ticket if the client
7046 presented a ticket with an expired session.
7047 - Try to set the ticket lifetime hint to something reasonable.
7048 - Make tickets shorter by excluding irrelevant information.
7049 - On the client side, don't ignore renewed tickets.
7051 *Adam Langley, Bodo Moeller (Google)*
7053 * Fix PSK session representation.
7057 * Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
7059 This work was sponsored by Intel.
7063 * Add GCM support to TLS library. Some custom code is needed to split
7064 the IV between the fixed (from PRF) and explicit (from TLS record)
7065 portions. This adds all GCM ciphersuites supported by RFC5288 and
7066 RFC5289. Generalise some `AES*` cipherstrings to include GCM and
7067 add a special AESGCM string for GCM only.
7071 * Expand range of ctrls for AES GCM. Permit setting invocation
7072 field on decrypt and retrieval of invocation field only on encrypt.
7076 * Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
7077 As required by RFC5289 these ciphersuites cannot be used if for
7078 versions of TLS earlier than 1.2.
7082 * For FIPS capable OpenSSL interpret a NULL default public key method
7083 as unset and return the appropriate default but do *not* set the default.
7084 This means we can return the appropriate method in applications that
7085 switch between FIPS and non-FIPS modes.
7089 * Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
7090 ENGINE is used then we cannot handle that in the FIPS module so we
7091 keep original code iff non-FIPS operations are allowed.
7095 * Add -attime option to openssl utilities.
7097 *Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson*
7099 * Redirect DSA and DH operations to FIPS module in FIPS mode.
7103 * Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
7104 FIPS EC methods unconditionally for now.
7108 * New build option no-ec2m to disable characteristic 2 code.
7112 * Backport libcrypto audit of return value checking from 1.1.0-dev; not
7113 all cases can be covered as some introduce binary incompatibilities.
7117 * Redirect RSA operations to FIPS module including keygen,
7118 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
7122 * Add similar low level API blocking to ciphers.
7126 * Low level digest APIs are not approved in FIPS mode: any attempt
7127 to use these will cause a fatal error. Applications that *really* want
7128 to use them can use the `private_*` version instead.
7132 * Redirect cipher operations to FIPS module for FIPS builds.
7136 * Redirect digest operations to FIPS module for FIPS builds.
7140 * Update build system to add "fips" flag which will link in fipscanister.o
7141 for static and shared library builds embedding a signature if needed.
7145 * Output TLS supported curves in preference order instead of numerical
7146 order. This is currently hardcoded for the highest order curves first.
7147 This should be configurable so applications can judge speed vs strength.
7151 * Add TLS v1.2 server support for client authentication.
7155 * Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
7160 * Functions FIPS_mode_set() and FIPS_mode() which call the underlying
7161 FIPS modules versions.
7165 * Add TLS v1.2 client side support for client authentication. Keep cache
7166 of handshake records longer as we don't know the hash algorithm to use
7167 until after the certificate request message is received.
7171 * Initial TLS v1.2 client support. Add a default signature algorithms
7172 extension including all the algorithms we support. Parse new signature
7173 format in client key exchange. Relax some ECC signing restrictions for
7174 TLS v1.2 as indicated in RFC5246.
7178 * Add server support for TLS v1.2 signature algorithms extension. Switch
7179 to new signature format when needed using client digest preference.
7180 All server ciphersuites should now work correctly in TLS v1.2. No client
7181 support yet and no support for client certificates.
7185 * Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
7186 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
7187 ciphersuites. At present only RSA key exchange ciphersuites work with
7188 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
7189 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
7190 and version checking.
7194 * New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
7195 with this defined it will not be affected by any changes to ssl internal
7196 structures. Add several utility functions to allow openssl application
7197 to work with OPENSSL_NO_SSL_INTERN defined.
7201 * A long standing patch to add support for SRP from EdelWeb (Peter
7202 Sylvester and Christophe Renou) was integrated.
7203 *Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
7204 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
7207 * Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
7211 * Permit abbreviated handshakes when renegotiating using the function
7212 SSL_renegotiate_abbreviated().
7214 *Robin Seggelmann <seggelmann@fh-muenster.de>*
7216 * Add call to ENGINE_register_all_complete() to
7217 ENGINE_load_builtin_engines(), so some implementations get used
7218 automatically instead of needing explicit application support.
7222 * Add support for TLS key exporter as described in RFC5705.
7224 *Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson*
7226 * Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
7227 a few changes are required:
7229 Add SSL_OP_NO_TLSv1_1 flag.
7230 Add TLSv1_1 methods.
7231 Update version checking logic to handle version 1.1.
7232 Add explicit IV handling (ported from DTLS code).
7233 Add command line options to s_client/s_server.
7240 ### Changes between 1.0.0s and 1.0.0t [3 Dec 2015]
7242 * X509_ATTRIBUTE memory leak
7244 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
7245 memory. This structure is used by the PKCS#7 and CMS routines so any
7246 application which reads PKCS#7 or CMS data from untrusted sources is
7247 affected. SSL/TLS is not affected.
7249 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
7255 * Race condition handling PSK identify hint
7257 If PSK identity hints are received by a multi-threaded client then
7258 the values are wrongly updated in the parent SSL_CTX structure. This can
7259 result in a race condition potentially leading to a double free of the
7265 ### Changes between 1.0.0r and 1.0.0s [11 Jun 2015]
7267 * Malformed ECParameters causes infinite loop
7269 When processing an ECParameters structure OpenSSL enters an infinite loop
7270 if the curve specified is over a specially malformed binary polynomial
7273 This can be used to perform denial of service against any
7274 system which processes public keys, certificate requests or
7275 certificates. This includes TLS clients and TLS servers with
7276 client authentication enabled.
7278 This issue was reported to OpenSSL by Joseph Barr-Pixton.
7283 * Exploitable out-of-bounds read in X509_cmp_time
7285 X509_cmp_time does not properly check the length of the ASN1_TIME
7286 string and can read a few bytes out of bounds. In addition,
7287 X509_cmp_time accepts an arbitrary number of fractional seconds in the
7290 An attacker can use this to craft malformed certificates and CRLs of
7291 various sizes and potentially cause a segmentation fault, resulting in
7292 a DoS on applications that verify certificates or CRLs. TLS clients
7293 that verify CRLs are affected. TLS clients and servers with client
7294 authentication enabled may be affected if they use custom verification
7297 This issue was reported to OpenSSL by Robert Swiecki (Google), and
7298 independently by Hanno Böck.
7303 * PKCS7 crash with missing EnvelopedContent
7305 The PKCS#7 parsing code does not handle missing inner EncryptedContent
7306 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
7307 with missing content and trigger a NULL pointer dereference on parsing.
7309 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
7310 structures from untrusted sources are affected. OpenSSL clients and
7311 servers are not affected.
7313 This issue was reported to OpenSSL by Michal Zalewski (Google).
7318 * CMS verify infinite loop with unknown hash function
7320 When verifying a signedData message the CMS code can enter an infinite loop
7321 if presented with an unknown hash function OID. This can be used to perform
7322 denial of service against any system which verifies signedData messages using
7324 This issue was reported to OpenSSL by Johannes Bauer.
7329 * Race condition handling NewSessionTicket
7331 If a NewSessionTicket is received by a multi-threaded client when attempting to
7332 reuse a previous ticket then a race condition can occur potentially leading to
7333 a double free of the ticket data.
7338 ### Changes between 1.0.0q and 1.0.0r [19 Mar 2015]
7340 * Segmentation fault in ASN1_TYPE_cmp fix
7342 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
7343 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
7344 certificate signature algorithm consistency this can be used to crash any
7345 certificate verification operation and exploited in a DoS attack. Any
7346 application which performs certificate verification is vulnerable including
7347 OpenSSL clients and servers which enable client authentication.
7352 * ASN.1 structure reuse memory corruption fix
7354 Reusing a structure in ASN.1 parsing may allow an attacker to cause
7355 memory corruption via an invalid write. Such reuse is and has been
7356 strongly discouraged and is believed to be rare.
7358 Applications that parse structures containing CHOICE or ANY DEFINED BY
7359 components may be affected. Certificate parsing (d2i_X509 and related
7360 functions) are however not affected. OpenSSL clients and servers are
7366 * PKCS7 NULL pointer dereferences fix
7368 The PKCS#7 parsing code does not handle missing outer ContentInfo
7369 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
7370 missing content and trigger a NULL pointer dereference on parsing.
7372 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
7373 otherwise parse PKCS#7 structures from untrusted sources are
7374 affected. OpenSSL clients and servers are not affected.
7376 This issue was reported to OpenSSL by Michal Zalewski (Google).
7381 * DoS via reachable assert in SSLv2 servers fix
7383 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
7384 servers that both support SSLv2 and enable export cipher suites by sending
7385 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
7387 This issue was discovered by Sean Burford (Google) and Emilia Käsper
7388 (OpenSSL development team).
7393 * Use After Free following d2i_ECPrivatekey error fix
7395 A malformed EC private key file consumed via the d2i_ECPrivateKey function
7396 could cause a use after free condition. This, in turn, could cause a double
7397 free in several private key parsing functions (such as d2i_PrivateKey
7398 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
7399 for applications that receive EC private keys from untrusted
7400 sources. This scenario is considered rare.
7402 This issue was discovered by the BoringSSL project and fixed in their
7408 * X509_to_X509_REQ NULL pointer deref fix
7410 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
7411 the certificate key is invalid. This function is rarely used in practice.
7413 This issue was discovered by Brian Carpenter.
7418 * Removed the export ciphers from the DEFAULT ciphers
7422 ### Changes between 1.0.0p and 1.0.0q [15 Jan 2015]
7424 * Build fixes for the Windows and OpenVMS platforms
7426 *Matt Caswell and Richard Levitte*
7428 ### Changes between 1.0.0o and 1.0.0p [8 Jan 2015]
7430 * Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
7431 message can cause a segmentation fault in OpenSSL due to a NULL pointer
7432 dereference. This could lead to a Denial Of Service attack. Thanks to
7433 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
7438 * Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
7439 dtls1_buffer_record function under certain conditions. In particular this
7440 could occur if an attacker sent repeated DTLS records with the same
7441 sequence number but for the next epoch. The memory leak could be exploited
7442 by an attacker in a Denial of Service attack through memory exhaustion.
7443 Thanks to Chris Mueller for reporting this issue.
7448 * Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
7449 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
7450 method would be set to NULL which could later result in a NULL pointer
7451 dereference. Thanks to Frank Schmirler for reporting this issue.
7456 * Abort handshake if server key exchange message is omitted for ephemeral
7459 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
7460 reporting this issue.
7465 * Remove non-export ephemeral RSA code on client and server. This code
7466 violated the TLS standard by allowing the use of temporary RSA keys in
7467 non-export ciphersuites and could be used by a server to effectively
7468 downgrade the RSA key length used to a value smaller than the server
7469 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
7470 INRIA or reporting this issue.
7475 * Fixed issue where DH client certificates are accepted without verification.
7476 An OpenSSL server will accept a DH certificate for client authentication
7477 without the certificate verify message. This effectively allows a client to
7478 authenticate without the use of a private key. This only affects servers
7479 which trust a client certificate authority which issues certificates
7480 containing DH keys: these are extremely rare and hardly ever encountered.
7481 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
7487 * Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
7488 results on some platforms, including x86_64. This bug occurs at random
7489 with a very low probability, and is not known to be exploitable in any
7490 way, though its exact impact is difficult to determine. Thanks to Pieter
7491 Wuille (Blockstream) who reported this issue and also suggested an initial
7492 fix. Further analysis was conducted by the OpenSSL development team and
7493 Adam Langley of Google. The final fix was developed by Andy Polyakov of
7494 the OpenSSL core team.
7499 * Fix various certificate fingerprint issues.
7501 By using non-DER or invalid encodings outside the signed portion of a
7502 certificate the fingerprint can be changed without breaking the signature.
7503 Although no details of the signed portion of the certificate can be changed
7504 this can cause problems with some applications: e.g. those using the
7505 certificate fingerprint for blacklists.
7507 1. Reject signatures with non zero unused bits.
7509 If the BIT STRING containing the signature has non zero unused bits reject
7510 the signature. All current signature algorithms require zero unused bits.
7512 2. Check certificate algorithm consistency.
7514 Check the AlgorithmIdentifier inside TBS matches the one in the
7515 certificate signature. NB: this will result in signature failure
7516 errors for some broken certificates.
7518 Thanks to Konrad Kraszewski from Google for reporting this issue.
7520 3. Check DSA/ECDSA signatures use DER.
7522 Reencode DSA/ECDSA signatures and compare with the original received
7523 signature. Return an error if there is a mismatch.
7525 This will reject various cases including garbage after signature
7526 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
7527 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
7528 (negative or with leading zeroes).
7530 Further analysis was conducted and fixes were developed by Stephen Henson
7531 of the OpenSSL core team.
7537 ### Changes between 1.0.0n and 1.0.0o [15 Oct 2014]
7539 * Session Ticket Memory Leak.
7541 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
7542 integrity of that ticket is first verified. In the event of a session
7543 ticket integrity check failing, OpenSSL will fail to free memory
7544 causing a memory leak. By sending a large number of invalid session
7545 tickets an attacker could exploit this issue in a Denial Of Service
7551 * Build option no-ssl3 is incomplete.
7553 When OpenSSL is configured with "no-ssl3" as a build option, servers
7554 could accept and complete a SSL 3.0 handshake, and clients could be
7555 configured to send them.
7558 *Akamai and the OpenSSL team*
7560 * Add support for TLS_FALLBACK_SCSV.
7561 Client applications doing fallback retries should call
7562 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
7565 *Adam Langley, Bodo Moeller*
7567 * Add additional DigestInfo checks.
7569 Reencode DigestInto in DER and check against the original when
7570 verifying RSA signature: this will reject any improperly encoded
7571 DigestInfo structures.
7573 Note: this is a precautionary measure and no attacks are currently known.
7577 ### Changes between 1.0.0m and 1.0.0n [6 Aug 2014]
7579 * OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
7580 to a denial of service attack. A malicious server can crash the client
7581 with a null pointer dereference (read) by specifying an anonymous (EC)DH
7582 ciphersuite and sending carefully crafted handshake messages.
7584 Thanks to Felix Gröbert (Google) for discovering and researching this
7590 * By sending carefully crafted DTLS packets an attacker could cause openssl
7591 to leak memory. This can be exploited through a Denial of Service attack.
7592 Thanks to Adam Langley for discovering and researching this issue.
7597 * An attacker can force openssl to consume large amounts of memory whilst
7598 processing DTLS handshake messages. This can be exploited through a
7599 Denial of Service attack.
7600 Thanks to Adam Langley for discovering and researching this issue.
7605 * An attacker can force an error condition which causes openssl to crash
7606 whilst processing DTLS packets due to memory being freed twice. This
7607 can be exploited through a Denial of Service attack.
7608 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
7614 * If a multithreaded client connects to a malicious server using a resumed
7615 session and the server sends an ec point format extension it could write
7616 up to 255 bytes to freed memory.
7618 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
7624 * A flaw in OBJ_obj2txt may cause pretty printing functions such as
7625 X509_name_oneline, X509_name_print_ex et al. to leak some information
7626 from the stack. Applications may be affected if they echo pretty printing
7627 output to the attacker.
7629 Thanks to Ivan Fratric (Google) for discovering this issue.
7632 *Emilia Käsper, and Steve Henson*
7634 * Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
7635 for corner cases. (Certain input points at infinity could lead to
7636 bogus results, with non-infinity inputs mapped to infinity too.)
7640 ### Changes between 1.0.0l and 1.0.0m [5 Jun 2014]
7642 * Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
7643 handshake can force the use of weak keying material in OpenSSL
7644 SSL/TLS clients and servers.
7646 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
7647 researching this issue. ([CVE-2014-0224])
7649 *KIKUCHI Masashi, Steve Henson*
7651 * Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
7652 OpenSSL DTLS client the code can be made to recurse eventually crashing
7655 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
7658 *Imre Rad, Steve Henson*
7660 * Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
7661 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
7662 client or server. This is potentially exploitable to run arbitrary
7663 code on a vulnerable client or server.
7665 Thanks to Jüri Aedla for reporting this issue. ([CVE-2014-0195])
7667 *Jüri Aedla, Steve Henson*
7669 * Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
7670 are subject to a denial of service attack.
7672 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
7673 this issue. ([CVE-2014-3470])
7675 *Felix Gröbert, Ivan Fratric, Steve Henson*
7677 * Harmonize version and its documentation. -f flag is used to display
7680 *mancha <mancha1@zoho.com>*
7682 * Fix eckey_priv_encode so it immediately returns an error upon a failure
7683 in i2d_ECPrivateKey.
7685 *mancha <mancha1@zoho.com>*
7687 * Fix some double frees. These are not thought to be exploitable.
7689 *mancha <mancha1@zoho.com>*
7691 * Fix for the attack described in the paper "Recovering OpenSSL
7692 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
7693 by Yuval Yarom and Naomi Benger. Details can be obtained from:
7694 <http://eprint.iacr.org/2014/140>
7696 Thanks to Yuval Yarom and Naomi Benger for discovering this
7697 flaw and to Yuval Yarom for supplying a fix ([CVE-2014-0076])
7699 *Yuval Yarom and Naomi Benger*
7701 ### Changes between 1.0.0k and 1.0.0l [6 Jan 2014]
7703 * Keep original DTLS digest and encryption contexts in retransmission
7704 structures so we can use the previous session parameters if they need
7705 to be resent. ([CVE-2013-6450])
7709 * Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
7710 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
7711 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
7712 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
7713 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
7714 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
7716 *Rob Stradling, Adam Langley*
7718 ### Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
7720 * Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
7722 This addresses the flaw in CBC record processing discovered by
7723 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
7724 at: <http://www.isg.rhul.ac.uk/tls/>
7726 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7727 Security Group at Royal Holloway, University of London
7728 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
7729 Emilia Käsper for the initial patch.
7732 *Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson*
7734 * Return an error when checking OCSP signatures when key is NULL.
7735 This fixes a DoS attack. ([CVE-2013-0166])
7739 * Call OCSP Stapling callback after ciphersuite has been chosen, so
7740 the right response is stapled. Also change SSL_get_certificate()
7741 so it returns the certificate actually sent.
7742 See <http://rt.openssl.org/Ticket/Display.html?id=2836>.
7743 (This is a backport)
7745 *Rob Stradling <rob.stradling@comodo.com>*
7747 * Fix possible deadlock when decoding public keys.
7751 ### Changes between 1.0.0i and 1.0.0j [10 May 2012]
7753 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
7756 * Sanity check record length before skipping explicit IV in DTLS
7759 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
7760 fuzzing as a service testing platform.
7765 * Initialise tkeylen properly when encrypting CMS messages.
7766 Thanks to Solar Designer of Openwall for reporting this issue.
7770 ### Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
7772 * Check for potentially exploitable overflows in asn1_d2i_read_bio
7773 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
7774 in CRYPTO_realloc_clean.
7776 Thanks to Tavis Ormandy, Google Security Team, for discovering this
7777 issue and to Adam Langley <agl@chromium.org> for fixing it.
7780 *Adam Langley (Google), Tavis Ormandy, Google Security Team*
7782 ### Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
7784 * Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
7785 in CMS and PKCS7 code. When RSA decryption fails use a random key for
7786 content decryption and always return the same error. Note: this attack
7787 needs on average 2^20 messages so it only affects automated senders. The
7788 old behaviour can be re-enabled in the CMS code by setting the
7789 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
7790 an MMA defence is not necessary.
7791 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
7792 this issue. ([CVE-2012-0884])
7796 * Fix CVE-2011-4619: make sure we really are receiving a
7797 client hello before rejecting multiple SGC restarts. Thanks to
7798 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
7802 ### Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
7804 * Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
7805 Thanks to Antonio Martin, Enterprise Secure Access Research and
7806 Development, Cisco Systems, Inc. for discovering this bug and
7807 preparing a fix. ([CVE-2012-0050])
7811 ### Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
7813 * Nadhem Alfardan and Kenny Paterson have discovered an extension
7814 of the Vaudenay padding oracle attack on CBC mode encryption
7815 which enables an efficient plaintext recovery attack against
7816 the OpenSSL implementation of DTLS. Their attack exploits timing
7817 differences arising during decryption processing. A research
7818 paper describing this attack can be found at:
7819 <http://www.isg.rhul.ac.uk/~kp/dtls.pdf>
7820 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
7821 Security Group at Royal Holloway, University of London
7822 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
7823 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
7824 for preparing the fix. ([CVE-2011-4108])
7826 *Robin Seggelmann, Michael Tuexen*
7828 * Clear bytes used for block padding of SSL 3.0 records.
7831 *Adam Langley (Google)*
7833 * Only allow one SGC handshake restart for SSL/TLS. Thanks to George
7834 Kadianakis <desnacked@gmail.com> for discovering this issue and
7835 Adam Langley for preparing the fix. ([CVE-2011-4619])
7837 *Adam Langley (Google)*
7839 * Check parameters are not NULL in GOST ENGINE. ([CVE-2012-0027])
7841 *Andrey Kulikov <amdeich@gmail.com>*
7843 * Prevent malformed RFC3779 data triggering an assertion failure.
7844 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
7845 and Rob Austein <sra@hactrn.net> for fixing it. ([CVE-2011-4577])
7847 *Rob Austein <sra@hactrn.net>*
7849 * Improved PRNG seeding for VOS.
7851 *Paul Green <Paul.Green@stratus.com>*
7853 * Fix ssl_ciph.c set-up race.
7855 *Adam Langley (Google)*
7857 * Fix spurious failures in ecdsatest.c.
7859 *Emilia Käsper (Google)*
7861 * Fix the BIO_f_buffer() implementation (which was mixing different
7862 interpretations of the `..._len` fields).
7864 *Adam Langley (Google)*
7866 * Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
7867 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
7868 threads won't reuse the same blinding coefficients.
7870 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
7871 lock to call BN_BLINDING_invert_ex, and avoids one use of
7872 BN_BLINDING_update for each BN_BLINDING structure (previously,
7873 the last update always remained unused).
7875 *Emilia Käsper (Google)*
7877 * In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
7879 *Bob Buckholz (Google)*
7881 ### Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
7883 * Fix bug where CRLs with nextUpdate in the past are sometimes accepted
7884 by initialising X509_STORE_CTX properly. ([CVE-2011-3207])
7886 *Kaspar Brand <ossl@velox.ch>*
7888 * Fix SSL memory handling for (EC)DH ciphersuites, in particular
7889 for multi-threaded use of ECDH. ([CVE-2011-3210])
7891 *Adam Langley (Google)*
7893 * Fix x509_name_ex_d2i memory leak on bad inputs.
7897 * Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
7898 signature public key algorithm by using OID xref utilities instead.
7899 Before this you could only use some ECC ciphersuites with SHA1 only.
7903 * Add protection against ECDSA timing attacks as mentioned in the paper
7904 by Billy Bob Brumley and Nicola Tuveri, see:
7905 <http://eprint.iacr.org/2011/232.pdf>
7907 *Billy Bob Brumley and Nicola Tuveri*
7909 ### Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
7911 * Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
7913 *Neel Mehta, Adam Langley, Bodo Moeller (Google)*
7915 * Fix bug in string printing code: if *any* escaping is enabled we must
7916 escape the escape character (backslash) or the resulting string is
7921 ### Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
7923 * Disable code workaround for ancient and obsolete Netscape browsers
7924 and servers: an attacker can use it in a ciphersuite downgrade attack.
7925 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
7929 * Fixed J-PAKE implementation error, originally discovered by
7930 Sebastien Martini, further info and confirmation from Stefan
7931 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
7935 ### Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
7937 * Fix extension code to avoid race conditions which can result in a buffer
7938 overrun vulnerability: resumed sessions must not be modified as they can
7939 be shared by multiple threads. CVE-2010-3864
7943 * Fix WIN32 build system to correctly link an ENGINE directory into
7948 ### Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
7950 * Check return value of int_rsa_verify in pkey_rsa_verifyrecover
7953 *Steve Henson, Peter-Michael Hager <hager@dortmund.net>*
7955 ### Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
7957 * Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
7958 context. The operation can be customised via the ctrl mechanism in
7959 case ENGINEs want to include additional functionality.
7963 * Tolerate yet another broken PKCS#8 key format: private key value negative.
7967 * Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
7968 output hashes compatible with older versions of OpenSSL.
7970 *Willy Weisz <weisz@vcpc.univie.ac.at>*
7972 * Fix compression algorithm handling: if resuming a session use the
7973 compression algorithm of the resumed session instead of determining
7974 it from client hello again. Don't allow server to change algorithm.
7978 * Add load_crls() function to apps tidying load_certs() too. Add option
7979 to verify utility to allow additional CRLs to be included.
7983 * Update OCSP request code to permit adding custom headers to the request:
7984 some responders need this.
7988 * The function EVP_PKEY_sign() returns <=0 on error: check return code
7991 *Julia Lawall <julia@diku.dk>*
7993 * Update verify callback code in apps/s_cb.c and apps/verify.c, it
7994 needlessly dereferenced structures, used obsolete functions and
7995 didn't handle all updated verify codes correctly.
7999 * Disable MD2 in the default configuration.
8003 * In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
8004 indicate the initial BIO being pushed or popped. This makes it possible
8005 to determine whether the BIO is the one explicitly called or as a result
8006 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
8007 it handles reference counts correctly and doesn't zero out the I/O bio
8008 when it is not being explicitly popped. WARNING: applications which
8009 included workarounds for the old buggy behaviour will need to be modified
8010 or they could free up already freed BIOs.
8014 * Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
8015 renaming to all platforms (within the 0.9.8 branch, this was
8016 done conditionally on Netware platforms to avoid a name clash).
8018 *Guenter <lists@gknw.net>*
8020 * Add ECDHE and PSK support to DTLS.
8022 *Michael Tuexen <tuexen@fh-muenster.de>*
8024 * Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
8029 * Add "missing" function EVP_MD_flags() (without this the only way to
8030 retrieve a digest flags is by accessing the structure directly. Update
8031 `EVP_MD_do_all*()` and `EVP_CIPHER_do_all*()` to include the name a digest
8032 or cipher is registered as in the "from" argument. Print out all
8033 registered digests in the dgst usage message instead of manually
8034 attempting to work them out.
8038 * If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
8039 this allows the use of compression and extensions. Change default cipher
8040 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
8041 by default unless an application cipher string requests it.
8045 * Alter match criteria in PKCS12_parse(). It used to try to use local
8046 key ids to find matching certificates and keys but some PKCS#12 files
8047 don't follow the (somewhat unwritten) rules and this strategy fails.
8048 Now just gather all certificates together and the first private key
8049 then look for the first certificate that matches the key.
8053 * Support use of registered digest and cipher names for dgst and cipher
8054 commands instead of having to add each one as a special case. So now
8061 openssl dgst -sha256 foo
8063 and this works for ENGINE based algorithms too.
8067 * Update Gost ENGINE to support parameter files.
8069 *Victor B. Wagner <vitus@cryptocom.ru>*
8071 * Support GeneralizedTime in ca utility.
8073 *Oliver Martin <oliver@volatilevoid.net>, Steve Henson*
8075 * Enhance the hash format used for certificate directory links. The new
8076 form uses the canonical encoding (meaning equivalent names will work
8077 even if they aren't identical) and uses SHA1 instead of MD5. This form
8078 is incompatible with the older format and as a result c_rehash should
8079 be used to rebuild symbolic links.
8083 * Make PKCS#8 the default write format for private keys, replacing the
8084 traditional format. This form is standardised, more secure and doesn't
8085 include an implicit MD5 dependency.
8089 * Add a $gcc_devteam_warn option to Configure. The idea is that any code
8090 committed to OpenSSL should pass this lot as a minimum.
8094 * Add session ticket override functionality for use by EAP-FAST.
8096 *Jouni Malinen <j@w1.fi>*
8098 * Modify HMAC functions to return a value. Since these can be implemented
8099 in an ENGINE errors can occur.
8103 * Type-checked OBJ_bsearch_ex.
8107 * Type-checked OBJ_bsearch. Also some constification necessitated
8108 by type-checking. Still to come: TXT_DB, bsearch(?),
8109 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
8114 * New function OPENSSL_gmtime_adj() to add a specific number of days and
8115 seconds to a tm structure directly, instead of going through OS
8116 specific date routines. This avoids any issues with OS routines such
8117 as the year 2038 bug. New `*_adj()` functions for ASN1 time structures
8118 and X509_time_adj_ex() to cover the extended range. The existing
8119 X509_time_adj() is still usable and will no longer have any date issues.
8123 * Delta CRL support. New use deltas option which will attempt to locate
8124 and search any appropriate delta CRLs available.
8126 This work was sponsored by Google.
8130 * Support for CRLs partitioned by reason code. Reorganise CRL processing
8131 code and add additional score elements. Validate alternate CRL paths
8132 as part of the CRL checking and indicate a new error "CRL path validation
8133 error" in this case. Applications wanting additional details can use
8134 the verify callback and check the new "parent" field. If this is not
8135 NULL CRL path validation is taking place. Existing applications won't
8136 see this because it requires extended CRL support which is off by
8139 This work was sponsored by Google.
8143 * Support for freshest CRL extension.
8145 This work was sponsored by Google.
8149 * Initial indirect CRL support. Currently only supported in the CRLs
8150 passed directly and not via lookup. Process certificate issuer
8151 CRL entry extension and lookup CRL entries by bother issuer name
8152 and serial number. Check and process CRL issuer entry in IDP extension.
8154 This work was sponsored by Google.
8158 * Add support for distinct certificate and CRL paths. The CRL issuer
8159 certificate is validated separately in this case. Only enabled if
8160 an extended CRL support flag is set: this flag will enable additional
8161 CRL functionality in future.
8163 This work was sponsored by Google.
8167 * Add support for policy mappings extension.
8169 This work was sponsored by Google.
8173 * Fixes to pathlength constraint, self issued certificate handling,
8174 policy processing to align with RFC3280 and PKITS tests.
8176 This work was sponsored by Google.
8180 * Support for name constraints certificate extension. DN, email, DNS
8181 and URI types are currently supported.
8183 This work was sponsored by Google.
8187 * To cater for systems that provide a pointer-based thread ID rather
8188 than numeric, deprecate the current numeric thread ID mechanism and
8189 replace it with a structure and associated callback type. This
8190 mechanism allows a numeric "hash" to be extracted from a thread ID in
8191 either case, and on platforms where pointers are larger than 'long',
8192 mixing is done to help ensure the numeric 'hash' is usable even if it
8193 can't be guaranteed unique. The default mechanism is to use "&errno"
8194 as a pointer-based thread ID to distinguish between threads.
8196 Applications that want to provide their own thread IDs should now use
8197 CRYPTO_THREADID_set_callback() to register a callback that will call
8198 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
8200 Note that ERR_remove_state() is now deprecated, because it is tied
8201 to the assumption that thread IDs are numeric. ERR_remove_state(0)
8202 to free the current thread's error state should be replaced by
8203 ERR_remove_thread_state(NULL).
8205 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
8206 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
8207 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
8208 application was previously providing a numeric thread callback that
8209 was inappropriate for distinguishing threads, then uniqueness might
8210 have been obtained with &errno that happened immediately in the
8211 intermediate development versions of OpenSSL; this is no longer the
8212 case, the numeric thread callback will now override the automatic use
8215 *Geoff Thorpe, with help from Bodo Moeller*
8217 * Initial support for different CRL issuing certificates. This covers a
8218 simple case where the self issued certificates in the chain exist and
8219 the real CRL issuer is higher in the existing chain.
8221 This work was sponsored by Google.
8225 * Removed effectively defunct crypto/store from the build.
8229 * Revamp of STACK to provide stronger type-checking. Still to come:
8230 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
8231 ASN1_STRING, CONF_VALUE.
8235 * Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
8236 RAM on SSL connections. This option can save about 34k per idle SSL.
8240 * Revamp of LHASH to provide stronger type-checking. Still to come:
8241 STACK, TXT_DB, bsearch, qsort.
8245 * Initial support for Cryptographic Message Syntax (aka CMS) based
8246 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
8247 support for data, signedData, compressedData, digestedData and
8248 encryptedData, envelopedData types included. Scripts to check against
8249 RFC4134 examples draft and interop and consistency checks of many
8250 content types and variants.
8254 * Add options to enc utility to support use of zlib compression BIO.
8258 * Extend mk1mf to support importing of options and assembly language
8259 files from Configure script, currently only included in VC-WIN32.
8260 The assembly language rules can now optionally generate the source
8261 files from the associated perl scripts.
8265 * Implement remaining functionality needed to support GOST ciphersuites.
8266 Interop testing has been performed using CryptoPro implementations.
8268 *Victor B. Wagner <vitus@cryptocom.ru>*
8270 * s390x assembler pack.
8274 * ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
8279 * Implement Opaque PRF Input TLS extension as specified in
8280 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
8281 official specification yet and no extension type assignment by
8282 IANA exists, this extension (for now) will have to be explicitly
8283 enabled when building OpenSSL by providing the extension number
8284 to use. For example, specify an option
8286 -DTLSEXT_TYPE_opaque_prf_input=0x9527
8288 to the "config" or "Configure" script to enable the extension,
8289 assuming extension number 0x9527 (which is a completely arbitrary
8290 and unofficial assignment based on the MD5 hash of the Internet
8291 Draft). Note that by doing so, you potentially lose
8292 interoperability with other TLS implementations since these might
8293 be using the same extension number for other purposes.
8295 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
8296 opaque PRF input value to use in the handshake. This will create
8297 an internal copy of the length-'len' string at 'src', and will
8298 return non-zero for success.
8300 To get more control and flexibility, provide a callback function
8303 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
8304 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
8308 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
8311 Callback function 'cb' will be called in handshakes, and is
8312 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
8313 Argument 'arg' is for application purposes (the value as given to
8314 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
8315 be provided to the callback function). The callback function
8316 has to return non-zero to report success: usually 1 to use opaque
8317 PRF input just if possible, or 2 to enforce use of the opaque PRF
8318 input. In the latter case, the library will abort the handshake
8319 if opaque PRF input is not successfully negotiated.
8321 Arguments 'peerinput' and 'len' given to the callback function
8322 will always be NULL and 0 in the case of a client. A server will
8323 see the client's opaque PRF input through these variables if
8324 available (NULL and 0 otherwise). Note that if the server
8325 provides an opaque PRF input, the length must be the same as the
8326 length of the client's opaque PRF input.
8328 Note that the callback function will only be called when creating
8329 a new session (session resumption can resume whatever was
8330 previously negotiated), and will not be called in SSL 2.0
8331 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
8332 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
8333 for applications that need to enforce opaque PRF input.
8337 * Update ssl code to support digests other than SHA1+MD5 for handshake
8340 *Victor B. Wagner <vitus@cryptocom.ru>*
8342 * Add RFC4507 support to OpenSSL. This includes the corrections in
8343 RFC4507bis. The encrypted ticket format is an encrypted encoded
8344 SSL_SESSION structure, that way new session features are automatically
8347 If a client application caches session in an SSL_SESSION structure
8348 support is transparent because tickets are now stored in the encoded
8351 The SSL_CTX structure automatically generates keys for ticket
8352 protection in servers so again support should be possible
8353 with no application modification.
8355 If a client or server wishes to disable RFC4507 support then the option
8356 SSL_OP_NO_TICKET can be set.
8358 Add a TLS extension debugging callback to allow the contents of any client
8359 or server extensions to be examined.
8361 This work was sponsored by Google.
8365 * Final changes to avoid use of pointer pointer casts in OpenSSL.
8366 OpenSSL should now compile cleanly on gcc 4.2
8368 *Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson*
8370 * Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
8371 support including streaming MAC support: this is required for GOST
8372 ciphersuite support.
8374 *Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson*
8376 * Add option -stream to use PKCS#7 streaming in smime utility. New
8377 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
8378 to output in BER and PEM format.
8382 * Experimental support for use of HMAC via EVP_PKEY interface. This
8383 allows HMAC to be handled via the `EVP_DigestSign*()` interface. The
8384 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
8385 ENGINE support for HMAC keys which are unextractable. New -mac and
8386 -macopt options to dgst utility.
8390 * New option -sigopt to dgst utility. Update dgst to use
8391 `EVP_Digest{Sign,Verify}*`. These two changes make it possible to use
8392 alternative signing parameters such as X9.31 or PSS in the dgst
8397 * Change ssl_cipher_apply_rule(), the internal function that does
8398 the work each time a ciphersuite string requests enabling
8399 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
8400 removing ("!foo+bar") a class of ciphersuites: Now it maintains
8401 the order of disabled ciphersuites such that those ciphersuites
8402 that most recently went from enabled to disabled not only stay
8403 in order with respect to each other, but also have higher priority
8404 than other disabled ciphersuites the next time ciphersuites are
8407 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
8408 the same ciphersuites as with "HIGH" alone, but in a specific
8409 order where the PSK ciphersuites come first (since they are the
8410 most recently disabled ciphersuites when "HIGH" is parsed).
8412 Also, change ssl_create_cipher_list() (using this new
8413 functionality) such that between otherwise identical
8414 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
8419 * Change ssl_create_cipher_list() so that it automatically
8420 arranges the ciphersuites in reasonable order before starting
8421 to process the rule string. Thus, the definition for "DEFAULT"
8422 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
8423 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
8424 This makes it much easier to arrive at a reasonable default order
8425 in applications for which anonymous ciphers are OK (meaning
8426 that you can't actually use DEFAULT).
8428 *Bodo Moeller; suggested by Victor Duchovni*
8430 * Split the SSL/TLS algorithm mask (as used for ciphersuite string
8431 processing) into multiple integers instead of setting
8432 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
8433 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
8434 (These masks as well as the individual bit definitions are hidden
8435 away into the non-exported interface ssl/ssl_locl.h, so this
8436 change to the definition of the SSL_CIPHER structure shouldn't
8437 affect applications.) This give us more bits for each of these
8438 categories, so there is no longer a need to coagulate AES128 and
8439 AES256 into a single algorithm bit, and to coagulate Camellia128
8440 and Camellia256 into a single algorithm bit, which has led to all
8443 Thus, among other things, the kludge introduced in 0.9.7m and
8444 0.9.8e for masking out AES256 independently of AES128 or masking
8445 out Camellia256 independently of AES256 is not needed here in 0.9.9.
8447 With the change, we also introduce new ciphersuite aliases that
8448 so far were missing: "AES128", "AES256", "CAMELLIA128", and
8453 * Add support for dsa-with-SHA224 and dsa-with-SHA256.
8454 Use the leftmost N bytes of the signature input if the input is
8455 larger than the prime q (with N being the size in bytes of q).
8459 * Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
8460 it yet and it is largely untested.
8464 * Add support for the ecdsa-with-SHA224/256/384/512 signature types.
8468 * Initial incomplete changes to avoid need for function casts in OpenSSL
8469 some compilers (gcc 4.2 and later) reject their use. Safestack is
8470 reimplemented. Update ASN1 to avoid use of legacy functions.
8474 * Win32/64 targets are linked with Winsock2.
8478 * Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
8479 to external functions. This can be used to increase CRL handling
8480 efficiency especially when CRLs are very large by (for example) storing
8481 the CRL revoked certificates in a database.
8485 * Overhaul of by_dir code. Add support for dynamic loading of CRLs so
8486 new CRLs added to a directory can be used. New command line option
8487 -verify_return_error to s_client and s_server. This causes real errors
8488 to be returned by the verify callback instead of carrying on no matter
8489 what. This reflects the way a "real world" verify callback would behave.
8493 * GOST engine, supporting several GOST algorithms and public key formats.
8494 Kindly donated by Cryptocom.
8498 * Partial support for Issuing Distribution Point CRL extension. CRLs
8499 partitioned by DP are handled but no indirect CRL or reason partitioning
8500 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
8501 selected via a scoring technique which handles IDP and AKID in CRLs.
8505 * New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
8506 will ultimately be used for all verify operations: this will remove the
8507 X509_STORE dependency on certificate verification and allow alternative
8508 lookup methods. X509_STORE based implementations of these two callbacks.
8512 * Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
8513 Modify get_crl() to find a valid (unexpired) CRL if possible.
8517 * New function X509_CRL_match() to check if two CRLs are identical. Normally
8518 this would be called X509_CRL_cmp() but that name is already used by
8519 a function that just compares CRL issuer names. Cache several CRL
8520 extensions in X509_CRL structure and cache CRLDP in X509.
8524 * Store a "canonical" representation of X509_NAME structure (ASN1 Name)
8525 this maps equivalent X509_NAME structures into a consistent structure.
8526 Name comparison can then be performed rapidly using memcmp().
8530 * Non-blocking OCSP request processing. Add -timeout option to ocsp
8535 * Allow digests to supply their own micalg string for S/MIME type using
8536 the ctrl EVP_MD_CTRL_MICALG.
8540 * During PKCS7 signing pass the PKCS7 SignerInfo structure to the
8541 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
8542 ctrl. It can then customise the structure before and/or after signing
8547 * New function OBJ_add_sigid() to allow application defined signature OIDs
8548 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
8549 to free up any added signature OIDs.
8553 * New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
8554 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
8555 digest and cipher tables. New options added to openssl utility:
8556 list-message-digest-algorithms and list-cipher-algorithms.
8560 * Change the array representation of binary polynomials: the list
8561 of degrees of non-zero coefficients is now terminated with -1.
8562 Previously it was terminated with 0, which was also part of the
8563 value; thus, the array representation was not applicable to
8564 polynomials where t^0 has coefficient zero. This change makes
8565 the array representation useful in a more general context.
8569 * Various modifications and fixes to SSL/TLS cipher string
8570 handling. For ECC, the code now distinguishes between fixed ECDH
8571 with RSA certificates on the one hand and with ECDSA certificates
8572 on the other hand, since these are separate ciphersuites. The
8573 unused code for Fortezza ciphersuites has been removed.
8575 For consistency with EDH, ephemeral ECDH is now called "EECDH"
8576 (not "ECDHE"). For consistency with the code for DH
8577 certificates, use of ECDH certificates is now considered ECDH
8578 authentication, not RSA or ECDSA authentication (the latter is
8579 merely the CA's signing algorithm and not actively used in the
8582 The temporary ciphersuite alias "ECCdraft" is no longer
8583 available, and ECC ciphersuites are no longer excluded from "ALL"
8584 and "DEFAULT". The following aliases now exist for RFC 4492
8585 ciphersuites, most of these by analogy with the DH case:
8587 kECDHr - ECDH cert, signed with RSA
8588 kECDHe - ECDH cert, signed with ECDSA
8589 kECDH - ECDH cert (signed with either RSA or ECDSA)
8590 kEECDH - ephemeral ECDH
8591 ECDH - ECDH cert or ephemeral ECDH
8597 AECDH - anonymous ECDH
8598 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
8602 * Add additional S/MIME capabilities for AES and GOST ciphers if supported.
8603 Use correct micalg parameters depending on digest(s) in signed message.
8607 * Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
8608 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
8612 * Initial engine support for EVP_PKEY_METHOD. New functions to permit
8613 an engine to register a method. Add ENGINE lookups for methods and
8614 functional reference processing.
8618 * New functions `EVP_Digest{Sign,Verify)*`. These are enhanced versions of
8619 `EVP_{Sign,Verify}*` which allow an application to customise the signature
8624 * New -resign option to smime utility. This adds one or more signers
8625 to an existing PKCS#7 signedData structure. Also -md option to use an
8626 alternative message digest algorithm for signing.
8630 * Tidy up PKCS#7 routines and add new functions to make it easier to
8631 create PKCS7 structures containing multiple signers. Update smime
8632 application to support multiple signers.
8636 * New -macalg option to pkcs12 utility to allow setting of an alternative
8641 * Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
8642 Reorganize PBE internals to lookup from a static table using NIDs,
8643 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
8644 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
8645 PRF which will be automatically used with PBES2.
8649 * Replace the algorithm specific calls to generate keys in "req" with the
8654 * Update PKCS#7 enveloped data routines to use new API. This is now
8655 supported by any public key method supporting the encrypt operation. A
8656 ctrl is added to allow the public key algorithm to examine or modify
8657 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
8662 * Add a ctrl to asn1 method to allow a public key algorithm to express
8663 a default digest type to use. In most cases this will be SHA1 but some
8664 algorithms (such as GOST) need to specify an alternative digest. The
8665 return value indicates how strong the preference is 1 means optional and
8666 2 is mandatory (that is it is the only supported type). Modify
8667 ASN1_item_sign() to accept a NULL digest argument to indicate it should
8668 use the default md. Update openssl utilities to use the default digest
8669 type for signing if it is not explicitly indicated.
8673 * Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
8674 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
8675 signing method from the key type. This effectively removes the link
8676 between digests and public key types.
8680 * Add an OID cross reference table and utility functions. Its purpose is to
8681 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
8682 rsaEncryption. This will allow some of the algorithm specific hackery
8683 needed to use the correct OID to be removed.
8687 * Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
8688 structures for PKCS7_sign(). They are now set up by the relevant public
8693 * Add provisional EC pkey method with support for ECDSA and ECDH.
8697 * Add support for key derivation (agreement) in the API, DH method and
8702 * Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
8703 public and private key formats. As a side effect these add additional
8704 command line functionality not previously available: DSA signatures can be
8705 generated and verified using pkeyutl and DH key support and generation in
8712 *Oliver Tappe <zooey@hirschkaefer.de>*
8714 * New make target "install_html_docs" installs HTML renditions of the
8717 *Oliver Tappe <zooey@hirschkaefer.de>*
8719 * New utility "genpkey" this is analogous to "genrsa" etc except it can
8720 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
8721 support key and parameter generation and add initial key generation
8722 functionality for RSA.
8726 * Add functions for main EVP_PKEY_method operations. The undocumented
8727 functions `EVP_PKEY_{encrypt,decrypt}` have been renamed to
8728 `EVP_PKEY_{encrypt,decrypt}_old`.
8732 * Initial definitions for EVP_PKEY_METHOD. This will be a high level public
8733 key API, doesn't do much yet.
8737 * New function EVP_PKEY_asn1_get0_info() to retrieve information about
8738 public key algorithms. New option to openssl utility:
8739 "list-public-key-algorithms" to print out info.
8743 * Implement the Supported Elliptic Curves Extension for
8744 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8748 * Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8749 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
8753 * New utilities pkey and pkeyparam. These are similar to algorithm specific
8754 utilities such as rsa, dsa, dsaparam etc except they process any key
8759 * Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
8760 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
8761 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
8766 * Initial support for pluggable public key ASN1.
8767 De-spaghettify the public key ASN1 handling. Move public and private
8768 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
8769 algorithm specific handling to a single module within the relevant
8770 algorithm directory. Add functions to allow (near) opaque processing
8771 of public and private key structures.
8775 * Implement the Supported Point Formats Extension for
8776 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
8780 * Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
8781 for the psk identity [hint] and the psk callback functions to the
8782 SSL_SESSION, SSL and SSL_CTX structure.
8785 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
8789 SSL_CTX_use_psk_identity_hint
8790 SSL_get_psk_identity_hint
8791 SSL_get_psk_identity
8792 SSL_use_psk_identity_hint
8794 *Mika Kousa and Pasi Eronen of Nokia Corporation*
8796 * Add RFC 3161 compliant time stamp request creation, response generation
8797 and response verification functionality.
8799 *Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project*
8801 * Add initial support for TLS extensions, specifically for the server_name
8802 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
8803 have new members for a host name. The SSL data structure has an
8804 additional member `SSL_CTX *initial_ctx` so that new sessions can be
8805 stored in that context to allow for session resumption, even after the
8806 SSL has been switched to a new SSL_CTX in reaction to a client's
8807 server_name extension.
8809 New functions (subject to change):
8811 SSL_get_servername()
8812 SSL_get_servername_type()
8815 New CTRL codes and macros (subject to change):
8817 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
8818 - SSL_CTX_set_tlsext_servername_callback()
8819 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
8820 - SSL_CTX_set_tlsext_servername_arg()
8821 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
8823 openssl s_client has a new '-servername ...' option.
8825 openssl s_server has new options '-servername_host ...', '-cert2 ...',
8826 '-key2 ...', '-servername_fatal' (subject to change). This allows
8827 testing the HostName extension for a specific single host name ('-cert'
8828 and '-key' remain fallbacks for handshakes without HostName
8829 negotiation). If the unrecognized_name alert has to be sent, this by
8830 default is a warning; it becomes fatal with the '-servername_fatal'
8833 *Peter Sylvester, Remy Allais, Christophe Renou*
8835 * Whirlpool hash implementation is added.
8839 * BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
8840 bn(64,32). Because of instruction set limitations it doesn't have
8841 any negative impact on performance. This was done mostly in order
8842 to make it possible to share assembler modules, such as bn_mul_mont
8843 implementations, between 32- and 64-bit builds without hassle.
8847 * Move code previously exiled into file crypto/ec/ec2_smpt.c
8848 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
8853 * New candidate for BIGNUM assembler implementation, bn_mul_mont,
8854 dedicated Montgomery multiplication procedure, is introduced.
8855 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
8856 "64-bit" performance on certain 32-bit targets.
8860 * New option SSL_OP_NO_COMP to disable use of compression selectively
8861 in SSL structures. New SSL ctrl to set maximum send fragment size.
8862 Save memory by setting the I/O buffer sizes dynamically instead of
8863 using the maximum available value.
8867 * New option -V for 'openssl ciphers'. This prints the ciphersuite code
8868 in addition to the text details.
8872 * Very, very preliminary EXPERIMENTAL support for printing of general
8873 ASN1 structures. This currently produces rather ugly output and doesn't
8874 handle several customised structures at all.
8878 * Integrated support for PVK file format and some related formats such
8879 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
8880 these in the 'rsa' and 'dsa' utilities.
8884 * Support for PKCS#1 RSAPublicKey format on rsa utility command line.
8888 * Remove the ancient ASN1_METHOD code. This was only ever used in one
8889 place for the (very old) "NETSCAPE" format certificates which are now
8890 handled using new ASN1 code equivalents.
8894 * Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
8895 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
8896 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
8900 * Modify CRL distribution points extension code to print out previously
8901 unsupported fields. Enhance extension setting code to allow setting of
8906 * Add print and set support for Issuing Distribution Point CRL extension.
8910 * Change 'Configure' script to enable Camellia by default.
8917 ### Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
8919 * When rejecting SSL/TLS records due to an incorrect version number, never
8920 update s->server with a new major version number. As of
8921 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
8922 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
8923 the previous behavior could result in a read attempt at NULL when
8924 receiving specific incorrect SSL/TLS records once record payload
8925 protection is active. ([CVE-2010-0740])
8927 *Bodo Moeller, Adam Langley <agl@chromium.org>*
8929 * Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
8930 could be crashed if the relevant tables were not present (e.g. chrooted).
8932 *Tomas Hoger <thoger@redhat.com>*
8934 ### Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
8936 * Always check bn_wexpand() return values for failure. ([CVE-2009-3245])
8938 *Martin Olsson, Neel Mehta*
8940 * Fix X509_STORE locking: Every 'objs' access requires a lock (to
8941 accommodate for stack sorting, always a write lock!).
8945 * On some versions of WIN32 Heap32Next is very slow. This can cause
8946 excessive delays in the RAND_poll(): over a minute. As a workaround
8947 include a time check in the inner Heap32Next loop too.
8951 * The code that handled flushing of data in SSL/TLS originally used the
8952 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
8953 the problem outlined in PR#1949. The fix suggested there however can
8954 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
8955 of Apache). So instead simplify the code to flush unconditionally.
8956 This should be fine since flushing with no data to flush is a no op.
8960 * Handle TLS versions 2.0 and later properly and correctly use the
8961 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
8962 off ancient servers have a habit of sticking around for a while...
8966 * Modify compression code so it frees up structures without using the
8967 ex_data callbacks. This works around a problem where some applications
8968 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
8969 restarting) then use compression (e.g. SSL with compression) later.
8970 This results in significant per-connection memory leaks and
8971 has caused some security issues including CVE-2008-1678 and
8976 * Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
8977 change when encrypting or decrypting.
8981 * Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
8982 connect and renegotiate with servers which do not support RI.
8983 Until RI is more widely deployed this option is enabled by default.
8987 * Add "missing" ssl ctrls to clear options and mode.
8991 * If client attempts to renegotiate and doesn't support RI respond with
8992 a no_renegotiation alert as required by RFC5746. Some renegotiating
8993 TLS clients will continue a connection gracefully when they receive
8994 the alert. Unfortunately OpenSSL mishandled this alert and would hang
8995 waiting for a server hello which it will never receive. Now we treat a
8996 received no_renegotiation alert as a fatal error. This is because
8997 applications requesting a renegotiation might well expect it to succeed
8998 and would have no code in place to handle the server denying it so the
8999 only safe thing to do is to terminate the connection.
9003 * Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
9004 peer supports secure renegotiation and 0 otherwise. Print out peer
9005 renegotiation support in s_client/s_server.
9009 * Replace the highly broken and deprecated SPKAC certification method with
9010 the updated NID creation version. This should correctly handle UTF8.
9014 * Implement RFC5746. Re-enable renegotiation but require the extension
9015 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
9016 turns out to be a bad idea. It has been replaced by
9017 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
9018 SSL_CTX_set_options(). This is really not recommended unless you
9019 know what you are doing.
9021 *Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson*
9023 * Fixes to stateless session resumption handling. Use initial_ctx when
9024 issuing and attempting to decrypt tickets in case it has changed during
9025 servername handling. Use a non-zero length session ID when attempting
9026 stateless session resumption: this makes it possible to determine if
9027 a resumption has occurred immediately after receiving server hello
9028 (several places in OpenSSL subtly assume this) instead of later in
9033 * The functions ENGINE_ctrl(), OPENSSL_isservice(),
9034 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
9035 fixes for a few places where the return code is not checked
9038 *Julia Lawall <julia@diku.dk>*
9040 * Add --strict-warnings option to Configure script to include devteam
9041 warnings in other configurations.
9045 * Add support for --libdir option and LIBDIR variable in makefiles. This
9046 makes it possible to install openssl libraries in locations which
9047 have names other than "lib", for example "/usr/lib64" which some
9050 *Steve Henson, based on patch from Jeremy Utley*
9052 * Don't allow the use of leading 0x80 in OIDs. This is a violation of
9053 X690 8.9.12 and can produce some misleading textual output of OIDs.
9055 *Steve Henson, reported by Dan Kaminsky*
9057 * Delete MD2 from algorithm tables. This follows the recommendation in
9058 several standards that it is not used in new applications due to
9059 several cryptographic weaknesses. For binary compatibility reasons
9060 the MD2 API is still compiled in by default.
9064 * Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
9069 * Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
9070 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
9073 *Guenter <lists@gknw.net>*
9075 * Fix the server certificate chain building code to use X509_verify_cert(),
9076 it used to have an ad-hoc builder which was unable to cope with anything
9077 other than a simple chain.
9079 *David Woodhouse <dwmw2@infradead.org>, Steve Henson*
9081 * Don't check self signed certificate signatures in X509_verify_cert()
9082 by default (a flag can override this): it just wastes time without
9083 adding any security. As a useful side effect self signed root CAs
9084 with non-FIPS digests are now usable in FIPS mode.
9088 * In dtls1_process_out_of_seq_message() the check if the current message
9089 is already buffered was missing. For every new message was memory
9090 allocated, allowing an attacker to perform an denial of service attack
9091 with sending out of seq handshake messages until there is no memory
9092 left. Additionally every future message was buffered, even if the
9093 sequence number made no sense and would be part of another handshake.
9094 So only messages with sequence numbers less than 10 in advance will be
9095 buffered. ([CVE-2009-1378])
9097 *Robin Seggelmann, discovered by Daniel Mentz*
9099 * Records are buffered if they arrive with a future epoch to be
9100 processed after finishing the corresponding handshake. There is
9101 currently no limitation to this buffer allowing an attacker to perform
9102 a DOS attack with sending records with future epochs until there is no
9103 memory left. This patch adds the pqueue_size() function to determine
9104 the size of a buffer and limits the record buffer to 100 entries.
9107 *Robin Seggelmann, discovered by Daniel Mentz*
9109 * Keep a copy of frag->msg_header.frag_len so it can be used after the
9110 parent structure is freed. ([CVE-2009-1379])
9114 * Handle non-blocking I/O properly in SSL_shutdown() call.
9116 *Darryl Miles <darryl-mailinglists@netbauds.net>*
9118 * Add `2.5.4.*` OIDs
9120 *Ilya O. <vrghost@gmail.com>*
9122 ### Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
9124 * Disable renegotiation completely - this fixes a severe security
9125 problem ([CVE-2009-3555]) at the cost of breaking all
9126 renegotiation. Renegotiation can be re-enabled by setting
9127 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
9128 run-time. This is really not recommended unless you know what
9133 ### Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
9135 * Don't set val to NULL when freeing up structures, it is freed up by
9136 underlying code. If `sizeof(void *) > sizeof(long)` this can result in
9137 zeroing past the valid field. ([CVE-2009-0789])
9139 *Paolo Ganci <Paolo.Ganci@AdNovum.CH>*
9141 * Fix bug where return value of CMS_SignerInfo_verify_content() was not
9142 checked correctly. This would allow some invalid signed attributes to
9143 appear to verify correctly. ([CVE-2009-0591])
9145 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9147 * Reject UniversalString and BMPString types with invalid lengths. This
9148 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
9149 a legal length. ([CVE-2009-0590])
9153 * Set S/MIME signing as the default purpose rather than setting it
9154 unconditionally. This allows applications to override it at the store
9159 * Permit restricted recursion of ASN1 strings. This is needed in practice
9160 to handle some structures.
9164 * Improve efficiency of mem_gets: don't search whole buffer each time
9167 *Jeremy Shapiro <jnshapir@us.ibm.com>*
9169 * New -hex option for openssl rand.
9173 * Print out UTF8String and NumericString when parsing ASN1.
9177 * Support NumericString type for name components.
9181 * Allow CC in the environment to override the automatically chosen
9182 compiler. Note that nothing is done to ensure flags work with the
9187 ### Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
9189 * Properly check EVP_VerifyFinal() and similar return values
9192 *Ben Laurie, Bodo Moeller, Google Security Team*
9194 * Enable TLS extensions by default.
9198 * Allow the CHIL engine to be loaded, whether the application is
9199 multithreaded or not. (This does not release the developer from the
9200 obligation to set up the dynamic locking callbacks.)
9202 *Sander Temme <sander@temme.net>*
9204 * Use correct exit code if there is an error in dgst command.
9206 *Steve Henson; problem pointed out by Roland Dirlewanger*
9208 * Tweak Configure so that you need to say "experimental-jpake" to enable
9209 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
9213 * Add experimental JPAKE support, including demo authentication in
9214 s_client and s_server.
9218 * Set the comparison function in v3_addr_canonize().
9220 *Rob Austein <sra@hactrn.net>*
9222 * Add support for XMPP STARTTLS in s_client.
9224 *Philip Paeps <philip@freebsd.org>*
9226 * Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
9227 to ensure that even with this option, only ciphersuites in the
9228 server's preference list will be accepted. (Note that the option
9229 applies only when resuming a session, so the earlier behavior was
9230 just about the algorithm choice for symmetric cryptography.)
9234 ### Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
9236 * Fix NULL pointer dereference if a DTLS server received
9237 ChangeCipherSpec as first record ([CVE-2009-1386]).
9241 * Fix a state transition in s3_srvr.c and d1_srvr.c
9242 (was using SSL3_ST_CW_CLNT_HELLO_B, should be `..._ST_SW_SRVR_...`).
9246 * The fix in 0.9.8c that supposedly got rid of unsafe
9247 double-checked locking was incomplete for RSA blinding,
9248 addressing just one layer of what turns out to have been
9249 doubly unsafe triple-checked locking.
9251 So now fix this for real by retiring the MONT_HELPER macro
9252 in crypto/rsa/rsa_eay.c.
9254 *Bodo Moeller; problem pointed out by Marius Schilder*
9256 * Various precautionary measures:
9258 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
9260 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
9261 (NB: This would require knowledge of the secret session ticket key
9262 to exploit, in which case you'd be SOL either way.)
9264 - Change bn_nist.c so that it will properly handle input BIGNUMs
9265 outside the expected range.
9267 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
9270 *Neel Mehta, Bodo Moeller*
9272 * Allow engines to be "soft loaded" - i.e. optionally don't die if
9273 the load fails. Useful for distros.
9275 *Ben Laurie and the FreeBSD team*
9277 * Add support for Local Machine Keyset attribute in PKCS#12 files.
9281 * Fix BN_GF2m_mod_arr() top-bit cleanup code.
9285 * Expand ENGINE to support engine supplied SSL client certificate functions.
9287 This work was sponsored by Logica.
9291 * Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
9292 keystores. Support for SSL/TLS client authentication too.
9293 Not compiled unless enable-capieng specified to Configure.
9295 This work was sponsored by Logica.
9299 * Fix bug in X509_ATTRIBUTE creation: don't set attribute using
9300 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
9301 attribute creation routines such as certificate requests and PKCS#12
9306 ### Changes between 0.9.8g and 0.9.8h [28 May 2008]
9308 * Fix flaw if 'Server Key exchange message' is omitted from a TLS
9309 handshake which could lead to a client crash as found using the
9310 Codenomicon TLS test suite ([CVE-2008-1672])
9312 *Steve Henson, Mark Cox*
9314 * Fix double free in TLS server name extensions which could lead to
9315 a remote crash found by Codenomicon TLS test suite ([CVE-2008-0891])
9319 * Clear error queue in SSL_CTX_use_certificate_chain_file()
9321 Clear the error queue to ensure that error entries left from
9322 older function calls do not interfere with the correct operation.
9324 *Lutz Jaenicke, Erik de Castro Lopo*
9326 * Remove root CA certificates of commercial CAs:
9328 The OpenSSL project does not recommend any specific CA and does not
9329 have any policy with respect to including or excluding any CA.
9330 Therefore it does not make any sense to ship an arbitrary selection
9331 of root CA certificates with the OpenSSL software.
9335 * RSA OAEP patches to fix two separate invalid memory reads.
9336 The first one involves inputs when 'lzero' is greater than
9337 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
9338 before the beginning of from). The second one involves inputs where
9339 the 'db' section contains nothing but zeroes (there is a one-byte
9340 invalid read after the end of 'db').
9342 *Ivan Nestlerode <inestlerode@us.ibm.com>*
9344 * Partial backport from 0.9.9-dev:
9346 Introduce bn_mul_mont (dedicated Montgomery multiplication
9347 procedure) as a candidate for BIGNUM assembler implementation.
9348 While 0.9.9-dev uses assembler for various architectures, only
9349 x86_64 is available by default here in the 0.9.8 branch, and
9350 32-bit x86 is available through a compile-time setting.
9352 To try the 32-bit x86 assembler implementation, use Configure
9353 option "enable-montasm" (which exists only for this backport).
9355 As "enable-montasm" for 32-bit x86 disclaims code stability
9356 anyway, in this constellation we activate additional code
9357 backported from 0.9.9-dev for further performance improvements,
9358 namely BN_from_montgomery_word. (To enable this otherwise,
9359 e.g. x86_64, try `-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD`.)
9361 *Andy Polyakov (backport partially by Bodo Moeller)*
9363 * Add TLS session ticket callback. This allows an application to set
9364 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
9365 values. This is useful for key rollover for example where several key
9366 sets may exist with different names.
9370 * Reverse ENGINE-internal logic for caching default ENGINE handles.
9371 This was broken until now in 0.9.8 releases, such that the only way
9372 a registered ENGINE could be used (assuming it initialises
9373 successfully on the host) was to explicitly set it as the default
9374 for the relevant algorithms. This is in contradiction with 0.9.7
9375 behaviour and the documentation. With this fix, when an ENGINE is
9376 registered into a given algorithm's table of implementations, the
9377 'uptodate' flag is reset so that auto-discovery will be used next
9378 time a new context for that algorithm attempts to select an
9381 *Ian Lister (tweaked by Geoff Thorpe)*
9383 * Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
9384 implementation in the following ways:
9386 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
9389 Lack of BER streaming support means one pass streaming processing is
9390 only supported if data is detached: setting the streaming flag is
9391 ignored for embedded content.
9393 CMS support is disabled by default and must be explicitly enabled
9394 with the enable-cms configuration option.
9398 * Update the GMP engine glue to do direct copies between BIGNUM and
9399 mpz_t when openssl and GMP use the same limb size. Otherwise the
9400 existing "conversion via a text string export" trick is still used.
9402 *Paul Sheer <paulsheer@gmail.com>*
9404 * Zlib compression BIO. This is a filter BIO which compressed and
9405 uncompresses any data passed through it.
9409 * Add AES_wrap_key() and AES_unwrap_key() functions to implement
9410 RFC3394 compatible AES key wrapping.
9414 * Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
9415 sets string data without copying. X509_ALGOR_set0() and
9416 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
9417 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
9418 from an X509_ATTRIBUTE structure optionally checking it occurs only
9419 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
9424 * Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
9425 to get the expected BN_FLG_CONSTTIME behavior.
9427 *Bodo Moeller (Google)*
9431 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
9432 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
9433 - added some more tests to do_tests.pl
9434 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
9435 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
9436 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
9437 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
9438 - various changes to netware.pl to enable gcc-cross builds on Win32
9440 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
9441 - various changes to fix missing prototype warnings
9442 - fixed x86nasm.pl to create correct asm files for NASM COFF output
9443 - added AES, WHIRLPOOL and CPUID assembler code to build files
9444 - added missing AES assembler make rules to mk1mf.pl
9445 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
9447 *Guenter Knauf <eflash@gmx.net>*
9449 * Implement certificate status request TLS extension defined in RFC3546.
9450 A client can set the appropriate parameters and receive the encoded
9451 OCSP response via a callback. A server can query the supplied parameters
9452 and set the encoded OCSP response in the callback. Add simplified examples
9453 to s_client and s_server.
9457 ### Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
9460 + Binary incompatibility of ssl_ctx_st structure
9461 + DTLS interoperation with non-compliant servers
9462 + Don't call get_session_cb() without proposed session
9463 + Fix ia64 assembler code
9465 *Andy Polyakov, Steve Henson*
9467 ### Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
9469 * DTLS Handshake overhaul. There were longstanding issues with
9470 OpenSSL DTLS implementation, which were making it impossible for
9471 RFC 4347 compliant client to communicate with OpenSSL server.
9472 Unfortunately just fixing these incompatibilities would "cut off"
9473 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
9474 server keeps tolerating non RFC compliant syntax. The opposite is
9475 not true, 0.9.8f client can not communicate with earlier server.
9476 This update even addresses CVE-2007-4995.
9480 * Changes to avoid need for function casts in OpenSSL: some compilers
9481 (gcc 4.2 and later) reject their use.
9482 *Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
9485 * Add RFC4507 support to OpenSSL. This includes the corrections in
9486 RFC4507bis. The encrypted ticket format is an encrypted encoded
9487 SSL_SESSION structure, that way new session features are automatically
9490 If a client application caches session in an SSL_SESSION structure
9491 support is transparent because tickets are now stored in the encoded
9494 The SSL_CTX structure automatically generates keys for ticket
9495 protection in servers so again support should be possible
9496 with no application modification.
9498 If a client or server wishes to disable RFC4507 support then the option
9499 SSL_OP_NO_TICKET can be set.
9501 Add a TLS extension debugging callback to allow the contents of any client
9502 or server extensions to be examined.
9504 This work was sponsored by Google.
9508 * Add initial support for TLS extensions, specifically for the server_name
9509 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9510 have new members for a host name. The SSL data structure has an
9511 additional member `SSL_CTX *initial_ctx` so that new sessions can be
9512 stored in that context to allow for session resumption, even after the
9513 SSL has been switched to a new SSL_CTX in reaction to a client's
9514 server_name extension.
9516 New functions (subject to change):
9518 SSL_get_servername()
9519 SSL_get_servername_type()
9522 New CTRL codes and macros (subject to change):
9524 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
9525 - SSL_CTX_set_tlsext_servername_callback()
9526 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
9527 - SSL_CTX_set_tlsext_servername_arg()
9528 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
9530 openssl s_client has a new '-servername ...' option.
9532 openssl s_server has new options '-servername_host ...', '-cert2 ...',
9533 '-key2 ...', '-servername_fatal' (subject to change). This allows
9534 testing the HostName extension for a specific single host name ('-cert'
9535 and '-key' remain fallbacks for handshakes without HostName
9536 negotiation). If the unrecognized_name alert has to be sent, this by
9537 default is a warning; it becomes fatal with the '-servername_fatal'
9540 *Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson*
9542 * Add AES and SSE2 assembly language support to VC++ build.
9546 * Mitigate attack on final subtraction in Montgomery reduction.
9550 * Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
9551 (which previously caused an internal error).
9555 * Squeeze another 10% out of IGE mode when in != out.
9559 * AES IGE mode speedup.
9561 *Dean Gaudet (Google)*
9563 * Add the Korean symmetric 128-bit cipher SEED (see
9564 <http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp>) and
9565 add SEED ciphersuites from RFC 4162:
9567 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
9568 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
9569 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
9570 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
9572 To minimize changes between patchlevels in the OpenSSL 0.9.8
9573 series, SEED remains excluded from compilation unless OpenSSL
9574 is configured with 'enable-seed'.
9576 *KISA, Bodo Moeller*
9578 * Mitigate branch prediction attacks, which can be practical if a
9579 single processor is shared, allowing a spy process to extract
9580 information. For detailed background information, see
9581 <http://eprint.iacr.org/2007/039> (O. Aciicmez, S. Gueron,
9582 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
9583 and Necessary Software Countermeasures"). The core of the change
9584 are new versions BN_div_no_branch() and
9585 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
9586 respectively, which are slower, but avoid the security-relevant
9587 conditional branches. These are automatically called by BN_div()
9588 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
9589 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
9590 remove a conditional branch.
9592 BN_FLG_CONSTTIME is the new name for the previous
9593 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
9594 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
9595 in the exponent causes BN_mod_exp_mont() to use the alternative
9596 implementation in BN_mod_exp_mont_consttime().) The old name
9597 remains as a deprecated alias.
9599 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
9600 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
9601 constant-time implementations for more than just exponentiation.
9602 Here too the old name is kept as a deprecated alias.
9604 BN_BLINDING_new() will now use BN_dup() for the modulus so that
9605 the BN_BLINDING structure gets an independent copy of the
9606 modulus. This means that the previous `BIGNUM *m` argument to
9607 BN_BLINDING_new() and to BN_BLINDING_create_param() now
9608 essentially becomes `const BIGNUM *m`, although we can't actually
9609 change this in the header file before 0.9.9. It allows
9610 RSA_setup_blinding() to use BN_with_flags() on the modulus to
9611 enable BN_FLG_CONSTTIME.
9613 *Matthew D Wood (Intel Corp)*
9615 * In the SSL/TLS server implementation, be strict about session ID
9616 context matching (which matters if an application uses a single
9617 external cache for different purposes). Previously,
9618 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
9619 set. This did ensure strict client verification, but meant that,
9620 with applications using a single external cache for quite
9621 different requirements, clients could circumvent ciphersuite
9622 restrictions for a given session ID context by starting a session
9623 in a different context.
9627 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
9628 a ciphersuite string such as "DEFAULT:RSA" cannot enable
9629 authentication-only ciphersuites.
9633 * Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
9634 not complete and could lead to a possible single byte overflow
9635 ([CVE-2007-5135]) [Ben Laurie]
9637 ### Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
9639 * Since AES128 and AES256 (and similarly Camellia128 and
9640 Camellia256) share a single mask bit in the logic of
9641 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
9642 kludge to work properly if AES128 is available and AES256 isn't
9643 (or if Camellia128 is available and Camellia256 isn't).
9647 * Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
9648 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
9649 When a point or a seed is encoded in a BIT STRING, we need to
9650 prevent the removal of trailing zero bits to get the proper DER
9651 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
9652 of a NamedBitList, for which trailing 0 bits need to be removed.)
9656 * Have SSL/TLS server implementation tolerate "mismatched" record
9657 protocol version while receiving ClientHello even if the
9658 ClientHello is fragmented. (The server can't insist on the
9659 particular protocol version it has chosen before the ServerHello
9660 message has informed the client about his choice.)
9664 * Add RFC 3779 support.
9666 *Rob Austein for ARIN, Ben Laurie*
9668 * Load error codes if they are not already present instead of using a
9669 static variable. This allows them to be cleanly unloaded and reloaded.
9670 Improve header file function name parsing.
9674 * extend SMTP and IMAP protocol emulation in s_client to use EHLO
9675 or CAPABILITY handshake as required by RFCs.
9679 ### Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
9681 * Introduce limits to prevent malicious keys being able to
9682 cause a denial of service. ([CVE-2006-2940])
9684 *Steve Henson, Bodo Moeller*
9686 * Fix ASN.1 parsing of certain invalid structures that can result
9687 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
9689 * Fix buffer overflow in SSL_get_shared_ciphers() function.
9690 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
9692 * Fix SSL client code which could crash if connecting to a
9693 malicious SSLv2 server. ([CVE-2006-4343])
9695 *Tavis Ormandy and Will Drewry, Google Security Team*
9697 * Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
9698 match only those. Before that, "AES256-SHA" would be interpreted
9699 as a pattern and match "AES128-SHA" too (since AES128-SHA got
9700 the same strength classification in 0.9.7h) as we currently only
9701 have a single AES bit in the ciphersuite description bitmap.
9702 That change, however, also applied to ciphersuite strings such as
9703 "RC4-MD5" that intentionally matched multiple ciphersuites --
9704 namely, SSL 2.0 ciphersuites in addition to the more common ones
9705 from SSL 3.0/TLS 1.0.
9707 So we change the selection algorithm again: Naming an explicit
9708 ciphersuite selects this one ciphersuite, and any other similar
9709 ciphersuite (same bitmap) from *other* protocol versions.
9710 Thus, "RC4-MD5" again will properly select both the SSL 2.0
9711 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
9713 Since SSL 2.0 does not have any ciphersuites for which the
9714 128/256 bit distinction would be relevant, this works for now.
9715 The proper fix will be to use different bits for AES128 and
9716 AES256, which would have avoided the problems from the beginning;
9717 however, bits are scarce, so we can only do this in a new release
9718 (not just a patchlevel) when we can change the SSL_CIPHER
9719 definition to split the single 'unsigned long mask' bitmap into
9720 multiple values to extend the available space.
9724 ### Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
9726 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
9727 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
9729 * Add AES IGE and biIGE modes.
9733 * Change the Unix randomness entropy gathering to use poll() when
9734 possible instead of select(), since the latter has some
9735 undesirable limitations.
9737 *Darryl Miles via Richard Levitte and Bodo Moeller*
9739 * Disable "ECCdraft" ciphersuites more thoroughly. Now special
9740 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
9741 cannot be implicitly activated as part of, e.g., the "AES" alias.
9742 However, please upgrade to OpenSSL 0.9.9[-dev] for
9743 non-experimental use of the ECC ciphersuites to get TLS extension
9744 support, which is required for curve and point format negotiation
9745 to avoid potential handshake problems.
9749 * Disable rogue ciphersuites:
9751 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
9752 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
9753 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
9755 The latter two were purportedly from
9756 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
9759 Also deactivate the remaining ciphersuites from
9760 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
9761 unofficial, and the ID has long expired.
9765 * Fix RSA blinding Heisenbug (problems sometimes occurred on
9766 dual-core machines) and other potential thread-safety issues.
9770 * Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
9771 versions), which is now available for royalty-free use
9772 (see <http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html>).
9773 Also, add Camellia TLS ciphersuites from RFC 4132.
9775 To minimize changes between patchlevels in the OpenSSL 0.9.8
9776 series, Camellia remains excluded from compilation unless OpenSSL
9777 is configured with 'enable-camellia'.
9781 * Disable the padding bug check when compression is in use. The padding
9782 bug check assumes the first packet is of even length, this is not
9783 necessarily true if compression is enabled and can result in false
9784 positives causing handshake failure. The actual bug test is ancient
9785 code so it is hoped that implementations will either have fixed it by
9786 now or any which still have the bug do not support compression.
9790 ### Changes between 0.9.8a and 0.9.8b [04 May 2006]
9792 * When applying a cipher rule check to see if string match is an explicit
9793 cipher suite and only match that one cipher suite if it is.
9797 * Link in manifests for VC++ if needed.
9799 *Austin Ziegler <halostatue@gmail.com>*
9801 * Update support for ECC-based TLS ciphersuites according to
9802 draft-ietf-tls-ecc-12.txt with proposed changes (but without
9803 TLS extensions, which are supported starting with the 0.9.9
9804 branch, not in the OpenSSL 0.9.8 branch).
9808 * New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
9809 opaque EVP_CIPHER_CTX handling.
9813 * Fixes and enhancements to zlib compression code. We now only use
9814 "zlib1.dll" and use the default `__cdecl` calling convention on Win32
9815 to conform with the standards mentioned here:
9816 <http://www.zlib.net/DLL_FAQ.txt>
9817 Static zlib linking now works on Windows and the new --with-zlib-include
9818 --with-zlib-lib options to Configure can be used to supply the location
9819 of the headers and library. Gracefully handle case where zlib library
9824 * Several fixes and enhancements to the OID generation code. The old code
9825 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
9826 handle numbers larger than ULONG_MAX, truncated printing and had a
9827 non standard OBJ_obj2txt() behaviour.
9831 * Add support for building of engines under engine/ as shared libraries
9832 under VC++ build system.
9836 * Corrected the numerous bugs in the Win32 path splitter in DSO.
9837 Hopefully, we will not see any false combination of paths any more.
9841 ### Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
9843 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
9844 (part of SSL_OP_ALL). This option used to disable the
9845 countermeasure against man-in-the-middle protocol-version
9846 rollback in the SSL 2.0 server implementation, which is a bad
9847 idea. ([CVE-2005-2969])
9849 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
9850 for Information Security, National Institute of Advanced Industrial
9851 Science and Technology [AIST], Japan)*
9853 * Add two function to clear and return the verify parameter flags.
9857 * Keep cipherlists sorted in the source instead of sorting them at
9858 runtime, thus removing the need for a lock.
9862 * Avoid some small subgroup attacks in Diffie-Hellman.
9864 *Nick Mathewson and Ben Laurie*
9866 * Add functions for well-known primes.
9870 * Extended Windows CE support.
9872 *Satoshi Nakamura and Andy Polyakov*
9874 * Initialize SSL_METHOD structures at compile time instead of during
9875 runtime, thus removing the need for a lock.
9879 * Make PKCS7_decrypt() work even if no certificate is supplied by
9880 attempting to decrypt each encrypted key in turn. Add support to
9885 ### Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
9887 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
9890 * Add libcrypto.pc and libssl.pc for those who feel they need them.
9894 * Change CA.sh and CA.pl so they don't bundle the CSR and the private
9895 key into the same file any more.
9899 * Add initial support for Win64, both IA64 and AMD64/x64 flavors.
9903 * Add -utf8 command line and config file option to 'ca'.
9905 *Stefan <stf@udoma.org*
9907 * Removed the macro des_crypt(), as it seems to conflict with some
9908 libraries. Use DES_crypt().
9912 * Correct naming of the 'chil' and '4758cca' ENGINEs. This
9913 involves renaming the source and generated shared-libs for
9914 both. The engines will accept the corrected or legacy ids
9915 ('ncipher' and '4758_cca' respectively) when binding. NB,
9916 this only applies when building 'shared'.
9918 *Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe*
9920 * Add attribute functions to EVP_PKEY structure. Modify
9921 PKCS12_create() to recognize a CSP name attribute and
9922 use it. Make -CSP option work again in pkcs12 utility.
9926 * Add new functionality to the bn blinding code:
9927 - automatic re-creation of the BN_BLINDING parameters after
9928 a fixed number of uses (currently 32)
9929 - add new function for parameter creation
9930 - introduce flags to control the update behaviour of the
9931 BN_BLINDING parameters
9932 - hide BN_BLINDING structure
9933 Add a second BN_BLINDING slot to the RSA structure to improve
9934 performance when a single RSA object is shared among several
9939 * Add support for DTLS.
9941 *Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie*
9943 * Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
9944 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
9948 * Remove buggy and incomplete DH cert support from
9949 ssl/ssl_rsa.c and ssl/s3_both.c
9953 * Use SHA-1 instead of MD5 as the default digest algorithm for
9954 the apps/openssl applications.
9958 * Compile clean with "-Wall -Wmissing-prototypes
9959 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9960 DEBUG_SAFESTACK must also be set.
9964 * Change ./Configure so that certain algorithms can be disabled by default.
9965 The new counterpiece to "no-xxx" is "enable-xxx".
9967 The patented RC5 and MDC2 algorithms will now be disabled unless
9968 "enable-rc5" and "enable-mdc2", respectively, are specified.
9970 (IDEA remains enabled despite being patented. This is because IDEA
9971 is frequently required for interoperability, and there is no license
9972 fee for non-commercial use. As before, "no-idea" can be used to
9973 avoid this algorithm.)
9977 * Add processing of proxy certificates (see RFC 3820). This work was
9978 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
9979 EGEE (Enabling Grids for E-science in Europe).
9983 * RC4 performance overhaul on modern architectures/implementations, such
9984 as Intel P4, IA-64 and AMD64.
9988 * New utility extract-section.pl. This can be used specify an alternative
9989 section number in a pod file instead of having to treat each file as
9990 a separate case in Makefile. This can be done by adding two lines to the
9993 =for comment openssl_section:XXX
9995 The blank line is mandatory.
9999 * New arguments -certform, -keyform and -pass for s_client and s_server
10000 to allow alternative format key and certificate files and passphrase
10005 * New structure X509_VERIFY_PARAM which combines current verify parameters,
10006 update associated structures and add various utility functions.
10008 Add new policy related verify parameters, include policy checking in
10009 standard verify code. Enhance 'smime' application with extra parameters
10010 to support policy checking and print out.
10014 * Add a new engine to support VIA PadLock ACE extensions in the VIA C3
10015 Nehemiah processors. These extensions support AES encryption in hardware
10016 as well as RNG (though RNG support is currently disabled).
10018 *Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov*
10020 * Deprecate `BN_[get|set]_params()` functions (they were ignored internally).
10024 * New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
10026 *Andy Polyakov and a number of other people*
10028 * Improved PowerPC platform support. Most notably BIGNUM assembler
10029 implementation contributed by IBM.
10031 *Suresh Chari, Peter Waltenberg, Andy Polyakov*
10033 * The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
10034 exponent rather than 'unsigned long'. There is a corresponding change to
10035 the new 'rsa_keygen' element of the RSA_METHOD structure.
10037 *Jelte Jansen, Geoff Thorpe*
10039 * Functionality for creating the initial serial number file is now
10040 moved from CA.pl to the 'ca' utility with a new option -create_serial.
10042 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
10043 number file to 1, which is bound to cause problems. To avoid
10044 the problems while respecting compatibility between different 0.9.7
10045 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
10046 CA.pl for serial number initialization. With the new release 0.9.8,
10047 we can fix the problem directly in the 'ca' utility.)
10051 * Reduced header interdependencies by declaring more opaque objects in
10052 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
10053 give fewer recursive includes, which could break lazy source code - so
10054 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
10055 developers should define this symbol when building and using openssl to
10056 ensure they track the recommended behaviour, interfaces, [etc], but
10057 backwards-compatible behaviour prevails when this isn't defined.
10061 * New function X509_POLICY_NODE_print() which prints out policy nodes.
10065 * Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
10066 This will generate a random key of the appropriate length based on the
10067 cipher context. The EVP_CIPHER can provide its own random key generation
10068 routine to support keys of a specific form. This is used in the des and
10069 3des routines to generate a key of the correct parity. Update S/MIME
10070 code to use new functions and hence generate correct parity DES keys.
10071 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
10072 valid (weak or incorrect parity).
10076 * Add a local set of CRLs that can be used by X509_verify_cert() as well
10077 as looking them up. This is useful when the verified structure may contain
10078 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
10079 present unless the new PKCS7_NO_CRL flag is asserted.
10083 * Extend ASN1 oid configuration module. It now additionally accepts the
10086 shortName = some long name, 1.2.3.4
10090 * Reimplemented the BN_CTX implementation. There is now no more static
10091 limitation on the number of variables it can handle nor the depth of the
10092 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10093 information can now expand as required, and rather than having a single
10094 static array of bignums, BN_CTX now uses a linked-list of such arrays
10095 allowing it to expand on demand whilst maintaining the usefulness of
10096 BN_CTX's "bundling".
10100 * Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
10101 to allow all RSA operations to function using a single BN_CTX.
10105 * Preliminary support for certificate policy evaluation and checking. This
10106 is initially intended to pass the tests outlined in "Conformance Testing
10107 of Relying Party Client Certificate Path Processing Logic" v1.07.
10111 * bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
10112 remained unused and not that useful. A variety of other little bignum
10113 tweaks and fixes have also been made continuing on from the audit (see
10118 * Constify all or almost all d2i, c2i, s2i and r2i functions, along with
10119 associated ASN1, EVP and SSL functions and old ASN1 macros.
10123 * BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
10124 and this should never fail. So the return value from the use of
10125 BN_set_word() (which can fail due to needless expansion) is now deprecated;
10126 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
10130 * BN_CTX_get() should return zero-valued bignums, providing the same
10131 initialised value as BN_new().
10133 *Geoff Thorpe, suggested by Ulf Möller*
10135 * Support for inhibitAnyPolicy certificate extension.
10139 * An audit of the BIGNUM code is underway, for which debugging code is
10140 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
10141 is considered valid when processing BIGNUMs, and causes execution to
10142 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
10143 further steps are taken to deliberately pollute unused data in BIGNUM
10144 structures to try and expose faulty code further on. For now, openssl will
10145 (in its default mode of operation) continue to tolerate the inconsistent
10146 forms that it has tolerated in the past, but authors and packagers should
10147 consider trying openssl and their own applications when compiled with
10148 these debugging symbols defined. It will help highlight potential bugs in
10149 their own code, and will improve the test coverage for OpenSSL itself. At
10150 some point, these tighter rules will become openssl's default to improve
10151 maintainability, though the assert()s and other overheads will remain only
10152 in debugging configurations. See bn.h for more details.
10154 *Geoff Thorpe, Nils Larsch, Ulf Möller*
10156 * BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
10157 that can only be obtained through BN_CTX_new() (which implicitly
10158 initialises it). The presence of this function only made it possible
10159 to overwrite an existing structure (and cause memory leaks).
10163 * Because of the callback-based approach for implementing LHASH as a
10164 template type, lh_insert() adds opaque objects to hash-tables and
10165 lh_doall() or lh_doall_arg() are typically used with a destructor callback
10166 to clean up those corresponding objects before destroying the hash table
10167 (and losing the object pointers). So some over-zealous constifications in
10168 LHASH have been relaxed so that lh_insert() does not take (nor store) the
10169 objects as "const" and the `lh_doall[_arg]` callback wrappers are not
10170 prototyped to have "const" restrictions on the object pointers they are
10171 given (and so aren't required to cast them away any more).
10175 * The tmdiff.h API was so ugly and minimal that our own timing utility
10176 (speed) prefers to use its own implementation. The two implementations
10177 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
10178 its object type properly exposed (MS_TM) instead of casting to/from
10179 `char *`. This may still change yet if someone realises MS_TM and
10181 aren't necessarily the greatest nomenclatures - but this is what was used
10182 internally to the implementation so I've used that for now.
10186 * Ensure that deprecated functions do not get compiled when
10187 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
10188 the self-tests were still using deprecated key-generation functions so
10189 these have been updated also.
10193 * Reorganise PKCS#7 code to separate the digest location functionality
10194 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
10195 New function PKCS7_set_digest() to set the digest type for PKCS#7
10196 digestedData type. Add additional code to correctly generate the
10197 digestedData type and add support for this type in PKCS7 initialization
10202 * New function PKCS7_set0_type_other() this initializes a PKCS7
10203 structure of type "other".
10207 * Fix prime generation loop in crypto/bn/bn_prime.pl by making
10208 sure the loop does correctly stop and breaking ("division by zero")
10209 modulus operations are not performed. The (pre-generated) prime
10210 table crypto/bn/bn_prime.h was already correct, but it could not be
10211 re-generated on some platforms because of the "division by zero"
10212 situation in the script.
10214 *Ralf S. Engelschall*
10216 * Update support for ECC-based TLS ciphersuites according to
10217 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
10218 SHA-1 now is only used for "small" curves (where the
10219 representation of a field element takes up to 24 bytes); for
10220 larger curves, the field element resulting from ECDH is directly
10221 used as premaster secret.
10223 *Douglas Stebila (Sun Microsystems Laboratories)*
10225 * Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
10226 curve secp160r1 to the tests.
10228 *Douglas Stebila (Sun Microsystems Laboratories)*
10230 * Add the possibility to load symbols globally with DSO.
10232 *Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte*
10234 * Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
10235 control of the error stack.
10239 * Add support for STORE in ENGINE.
10243 * Add the STORE type. The intention is to provide a common interface
10244 to certificate and key stores, be they simple file-based stores, or
10245 HSM-type store, or LDAP stores, or...
10246 NOTE: The code is currently UNTESTED and isn't really used anywhere.
10250 * Add a generic structure called OPENSSL_ITEM. This can be used to
10251 pass a list of arguments to any function as well as provide a way
10252 for a function to pass data back to the caller.
10256 * Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
10257 works like BUF_strdup() but can be used to duplicate a portion of
10258 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
10263 * Add the function sk_find_ex() which works like sk_find(), but will
10264 return an index to an element even if an exact match couldn't be
10265 found. The index is guaranteed to point at the element where the
10266 searched-for key would be inserted to preserve sorting order.
10270 * Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
10271 takes an extra flags argument for optional functionality. Currently,
10272 the following flags are defined:
10274 OBJ_BSEARCH_VALUE_ON_NOMATCH
10275 This one gets OBJ_bsearch_ex() to return a pointer to the first
10276 element where the comparing function returns a negative or zero
10279 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
10280 This one gets OBJ_bsearch_ex() to return a pointer to the first
10281 element where the comparing function returns zero. This is useful
10282 if there are more than one element where the comparing function
10287 * Make it possible to create self-signed certificates with 'openssl ca'
10288 in such a way that the self-signed certificate becomes part of the
10289 CA database and uses the same mechanisms for serial number generation
10290 as all other certificate signing. The new flag '-selfsign' enables
10291 this functionality. Adapt CA.sh and CA.pl.in.
10295 * Add functionality to check the public key of a certificate request
10296 against a given private. This is useful to check that a certificate
10297 request can be signed by that key (self-signing).
10301 * Make it possible to have multiple active certificates with the same
10302 subject in the CA index file. This is done only if the keyword
10303 'unique_subject' is set to 'no' in the main CA section (default
10304 if 'CA_default') of the configuration file. The value is saved
10305 with the database itself in a separate index attribute file,
10306 named like the index file with '.attr' appended to the name.
10310 * Generate multi-valued AVAs using '+' notation in config files for
10315 * Support for nameConstraints certificate extension.
10319 * Support for policyConstraints certificate extension.
10323 * Support for policyMappings certificate extension.
10327 * Make sure the default DSA_METHOD implementation only uses its
10328 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
10329 and change its own handlers to be NULL so as to remove unnecessary
10330 indirection. This lets alternative implementations fallback to the
10331 default implementation more easily.
10335 * Support for directoryName in GeneralName related extensions
10340 * Make it possible to link applications using Makefile.shared.
10341 Make that possible even when linking against static libraries!
10345 * Support for single pass processing for S/MIME signing. This now
10346 means that S/MIME signing can be done from a pipe, in addition
10347 cleartext signing (multipart/signed type) is effectively streaming
10348 and the signed data does not need to be all held in memory.
10350 This is done with a new flag PKCS7_STREAM. When this flag is set
10351 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
10352 is done after the data is output (and digests calculated) in
10353 SMIME_write_PKCS7().
10357 * Add full support for -rpath/-R, both in shared libraries and
10358 applications, at least on the platforms where it's known how
10363 * In crypto/ec/ec_mult.c, implement fast point multiplication with
10364 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
10365 will now compute a table of multiples of the generator that
10366 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
10367 faster (notably in the case of a single point multiplication,
10368 scalar * generator).
10370 *Nils Larsch, Bodo Moeller*
10372 * IPv6 support for certificate extensions. The various extensions
10373 which use the IP:a.b.c.d can now take IPv6 addresses using the
10374 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
10379 * Added an ENGINE that implements RSA by performing private key
10380 exponentiations with the GMP library. The conversions to and from
10381 GMP's mpz_t format aren't optimised nor are any montgomery forms
10382 cached, and on x86 it appears OpenSSL's own performance has caught up.
10383 However there are likely to be other architectures where GMP could
10384 provide a boost. This ENGINE is not built in by default, but it can be
10385 specified at Configure time and should be accompanied by the necessary
10386 linker additions, eg;
10387 ./config -DOPENSSL_USE_GMP -lgmp
10391 * "openssl engine" will not display ENGINE/DSO load failure errors when
10392 testing availability of engines with "-t" - the old behaviour is
10393 produced by increasing the feature's verbosity with "-tt".
10397 * ECDSA routines: under certain error conditions uninitialized BN objects
10398 could be freed. Solution: make sure initialization is performed early
10399 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
10404 * Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
10405 and DH_METHOD (eg. by ENGINE implementations) to override the normal
10406 software implementations. For DSA and DH, parameter generation can
10407 also be overridden by providing the appropriate method callbacks.
10411 * Change the "progress" mechanism used in key-generation and
10412 primality testing to functions that take a new BN_GENCB pointer in
10413 place of callback/argument pairs. The new API functions have `_ex`
10414 postfixes and the older functions are reimplemented as wrappers for
10415 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
10416 declarations of the old functions to help (graceful) attempts to
10417 migrate to the new functions. Also, the new key-generation API
10418 functions operate on a caller-supplied key-structure and return
10419 success/failure rather than returning a key or NULL - this is to
10420 help make "keygen" another member function of RSA_METHOD etc.
10422 Example for using the new callback interface:
10424 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
10425 void *my_arg = ...;
10428 BN_GENCB_set(&my_cb, my_callback, my_arg);
10430 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
10431 /* For the meaning of a, b in calls to my_callback(), see the
10432 * documentation of the function that calls the callback.
10433 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
10434 * my_callback should return 1 if it wants BN_is_prime_ex()
10435 * to continue, or 0 to stop.
10440 * Change the ZLIB compression method to be stateful, and make it
10441 available to TLS with the number defined in
10442 draft-ietf-tls-compression-04.txt.
10446 * Add the ASN.1 structures and functions for CertificatePair, which
10447 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
10449 CertificatePair ::= SEQUENCE {
10450 forward [0] Certificate OPTIONAL,
10451 reverse [1] Certificate OPTIONAL,
10452 -- at least one of the pair shall be present -- }
10454 Also implement the PEM functions to read and write certificate
10455 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
10457 This needed to be defined, mostly for the sake of the LDAP
10458 attribute crossCertificatePair, but may prove useful elsewhere as
10463 * Make it possible to inhibit symlinking of shared libraries in
10464 Makefile.shared, for Cygwin's sake.
10468 * Extend the BIGNUM API by creating a function
10469 void BN_set_negative(BIGNUM *a, int neg);
10470 and a macro that behave like
10471 int BN_is_negative(const BIGNUM *a);
10473 to avoid the need to access 'a->neg' directly in applications.
10477 * Implement fast modular reduction for pseudo-Mersenne primes
10478 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
10479 EC_GROUP_new_curve_GFp() will now automatically use this
10482 *Nils Larsch <nla@trustcenter.de>*
10484 * Add new lock type (CRYPTO_LOCK_BN).
10488 * Change the ENGINE framework to automatically load engines
10489 dynamically from specific directories unless they could be
10490 found to already be built in or loaded. Move all the
10491 current engines except for the cryptodev one to a new
10492 directory engines/.
10493 The engines in engines/ are built as shared libraries if
10494 the "shared" options was given to ./Configure or ./config.
10495 Otherwise, they are inserted in libcrypto.a.
10496 /usr/local/ssl/engines is the default directory for dynamic
10497 engines, but that can be overridden at configure time through
10498 the usual use of --prefix and/or --openssldir, and at run
10499 time with the environment variable OPENSSL_ENGINES.
10501 *Geoff Thorpe and Richard Levitte*
10503 * Add Makefile.shared, a helper makefile to build shared
10504 libraries. Adapt Makefile.org.
10508 * Add version info to Win32 DLLs.
10510 *Peter 'Luna' Runestig" <peter@runestig.com>*
10512 * Add new 'medium level' PKCS#12 API. Certificates and keys
10513 can be added using this API to created arbitrary PKCS#12
10514 files while avoiding the low level API.
10516 New options to PKCS12_create(), key or cert can be NULL and
10517 will then be omitted from the output file. The encryption
10518 algorithm NIDs can be set to -1 for no encryption, the mac
10519 iteration count can be set to 0 to omit the mac.
10521 Enhance pkcs12 utility by making the -nokeys and -nocerts
10522 options work when creating a PKCS#12 file. New option -nomac
10523 to omit the mac, NONE can be set for an encryption algorithm.
10524 New code is modified to use the enhanced PKCS12_create()
10525 instead of the low level API.
10529 * Extend ASN1 encoder to support indefinite length constructed
10530 encoding. This can output sequences tags and octet strings in
10531 this form. Modify pk7_asn1.c to support indefinite length
10532 encoding. This is experimental and needs additional code to
10533 be useful, such as an ASN1 bio and some enhanced streaming
10536 Extend template encode functionality so that tagging is passed
10537 down to the template encoder.
10541 * Let 'openssl req' fail if an argument to '-newkey' is not
10542 recognized instead of using RSA as a default.
10546 * Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
10547 As these are not official, they are not included in "ALL";
10548 the "ECCdraft" ciphersuite group alias can be used to select them.
10550 *Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)*
10552 * Add ECDH engine support.
10554 *Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)*
10556 * Add ECDH in new directory crypto/ecdh/.
10558 *Douglas Stebila (Sun Microsystems Laboratories)*
10560 * Let BN_rand_range() abort with an error after 100 iterations
10561 without success (which indicates a broken PRNG).
10565 * Change BN_mod_sqrt() so that it verifies that the input value
10566 is really the square of the return value. (Previously,
10567 BN_mod_sqrt would show GIGO behaviour.)
10571 * Add named elliptic curves over binary fields from X9.62, SECG,
10572 and WAP/WTLS; add OIDs that were still missing.
10574 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10576 * Extend the EC library for elliptic curves over binary fields
10577 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
10580 EC_GF2m_simple_method
10584 EC_GROUP_new_curve_GF2m
10585 EC_GROUP_set_curve_GF2m
10586 EC_GROUP_get_curve_GF2m
10587 EC_POINT_set_affine_coordinates_GF2m
10588 EC_POINT_get_affine_coordinates_GF2m
10589 EC_POINT_set_compressed_coordinates_GF2m
10591 Point compression for binary fields is disabled by default for
10592 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
10595 As binary polynomials are represented as BIGNUMs, various members
10596 of the EC_GROUP and EC_POINT data structures can be shared
10597 between the implementations for prime fields and binary fields;
10598 the above `..._GF2m functions` (except for EX_GROUP_new_curve_GF2m)
10599 are essentially identical to their `..._GFp` counterparts.
10600 (For simplicity, the `..._GFp` prefix has been dropped from
10601 various internal method names.)
10603 An internal 'field_div' method (similar to 'field_mul' and
10604 'field_sqr') has been added; this is used only for binary fields.
10606 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10608 * Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
10609 through methods ('mul', 'precompute_mult').
10611 The generic implementations (now internally called 'ec_wNAF_mul'
10612 and 'ec_wNAF_precomputed_mult') remain the default if these
10613 methods are undefined.
10615 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10617 * New function EC_GROUP_get_degree, which is defined through
10618 EC_METHOD. For curves over prime fields, this returns the bit
10619 length of the modulus.
10621 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10623 * New functions EC_GROUP_dup, EC_POINT_dup.
10624 (These simply call ..._new and ..._copy).
10626 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10628 * Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
10629 Polynomials are represented as BIGNUMs (where the sign bit is not
10630 used) in the following functions [macros]:
10633 BN_GF2m_sub [= BN_GF2m_add]
10634 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
10635 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
10636 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
10638 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
10639 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
10640 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
10641 BN_GF2m_cmp [= BN_ucmp]
10643 (Note that only the 'mod' functions are actually for fields GF(2^m).
10644 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
10646 For some functions, an the irreducible polynomial defining a
10647 field can be given as an 'unsigned int[]' with strictly
10648 decreasing elements giving the indices of those bits that are set;
10649 i.e., p[] represents the polynomial
10650 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
10652 p[0] > p[1] > ... > p[k] = 0.
10653 This applies to the following functions:
10656 BN_GF2m_mod_mul_arr
10657 BN_GF2m_mod_sqr_arr
10658 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
10659 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
10660 BN_GF2m_mod_exp_arr
10661 BN_GF2m_mod_sqrt_arr
10662 BN_GF2m_mod_solve_quad_arr
10666 Conversion can be performed by the following functions:
10671 bntest.c has additional tests for binary polynomial arithmetic.
10673 Two implementations for BN_GF2m_mod_div() are available.
10674 The default algorithm simply uses BN_GF2m_mod_inv() and
10675 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
10676 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
10677 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
10679 *Sheueling Chang Shantz and Douglas Stebila (Sun Microsystems Laboratories)*
10681 * Add new error code 'ERR_R_DISABLED' that can be used when some
10682 functionality is disabled at compile-time.
10684 *Douglas Stebila <douglas.stebila@sun.com>*
10686 * Change default behaviour of 'openssl asn1parse' so that more
10687 information is visible when viewing, e.g., a certificate:
10689 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
10690 mode the content of non-printable OCTET STRINGs is output in a
10691 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
10692 avoid the appearance of a printable string.
10694 *Nils Larsch <nla@trustcenter.de>*
10696 * Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
10698 EC_GROUP_set_asn1_flag()
10699 EC_GROUP_get_asn1_flag()
10700 EC_GROUP_set_point_conversion_form()
10701 EC_GROUP_get_point_conversion_form()
10702 These control ASN1 encoding details:
10703 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
10704 has been set to OPENSSL_EC_NAMED_CURVE.
10705 - Points are encoded in uncompressed form by default; options for
10706 asn1_for are as for point2oct, namely
10707 POINT_CONVERSION_COMPRESSED
10708 POINT_CONVERSION_UNCOMPRESSED
10709 POINT_CONVERSION_HYBRID
10711 Also add 'seed' and 'seed_len' members to EC_GROUP with access
10713 EC_GROUP_set_seed()
10714 EC_GROUP_get0_seed()
10715 EC_GROUP_get_seed_len()
10716 This is used only for ASN1 purposes (so far).
10718 *Nils Larsch <nla@trustcenter.de>*
10720 * Add 'field_type' member to EC_METHOD, which holds the NID
10721 of the appropriate field type OID. The new function
10722 EC_METHOD_get_field_type() returns this value.
10724 *Nils Larsch <nla@trustcenter.de>*
10727 EC_POINT_point2bn()
10728 EC_POINT_bn2point()
10729 EC_POINT_point2hex()
10730 EC_POINT_hex2point()
10731 providing useful interfaces to EC_POINT_point2oct() and
10732 EC_POINT_oct2point().
10734 *Nils Larsch <nla@trustcenter.de>*
10736 * Change internals of the EC library so that the functions
10737 EC_GROUP_set_generator()
10738 EC_GROUP_get_generator()
10739 EC_GROUP_get_order()
10740 EC_GROUP_get_cofactor()
10741 are implemented directly in crypto/ec/ec_lib.c and not dispatched
10742 to methods, which would lead to unnecessary code duplication when
10743 adding different types of curves.
10745 *Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller*
10747 * Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
10748 arithmetic, and such that modified wNAFs are generated
10749 (which avoid length expansion in many cases).
10753 * Add a function EC_GROUP_check_discriminant() (defined via
10754 EC_METHOD) that verifies that the curve discriminant is non-zero.
10756 Add a function EC_GROUP_check() that makes some sanity tests
10757 on a EC_GROUP, its generator and order. This includes
10758 EC_GROUP_check_discriminant().
10760 *Nils Larsch <nla@trustcenter.de>*
10762 * Add ECDSA in new directory crypto/ecdsa/.
10764 Add applications 'openssl ecparam' and 'openssl ecdsa'
10765 (these are based on 'openssl dsaparam' and 'openssl dsa').
10767 ECDSA support is also included in various other files across the
10768 library. Most notably,
10769 - 'openssl req' now has a '-newkey ecdsa:file' option;
10770 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
10771 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
10772 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
10773 them suitable for ECDSA where domain parameters must be
10774 extracted before the specific public key;
10775 - ECDSA engine support has been added.
10777 *Nils Larsch <nla@trustcenter.de>*
10779 * Include some named elliptic curves, and add OIDs from X9.62,
10780 SECG, and WAP/WTLS. Each curve can be obtained from the new
10782 EC_GROUP_new_by_curve_name(),
10783 and the list of available named curves can be obtained with
10784 EC_get_builtin_curves().
10785 Also add a 'curve_name' member to EC_GROUP objects, which can be
10787 EC_GROUP_set_curve_name()
10788 EC_GROUP_get_curve_name()
10790 *Nils Larsch <larsch@trustcenter.de, Bodo Moeller*
10792 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
10793 was actually never needed) and in BN_mul(). The removal in BN_mul()
10794 required a small change in bn_mul_part_recursive() and the addition
10795 of the functions bn_cmp_part_words(), bn_sub_part_words() and
10796 bn_add_part_words(), which do the same thing as bn_cmp_words(),
10797 bn_sub_words() and bn_add_words() except they take arrays with
10802 ### Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
10804 * Cleanse PEM buffers before freeing them since they may contain
10807 *Benjamin Bennett <ben@psc.edu>*
10809 * Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
10810 a ciphersuite string such as "DEFAULT:RSA" cannot enable
10811 authentication-only ciphersuites.
10815 * Since AES128 and AES256 share a single mask bit in the logic of
10816 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
10817 kludge to work properly if AES128 is available and AES256 isn't.
10821 * Expand security boundary to match 1.1.1 module.
10825 * Remove redundant features: hash file source, editing of test vectors
10826 modify fipsld to use external fips_premain.c signature.
10830 * New perl script mkfipsscr.pl to create shell scripts or batch files to
10831 run algorithm test programs.
10835 * Make algorithm test programs more tolerant of whitespace.
10839 * Have SSL/TLS server implementation tolerate "mismatched" record
10840 protocol version while receiving ClientHello even if the
10841 ClientHello is fragmented. (The server can't insist on the
10842 particular protocol version it has chosen before the ServerHello
10843 message has informed the client about his choice.)
10847 * Load error codes if they are not already present instead of using a
10848 static variable. This allows them to be cleanly unloaded and reloaded.
10852 ### Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
10854 * Introduce limits to prevent malicious keys being able to
10855 cause a denial of service. ([CVE-2006-2940])
10857 *Steve Henson, Bodo Moeller*
10859 * Fix ASN.1 parsing of certain invalid structures that can result
10860 in a denial of service. ([CVE-2006-2937]) [Steve Henson]
10862 * Fix buffer overflow in SSL_get_shared_ciphers() function.
10863 ([CVE-2006-3738]) [Tavis Ormandy and Will Drewry, Google Security Team]
10865 * Fix SSL client code which could crash if connecting to a
10866 malicious SSLv2 server. ([CVE-2006-4343])
10868 *Tavis Ormandy and Will Drewry, Google Security Team*
10870 * Change ciphersuite string processing so that an explicit
10871 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
10872 will no longer include "AES128-SHA"), and any other similar
10873 ciphersuite (same bitmap) from *other* protocol versions (so that
10874 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
10875 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
10876 changes from 0.9.8b and 0.9.8d.
10880 ### Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
10882 * Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
10883 ([CVE-2006-4339]) [Ben Laurie and Google Security Team]
10885 * Change the Unix randomness entropy gathering to use poll() when
10886 possible instead of select(), since the latter has some
10887 undesirable limitations.
10889 *Darryl Miles via Richard Levitte and Bodo Moeller*
10891 * Disable rogue ciphersuites:
10893 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
10894 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
10895 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
10897 The latter two were purportedly from
10898 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
10901 Also deactivate the remaining ciphersuites from
10902 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
10903 unofficial, and the ID has long expired.
10907 * Fix RSA blinding Heisenbug (problems sometimes occurred on
10908 dual-core machines) and other potential thread-safety issues.
10912 ### Changes between 0.9.7i and 0.9.7j [04 May 2006]
10914 * Adapt fipsld and the build system to link against the validated FIPS
10915 module in FIPS mode.
10919 * Fixes for VC++ 2005 build under Windows.
10923 * Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
10924 from a Windows bash shell such as MSYS. It is autodetected from the
10925 "config" script when run from a VC++ environment. Modify standard VC++
10926 build to use fipscanister.o from the GNU make build.
10930 ### Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
10932 * Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
10933 The value now differs depending on if you build for FIPS or not.
10934 BEWARE! A program linked with a shared FIPSed libcrypto can't be
10935 safely run with a non-FIPSed libcrypto, as it may crash because of
10936 the difference induced by this change.
10940 ### Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
10942 * Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
10943 (part of SSL_OP_ALL). This option used to disable the
10944 countermeasure against man-in-the-middle protocol-version
10945 rollback in the SSL 2.0 server implementation, which is a bad
10946 idea. ([CVE-2005-2969])
10948 *Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
10949 for Information Security, National Institute of Advanced Industrial
10950 Science and Technology [AIST, Japan)]*
10952 * Minimal support for X9.31 signatures and PSS padding modes. This is
10953 mainly for FIPS compliance and not fully integrated at this stage.
10957 * For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
10958 the exponentiation using a fixed-length exponent. (Otherwise,
10959 the information leaked through timing could expose the secret key
10960 after many signatures; cf. Bleichenbacher's attack on DSA with
10965 * Make a new fixed-window mod_exp implementation the default for
10966 RSA, DSA, and DH private-key operations so that the sequence of
10967 squares and multiplies and the memory access pattern are
10968 independent of the particular secret key. This will mitigate
10969 cache-timing and potential related attacks.
10971 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
10972 and this is automatically used by BN_mod_exp_mont() if the new flag
10973 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
10974 will use this BN flag for private exponents unless the flag
10975 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
10976 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
10978 *Matthew D Wood (Intel Corp), with some changes by Bodo Moeller*
10980 * Change the client implementation for SSLv23_method() and
10981 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
10982 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
10983 (Previously, the SSL 2.0 backwards compatible Client Hello
10984 message format would be used even with SSL_OP_NO_SSLv2.)
10988 * Add support for smime-type MIME parameter in S/MIME messages which some
10993 * New function BN_MONT_CTX_set_locked() to set montgomery parameters in
10994 a threadsafe manner. Modify rsa code to use new function and add calls
10995 to dsa and dh code (which had race conditions before).
10999 * Include the fixed error library code in the C error file definitions
11000 instead of fixing them up at runtime. This keeps the error code
11001 structures constant.
11005 ### Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
11007 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
11010 * Fixes for newer kerberos headers. NB: the casts are needed because
11011 the 'length' field is signed on one version and unsigned on another
11012 with no (?) obvious way to tell the difference, without these VC++
11013 complains. Also the "definition" of FAR (blank) is no longer included
11014 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
11015 some needed definitions.
11019 * Undo Cygwin change.
11023 * Added support for proxy certificates according to RFC 3820.
11024 Because they may be a security thread to unaware applications,
11025 they must be explicitly allowed in run-time. See
11026 docs/HOWTO/proxy_certificates.txt for further information.
11030 ### Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
11032 * Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
11033 server and client random values. Previously
11034 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
11035 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
11037 This change has negligible security impact because:
11039 1. Server and client random values still have 24 bytes of pseudo random
11042 2. Server and client random values are sent in the clear in the initial
11045 3. The master secret is derived using the premaster secret (48 bytes in
11046 size for static RSA ciphersuites) as well as client server and random
11049 The OpenSSL team would like to thank the UK NISCC for bringing this issue
11052 *Stephen Henson, reported by UK NISCC*
11054 * Use Windows randomness collection on Cygwin.
11058 * Fix hang in EGD/PRNGD query when communication socket is closed
11059 prematurely by EGD/PRNGD.
11061 *Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014*
11063 * Prompt for pass phrases when appropriate for PKCS12 input format.
11067 * Back-port of selected performance improvements from development
11068 branch, as well as improved support for PowerPC platforms.
11072 * Add lots of checks for memory allocation failure, error codes to indicate
11073 failure and freeing up memory if a failure occurs.
11075 *Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson*
11077 * Add new -passin argument to dgst.
11081 * Perform some character comparisons of different types in X509_NAME_cmp:
11082 this is needed for some certificates that re-encode DNs into UTF8Strings
11083 (in violation of RFC3280) and can't or won't issue name rollover
11088 * Make an explicit check during certificate validation to see that
11089 the CA setting in each certificate on the chain is correct. As a
11090 side effect always do the following basic checks on extensions,
11091 not just when there's an associated purpose to the check:
11093 - if there is an unhandled critical extension (unless the user
11094 has chosen to ignore this fault)
11095 - if the path length has been exceeded (if one is set at all)
11096 - that certain extensions fit the associated purpose (if one has
11101 ### Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
11103 * Avoid a race condition when CRLs are checked in a multi threaded
11104 environment. This would happen due to the reordering of the revoked
11105 entries during signature checking and serial number lookup. Now the
11106 encoding is cached and the serial number sort performed under a lock.
11107 Add new STACK function sk_is_sorted().
11111 * Add Delta CRL to the extension code.
11115 * Various fixes to s3_pkt.c so alerts are sent properly.
11117 *David Holmes <d.holmes@f5.com>*
11119 * Reduce the chances of duplicate issuer name and serial numbers (in
11120 violation of RFC3280) using the OpenSSL certificate creation utilities.
11121 This is done by creating a random 64 bit value for the initial serial
11122 number when a serial number file is created or when a self signed
11123 certificate is created using 'openssl req -x509'. The initial serial
11124 number file is created using 'openssl x509 -next_serial' in CA.pl
11125 rather than being initialized to 1.
11129 ### Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
11131 * Fix null-pointer assignment in do_change_cipher_spec() revealed
11132 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
11134 *Joe Orton, Steve Henson*
11136 * Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
11139 *Joe Orton, Steve Henson*
11141 * Make it possible to have multiple active certificates with the same
11142 subject in the CA index file. This is done only if the keyword
11143 'unique_subject' is set to 'no' in the main CA section (default
11144 if 'CA_default') of the configuration file. The value is saved
11145 with the database itself in a separate index attribute file,
11146 named like the index file with '.attr' appended to the name.
11150 * X509 verify fixes. Disable broken certificate workarounds when
11151 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
11152 keyUsage extension present. Don't accept CRLs with unhandled critical
11153 extensions: since verify currently doesn't process CRL extensions this
11154 rejects a CRL with *any* critical extensions. Add new verify error codes
11159 * When creating an OCSP nonce use an OCTET STRING inside the extnValue.
11160 A clarification of RFC2560 will require the use of OCTET STRINGs and
11161 some implementations cannot handle the current raw format. Since OpenSSL
11162 copies and compares OCSP nonces as opaque blobs without any attempt at
11163 parsing them this should not create any compatibility issues.
11167 * New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
11168 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
11169 this HMAC (and other) operations are several times slower than OpenSSL
11174 * Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
11176 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11178 * Use the correct content when signing type "other".
11182 ### Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
11184 * Fix various bugs revealed by running the NISCC test suite:
11186 Stop out of bounds reads in the ASN1 code when presented with
11187 invalid tags (CVE-2003-0543 and CVE-2003-0544).
11189 Free up ASN1_TYPE correctly if ANY type is invalid ([CVE-2003-0545]).
11191 If verify callback ignores invalid public key errors don't try to check
11192 certificate signature with the NULL public key.
11196 * New -ignore_err option in ocsp application to stop the server
11197 exiting on the first error in a request.
11201 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
11202 if the server requested one: as stated in TLS 1.0 and SSL 3.0
11207 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
11208 extra data after the compression methods not only for TLS 1.0
11209 but also for SSL 3.0 (as required by the specification).
11211 *Bodo Moeller; problem pointed out by Matthias Loepfe*
11213 * Change X509_certificate_type() to mark the key as exported/exportable
11214 when it's 512 *bits* long, not 512 bytes.
11218 * Change AES_cbc_encrypt() so it outputs exact multiple of
11219 blocks during encryption.
11223 * Various fixes to base64 BIO and non blocking I/O. On write
11224 flushes were not handled properly if the BIO retried. On read
11225 data was not being buffered properly and had various logic bugs.
11226 This also affects blocking I/O when the data being decoded is a
11231 * Various S/MIME bugfixes and compatibility changes:
11232 output correct application/pkcs7 MIME type if
11233 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
11234 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
11235 of files as .eml work). Correctly handle very long lines in MIME
11240 ### Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
11242 * Countermeasure against the Klima-Pokorny-Rosa extension of
11243 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
11244 a protocol version number mismatch like a decryption error
11245 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
11249 * Turn on RSA blinding by default in the default implementation
11250 to avoid a timing attack. Applications that don't want it can call
11251 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
11252 They would be ill-advised to do so in most cases.
11254 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
11256 * Change RSA blinding code so that it works when the PRNG is not
11257 seeded (in this case, the secret RSA exponent is abused as
11258 an unpredictable seed -- if it is not unpredictable, there
11259 is no point in blinding anyway). Make RSA blinding thread-safe
11260 by remembering the creator's thread ID in rsa->blinding and
11261 having all other threads use local one-time blinding factors
11262 (this requires more computation than sharing rsa->blinding, but
11263 avoids excessive locking; and if an RSA object is not shared
11264 between threads, blinding will still be very fast).
11268 * Fixed a typo bug that would cause ENGINE_set_default() to set an
11269 ENGINE as defaults for all supported algorithms irrespective of
11270 the 'flags' parameter. 'flags' is now honoured, so applications
11271 should make sure they are passing it correctly.
11275 * Target "mingw" now allows native Windows code to be generated in
11276 the Cygwin environment as well as with the MinGW compiler.
11280 ### Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
11282 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
11283 via timing by performing a MAC computation even if incorrect
11284 block cipher padding has been found. This is a countermeasure
11285 against active attacks where the attacker has to distinguish
11286 between bad padding and a MAC verification error. ([CVE-2003-0078])
11288 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
11289 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
11290 Martin Vuagnoux (EPFL, Ilion)*
11292 * Make the no-err option work as intended. The intention with no-err
11293 is not to have the whole error stack handling routines removed from
11294 libcrypto, it's only intended to remove all the function name and
11295 reason texts, thereby removing some of the footprint that may not
11296 be interesting if those errors aren't displayed anyway.
11298 NOTE: it's still possible for any application or module to have its
11299 own set of error texts inserted. The routines are there, just not
11300 used by default when no-err is given.
11304 * Add support for FreeBSD on IA64.
11306 *dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454*
11308 * Adjust DES_cbc_cksum() so it returns the same value as the MIT
11309 Kerberos function mit_des_cbc_cksum(). Before this change,
11310 the value returned by DES_cbc_cksum() was like the one from
11311 mit_des_cbc_cksum(), except the bytes were swapped.
11313 *Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte*
11315 * Allow an application to disable the automatic SSL chain building.
11316 Before this a rather primitive chain build was always performed in
11317 ssl3_output_cert_chain(): an application had no way to send the
11318 correct chain if the automatic operation produced an incorrect result.
11320 Now the chain builder is disabled if either:
11322 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
11324 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
11326 The reasoning behind this is that an application would not want the
11327 auto chain building to take place if extra chain certificates are
11328 present and it might also want a means of sending no additional
11329 certificates (for example the chain has two certificates and the
11334 * Add the possibility to build without the ENGINE framework.
11336 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11338 * Under Win32 gmtime() can return NULL: check return value in
11339 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
11343 * DSA routines: under certain error conditions uninitialized BN objects
11344 could be freed. Solution: make sure initialization is performed early
11345 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
11346 Nils Larsch <nla@trustcenter.de> via PR#459)
11350 * Another fix for SSLv2 session ID handling: the session ID was incorrectly
11351 checked on reconnect on the client side, therefore session resumption
11352 could still fail with a "ssl session id is different" error. This
11353 behaviour is masked when SSL_OP_ALL is used due to
11354 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
11355 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11356 followup to PR #377.
11360 * IA-32 assembler support enhancements: unified ELF targets, support
11361 for SCO/Caldera platforms, fix for Cygwin shared build.
11365 * Add support for FreeBSD on sparc64. As a consequence, support for
11366 FreeBSD on non-x86 processors is separate from x86 processors on
11367 the config script, much like the NetBSD support.
11369 *Richard Levitte & Kris Kennaway <kris@obsecurity.org>*
11371 ### Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
11373 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
11376 * Fix session ID handling in SSLv2 client code: the SERVER FINISHED
11377 code (06) was taken as the first octet of the session ID and the last
11378 octet was ignored consequently. As a result SSLv2 client side session
11379 caching could not have worked due to the session ID mismatch between
11381 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
11386 * Change the declaration of needed Kerberos libraries to use EX_LIBS
11387 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
11392 * The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
11393 seems that in spite of existing for more than a year, many application
11394 author have done nothing to provide the necessary callbacks, which
11395 means that this particular engine will not work properly anywhere.
11396 This is a very unfortunate situation which forces us, in the name
11397 of usability, to give the hw_ncipher.c a static lock, which is part
11399 NOTE: This is for the 0.9.7 series ONLY. This hack will never
11400 appear in 0.9.8 or later. We EXPECT application authors to have
11401 dealt properly with this when 0.9.8 is released (unless we actually
11402 make such changes in the libcrypto locking code that changes will
11403 have to be made anyway).
11407 * In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
11408 octets have been read, EOF or an error occurs. Without this change
11409 some truncated ASN1 structures will not produce an error.
11413 * Disable Heimdal support, since it hasn't been fully implemented.
11414 Still give the possibility to force the use of Heimdal, but with
11415 warnings and a request that patches get sent to openssl-dev.
11419 * Add the VC-CE target, introduce the WINCE sysname, and add
11420 INSTALL.WCE and appropriate conditionals to make it build.
11422 *Steven Reddie <smr@essemer.com.au> via Richard Levitte*
11424 * Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
11425 cygssl-x.y.z.dll, where x, y and z are the major, minor and
11426 edit numbers of the version.
11428 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
11430 * Introduce safe string copy and catenation functions
11431 (BUF_strlcpy() and BUF_strlcat()).
11433 *Ben Laurie (CHATS) and Richard Levitte*
11435 * Avoid using fixed-size buffers for one-line DNs.
11437 *Ben Laurie (CHATS)*
11439 * Add BUF_MEM_grow_clean() to avoid information leakage when
11440 resizing buffers containing secrets, and use where appropriate.
11442 *Ben Laurie (CHATS)*
11444 * Avoid using fixed size buffers for configuration file location.
11446 *Ben Laurie (CHATS)*
11448 * Avoid filename truncation for various CA files.
11450 *Ben Laurie (CHATS)*
11452 * Use sizeof in preference to magic numbers.
11454 *Ben Laurie (CHATS)*
11456 * Avoid filename truncation in cert requests.
11458 *Ben Laurie (CHATS)*
11460 * Add assertions to check for (supposedly impossible) buffer
11463 *Ben Laurie (CHATS)*
11465 * Don't cache truncated DNS entries in the local cache (this could
11466 potentially lead to a spoofing attack).
11468 *Ben Laurie (CHATS)*
11470 * Fix various buffers to be large enough for hex/decimal
11471 representations in a platform independent manner.
11473 *Ben Laurie (CHATS)*
11475 * Add CRYPTO_realloc_clean() to avoid information leakage when
11476 resizing buffers containing secrets, and use where appropriate.
11478 *Ben Laurie (CHATS)*
11480 * Add BIO_indent() to avoid much slightly worrying code to do
11483 *Ben Laurie (CHATS)*
11485 * Convert sprintf()/BIO_puts() to BIO_printf().
11487 *Ben Laurie (CHATS)*
11489 * buffer_gets() could terminate with the buffer only half
11492 *Ben Laurie (CHATS)*
11494 * Add assertions to prevent user-supplied crypto functions from
11495 overflowing internal buffers by having large block sizes, etc.
11497 *Ben Laurie (CHATS)*
11499 * New OPENSSL_assert() macro (similar to assert(), but enabled
11502 *Ben Laurie (CHATS)*
11504 * Eliminate unused copy of key in RC4.
11506 *Ben Laurie (CHATS)*
11508 * Eliminate unused and incorrectly sized buffers for IV in pem.h.
11510 *Ben Laurie (CHATS)*
11512 * Fix off-by-one error in EGD path.
11514 *Ben Laurie (CHATS)*
11516 * If RANDFILE path is too long, ignore instead of truncating.
11518 *Ben Laurie (CHATS)*
11520 * Eliminate unused and incorrectly sized X.509 structure
11523 *Ben Laurie (CHATS)*
11525 * Eliminate unused and dangerous function knumber().
11527 *Ben Laurie (CHATS)*
11529 * Eliminate unused and dangerous structure, KSSL_ERR.
11531 *Ben Laurie (CHATS)*
11533 * Protect against overlong session ID context length in an encoded
11534 session object. Since these are local, this does not appear to be
11537 *Ben Laurie (CHATS)*
11539 * Change from security patch (see 0.9.6e below) that did not affect
11540 the 0.9.6 release series:
11542 Remote buffer overflow in SSL3 protocol - an attacker could
11543 supply an oversized master key in Kerberos-enabled versions.
11546 *Ben Laurie (CHATS)*
11548 * Change the SSL kerb5 codes to match RFC 2712.
11552 * Make -nameopt work fully for req and add -reqopt switch.
11554 *Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson*
11556 * The "block size" for block ciphers in CFB and OFB mode should be 1.
11558 *Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>*
11560 * Make sure tests can be performed even if the corresponding algorithms
11561 have been removed entirely. This was also the last step to make
11562 OpenSSL compilable with DJGPP under all reasonable conditions.
11564 *Richard Levitte, Doug Kaufman <dkaufman@rahul.net>*
11566 * Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
11567 to allow version independent disabling of normally unselected ciphers,
11568 which may be activated as a side-effect of selecting a single cipher.
11570 (E.g., cipher list string "RSA" enables ciphersuites that are left
11571 out of "ALL" because they do not provide symmetric encryption.
11572 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
11574 *Lutz Jaenicke, Bodo Moeller*
11576 * Add appropriate support for separate platform-dependent build
11577 directories. The recommended way to make a platform-dependent
11578 build directory is the following (tested on Linux), maybe with
11581 # Place yourself outside of the OpenSSL source tree. In
11582 # this example, the environment variable OPENSSL_SOURCE
11583 # is assumed to contain the absolute OpenSSL source directory.
11584 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
11585 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
11586 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
11587 mkdir -p `dirname $F`
11588 ln -s $OPENSSL_SOURCE/$F $F
11591 To be absolutely sure not to disturb the source tree, a "make clean"
11592 is a good thing. If it isn't successful, don't worry about it,
11593 it probably means the source directory is very clean.
11597 * Make sure any ENGINE control commands make local copies of string
11598 pointers passed to them whenever necessary. Otherwise it is possible
11599 the caller may have overwritten (or deallocated) the original string
11600 data when a later ENGINE operation tries to use the stored values.
11602 *Götz Babin-Ebell <babinebell@trustcenter.de>*
11604 * Improve diagnostics in file reading and command-line digests.
11606 *Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>*
11608 * Add AES modes CFB and OFB to the object database. Correct an
11609 error in AES-CFB decryption.
11613 * Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
11614 allows existing EVP_CIPHER_CTX structures to be reused after
11615 calling `EVP_*Final()`. This behaviour is used by encryption
11616 BIOs and some applications. This has the side effect that
11617 applications must explicitly clean up cipher contexts with
11618 EVP_CIPHER_CTX_cleanup() or they will leak memory.
11622 * Check the values of dna and dnb in bn_mul_recursive before calling
11623 bn_mul_comba (a non zero value means the a or b arrays do not contain
11624 n2 elements) and fallback to bn_mul_normal if either is not zero.
11628 * Fix escaping of non-ASCII characters when using the -subj option
11629 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
11633 * Make object definitions compliant to LDAP (RFC2256): SN is the short
11634 form for "surname", serialNumber has no short form.
11635 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
11636 therefore remove "mail" short name for "internet 7".
11637 The OID for unique identifiers in X509 certificates is
11638 x500UniqueIdentifier, not uniqueIdentifier.
11639 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
11643 * Add an "init" command to the ENGINE config module and auto initialize
11644 ENGINEs. Without any "init" command the ENGINE will be initialized
11645 after all ctrl commands have been executed on it. If init=1 the
11646 ENGINE is initialized at that point (ctrls before that point are run
11647 on the uninitialized ENGINE and after on the initialized one). If
11648 init=0 then the ENGINE will not be initialized at all.
11652 * Fix the 'app_verify_callback' interface so that the user-defined
11653 argument is actually passed to the callback: In the
11654 SSL_CTX_set_cert_verify_callback() prototype, the callback
11655 declaration has been changed from
11658 int (*cb)(X509_STORE_CTX *,void *);
11659 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
11660 i=s->ctx->app_verify_callback(&ctx)
11661 has been changed into
11662 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
11664 To update applications using SSL_CTX_set_cert_verify_callback(),
11665 a dummy argument can be added to their callback functions.
11667 *D. K. Smetters <smetters@parc.xerox.com>*
11669 * Added the '4758cca' ENGINE to support IBM 4758 cards.
11671 *Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe*
11673 * Add and OPENSSL_LOAD_CONF define which will cause
11674 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
11675 This allows older applications to transparently support certain
11676 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
11677 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
11678 load the config file and OPENSSL_add_all_algorithms_conf() which will
11679 always load it have also been added.
11683 * Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
11684 Adjust NIDs and EVP layer.
11686 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11688 * Config modules support in openssl utility.
11690 Most commands now load modules from the config file,
11691 though in a few (such as version) this isn't done
11692 because it couldn't be used for anything.
11694 In the case of ca and req the config file used is
11695 the same as the utility itself: that is the -config
11696 command line option can be used to specify an
11701 * Move default behaviour from OPENSSL_config(). If appname is NULL
11702 use "openssl_conf" if filename is NULL use default openssl config file.
11706 * Add an argument to OPENSSL_config() to allow the use of an alternative
11707 config section name. Add a new flag to tolerate a missing config file
11708 and move code to CONF_modules_load_file().
11712 * Support for crypto accelerator cards from Accelerated Encryption
11713 Processing, www.aep.ie. (Use engine 'aep')
11714 The support was copied from 0.9.6c [engine] and adapted/corrected
11715 to work with the new engine framework.
11717 *AEP Inc. and Richard Levitte*
11719 * Support for SureWare crypto accelerator cards from Baltimore
11720 Technologies. (Use engine 'sureware')
11721 The support was copied from 0.9.6c [engine] and adapted
11722 to work with the new engine framework.
11726 * Have the CHIL engine fork-safe (as defined by nCipher) and actually
11727 make the newer ENGINE framework commands for the CHIL engine work.
11729 *Toomas Kiisk <vix@cyber.ee> and Richard Levitte*
11731 * Make it possible to produce shared libraries on ReliantUNIX.
11733 *Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte*
11735 * Add the configuration target debug-linux-ppro.
11736 Make 'openssl rsa' use the general key loading routines
11737 implemented in apps.c, and make those routines able to
11738 handle the key format FORMAT_NETSCAPE and the variant
11741 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11743 * Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
11745 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
11747 * Add -keyform to rsautl, and document -engine.
11749 *Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>*
11751 * Change BIO_new_file (crypto/bio/bss_file.c) to use new
11752 BIO_R_NO_SUCH_FILE error code rather than the generic
11753 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
11757 * Add new functions
11758 ERR_peek_last_error
11759 ERR_peek_last_error_line
11760 ERR_peek_last_error_line_data.
11761 These are similar to
11763 ERR_peek_error_line
11764 ERR_peek_error_line_data,
11765 but report on the latest error recorded rather than the first one
11766 still in the error queue.
11768 *Ben Laurie, Bodo Moeller*
11770 * default_algorithms option in ENGINE config module. This allows things
11772 default_algorithms = ALL
11773 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
11777 * Preliminary ENGINE config module.
11781 * New experimental application configuration code.
11785 * Change the AES code to follow the same name structure as all other
11786 symmetric ciphers, and behave the same way. Move everything to
11787 the directory crypto/aes, thereby obsoleting crypto/rijndael.
11789 *Stephen Sprunk <stephen@sprunk.org> and Richard Levitte*
11791 * SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
11793 *Ben Laurie and Theo de Raadt*
11795 * Add option to output public keys in req command.
11797 *Massimiliano Pala madwolf@openca.org*
11799 * Use wNAFs in EC_POINTs_mul() for improved efficiency
11800 (up to about 10% better than before for P-192 and P-224).
11804 * New functions/macros
11806 SSL_CTX_set_msg_callback(ctx, cb)
11807 SSL_CTX_set_msg_callback_arg(ctx, arg)
11808 SSL_set_msg_callback(ssl, cb)
11809 SSL_set_msg_callback_arg(ssl, arg)
11811 to request calling a callback function
11813 void cb(int write_p, int version, int content_type,
11814 const void *buf, size_t len, SSL *ssl, void *arg)
11816 whenever a protocol message has been completely received
11817 (write_p == 0) or sent (write_p == 1). Here 'version' is the
11818 protocol version according to which the SSL library interprets
11819 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
11820 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
11821 the content type as defined in the SSL 3.0/TLS 1.0 protocol
11822 specification (change_cipher_spec(20), alert(21), handshake(22)).
11823 'buf' and 'len' point to the actual message, 'ssl' to the
11824 SSL object, and 'arg' is the application-defined value set by
11825 SSL[_CTX]_set_msg_callback_arg().
11827 'openssl s_client' and 'openssl s_server' have new '-msg' options
11828 to enable a callback that displays all protocol messages.
11832 * Change the shared library support so shared libraries are built as
11833 soon as the corresponding static library is finished, and thereby get
11834 openssl and the test programs linked against the shared library.
11835 This still only happens when the keyword "shard" has been given to
11836 the configuration scripts.
11838 NOTE: shared library support is still an experimental thing, and
11839 backward binary compatibility is still not guaranteed.
11841 *"Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte*
11843 * Add support for Subject Information Access extension.
11845 *Peter Sylvester <Peter.Sylvester@EdelWeb.fr>*
11847 * Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
11848 additional bytes when new memory had to be allocated, not just
11849 when reusing an existing buffer.
11853 * New command line and configuration option 'utf8' for the req command.
11854 This allows field values to be specified as UTF8 strings.
11858 * Add -multi and -mr options to "openssl speed" - giving multiple parallel
11859 runs for the former and machine-readable output for the latter.
11863 * Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
11864 of the e-mail address in the DN (i.e., it will go into a certificate
11865 extension only). The new configuration file option 'email_in_dn = no'
11866 has the same effect.
11868 *Massimiliano Pala madwolf@openca.org*
11870 * Change all functions with names starting with `des_` to be starting
11871 with `DES_` instead. Add wrappers that are compatible with libdes,
11872 but are named `_ossl_old_des_*`. Finally, add macros that map the
11873 `des_*` symbols to the corresponding `_ossl_old_des_*` if libdes
11874 compatibility is desired. If OpenSSL 0.9.6c compatibility is
11875 desired, the `des_*` symbols will be mapped to `DES_*`, with one
11878 Since we provide two compatibility mappings, the user needs to
11879 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
11880 compatibility is desired. The default (i.e., when that macro
11881 isn't defined) is OpenSSL 0.9.6c compatibility.
11883 There are also macros that enable and disable the support of old
11884 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
11885 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
11886 are defined, the default will apply: to support the old des routines.
11888 In either case, one must include openssl/des.h to get the correct
11889 definitions. Do not try to just include openssl/des_old.h, that
11892 NOTE: This is a major break of an old API into a new one. Software
11893 authors are encouraged to switch to the `DES_` style functions. Some
11894 time in the future, des_old.h and the libdes compatibility functions
11895 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
11896 default), and then completely removed.
11900 * Test for certificates which contain unsupported critical extensions.
11901 If such a certificate is found during a verify operation it is
11902 rejected by default: this behaviour can be overridden by either
11903 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
11904 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
11905 X509_supported_extension() has also been added which returns 1 if a
11906 particular extension is supported.
11910 * Modify the behaviour of EVP cipher functions in similar way to digests
11911 to retain compatibility with existing code.
11915 * Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
11916 compatibility with existing code. In particular the 'ctx' parameter does
11917 not have to be to be initialized before the call to EVP_DigestInit() and
11918 it is tidied up after a call to EVP_DigestFinal(). New function
11919 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
11920 EVP_MD_CTX_copy() changed to not require the destination to be
11921 initialized valid and new function EVP_MD_CTX_copy_ex() added which
11922 requires the destination to be valid.
11924 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
11925 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
11929 * Change ssl3_get_message (ssl/s3_both.c) and the functions using it
11930 so that complete 'Handshake' protocol structures are kept in memory
11931 instead of overwriting 'msg_type' and 'length' with 'body' data.
11935 * Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
11937 *Massimo Santin via Richard Levitte*
11939 * Major restructuring to the underlying ENGINE code. This includes
11940 reduction of linker bloat, separation of pure "ENGINE" manipulation
11941 (initialisation, etc) from functionality dealing with implementations
11942 of specific crypto interfaces. This change also introduces integrated
11943 support for symmetric ciphers and digest implementations - so ENGINEs
11944 can now accelerate these by providing EVP_CIPHER and EVP_MD
11945 implementations of their own. This is detailed in
11946 [crypto/engine/README.md](crypto/engine/README.md)
11947 as it couldn't be adequately described here. However, there are a few
11948 API changes worth noting - some RSA, DSA, DH, and RAND functions that
11949 were changed in the original introduction of ENGINE code have now
11950 reverted back - the hooking from this code to ENGINE is now a good
11951 deal more passive and at run-time, operations deal directly with
11952 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
11953 dereferencing through an ENGINE pointer any more. Also, the ENGINE
11954 functions dealing with `BN_MOD_EXP[_CRT]` handlers have been removed -
11955 they were not being used by the framework as there is no concept of a
11956 BIGNUM_METHOD and they could not be generalised to the new
11957 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
11958 ENGINE_cpy() has been removed as it cannot be consistently defined in
11963 * Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
11967 * Change mkdef.pl to sort symbols that get the same entry number,
11968 and make sure the automatically generated functions `ERR_load_*`
11969 become part of libeay.num as well.
11973 * New function SSL_renegotiate_pending(). This returns true once
11974 renegotiation has been requested (either SSL_renegotiate() call
11975 or HelloRequest/ClientHello received from the peer) and becomes
11976 false once a handshake has been completed.
11977 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
11978 sends a HelloRequest, but does not ensure that a handshake takes
11979 place. SSL_renegotiate_pending() is useful for checking if the
11980 client has followed the request.)
11984 * New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
11985 By default, clients may request session resumption even during
11986 renegotiation (if session ID contexts permit); with this option,
11987 session resumption is possible only in the first handshake.
11989 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
11990 more bits available for options that should not be part of
11991 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
11995 * Add some demos for certificate and certificate request creation.
11999 * Make maximum certificate chain size accepted from the peer application
12000 settable (`SSL*_get/set_max_cert_list()`), as proposed by
12001 "Douglas E. Engert" <deengert@anl.gov>.
12005 * Add support for shared libraries for Unixware-7
12006 (Boyd Lynn Gerber <gerberb@zenez.com>).
12010 * Add a "destroy" handler to ENGINEs that allows structural cleanup to
12011 be done prior to destruction. Use this to unload error strings from
12012 ENGINEs that load their own error strings. NB: This adds two new API
12013 functions to "get" and "set" this destroy handler in an ENGINE.
12017 * Alter all existing ENGINE implementations (except "openssl" and
12018 "openbsd") to dynamically instantiate their own error strings. This
12019 makes them more flexible to be built both as statically-linked ENGINEs
12020 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
12021 Also, add stub code to each that makes building them as self-contained
12022 shared-libraries easier (see [README-Engine.md](README-Engine.md)).
12026 * Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
12027 implementations into applications that are completely implemented in
12028 self-contained shared-libraries. The "dynamic" ENGINE exposes control
12029 commands that can be used to configure what shared-library to load and
12030 to control aspects of the way it is handled. Also, made an update to
12031 the [README-Engine.md](README-Engine.md) file
12032 that brings its information up-to-date and
12033 provides some information and instructions on the "dynamic" ENGINE
12034 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
12038 * Make it possible to unload ranges of ERR strings with a new
12039 "ERR_unload_strings" function.
12043 * Add a copy() function to EVP_MD.
12047 * Make EVP_MD routines take a context pointer instead of just the
12048 md_data void pointer.
12052 * Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
12053 that the digest can only process a single chunk of data
12054 (typically because it is provided by a piece of
12055 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
12056 is only going to provide a single chunk of data, and hence the
12057 framework needn't accumulate the data for oneshot drivers.
12061 * As with "ERR", make it possible to replace the underlying "ex_data"
12062 functions. This change also alters the storage and management of global
12063 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
12064 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
12065 index counters. The API functions that use this state have been changed
12066 to take a "class_index" rather than pointers to the class's local STACK
12067 and counter, and there is now an API function to dynamically create new
12068 classes. This centralisation allows us to (a) plug a lot of the
12069 thread-safety problems that existed, and (b) makes it possible to clean
12070 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
12071 such data would previously have always leaked in application code and
12072 workarounds were in place to make the memory debugging turn a blind eye
12073 to it. Application code that doesn't use this new function will still
12074 leak as before, but their memory debugging output will announce it now
12075 rather than letting it slide.
12077 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
12078 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
12079 has a return value to indicate success or failure.
12083 * Make it possible to replace the underlying "ERR" functions such that the
12084 global state (2 LHASH tables and 2 locks) is only used by the "default"
12085 implementation. This change also adds two functions to "get" and "set"
12086 the implementation prior to it being automatically set the first time
12087 any other ERR function takes place. Ie. an application can call "get",
12088 pass the return value to a module it has just loaded, and that module
12089 can call its own "set" function using that value. This means the
12090 module's "ERR" operations will use (and modify) the error state in the
12091 application and not in its own statically linked copy of OpenSSL code.
12095 * Give DH, DSA, and RSA types their own `*_up_ref()` function to increment
12096 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
12097 the operation, and provides a more encapsulated way for external code
12098 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
12099 to use these functions rather than manually incrementing the counts.
12101 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
12105 * Add EVP test program.
12109 * Add symmetric cipher support to ENGINE. Expect the API to change!
12113 * New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
12114 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
12115 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
12116 These allow a CRL to be built without having to access X509_CRL fields
12117 directly. Modify 'ca' application to use new functions.
12121 * Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
12122 bug workarounds. Rollback attack detection is a security feature.
12123 The problem will only arise on OpenSSL servers when TLSv1 is not
12124 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
12125 Software authors not wanting to support TLSv1 will have special reasons
12126 for their choice and can explicitly enable this option.
12128 *Bodo Moeller, Lutz Jaenicke*
12130 * Rationalise EVP so it can be extended: don't include a union of
12131 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
12132 (similar to those existing for EVP_CIPHER_CTX).
12137 EVP_MD_CTX_init(&md); /* new function call */
12138 EVP_DigestInit(&md, EVP_sha1());
12139 EVP_DigestUpdate(&md, in, len);
12140 EVP_DigestFinal(&md, out, NULL);
12141 EVP_MD_CTX_cleanup(&md); /* new function call */
12145 * Make DES key schedule conform to the usual scheme, as well as
12146 correcting its structure. This means that calls to DES functions
12147 now have to pass a pointer to a des_key_schedule instead of a
12148 plain des_key_schedule (which was actually always a pointer
12151 des_key_schedule ks;
12153 des_set_key_checked(..., &ks);
12154 des_ncbc_encrypt(..., &ks, ...);
12156 (Note that a later change renames 'des_...' into 'DES_...'.)
12160 * Initial reduction of linker bloat: the use of some functions, such as
12161 PEM causes large amounts of unused functions to be linked in due to
12162 poor organisation. For example pem_all.c contains every PEM function
12163 which has a knock on effect of linking in large amounts of (unused)
12164 ASN1 code. Grouping together similar functions and splitting unrelated
12165 functions prevents this.
12169 * Cleanup of EVP macros.
12173 * Change historical references to `{NID,SN,LN}_des_ede` and ede3 to add the
12174 correct `_ecb suffix`.
12178 * Add initial OCSP responder support to ocsp application. The
12179 revocation information is handled using the text based index
12180 use by the ca application. The responder can either handle
12181 requests generated internally, supplied in files (for example
12182 via a CGI script) or using an internal minimal server.
12186 * Add configuration choices to get zlib compression for TLS.
12190 * Changes to Kerberos SSL for RFC 2712 compliance:
12191 1. Implemented real KerberosWrapper, instead of just using
12192 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
12193 2. Implemented optional authenticator field of KerberosWrapper.
12195 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
12196 and authenticator structs; see crypto/krb5/.
12198 Generalized Kerberos calls to support multiple Kerberos libraries.
12199 *Vern Staats <staatsvr@asc.hpc.mil>, Jeffrey Altman <jaltman@columbia.edu>
12200 via Richard Levitte*
12202 * Cause 'openssl speed' to use fully hard-coded DSA keys as it
12203 already does with RSA. testdsa.h now has 'priv_key/pub_key'
12204 values for each of the key sizes rather than having just
12205 parameters (and 'speed' generating keys each time).
12209 * Speed up EVP routines.
12212 pe 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
12213 s-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
12214 s-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
12215 s-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
12217 s-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
12218 s-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
12219 s-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
12222 s-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
12224 s-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
12228 * Added the OS2-EMX target.
12230 *"Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte*
12232 * Rewrite apps to use NCONF routines instead of the old CONF. New functions
12233 to support NCONF routines in extension code. New function CONF_set_nconf()
12234 to allow functions which take an NCONF to also handle the old LHASH
12235 structure: this means that the old CONF compatible routines can be
12236 retained (in particular wrt extensions) without having to duplicate the
12237 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
12241 * Enhance the general user interface with mechanisms for inner control
12242 and with possibilities to have yes/no kind of prompts.
12246 * Change all calls to low level digest routines in the library and
12247 applications to use EVP. Add missing calls to HMAC_cleanup() and
12248 don't assume HMAC_CTX can be copied using memcpy().
12250 *Verdon Walker <VWalker@novell.com>, Steve Henson*
12252 * Add the possibility to control engines through control names but with
12253 arbitrary arguments instead of just a string.
12254 Change the key loaders to take a UI_METHOD instead of a callback
12255 function pointer. NOTE: this breaks binary compatibility with earlier
12256 versions of OpenSSL [engine].
12257 Adapt the nCipher code for these new conditions and add a card insertion
12262 * Enhance the general user interface with mechanisms to better support
12263 dialog box interfaces, application-defined prompts, the possibility
12264 to use defaults (for example default passwords from somewhere else)
12265 and interrupts/cancellations.
12269 * Tidy up PKCS#12 attribute handling. Add support for the CSP name
12270 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
12274 * Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
12275 tidy up some unnecessarily weird code in 'sk_new()').
12277 *Geoff, reported by Diego Tartara <dtartara@novamens.com>*
12279 * Change the key loading routines for ENGINEs to use the same kind
12280 callback (pem_password_cb) as all other routines that need this
12285 * Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
12286 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
12287 than this minimum value is recommended.
12291 * New random seeder for OpenVMS, using the system process statistics
12292 that are easily reachable.
12296 * Windows apparently can't transparently handle global
12297 variables defined in DLLs. Initialisations such as:
12299 const ASN1_ITEM *it = &ASN1_INTEGER_it;
12301 won't compile. This is used by the any applications that need to
12302 declare their own ASN1 modules. This was fixed by adding the option
12303 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
12304 needed for static libraries under Win32.
12308 * New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
12309 setting of purpose and trust fields. New X509_STORE trust and
12310 purpose functions and tidy up setting in other SSL functions.
12314 * Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
12315 structure. These are inherited by X509_STORE_CTX when it is
12316 initialised. This allows various defaults to be set in the
12317 X509_STORE structure (such as flags for CRL checking and custom
12318 purpose or trust settings) for functions which only use X509_STORE_CTX
12319 internally such as S/MIME.
12321 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
12322 trust settings if they are not set in X509_STORE. This allows X509_STORE
12323 purposes and trust (in S/MIME for example) to override any set by default.
12325 Add command line options for CRL checking to smime, s_client and s_server
12330 * Initial CRL based revocation checking. If the CRL checking flag(s)
12331 are set then the CRL is looked up in the X509_STORE structure and
12332 its validity and signature checked, then if the certificate is found
12333 in the CRL the verify fails with a revoked error.
12335 Various new CRL related callbacks added to X509_STORE_CTX structure.
12337 Command line options added to 'verify' application to support this.
12339 This needs some additional work, such as being able to handle multiple
12340 CRLs with different times, extension based lookup (rather than just
12341 by subject name) and ultimately more complete V2 CRL extension
12346 * Add a general user interface API (crypto/ui/). This is designed
12347 to replace things like des_read_password and friends (backward
12348 compatibility functions using this new API are provided).
12349 The purpose is to remove prompting functions from the DES code
12350 section as well as provide for prompting through dialog boxes in
12351 a window system and the like.
12355 * Add "ex_data" support to ENGINE so implementations can add state at a
12356 per-structure level rather than having to store it globally.
12360 * Make it possible for ENGINE structures to be copied when retrieved by
12361 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
12362 This causes the "original" ENGINE structure to act like a template,
12363 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
12364 operational state can be localised to each ENGINE structure, despite the
12365 fact they all share the same "methods". New ENGINE structures returned in
12366 this case have no functional references and the return value is the single
12367 structural reference. This matches the single structural reference returned
12368 by ENGINE_by_id() normally, when it is incremented on the pre-existing
12373 * Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
12374 needs to match any other type at all we need to manually clear the
12379 * Changes to the "openssl engine" utility to include;
12380 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
12381 about an ENGINE's available control commands.
12382 - executing control commands from command line arguments using the
12383 '-pre' and '-post' switches. '-post' is only used if '-t' is
12384 specified and the ENGINE is successfully initialised. The syntax for
12385 the individual commands are colon-separated, for example;
12386 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
12390 * New dynamic control command support for ENGINEs. ENGINEs can now
12391 declare their own commands (numbers), names (strings), descriptions,
12392 and input types for run-time discovery by calling applications. A
12393 subset of these commands are implicitly classed as "executable"
12394 depending on their input type, and only these can be invoked through
12395 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
12396 can be based on user input, config files, etc). The distinction is
12397 that "executable" commands cannot return anything other than a boolean
12398 result and can only support numeric or string input, whereas some
12399 discoverable commands may only be for direct use through
12400 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
12401 pointers, or other custom uses. The "executable" commands are to
12402 support parameterisations of ENGINE behaviour that can be
12403 unambiguously defined by ENGINEs and used consistently across any
12404 OpenSSL-based application. Commands have been added to all the
12405 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
12406 control over shared-library paths without source code alterations.
12410 * Changed all ENGINE implementations to dynamically allocate their
12411 ENGINEs rather than declaring them statically. Apart from this being
12412 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
12413 this also allows the implementations to compile without using the
12414 internal engine_int.h header.
12418 * Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
12419 'const' value. Any code that should be able to modify a RAND_METHOD
12420 should already have non-const pointers to it (ie. they should only
12421 modify their own ones).
12425 * Made a variety of little tweaks to the ENGINE code.
12426 - "atalla" and "ubsec" string definitions were moved from header files
12427 to C code. "nuron" string definitions were placed in variables
12428 rather than hard-coded - allowing parameterisation of these values
12429 later on via ctrl() commands.
12430 - Removed unused "#if 0"'d code.
12431 - Fixed engine list iteration code so it uses ENGINE_free() to release
12432 structural references.
12433 - Constified the RAND_METHOD element of ENGINE structures.
12434 - Constified various get/set functions as appropriate and added
12435 missing functions (including a catch-all ENGINE_cpy that duplicates
12436 all ENGINE values onto a new ENGINE except reference counts/state).
12437 - Removed NULL parameter checks in get/set functions. Setting a method
12438 or function to NULL is a way of cancelling out a previously set
12439 value. Passing a NULL ENGINE parameter is just plain stupid anyway
12440 and doesn't justify the extra error symbols and code.
12441 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
12442 flags from engine_int.h to engine.h.
12443 - Changed prototypes for ENGINE handler functions (init(), finish(),
12444 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
12448 * Implement binary inversion algorithm for BN_mod_inverse in addition
12449 to the algorithm using long division. The binary algorithm can be
12450 used only if the modulus is odd. On 32-bit systems, it is faster
12451 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
12452 roughly 5-15% for 256-bit moduli), so we use it only for moduli
12453 up to 450 bits. In 64-bit environments, the binary algorithm
12454 appears to be advantageous for much longer moduli; here we use it
12455 for moduli up to 2048 bits.
12459 * Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
12460 could not support the combine flag in choice fields.
12464 * Add a 'copy_extensions' option to the 'ca' utility. This copies
12465 extensions from a certificate request to the certificate.
12469 * Allow multiple 'certopt' and 'nameopt' options to be separated
12470 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
12471 file: this allows the display of the certificate about to be
12472 signed to be customised, to allow certain fields to be included
12473 or excluded and extension details. The old system didn't display
12474 multicharacter strings properly, omitted fields not in the policy
12475 and couldn't display additional details such as extensions.
12479 * Function EC_POINTs_mul for multiple scalar multiplication
12480 of an arbitrary number of elliptic curve points
12481 \sum scalars[i]*points[i],
12482 optionally including the generator defined for the EC_GROUP:
12483 scalar*generator + \sum scalars[i]*points[i].
12485 EC_POINT_mul is a simple wrapper function for the typical case
12486 that the point list has just one item (besides the optional
12491 * First EC_METHODs for curves over GF(p):
12493 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
12494 operations and provides various method functions that can also
12495 operate with faster implementations of modular arithmetic.
12497 EC_GFp_mont_method() reuses most functions that are part of
12498 EC_GFp_simple_method, but uses Montgomery arithmetic.
12500 *Bodo Moeller; point addition and point doubling
12501 implementation directly derived from source code provided by
12502 Lenka Fibikova <fibikova@exp-math.uni-essen.de>*
12504 * Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
12505 crypto/ec/ec_lib.c):
12507 Curves are EC_GROUP objects (with an optional group generator)
12508 based on EC_METHODs that are built into the library.
12510 Points are EC_POINT objects based on EC_GROUP objects.
12512 Most of the framework would be able to handle curves over arbitrary
12513 finite fields, but as there are no obvious types for fields other
12514 than GF(p), some functions are limited to that for now.
12518 * Add the -HTTP option to s_server. It is similar to -WWW, but requires
12519 that the file contains a complete HTTP response.
12523 * Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
12524 change the def and num file printf format specifier from "%-40sXXX"
12525 to "%-39s XXX". The latter will always guarantee a space after the
12526 field while the former will cause them to run together if the field
12527 is 40 of more characters long.
12531 * Constify the cipher and digest 'method' functions and structures
12532 and modify related functions to take constant EVP_MD and EVP_CIPHER
12537 * Hide BN_CTX structure details in bn_lcl.h instead of publishing them
12538 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
12542 * Modify `EVP_Digest*()` routines so they now return values. Although the
12543 internal software routines can never fail additional hardware versions
12548 * Clean up crypto/err/err.h and change some error codes to avoid conflicts:
12550 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
12551 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
12554 ERR_R_NESTED_ASN1_ERROR
12556 ERR_R_MISSING_ASN1_EOS
12557 were 4 .. 9, conflicting with
12558 ERR_LIB_RSA (= ERR_R_RSA_LIB)
12560 ERR_LIB_PEM (= ERR_R_PEM_LIB).
12561 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
12563 Add new error code 'ERR_R_INTERNAL_ERROR'.
12567 * Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
12572 * New option '-subj arg' for 'openssl req' and 'openssl ca'. This
12573 sets the subject name for a new request or supersedes the
12574 subject name in a given request. Formats that can be parsed are
12575 'CN=Some Name, OU=myOU, C=IT'
12577 'CN=Some Name/OU=myOU/C=IT'.
12579 Add options '-batch' and '-verbose' to 'openssl req'.
12581 *Massimiliano Pala <madwolf@hackmasters.net>*
12583 * Introduce the possibility to access global variables through
12584 functions on platform were that's the best way to handle exporting
12585 global variables in shared libraries. To enable this functionality,
12586 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
12587 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
12588 is normally done by Configure or something similar).
12590 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
12591 in the source file (foo.c) like this:
12593 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
12594 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
12596 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
12597 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
12599 OPENSSL_DECLARE_GLOBAL(int,foo);
12600 #define foo OPENSSL_GLOBAL_REF(foo)
12601 OPENSSL_DECLARE_GLOBAL(double,bar);
12602 #define bar OPENSSL_GLOBAL_REF(bar)
12604 The #defines are very important, and therefore so is including the
12605 header file everywhere where the defined globals are used.
12607 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
12608 of ASN.1 items, but that structure is a bit different.
12610 The largest change is in util/mkdef.pl which has been enhanced with
12611 better and easier to understand logic to choose which symbols should
12612 go into the Windows .def files as well as a number of fixes and code
12613 cleanup (among others, algorithm keywords are now sorted
12614 lexicographically to avoid constant rewrites).
12618 * In BN_div() keep a copy of the sign of 'num' before writing the
12619 result to 'rm' because if rm==num the value will be overwritten
12620 and produce the wrong result if 'num' is negative: this caused
12621 problems with BN_mod() and BN_nnmod().
12625 * Function OCSP_request_verify(). This checks the signature on an
12626 OCSP request and verifies the signer certificate. The signer
12627 certificate is just checked for a generic purpose and OCSP request
12632 * Add OCSP_check_validity() function to check the validity of OCSP
12633 responses. OCSP responses are prepared in real time and may only
12634 be a few seconds old. Simply checking that the current time lies
12635 between thisUpdate and nextUpdate max reject otherwise valid responses
12636 caused by either OCSP responder or client clock inaccuracy. Instead
12637 we allow thisUpdate and nextUpdate to fall within a certain period of
12638 the current time. The age of the response can also optionally be
12639 checked. Two new options -validity_period and -status_age added to
12644 * If signature or public key algorithm is unrecognized print out its
12645 OID rather that just UNKNOWN.
12649 * Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
12650 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
12651 ID to be generated from the issuer certificate alone which can then be
12652 passed to OCSP_id_issuer_cmp().
12656 * New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
12657 ASN1 modules to export functions returning ASN1_ITEM pointers
12658 instead of the ASN1_ITEM structures themselves. This adds several
12659 new macros which allow the underlying ASN1 function/structure to
12660 be accessed transparently. As a result code should not use ASN1_ITEM
12661 references directly (such as &X509_it) but instead use the relevant
12662 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
12663 use of the new ASN1 code on platforms where exporting structures
12664 is problematical (for example in shared libraries) but exporting
12665 functions returning pointers to structures is not.
12669 * Add support for overriding the generation of SSL/TLS session IDs.
12670 These callbacks can be registered either in an SSL_CTX or per SSL.
12671 The purpose of this is to allow applications to control, if they wish,
12672 the arbitrary values chosen for use as session IDs, particularly as it
12673 can be useful for session caching in multiple-server environments. A
12674 command-line switch for testing this (and any client code that wishes
12675 to use such a feature) has been added to "s_server".
12677 *Geoff Thorpe, Lutz Jaenicke*
12679 * Modify mkdef.pl to recognise and parse preprocessor conditionals
12680 of the form `#if defined(...) || defined(...) || ...` and
12681 `#if !defined(...) && !defined(...) && ...`. This also avoids
12682 the growing number of special cases it was previously handling.
12686 * Make all configuration macros available for application by making
12687 sure they are available in opensslconf.h, by giving them names starting
12688 with `OPENSSL_` to avoid conflicts with other packages and by making
12689 sure e_os2.h will cover all platform-specific cases together with
12691 Additionally, it is now possible to define configuration/platform-
12692 specific names (called "system identities"). In the C code, these
12693 are prefixed with `OPENSSL_SYSNAME_`. e_os2.h will create another
12694 macro with the name beginning with `OPENSSL_SYS_`, which is determined
12695 from `OPENSSL_SYSNAME_*` or compiler-specific macros depending on
12700 * New option -set_serial to 'req' and 'x509' this allows the serial
12701 number to use to be specified on the command line. Previously self
12702 signed certificates were hard coded with serial number 0 and the
12703 CA options of 'x509' had to use a serial number in a file which was
12708 * New options to 'ca' utility to support V2 CRL entry extensions.
12709 Currently CRL reason, invalidity date and hold instruction are
12710 supported. Add new CRL extensions to V3 code and some new objects.
12714 * New function EVP_CIPHER_CTX_set_padding() this is used to
12715 disable standard block padding (aka PKCS#5 padding) in the EVP
12716 API, which was previously mandatory. This means that the data is
12717 not padded in any way and so the total length much be a multiple
12718 of the block size, otherwise an error occurs.
12722 * Initial (incomplete) OCSP SSL support.
12726 * New function OCSP_parse_url(). This splits up a URL into its host,
12727 port and path components: primarily to parse OCSP URLs. New -url
12728 option to ocsp utility.
12732 * New nonce behavior. The return value of OCSP_check_nonce() now
12733 reflects the various checks performed. Applications can decide
12734 whether to tolerate certain situations such as an absent nonce
12735 in a response when one was present in a request: the ocsp application
12736 just prints out a warning. New function OCSP_add1_basic_nonce()
12737 this is to allow responders to include a nonce in a response even if
12738 the request is nonce-less.
12742 * Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
12743 skipped when using openssl x509 multiple times on a single input file,
12744 e.g. `(openssl x509 -out cert1; openssl x509 -out cert2) <certs`.
12748 * Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
12749 set string type: to handle setting ASN1_TIME structures. Fix ca
12750 utility to correctly initialize revocation date of CRLs.
12754 * New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
12755 the clients preferred ciphersuites and rather use its own preferences.
12756 Should help to work around M$ SGC (Server Gated Cryptography) bug in
12757 Internet Explorer by ensuring unchanged hash method during stepup.
12758 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
12762 * Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
12763 to aes and add a new 'exist' option to print out symbols that don't
12768 * Additional options to ocsp utility to allow flags to be set and
12769 additional certificates supplied.
12773 * Add the option -VAfile to 'openssl ocsp', so the user can give the
12774 OCSP client a number of certificate to only verify the response
12779 * Update Rijndael code to version 3.0 and change EVP AES ciphers to
12780 handle the new API. Currently only ECB, CBC modes supported. Add new
12783 Add TLS AES ciphersuites as described in RFC3268, "Advanced
12784 Encryption Standard (AES) Ciphersuites for Transport Layer
12785 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
12786 not enabled by default and were not part of the "ALL" ciphersuite
12787 alias because they were not yet official; they could be
12788 explicitly requested by specifying the "AESdraft" ciphersuite
12789 group alias. In the final release of OpenSSL 0.9.7, the group
12790 alias is called "AES" and is part of "ALL".)
12792 *Ben Laurie, Steve Henson, Bodo Moeller*
12794 * New function OCSP_copy_nonce() to copy nonce value (if present) from
12795 request to response.
12799 * Functions for OCSP responders. OCSP_request_onereq_count(),
12800 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
12801 extract information from a certificate request. OCSP_response_create()
12802 creates a response and optionally adds a basic response structure.
12803 OCSP_basic_add1_status() adds a complete single response to a basic
12804 response and returns the OCSP_SINGLERESP structure just added (to allow
12805 extensions to be included for example). OCSP_basic_add1_cert() adds a
12806 certificate to a basic response and OCSP_basic_sign() signs a basic
12807 response with various flags. New helper functions ASN1_TIME_check()
12808 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
12809 (converts ASN1_TIME to GeneralizedTime).
12813 * Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
12814 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
12815 structure from a certificate. X509_pubkey_digest() digests the public_key
12816 contents: this is used in various key identifiers.
12820 * Make sk_sort() tolerate a NULL argument.
12822 *Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>*
12824 * New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
12825 passed by the function are trusted implicitly. If any of them signed the
12826 response then it is assumed to be valid and is not verified.
12830 * In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
12831 to data. This was previously part of the PKCS7 ASN1 code. This
12832 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
12833 *Steve Henson, reported by Kenneth R. Robinette
12834 <support@securenetterm.com>*
12836 * Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
12837 routines: without these tracing memory leaks is very painful.
12838 Fix leaks in PKCS12 and PKCS7 routines.
12842 * Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
12843 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
12844 effectively meant GeneralizedTime would never be used. Now it
12845 is initialised to -1 but X509_time_adj() now has to check the value
12846 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
12847 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
12848 *Steve Henson, reported by Kenneth R. Robinette
12849 <support@securenetterm.com>*
12851 * Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
12852 result in a zero length in the ASN1_INTEGER structure which was
12853 not consistent with the structure when d2i_ASN1_INTEGER() was used
12854 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
12855 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
12856 where it did not print out a minus for negative ASN1_INTEGER.
12860 * Add summary printout to ocsp utility. The various functions which
12861 convert status values to strings have been renamed to:
12862 OCSP_response_status_str(), OCSP_cert_status_str() and
12863 OCSP_crl_reason_str() and are no longer static. New options
12864 to verify nonce values and to disable verification. OCSP response
12865 printout format cleaned up.
12869 * Add additional OCSP certificate checks. These are those specified
12870 in RFC2560. This consists of two separate checks: the CA of the
12871 certificate being checked must either be the OCSP signer certificate
12872 or the issuer of the OCSP signer certificate. In the latter case the
12873 OCSP signer certificate must contain the OCSP signing extended key
12874 usage. This check is performed by attempting to match the OCSP
12875 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
12876 in the OCSP_CERTID structures of the response.
12880 * Initial OCSP certificate verification added to OCSP_basic_verify()
12881 and related routines. This uses the standard OpenSSL certificate
12882 verify routines to perform initial checks (just CA validity) and
12883 to obtain the certificate chain. Then additional checks will be
12884 performed on the chain. Currently the root CA is checked to see
12885 if it is explicitly trusted for OCSP signing. This is used to set
12886 a root CA as a global signing root: that is any certificate that
12887 chains to that CA is an acceptable OCSP signing certificate.
12891 * New '-extfile ...' option to 'openssl ca' for reading X.509v3
12892 extensions from a separate configuration file.
12893 As when reading extensions from the main configuration file,
12894 the '-extensions ...' option may be used for specifying the
12897 *Massimiliano Pala <madwolf@comune.modena.it>*
12899 * New OCSP utility. Allows OCSP requests to be generated or
12900 read. The request can be sent to a responder and the output
12901 parsed, outputted or printed in text form. Not complete yet:
12902 still needs to check the OCSP response validity.
12906 * New subcommands for 'openssl ca':
12907 `openssl ca -status <serial>` prints the status of the cert with
12908 the given serial number (according to the index file).
12909 `openssl ca -updatedb` updates the expiry status of certificates
12912 *Massimiliano Pala <madwolf@comune.modena.it>*
12914 * New '-newreq-nodes' command option to CA.pl. This is like
12915 '-newreq', but calls 'openssl req' with the '-nodes' option
12916 so that the resulting key is not encrypted.
12918 *Damien Miller <djm@mindrot.org>*
12920 * New configuration for the GNU Hurd.
12922 *Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte*
12924 * Initial code to implement OCSP basic response verify. This
12925 is currently incomplete. Currently just finds the signer's
12926 certificate and verifies the signature on the response.
12930 * New SSLeay_version code SSLEAY_DIR to determine the compiled-in
12931 value of OPENSSLDIR. This is available via the new '-d' option
12932 to 'openssl version', and is also included in 'openssl version -a'.
12936 * Allowing defining memory allocation callbacks that will be given
12937 file name and line number information in additional arguments
12938 (a `const char*` and an int). The basic functionality remains, as
12939 well as the original possibility to just replace malloc(),
12940 realloc() and free() by functions that do not know about these
12941 additional arguments. To register and find out the current
12942 settings for extended allocation functions, the following
12943 functions are provided:
12945 CRYPTO_set_mem_ex_functions
12946 CRYPTO_set_locked_mem_ex_functions
12947 CRYPTO_get_mem_ex_functions
12948 CRYPTO_get_locked_mem_ex_functions
12950 These work the same way as CRYPTO_set_mem_functions and friends.
12951 `CRYPTO_get_[locked_]mem_functions` now writes 0 where such an
12952 extended allocation function is enabled.
12953 Similarly, `CRYPTO_get_[locked_]mem_ex_functions` writes 0 where
12954 a conventional allocation function is enabled.
12956 *Richard Levitte, Bodo Moeller*
12958 * Finish off removing the remaining LHASH function pointer casts.
12959 There should no longer be any prototype-casting required when using
12960 the LHASH abstraction, and any casts that remain are "bugs". See
12961 the callback types and macros at the head of lhash.h for details
12962 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
12966 * Add automatic query of EGD sockets in RAND_poll() for the unix variant.
12967 If /dev/[u]random devices are not available or do not return enough
12968 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
12970 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
12971 /etc/entropy will be queried once each in this sequence, querying stops
12972 when enough entropy was collected without querying more sockets.
12976 * Change the Unix RAND_poll() variant to be able to poll several
12977 random devices, as specified by DEVRANDOM, until a sufficient amount
12978 of data has been collected. We spend at most 10 ms on each file
12979 (select timeout) and read in non-blocking mode. DEVRANDOM now
12980 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
12981 (previously it was just the string "/dev/urandom"), so on typical
12982 platforms the 10 ms delay will never occur.
12983 Also separate out the Unix variant to its own file, rand_unix.c.
12984 For VMS, there's a currently-empty rand_vms.c.
12988 * Move OCSP client related routines to ocsp_cl.c. These
12989 provide utility functions which an application needing
12990 to issue a request to an OCSP responder and analyse the
12991 response will typically need: as opposed to those which an
12992 OCSP responder itself would need which will be added later.
12994 OCSP_request_sign() signs an OCSP request with an API similar
12995 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
12996 response. OCSP_response_get1_basic() extracts basic response
12997 from response. OCSP_resp_find_status(): finds and extracts status
12998 information from an OCSP_CERTID structure (which will be created
12999 when the request structure is built). These are built from lower
13000 level functions which work on OCSP_SINGLERESP structures but
13001 won't normally be used unless the application wishes to examine
13002 extensions in the OCSP response for example.
13004 Replace nonce routines with a pair of functions.
13005 OCSP_request_add1_nonce() adds a nonce value and optionally
13006 generates a random value. OCSP_check_nonce() checks the
13007 validity of the nonce in an OCSP response.
13011 * Change function OCSP_request_add() to OCSP_request_add0_id().
13012 This doesn't copy the supplied OCSP_CERTID and avoids the
13013 need to free up the newly created id. Change return type
13014 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
13015 This can then be used to add extensions to the request.
13016 Deleted OCSP_request_new(), since most of its functionality
13017 is now in OCSP_REQUEST_new() (and the case insensitive name
13018 clash) apart from the ability to set the request name which
13019 will be added elsewhere.
13023 * Update OCSP API. Remove obsolete extensions argument from
13024 various functions. Extensions are now handled using the new
13025 OCSP extension code. New simple OCSP HTTP function which
13026 can be used to send requests and parse the response.
13030 * Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
13031 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
13032 uses the special reorder version of SET OF to sort the attributes
13033 and reorder them to match the encoded order. This resolves a long
13034 standing problem: a verify on a PKCS7 structure just after signing
13035 it used to fail because the attribute order did not match the
13036 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
13037 it uses the received order. This is necessary to tolerate some broken
13038 software that does not order SET OF. This is handled by encoding
13039 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
13040 to produce the required SET OF.
13044 * Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
13045 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
13046 files to get correct declarations of the ASN.1 item variables.
13050 * Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
13051 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
13052 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
13053 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
13054 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
13055 ASN1_ITEM and no wrapper functions.
13059 * New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
13060 replace the old function pointer based I/O routines. Change most of
13061 the `*_d2i_bio()` and `*_d2i_fp()` functions to use these.
13065 * Enhance mkdef.pl to be more accepting about spacing in C preprocessor
13066 lines, recognize more "algorithms" that can be deselected, and make
13067 it complain about algorithm deselection that isn't recognised.
13071 * New ASN1 functions to handle dup, sign, verify, digest, pack and
13072 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
13073 to use new functions. Add NO_ASN1_OLD which can be set to remove
13074 some old style ASN1 functions: this can be used to determine if old
13075 code will still work when these eventually go away.
13079 * New extension functions for OCSP structures, these follow the
13080 same conventions as certificates and CRLs.
13084 * New function X509V3_add1_i2d(). This automatically encodes and
13085 adds an extension. Its behaviour can be customised with various
13086 flags to append, replace or delete. Various wrappers added for
13087 certificates and CRLs.
13091 * Fix to avoid calling the underlying ASN1 print routine when
13092 an extension cannot be parsed. Correct a typo in the
13093 OCSP_SERVICELOC extension. Tidy up print OCSP format.
13097 * Make mkdef.pl parse some of the ASN1 macros and add appropriate
13098 entries for variables.
13102 * Add functionality to apps/openssl.c for detecting locking
13103 problems: As the program is single-threaded, all we have
13104 to do is register a locking callback using an array for
13105 storing which locks are currently held by the program.
13109 * Use a lock around the call to CRYPTO_get_ex_new_index() in
13110 SSL_get_ex_data_X509_STORE_idx(), which is used in
13111 ssl_verify_cert_chain() and thus can be called at any time
13112 during TLS/SSL handshakes so that thread-safety is essential.
13113 Unfortunately, the ex_data design is not at all suited
13114 for multi-threaded use, so it probably should be abolished.
13118 * Added Broadcom "ubsec" ENGINE to OpenSSL.
13120 *Broadcom, tweaked and integrated by Geoff Thorpe*
13122 * Move common extension printing code to new function
13123 X509V3_print_extensions(). Reorganise OCSP print routines and
13124 implement some needed OCSP ASN1 functions. Add OCSP extensions.
13128 * New function X509_signature_print() to remove duplication in some
13133 * Add a special meaning when SET OF and SEQUENCE OF flags are both
13134 set (this was treated exactly the same as SET OF previously). This
13135 is used to reorder the STACK representing the structure to match the
13136 encoding. This will be used to get round a problem where a PKCS7
13137 structure which was signed could not be verified because the STACK
13138 order did not reflect the encoded order.
13142 * Reimplement the OCSP ASN1 module using the new code.
13146 * Update the X509V3 code to permit the use of an ASN1_ITEM structure
13147 for its ASN1 operations. The old style function pointers still exist
13148 for now but they will eventually go away.
13152 * Merge in replacement ASN1 code from the ASN1 branch. This almost
13153 completely replaces the old ASN1 functionality with a table driven
13154 encoder and decoder which interprets an ASN1_ITEM structure describing
13155 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
13156 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
13157 has also been converted to the new form.
13161 * Change BN_mod_exp_recp so that negative moduli are tolerated
13162 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
13163 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
13164 for negative moduli.
13168 * Fix BN_uadd and BN_usub: Always return non-negative results instead
13169 of not touching the result's sign bit.
13173 * BN_div bugfix: If the result is 0, the sign (res->neg) must not be
13178 * Changed the LHASH code to use prototypes for callbacks, and created
13179 macros to declare and implement thin (optionally static) functions
13180 that provide type-safety and avoid function pointer casting for the
13181 type-specific callbacks.
13185 * Added Kerberos Cipher Suites to be used with TLS, as written in
13187 *Veers Staats <staatsvr@asc.hpc.mil>,
13188 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte*
13190 * Reformat the FAQ so the different questions and answers can be divided
13191 in sections depending on the subject.
13195 * Have the zlib compression code load ZLIB.DLL dynamically under
13200 * New function BN_mod_sqrt for computing square roots modulo a prime
13201 (using the probabilistic Tonelli-Shanks algorithm unless
13202 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
13203 be handled deterministically).
13205 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13207 * Make BN_mod_inverse faster by explicitly handling small quotients
13208 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
13209 512 bits], about 30% for larger ones [1024 or 2048 bits].)
13213 * New function BN_kronecker.
13217 * Fix BN_gcd so that it works on negative inputs; the result is
13218 positive unless both parameters are zero.
13219 Previously something reasonably close to an infinite loop was
13220 possible because numbers could be growing instead of shrinking
13221 in the implementation of Euclid's algorithm.
13225 * Fix BN_is_word() and BN_is_one() macros to take into account the
13226 sign of the number in question.
13228 Fix BN_is_word(a,w) to work correctly for w == 0.
13230 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
13231 because its test if the absolute value of 'a' equals 'w'.
13232 Note that BN_abs_is_word does *not* handle w == 0 reliably;
13233 it exists mostly for use in the implementations of BN_is_zero(),
13234 BN_is_one(), and BN_is_word().
13238 * New function BN_swap.
13242 * Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
13243 the exponentiation functions are more likely to produce reasonable
13244 results on negative inputs.
13248 * Change BN_mod_mul so that the result is always non-negative.
13249 Previously, it could be negative if one of the factors was negative;
13250 I don't think anyone really wanted that behaviour.
13254 * Move `BN_mod_...` functions into new file `crypto/bn/bn_mod.c`
13255 (except for exponentiation, which stays in `crypto/bn/bn_exp.c`,
13256 and `BN_mod_mul_reciprocal`, which stays in `crypto/bn/bn_recp.c`)
13257 and add new functions:
13266 BN_mod_lshift1_quick
13268 BN_mod_lshift_quick
13270 These functions always generate non-negative results.
13272 `BN_nnmod` otherwise is `like BN_mod` (if `BN_mod` computes a remainder `r`
13273 such that `|m| < r < 0`, `BN_nnmod` will output `rem + |m|` instead).
13275 `BN_mod_XXX_quick(r, a, [b,] m)` generates the same result as
13276 `BN_mod_XXX(r, a, [b,] m, ctx)`, but requires that `a` [and `b`]
13277 be reduced modulo `m`.
13279 *Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller*
13282 The following entry accidentally appeared in the CHANGES file
13283 distributed with OpenSSL 0.9.7. The modifications described in
13284 it do *not* apply to OpenSSL 0.9.7.
13286 * Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
13287 was actually never needed) and in BN_mul(). The removal in BN_mul()
13288 required a small change in bn_mul_part_recursive() and the addition
13289 of the functions bn_cmp_part_words(), bn_sub_part_words() and
13290 bn_add_part_words(), which do the same thing as bn_cmp_words(),
13291 bn_sub_words() and bn_add_words() except they take arrays with
13297 * In 'openssl passwd', verify passwords read from the terminal
13298 unless the '-salt' option is used (which usually means that
13299 verification would just waste user's time since the resulting
13300 hash is going to be compared with some given password hash)
13301 or the new '-noverify' option is used.
13303 This is an incompatible change, but it does not affect
13304 non-interactive use of 'openssl passwd' (passwords on the command
13305 line, '-stdin' option, '-in ...' option) and thus should not
13306 cause any problems.
13310 * Remove all references to RSAref, since there's no more need for it.
13314 * Make DSO load along a path given through an environment variable
13315 (SHLIB_PATH) with shl_load().
13319 * Constify the ENGINE code as a result of BIGNUM constification.
13320 Also constify the RSA code and most things related to it. In a
13321 few places, most notable in the depth of the ASN.1 code, ugly
13322 casts back to non-const were required (to be solved at a later
13327 * Make it so the openssl application has all engines loaded by default.
13331 * Constify the BIGNUM routines a little more.
13335 * Add the following functions:
13337 ENGINE_load_cswift()
13339 ENGINE_load_atalla()
13340 ENGINE_load_nuron()
13341 ENGINE_load_builtin_engines()
13343 That way, an application can itself choose if external engines that
13344 are built-in in OpenSSL shall ever be used or not. The benefit is
13345 that applications won't have to be linked with libdl or other dso
13346 libraries unless it's really needed.
13348 Changed 'openssl engine' to load all engines on demand.
13349 Changed the engine header files to avoid the duplication of some
13350 declarations (they differed!).
13354 * 'openssl engine' can now list capabilities.
13358 * Better error reporting in 'openssl engine'.
13362 * Never call load_dh_param(NULL) in s_server.
13366 * Add engine application. It can currently list engines by name and
13367 identity, and test if they are actually available.
13371 * Improve RPM specification file by forcing symbolic linking and making
13372 sure the installed documentation is also owned by root.root.
13374 *Damien Miller <djm@mindrot.org>*
13376 * Give the OpenSSL applications more possibilities to make use of
13377 keys (public as well as private) handled by engines.
13381 * Add OCSP code that comes from CertCo.
13385 * Add VMS support for the Rijndael code.
13389 * Added untested support for Nuron crypto accelerator.
13393 * Add support for external cryptographic devices. This code was
13394 previously distributed separately as the "engine" branch.
13396 *Geoff Thorpe, Richard Levitte*
13398 * Rework the filename-translation in the DSO code. It is now possible to
13399 have far greater control over how a "name" is turned into a filename
13400 depending on the operating environment and any oddities about the
13401 different shared library filenames on each system.
13405 * Support threads on FreeBSD-elf in Configure.
13409 * Fix for SHA1 assembly problem with MASM: it produces
13410 warnings about corrupt line number information when assembling
13411 with debugging information. This is caused by the overlapping
13414 *Bernd Matthes <mainbug@celocom.de>, Steve Henson*
13417 NCONF_get_number() has no error checking at all. As a replacement,
13418 NCONF_get_number_e() is defined (`_e` for "error checking") and is
13419 promoted strongly. The old NCONF_get_number is kept around for
13420 binary backward compatibility.
13421 Make it possible for methods to load from something other than a BIO,
13422 by providing a function pointer that is given a name instead of a BIO.
13423 For example, this could be used to load configuration data from an
13428 * Fix for non blocking accept BIOs. Added new I/O special reason
13429 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
13430 with non blocking I/O was not possible because no retry code was
13431 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
13436 * Added the beginnings of Rijndael support.
13440 * Fix for bug in DirectoryString mask setting. Add support for
13441 X509_NAME_print_ex() in 'req' and X509_print_ex() function
13442 to allow certificate printing to more controllable, additional
13443 'certopt' option to 'x509' to allow new printing options to be
13448 * Clean old EAY MD5 hack from e_os.h.
13452 ### Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
13454 * Fix null-pointer assignment in do_change_cipher_spec() revealed
13455 by using the Codenomicon TLS Test Tool ([CVE-2004-0079])
13457 *Joe Orton, Steve Henson*
13459 ### Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
13461 * Fix additional bug revealed by the NISCC test suite:
13463 Stop bug triggering large recursion when presented with
13464 certain ASN.1 tags ([CVE-2003-0851])
13468 ### Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
13470 * Fix various bugs revealed by running the NISCC test suite:
13472 Stop out of bounds reads in the ASN1 code when presented with
13473 invalid tags (CVE-2003-0543 and CVE-2003-0544).
13475 If verify callback ignores invalid public key errors don't try to check
13476 certificate signature with the NULL public key.
13480 * In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
13481 if the server requested one: as stated in TLS 1.0 and SSL 3.0
13486 * In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
13487 extra data after the compression methods not only for TLS 1.0
13488 but also for SSL 3.0 (as required by the specification).
13490 *Bodo Moeller; problem pointed out by Matthias Loepfe*
13492 * Change X509_certificate_type() to mark the key as exported/exportable
13493 when it's 512 *bits* long, not 512 bytes.
13497 ### Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
13499 * Countermeasure against the Klima-Pokorny-Rosa extension of
13500 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
13501 a protocol version number mismatch like a decryption error
13502 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
13506 * Turn on RSA blinding by default in the default implementation
13507 to avoid a timing attack. Applications that don't want it can call
13508 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
13509 They would be ill-advised to do so in most cases.
13511 *Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller*
13513 * Change RSA blinding code so that it works when the PRNG is not
13514 seeded (in this case, the secret RSA exponent is abused as
13515 an unpredictable seed -- if it is not unpredictable, there
13516 is no point in blinding anyway). Make RSA blinding thread-safe
13517 by remembering the creator's thread ID in rsa->blinding and
13518 having all other threads use local one-time blinding factors
13519 (this requires more computation than sharing rsa->blinding, but
13520 avoids excessive locking; and if an RSA object is not shared
13521 between threads, blinding will still be very fast).
13525 ### Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
13527 * In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
13528 via timing by performing a MAC computation even if incorrect
13529 block cipher padding has been found. This is a countermeasure
13530 against active attacks where the attacker has to distinguish
13531 between bad padding and a MAC verification error. ([CVE-2003-0078])
13533 *Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
13534 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
13535 Martin Vuagnoux (EPFL, Ilion)*
13537 ### Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
13539 * New function OPENSSL_cleanse(), which is used to cleanse a section of
13540 memory from its contents. This is done with a counter that will
13541 place alternating values in each byte. This can be used to solve
13542 two issues: 1) the removal of calls to memset() by highly optimizing
13543 compilers, and 2) cleansing with other values than 0, since those can
13544 be read through on certain media, for example a swap space on disk.
13548 * Bugfix: client side session caching did not work with external caching,
13549 because the session->cipher setting was not restored when reloading
13550 from the external cache. This problem was masked, when
13551 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
13552 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
13556 * Fix client_certificate (ssl/s2_clnt.c): The permissible total
13557 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
13559 *Zeev Lieber <zeev-l@yahoo.com>*
13561 * Undo an undocumented change introduced in 0.9.6e which caused
13562 repeated calls to OpenSSL_add_all_ciphers() and
13563 OpenSSL_add_all_digests() to be ignored, even after calling
13568 * Change the default configuration reader to deal with last line not
13569 being properly terminated.
13573 * Change X509_NAME_cmp() so it applies the special rules on handling
13574 DN values that are of type PrintableString, as well as RDNs of type
13575 emailAddress where the value has the type ia5String.
13577 *stefank@valicert.com via Richard Levitte*
13579 * Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
13580 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
13581 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
13582 the bitwise-OR of the two for use by the majority of applications
13583 wanting this behaviour, and update the docs. The documented
13584 behaviour and actual behaviour were inconsistent and had been
13585 changing anyway, so this is more a bug-fix than a behavioural
13588 *Geoff Thorpe, diagnosed by Nadav Har'El*
13590 * Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
13591 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
13595 * Fix initialization code race conditions in
13596 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
13597 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
13598 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
13599 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
13600 ssl2_get_cipher_by_char(),
13601 ssl3_get_cipher_by_char().
13603 *Patrick McCormick <patrick@tellme.com>, Bodo Moeller*
13605 * Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
13606 the cached sessions are flushed, as the remove_cb() might use ex_data
13607 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
13608 (see [openssl.org #212]).
13610 *Geoff Thorpe, Lutz Jaenicke*
13612 * Fix typo in OBJ_txt2obj which incorrectly passed the content
13613 length, instead of the encoding length to d2i_ASN1_OBJECT.
13617 ### Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
13619 * [In 0.9.6g-engine release:]
13620 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use `_stdcall`).
13622 *Lynn Gazis <lgazis@rainbow.com>*
13624 ### Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
13626 * Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
13627 and get fix the header length calculation.
13628 *Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
13629 Alon Kantor <alonk@checkpoint.com> (and others), Steve Henson*
13631 * Use proper error handling instead of 'assertions' in buffer
13632 overflow checks added in 0.9.6e. This prevents DoS (the
13633 assertions could call abort()).
13635 *Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller*
13637 ### Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
13639 * Add various sanity checks to asn1_get_length() to reject
13640 the ASN1 length bytes if they exceed sizeof(long), will appear
13641 negative or the content length exceeds the length of the
13644 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13646 * Fix cipher selection routines: ciphers without encryption had no flags
13647 for the cipher strength set and where therefore not handled correctly
13648 by the selection routines (PR #130).
13652 * Fix EVP_dsa_sha macro.
13657 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
13658 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
13659 that was added in OpenSSL 0.9.6d.
13661 As the countermeasure turned out to be incompatible with some
13662 broken SSL implementations, the new option is part of SSL_OP_ALL.
13663 SSL_OP_ALL is usually employed when compatibility with weird SSL
13664 implementations is desired (e.g. '-bugs' option to 's_client' and
13665 's_server'), so the new option is automatically set in many
13670 * Changes in security patch:
13672 Changes marked "(CHATS)" were sponsored by the Defense Advanced
13673 Research Projects Agency (DARPA) and Air Force Research Laboratory,
13674 Air Force Materiel Command, USAF, under agreement number
13677 * Add various sanity checks to asn1_get_length() to reject
13678 the ASN1 length bytes if they exceed sizeof(long), will appear
13679 negative or the content length exceeds the length of the
13680 supplied buffer. ([CVE-2002-0659])
13682 *Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>*
13684 * Assertions for various potential buffer overflows, not known to
13685 happen in practice.
13687 *Ben Laurie (CHATS)*
13689 * Various temporary buffers to hold ASCII versions of integers were
13690 too small for 64 bit platforms. ([CVE-2002-0655])
13691 *Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>*
13693 * Remote buffer overflow in SSL3 protocol - an attacker could
13694 supply an oversized session ID to a client. ([CVE-2002-0656])
13696 *Ben Laurie (CHATS)*
13698 * Remote buffer overflow in SSL2 protocol - an attacker could
13699 supply an oversized client master key. ([CVE-2002-0656])
13701 *Ben Laurie (CHATS)*
13703 ### Changes between 0.9.6c and 0.9.6d [9 May 2002]
13705 * Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
13706 encoded as NULL) with id-dsa-with-sha1.
13708 *Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller*
13710 * Check various `X509_...()` return values in apps/req.c.
13712 *Nils Larsch <nla@trustcenter.de>*
13714 * Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
13715 an end-of-file condition would erroneously be flagged, when the CRLF
13716 was just at the end of a processed block. The bug was discovered when
13717 processing data through a buffering memory BIO handing the data to a
13718 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
13719 <ptsekov@syntrex.com> and Nedelcho Stanev.
13723 * Implement a countermeasure against a vulnerability recently found
13724 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
13725 before application data chunks to avoid the use of known IVs
13726 with data potentially chosen by the attacker.
13730 * Fix length checks in ssl3_get_client_hello().
13734 * TLS/SSL library bugfix: use s->s3->in_read_app_data differently
13735 to prevent ssl3_read_internal() from incorrectly assuming that
13736 ssl3_read_bytes() found application data while handshake
13737 processing was enabled when in fact s->s3->in_read_app_data was
13738 merely automatically cleared during the initial handshake.
13740 *Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>*
13742 * Fix object definitions for Private and Enterprise: they were not
13743 recognized in their shortname (=lowercase) representation. Extend
13744 obj_dat.pl to issue an error when using undefined keywords instead
13745 of silently ignoring the problem (Svenning Sorensen
13746 <sss@sss.dnsalias.net>).
13750 * Fix DH_generate_parameters() so that it works for 'non-standard'
13751 generators, i.e. generators other than 2 and 5. (Previously, the
13752 code did not properly initialise the 'add' and 'rem' values to
13753 BN_generate_prime().)
13755 In the new general case, we do not insist that 'generator' is
13756 actually a primitive root: This requirement is rather pointless;
13757 a generator of the order-q subgroup is just as good, if not
13762 * Map new X509 verification errors to alerts. Discovered and submitted by
13763 Tom Wu <tom@arcot.com>.
13767 * Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
13768 returning non-zero before the data has been completely received
13769 when using non-blocking I/O.
13771 *Bodo Moeller; problem pointed out by John Hughes*
13773 * Some of the ciphers missed the strength entry (SSL_LOW etc).
13775 *Ben Laurie, Lutz Jaenicke*
13777 * Fix bug in SSL_clear(): bad sessions were not removed (found by
13778 Yoram Zahavi <YoramZ@gilian.com>).
13782 * Add information about CygWin 1.3 and on, and preserve proper
13783 configuration for the versions before that.
13785 *Corinna Vinschen <vinschen@redhat.com> and Richard Levitte*
13787 * Make removal from session cache (SSL_CTX_remove_session()) more robust:
13788 check whether we deal with a copy of a session and do not delete from
13789 the cache in this case. Problem reported by "Izhar Shoshani Levi"
13790 <izhar@checkpoint.com>.
13794 * Do not store session data into the internal session cache, if it
13795 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
13796 flag is set). Proposed by Aslam <aslam@funk.com>.
13800 * Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
13805 * [In 0.9.6d-engine release:]
13806 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
13808 *Toomas Kiisk <vix@cyber.ee> via Richard Levitte*
13810 * Add the configuration target linux-s390x.
13812 *Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte*
13814 * The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
13815 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
13816 variable as an indication that a ClientHello message has been
13817 received. As the flag value will be lost between multiple
13818 invocations of ssl3_accept when using non-blocking I/O, the
13819 function may not be aware that a handshake has actually taken
13820 place, thus preventing a new session from being added to the
13823 To avoid this problem, we now set s->new_session to 2 instead of
13824 using a local variable.
13826 *Lutz Jaenicke, Bodo Moeller*
13828 * Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
13829 if the SSL_R_LENGTH_MISMATCH error is detected.
13831 *Geoff Thorpe, Bodo Moeller*
13833 * New 'shared_ldflag' column in Configure platform table.
13837 * Fix EVP_CIPHER_mode macro.
13839 *"Dan S. Camper" <dan@bti.net>*
13841 * Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
13842 type, we must throw them away by setting rr->length to 0.
13844 *D P Chang <dpc@qualys.com>*
13846 ### Changes between 0.9.6b and 0.9.6c [21 dec 2001]
13848 * Fix BN_rand_range bug pointed out by Dominikus Scherkl
13849 <Dominikus.Scherkl@biodata.com>. (The previous implementation
13850 worked incorrectly for those cases where range = `10..._2` and
13851 `3*range` is two bits longer than range.)
13855 * Only add signing time to PKCS7 structures if it is not already
13860 * Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
13861 OBJ_ld_ce should be OBJ_id_ce.
13862 Also some ip-pda OIDs in crypto/objects/objects.txt were
13863 incorrect (cf. RFC 3039).
13865 *Matt Cooper, Frederic Giudicelli, Bodo Moeller*
13867 * Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
13868 returns early because it has nothing to do.
13870 *Andy Schneider <andy.schneider@bjss.co.uk>*
13872 * [In 0.9.6c-engine release:]
13873 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
13875 *Andy Schneider <andy.schneider@bjss.co.uk>*
13877 * [In 0.9.6c-engine release:]
13878 Add support for Cryptographic Appliance's keyserver technology.
13879 (Use engine 'keyclient')
13881 *Cryptographic Appliances and Geoff Thorpe*
13883 * Add a configuration entry for OS/390 Unix. The C compiler 'c89'
13884 is called via tools/c89.sh because arguments have to be
13885 rearranged (all '-L' options must appear before the first object
13888 *Richard Shapiro <rshapiro@abinitio.com>*
13890 * [In 0.9.6c-engine release:]
13891 Add support for Broadcom crypto accelerator cards, backported
13894 *Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox*
13896 * [In 0.9.6c-engine release:]
13897 Add support for SureWare crypto accelerator cards from
13898 Baltimore Technologies. (Use engine 'sureware')
13900 *Baltimore Technologies and Mark Cox*
13902 * [In 0.9.6c-engine release:]
13903 Add support for crypto accelerator cards from Accelerated
13904 Encryption Processing, www.aep.ie. (Use engine 'aep')
13906 *AEP Inc. and Mark Cox*
13908 * Add a configuration entry for gcc on UnixWare.
13910 *Gary Benson <gbenson@redhat.com>*
13912 * Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
13913 messages are stored in a single piece (fixed-length part and
13914 variable-length part combined) and fix various bugs found on the way.
13918 * Disable caching in BIO_gethostbyname(), directly use gethostbyname()
13919 instead. BIO_gethostbyname() does not know what timeouts are
13920 appropriate, so entries would stay in cache even when they have
13922 *Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>*
13924 * Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
13925 faced with a pathologically small ClientHello fragment that does
13926 not contain client_version: Instead of aborting with an error,
13927 simply choose the highest available protocol version (i.e.,
13928 TLS 1.0 unless it is disabled). In practice, ClientHello
13929 messages are never sent like this, but this change gives us
13930 strictly correct behaviour at least for TLS.
13934 * Fix SSL handshake functions and SSL_clear() such that SSL_clear()
13935 never resets s->method to s->ctx->method when called from within
13936 one of the SSL handshake functions.
13938 *Bodo Moeller; problem pointed out by Niko Baric*
13940 * In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
13941 (sent using the client's version number) if client_version is
13942 smaller than the protocol version in use. Also change
13943 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
13944 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
13945 the client will at least see that alert.
13949 * Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
13954 * Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
13955 client receives HelloRequest while in a handshake.
13957 *Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>*
13959 * Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
13960 should end in 'break', not 'goto end' which circumvents various
13961 cleanups done in state SSL_ST_OK. But session related stuff
13962 must be disabled for SSL_ST_OK in the case that we just sent a
13965 Also avoid some overhead by not calling ssl_init_wbio_buffer()
13966 before just sending a HelloRequest.
13968 *Bodo Moeller, Eric Rescorla <ekr@rtfm.com>*
13970 * Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
13971 reveal whether illegal block cipher padding was found or a MAC
13972 verification error occurred. (Neither SSLerr() codes nor alerts
13973 are directly visible to potential attackers, but the information
13974 may leak via logfiles.)
13976 Similar changes are not required for the SSL 2.0 implementation
13977 because the number of padding bytes is sent in clear for SSL 2.0,
13978 and the extra bytes are just ignored. However ssl/s2_pkt.c
13979 failed to verify that the purported number of padding bytes is in
13984 * Add OpenUNIX-8 support including shared libraries
13985 (Boyd Lynn Gerber <gerberb@zenez.com>).
13989 * Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
13990 'wristwatch attack' using huge encoding parameters (cf.
13991 James H. Manger's CRYPTO 2001 paper). Note that the
13992 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
13993 encoding parameters and hence was not vulnerable.
13997 * BN_sqr() bug fix.
13999 *Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>*
14001 * Rabin-Miller test analyses assume uniformly distributed witnesses,
14002 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
14003 followed by modular reduction.
14005 *Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>*
14007 * Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
14008 equivalent based on BN_pseudo_rand() instead of BN_rand().
14012 * s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
14013 This function was broken, as the check for a new client hello message
14014 to handle SGC did not allow these large messages.
14015 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
14019 * Add alert descriptions for TLSv1 to `SSL_alert_desc_string[_long]()`.
14023 * Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
14024 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
14028 * Rework the configuration and shared library support for Tru64 Unix.
14029 The configuration part makes use of modern compiler features and
14030 still retains old compiler behavior for those that run older versions
14031 of the OS. The shared library support part includes a variant that
14032 uses the RPATH feature, and is available through the special
14033 configuration target "alpha-cc-rpath", which will never be selected
14036 *Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte*
14038 * In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
14039 with the same message size as in ssl3_get_certificate_request().
14040 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
14041 messages might inadvertently be reject as too long.
14043 *Petr Lampa <lampa@fee.vutbr.cz>*
14045 * Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
14049 * Modified SSL library such that the verify_callback that has been set
14050 specifically for an SSL object with SSL_set_verify() is actually being
14051 used. Before the change, a verify_callback set with this function was
14052 ignored and the verify_callback() set in the SSL_CTX at the time of
14053 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
14054 to allow the necessary settings.
14058 * Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
14059 explicitly to NULL, as at least on Solaris 8 this seems not always to be
14060 done automatically (in contradiction to the requirements of the C
14061 standard). This made problems when used from OpenSSH.
14065 * In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
14066 dh->length and always used
14068 BN_rand_range(priv_key, dh->p).
14070 BN_rand_range() is not necessary for Diffie-Hellman, and this
14071 specific range makes Diffie-Hellman unnecessarily inefficient if
14072 dh->length (recommended exponent length) is much smaller than the
14073 length of dh->p. We could use BN_rand_range() if the order of
14074 the subgroup was stored in the DH structure, but we only have
14079 BN_rand(priv_key, l, ...)
14081 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
14088 RSA_eay_public_encrypt
14089 RSA_eay_private_decrypt
14090 RSA_eay_private_encrypt (signing)
14091 RSA_eay_public_decrypt (signature verification)
14093 (default implementations for RSA_public_encrypt,
14094 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
14095 always reject numbers >= n.
14099 * In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
14100 to synchronize access to 'locking_thread'. This is necessary on
14101 systems where access to 'locking_thread' (an 'unsigned long'
14102 variable) is not atomic.
14106 * In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
14107 *before* setting the 'crypto_lock_rand' flag. The previous code had
14108 a race condition if 0 is a valid thread ID.
14110 *Travis Vitek <vitek@roguewave.com>*
14112 * Add support for shared libraries under Irix.
14114 *Albert Chin-A-Young <china@thewrittenword.com>*
14116 * Add configuration option to build on Linux on both big-endian and
14117 little-endian MIPS.
14119 *Ralf Baechle <ralf@uni-koblenz.de>*
14121 * Add the possibility to create shared libraries on HP-UX.
14125 ### Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
14127 * Change ssleay_rand_bytes (crypto/rand/md_rand.c)
14128 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
14129 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
14130 PRNG state recovery was possible based on the output of
14131 one PRNG request appropriately sized to gain knowledge on
14132 'md' followed by enough consecutive 1-byte PRNG requests
14133 to traverse all of 'state'.
14135 1. When updating 'md_local' (the current thread's copy of 'md')
14136 during PRNG output generation, hash all of the previous
14137 'md_local' value, not just the half used for PRNG output.
14139 2. Make the number of bytes from 'state' included into the hash
14140 independent from the number of PRNG bytes requested.
14142 The first measure alone would be sufficient to avoid
14143 Markku-Juhani's attack. (Actually it had never occurred
14144 to me that the half of 'md_local' used for chaining was the
14145 half from which PRNG output bytes were taken -- I had always
14146 assumed that the secret half would be used.) The second
14147 measure makes sure that additional data from 'state' is never
14148 mixed into 'md_local' in small portions; this heuristically
14149 further strengthens the PRNG.
14153 * Fix crypto/bn/asm/mips3.s.
14157 * When only the key is given to "enc", the IV is undefined. Print out
14158 an error message in this case.
14162 * Handle special case when X509_NAME is empty in X509 printing routines.
14166 * In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
14167 positive and less than q.
14171 * Don't change `*pointer` in CRYPTO_add_lock() is add_lock_callback is
14172 used: it isn't thread safe and the add_lock_callback should handle
14175 *Paul Rose <Paul.Rose@bridge.com>*
14177 * Verify that incoming data obeys the block size in
14178 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
14184 *Ulf Möller, Bodo Möller*
14186 * The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
14187 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
14188 when fixing the server behaviour for backwards-compatible 'client
14189 hello' messages. (Note that the attack is impractical against
14190 SSL 3.0 and TLS 1.0 anyway because length and version checking
14191 means that the probability of guessing a valid ciphertext is
14192 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
14195 Before 0.9.5, the countermeasure (hide the error by generating a
14196 random 'decryption result') did not work properly because
14197 ERR_clear_error() was missing, meaning that SSL_get_error() would
14198 detect the supposedly ignored error.
14200 Both problems are now fixed.
14204 * In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
14205 (previously it was 1024).
14209 * Fix for compatibility mode trust settings: ignore trust settings
14210 unless some valid trust or reject settings are present.
14214 * Fix for blowfish EVP: its a variable length cipher.
14218 * Fix various bugs related to DSA S/MIME verification. Handle missing
14219 parameters in DSA public key structures and return an error in the
14220 DSA routines if parameters are absent.
14224 * In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
14225 in the current directory if neither $RANDFILE nor $HOME was set.
14226 RAND_file_name() in 0.9.6a returned NULL in this case. This has
14227 caused some confusion to Windows users who haven't defined $HOME.
14228 Thus RAND_file_name() is changed again: e_os.h can define a
14229 DEFAULT_HOME, which will be used if $HOME is not set.
14230 For Windows, we use "C:"; on other platforms, we still require
14231 environment variables.
14233 * Move 'if (!initialized) RAND_poll()' into regions protected by
14234 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
14235 having multiple threads call RAND_poll() concurrently.
14239 * In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
14240 combination of a flag and a thread ID variable.
14241 Otherwise while one thread is in ssleay_rand_bytes (which sets the
14242 flag), *other* threads can enter ssleay_add_bytes without obeying
14243 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
14244 that they do not hold after the first thread unsets add_do_not_lock).
14248 * Change bctest again: '-x' expressions are not available in all
14249 versions of 'test'.
14253 ### Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
14255 * Fix a couple of memory leaks in PKCS7_dataDecode()
14257 *Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>*
14259 * Change Configure and Makefiles to provide EXE_EXT, which will contain
14260 the default extension for executables, if any. Also, make the perl
14261 scripts that use symlink() to test if it really exists and use "cp"
14262 if it doesn't. All this made OpenSSL compilable and installable in
14267 * Fix for asn1_GetSequence() for indefinite length constructed data.
14268 If SEQUENCE is length is indefinite just set c->slen to the total
14269 amount of data available.
14271 *Steve Henson, reported by shige@FreeBSD.org*
14273 *This change does not apply to 0.9.7.*
14275 * Change bctest to avoid here-documents inside command substitution
14276 (workaround for FreeBSD /bin/sh bug).
14277 For compatibility with Ultrix, avoid shell functions (introduced
14278 in the bctest version that searches along $PATH).
14282 * Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
14283 with des_encrypt() defined on some operating systems, like Solaris
14288 * Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
14289 On the Importance of Eliminating Errors in Cryptographic
14290 Computations, J. Cryptology 14 (2001) 2, 101-119,
14291 <http://theory.stanford.edu/~dabo/papers/faults.ps.gz>).
14295 * MIPS assembler BIGNUM division bug fix.
14299 * Disabled incorrect Alpha assembler code.
14303 * Fix PKCS#7 decode routines so they correctly update the length
14304 after reading an EOC for the EXPLICIT tag.
14308 *This change does not apply to 0.9.7.*
14310 * Fix bug in PKCS#12 key generation routines. This was triggered
14311 if a 3DES key was generated with a 0 initial byte. Include
14312 PKCS12_BROKEN_KEYGEN compilation option to retain the old
14313 (but broken) behaviour.
14317 * Enhance bctest to search for a working bc along $PATH and print
14320 *Tim Rice <tim@multitalents.net> via Richard Levitte*
14322 * Fix memory leaks in err.c: free err_data string if necessary;
14323 don't write to the wrong index in ERR_set_error_data.
14327 * Implement ssl23_peek (analogous to ssl23_read), which previously
14332 * Replace rdtsc with `_emit` statements for VC++ version 5.
14334 *Jeremy Cooper <jeremy@baymoo.org>*
14336 * Make it possible to reuse SSLv2 sessions.
14340 * In copy_email() check for >= 0 as a return value for
14341 X509_NAME_get_index_by_NID() since 0 is a valid index.
14343 *Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>*
14345 * Avoid coredump with unsupported or invalid public keys by checking if
14346 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
14347 PKCS7_verify() fails with non detached data.
14351 * Don't use getenv in library functions when run as setuid/setgid.
14352 New function OPENSSL_issetugid().
14356 * Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
14357 due to incorrect handling of multi-threading:
14359 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
14361 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
14363 3. Count how many times MemCheck_off() has been called so that
14364 nested use can be treated correctly. This also avoids
14365 inband-signalling in the previous code (which relied on the
14366 assumption that thread ID 0 is impossible).
14370 * Add "-rand" option also to s_client and s_server.
14374 * Fix CPU detection on Irix 6.x.
14375 *Kurt Hockenbury <khockenb@stevens-tech.edu> and
14376 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14378 * Fix X509_NAME bug which produced incorrect encoding if X509_NAME
14383 *This change does not apply to 0.9.7.*
14385 * Use the cached encoding of an X509_NAME structure rather than
14386 copying it. This is apparently the reason for the libsafe "errors"
14387 but the code is actually correct.
14391 * Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
14392 Bleichenbacher's DSA attack.
14393 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
14394 to be set and top=0 forces the highest bit to be set; top=-1 is new
14395 and leaves the highest bit random.
14397 *Ulf Moeller, Bodo Moeller*
14399 * In the `NCONF_...`-based implementations for `CONF_...` queries
14400 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
14401 a temporary CONF structure with the data component set to NULL
14402 (which gives segmentation faults in lh_retrieve).
14403 Instead, use NULL for the CONF pointer in CONF_get_string and
14404 CONF_get_number (which may use environment variables) and directly
14405 return NULL from CONF_get_section.
14409 * Fix potential buffer overrun for EBCDIC.
14413 * Tolerate nonRepudiation as being valid for S/MIME signing and certSign
14414 keyUsage if basicConstraints absent for a CA.
14418 * Make SMIME_write_PKCS7() write mail header values with a format that
14419 is more generally accepted (no spaces before the semicolon), since
14420 some programs can't parse those values properly otherwise. Also make
14421 sure BIO's that break lines after each write do not create invalid
14426 * Make the CRL encoding routines work with empty SEQUENCE OF. The
14427 macros previously used would not encode an empty SEQUENCE OF
14428 and break the signature.
14432 *This change does not apply to 0.9.7.*
14434 * Zero the premaster secret after deriving the master secret in
14439 * Add some EVP_add_digest_alias registrations (as found in
14440 OpenSSL_add_all_digests()) to SSL_library_init()
14441 aka OpenSSL_add_ssl_algorithms(). This provides improved
14442 compatibility with peers using X.509 certificates
14443 with unconventional AlgorithmIdentifier OIDs.
14447 * Fix for Irix with NO_ASM.
14449 *"Bruce W. Forsberg" <bruce.forsberg@baesystems.com>*
14451 * ./config script fixes.
14453 *Ulf Moeller, Richard Levitte*
14455 * Fix 'openssl passwd -1'.
14459 * Change PKCS12_key_gen_asc() so it can cope with non null
14460 terminated strings whose length is passed in the passlen
14461 parameter, for example from PEM callbacks. This was done
14462 by adding an extra length parameter to asc2uni().
14464 *Steve Henson, reported by <oddissey@samsung.co.kr>*
14466 * Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
14467 call failed, free the DSA structure.
14471 * Fix to uni2asc() to cope with zero length Unicode strings.
14472 These are present in some PKCS#12 files.
14476 * Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
14477 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
14478 when writing a 32767 byte record.
14480 *Bodo Moeller; problem reported by Eric Day <eday@concentric.net>*
14482 * In `RSA_eay_public_{en,ed}crypt` and RSA_eay_mod_exp (rsa_eay.c),
14483 obtain lock CRYPTO_LOCK_RSA before setting `rsa->_method_mod_{n,p,q}`.
14485 (RSA objects have a reference count access to which is protected
14486 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
14487 so they are meant to be shared between threads.)
14488 *Bodo Moeller, Geoff Thorpe; original patch submitted by
14489 "Reddie, Steven" <Steven.Reddie@ca.com>*
14491 * Fix a deadlock in CRYPTO_mem_leaks().
14495 * Use better test patterns in bntest.
14499 * rand_win.c fix for Borland C.
14503 * BN_rshift bugfix for n == 0.
14507 * Add a 'bctest' script that checks for some known 'bc' bugs
14508 so that 'make test' does not abort just because 'bc' is broken.
14512 * Store verify_result within SSL_SESSION also for client side to
14513 avoid potential security hole. (Re-used sessions on the client side
14514 always resulted in verify_result==X509_V_OK, not using the original
14515 result of the server certificate verification.)
14519 * Fix ssl3_pending: If the record in s->s3->rrec is not of type
14520 SSL3_RT_APPLICATION_DATA, return 0.
14521 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
14526 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
14527 releases, have been re-implemented by renaming the previous
14528 implementations of ssl2_read and ssl3_read to ssl2_read_internal
14529 and ssl3_read_internal, respectively, and adding 'peek' parameters
14530 to them. The new ssl[23]_{read,peek} functions are calls to
14531 ssl[23]_read_internal with the 'peek' flag set appropriately.
14532 A 'peek' parameter has also been added to ssl3_read_bytes, which
14533 does the actual work for ssl3_read_internal.
14537 * Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
14538 the method-specific "init()" handler. Also clean up ex_data after
14539 calling the method-specific "finish()" handler. Previously, this was
14540 happening the other way round.
14544 * Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
14545 The previous value, 12, was not always sufficient for BN_mod_exp().
14549 * Make sure that shared libraries get the internal name engine with
14550 the full version number and not just 0. This should mark the
14551 shared libraries as not backward compatible. Of course, this should
14552 be changed again when we can guarantee backward binary compatibility.
14556 * Fix typo in get_cert_by_subject() in by_dir.c
14558 *Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>*
14560 * Rework the system to generate shared libraries:
14562 - Make note of the expected extension for the shared libraries and
14563 if there is a need for symbolic links from for example libcrypto.so.0
14564 to libcrypto.so.0.9.7. There is extended info in Configure for
14567 - Make as few rebuilds of the shared libraries as possible.
14569 - Still avoid linking the OpenSSL programs with the shared libraries.
14571 - When installing, install the shared libraries separately from the
14576 * Fix SSL_CTX_set_read_ahead macro to actually use its argument.
14578 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
14579 and not in SSL_clear because the latter is also used by the
14580 accept/connect functions; previously, the settings made by
14581 SSL_set_read_ahead would be lost during the handshake.
14583 *Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>*
14585 * Correct util/mkdef.pl to be selective about disabled algorithms.
14586 Previously, it would create entries for disabled algorithms no
14591 * Added several new manual pages for SSL_* function.
14595 ### Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
14597 * In ssl23_get_client_hello, generate an error message when faced
14598 with an initial SSL 3.0/TLS record that is too small to contain the
14599 first two bytes of the ClientHello message, i.e. client_version.
14600 (Note that this is a pathologic case that probably has never happened
14601 in real life.) The previous approach was to use the version number
14602 from the record header as a substitute; but our protocol choice
14603 should not depend on that one because it is not authenticated
14604 by the Finished messages.
14608 * More robust randomness gathering functions for Windows.
14610 *Jeffrey Altman <jaltman@columbia.edu>*
14612 * For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
14613 not set then we don't setup the error code for issuer check errors
14614 to avoid possibly overwriting other errors which the callback does
14615 handle. If an application does set the flag then we assume it knows
14616 what it is doing and can handle the new informational codes
14621 * Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
14622 a general "ANY" type, as such it should be able to decode anything
14623 including tagged types. However it didn't check the class so it would
14624 wrongly interpret tagged types in the same way as their universal
14625 counterpart and unknown types were just rejected. Changed so that the
14626 tagged and unknown types are handled in the same way as a SEQUENCE:
14627 that is the encoding is stored intact. There is also a new type
14628 "V_ASN1_OTHER" which is used when the class is not universal, in this
14629 case we have no idea what the actual type is so we just lump them all
14634 * On VMS, stdout may very well lead to a file that is written to
14635 in a record-oriented fashion. That means that every write() will
14636 write a separate record, which will be read separately by the
14637 programs trying to read from it. This can be very confusing.
14639 The solution is to put a BIO filter in the way that will buffer
14640 text until a linefeed is reached, and then write everything a
14641 line at a time, so every record written will be an actual line,
14642 not chunks of lines and not (usually doesn't happen, but I've
14643 seen it once) several lines in one record. BIO_f_linebuffer() is
14646 Currently, it's a VMS-only method, because that's where it has
14647 been tested well enough.
14651 * Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
14652 it can return incorrect results.
14653 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
14654 but it was in 0.9.6-beta[12].)
14658 * Disable the check for content being present when verifying detached
14659 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
14660 include zero length content when signing messages.
14664 * New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
14665 BIO_ctrl (for BIO pairs).
14669 * Add DSO method for VMS.
14673 * Bug fix: Montgomery multiplication could produce results with the
14678 * Add RPM specification openssl.spec and modify it to build three
14679 packages. The default package contains applications, application
14680 documentation and run-time libraries. The devel package contains
14681 include files, static libraries and function documentation. The
14682 doc package contains the contents of the doc directory. The original
14683 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
14687 * Add a large number of documentation files for many SSL routines.
14689 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
14691 * Add a configuration entry for Sony News 4.
14693 *NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>*
14695 * Don't set the two most significant bits to one when generating a
14696 random number < q in the DSA library.
14700 * New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
14701 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
14702 the underlying transport is blocking) if a handshake took place.
14703 (The default behaviour is needed by applications such as s_client
14704 and s_server that use select() to determine when to use SSL_read;
14705 but for applications that know in advance when to expect data, it
14706 just makes things more complicated.)
14710 * Add RAND_egd_bytes(), which gives control over the number of bytes read
14715 * Add a few more EBCDIC conditionals that make `req` and `x509`
14716 work better on such systems.
14718 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
14720 * Add two demo programs for PKCS12_parse() and PKCS12_create().
14721 Update PKCS12_parse() so it copies the friendlyName and the
14722 keyid to the certificates aux info.
14726 * Fix bug in PKCS7_verify() which caused an infinite loop
14727 if there was more than one signature.
14729 *Sven Uszpelkat <su@celocom.de>*
14731 * Major change in util/mkdef.pl to include extra information
14732 about each symbol, as well as presenting variables as well
14733 as functions. This change means that there's n more need
14734 to rebuild the .num files when some algorithms are excluded.
14738 * Allow the verify time to be set by an application,
14739 rather than always using the current time.
14743 * Phase 2 verify code reorganisation. The certificate
14744 verify code now looks up an issuer certificate by a
14745 number of criteria: subject name, authority key id
14746 and key usage. It also verifies self signed certificates
14747 by the same criteria. The main comparison function is
14748 X509_check_issued() which performs these checks.
14750 Lot of changes were necessary in order to support this
14751 without completely rewriting the lookup code.
14753 Authority and subject key identifier are now cached.
14755 The LHASH 'certs' is X509_STORE has now been replaced
14756 by a STACK_OF(X509_OBJECT). This is mainly because an
14757 LHASH can't store or retrieve multiple objects with
14758 the same hash value.
14760 As a result various functions (which were all internal
14761 use only) have changed to handle the new X509_STORE
14762 structure. This will break anything that messed round
14763 with X509_STORE internally.
14765 The functions X509_STORE_add_cert() now checks for an
14766 exact match, rather than just subject name.
14768 The X509_STORE API doesn't directly support the retrieval
14769 of multiple certificates matching a given criteria, however
14770 this can be worked round by performing a lookup first
14771 (which will fill the cache with candidate certificates)
14772 and then examining the cache for matches. This is probably
14773 the best we can do without throwing out X509_LOOKUP
14774 entirely (maybe later...).
14776 The X509_VERIFY_CTX structure has been enhanced considerably.
14778 All certificate lookup operations now go via a get_issuer()
14779 callback. Although this currently uses an X509_STORE it
14780 can be replaced by custom lookups. This is a simple way
14781 to bypass the X509_STORE hackery necessary to make this
14782 work and makes it possible to use more efficient techniques
14783 in future. A very simple version which uses a simple
14784 STACK for its trusted certificate store is also provided
14785 using X509_STORE_CTX_trusted_stack().
14787 The verify_cb() and verify() callbacks now have equivalents
14788 in the X509_STORE_CTX structure.
14790 X509_STORE_CTX also has a 'flags' field which can be used
14791 to customise the verify behaviour.
14795 * Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
14796 excludes S/MIME capabilities.
14800 * When a certificate request is read in keep a copy of the
14801 original encoding of the signed data and use it when outputting
14802 again. Signatures then use the original encoding rather than
14803 a decoded, encoded version which may cause problems if the
14804 request is improperly encoded.
14808 * For consistency with other BIO_puts implementations, call
14809 buffer_write(b, ...) directly in buffer_puts instead of calling
14812 In BIO_puts, increment b->num_write as in BIO_write.
14814 *Peter.Sylvester@EdelWeb.fr*
14816 * Fix BN_mul_word for the case where the word is 0. (We have to use
14817 BN_zero, we may not return a BIGNUM with an array consisting of
14818 words set to zero.)
14822 * Avoid calling abort() from within the library when problems are
14823 detected, except if preprocessor symbols have been defined
14824 (such as REF_CHECK, BN_DEBUG etc.).
14828 * New openssl application 'rsautl'. This utility can be
14829 used for low level RSA operations. DER public key
14830 BIO/fp routines also added.
14834 * New Configure entry and patches for compiling on QNX 4.
14836 *Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>*
14838 * A demo state-machine implementation was sponsored by
14839 Nuron (<http://www.nuron.com/>) and is now available in
14840 demos/state_machine.
14844 * New options added to the 'dgst' utility for signature
14845 generation and verification.
14849 * Unrecognized PKCS#7 content types are now handled via a
14850 catch all ASN1_TYPE structure. This allows unsupported
14851 types to be stored as a "blob" and an application can
14852 encode and decode it manually.
14856 * Fix various signed/unsigned issues to make a_strex.c
14857 compile under VC++.
14859 *Oscar Jacobsson <oscar.jacobsson@celocom.com>*
14861 * ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
14862 length if passed a buffer. ASN1_INTEGER_to_BN failed
14863 if passed a NULL BN and its argument was negative.
14865 *Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>*
14867 * Modification to PKCS#7 encoding routines to output definite
14868 length encoding. Since currently the whole structures are in
14869 memory there's not real point in using indefinite length
14870 constructed encoding. However if OpenSSL is compiled with
14871 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
14875 * Added BIO_vprintf() and BIO_vsnprintf().
14879 * Added more prefixes to parse for in the strings written
14880 through a logging bio, to cover all the levels that are available
14881 through syslog. The prefixes are now:
14883 PANIC, EMERG, EMR => LOG_EMERG
14884 ALERT, ALR => LOG_ALERT
14885 CRIT, CRI => LOG_CRIT
14886 ERROR, ERR => LOG_ERR
14887 WARNING, WARN, WAR => LOG_WARNING
14888 NOTICE, NOTE, NOT => LOG_NOTICE
14889 INFO, INF => LOG_INFO
14890 DEBUG, DBG => LOG_DEBUG
14892 and as before, if none of those prefixes are present at the
14893 beginning of the string, LOG_ERR is chosen.
14895 On Win32, the `LOG_*` levels are mapped according to this:
14897 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
14898 LOG_WARNING => EVENTLOG_WARNING_TYPE
14899 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
14903 * Made it possible to reconfigure with just the configuration
14904 argument "reconf" or "reconfigure". The command line arguments
14905 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
14906 and are retrieved from there when reconfiguring.
14912 *Assar Westerlund <assar@sics.se>, Richard Levitte*
14914 * Add the arguments -CAfile and -CApath to the pkcs12 utility.
14918 * The obj_dat.pl script was messing up the sorting of object
14919 names. The reason was that it compared the quoted version
14920 of strings as a result "OCSP" > "OCSP Signing" because
14921 " > SPACE. Changed script to store unquoted versions of
14922 names and add quotes on output. It was also omitting some
14923 names from the lookup table if they were given a default
14924 value (that is if SN is missing it is given the same
14925 value as LN and vice versa), these are now added on the
14926 grounds that if an object has a name we should be able to
14927 look it up. Finally added warning output when duplicate
14928 short or long names are found.
14932 * Changes needed for Tandem NSK.
14934 *Scott Uroff <scott@xypro.com>*
14936 * Fix SSL 2.0 rollback checking: Due to an off-by-one error in
14937 RSA_padding_check_SSLv23(), special padding was never detected
14938 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
14939 version rollback attacks was not effective.
14941 In s23_clnt.c, don't use special rollback-attack detection padding
14942 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
14943 client; similarly, in s23_srvr.c, don't do the rollback check if
14944 SSL 2.0 is the only protocol enabled in the server.
14948 * Make it possible to get hexdumps of unprintable data with 'openssl
14949 asn1parse'. By implication, the functions ASN1_parse_dump() and
14950 BIO_dump_indent() are added.
14954 * New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
14955 these print out strings and name structures based on various
14956 flags including RFC2253 support and proper handling of
14957 multibyte characters. Added options to the 'x509' utility
14958 to allow the various flags to be set.
14962 * Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
14963 Also change the functions X509_cmp_current_time() and
14964 X509_gmtime_adj() work with an ASN1_TIME structure,
14965 this will enable certificates using GeneralizedTime in validity
14966 dates to be checked.
14970 * Make the NEG_PUBKEY_BUG code (which tolerates invalid
14971 negative public key encodings) on by default,
14972 NO_NEG_PUBKEY_BUG can be set to disable it.
14976 * New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
14977 content octets. An i2c_ASN1_OBJECT is unnecessary because
14978 the encoding can be trivially obtained from the structure.
14982 * crypto/err.c locking bugfix: Use write locks (`CRYPTO_w_[un]lock`),
14983 not read locks (`CRYPTO_r_[un]lock`).
14987 * A first attempt at creating official support for shared
14988 libraries through configuration. I've kept it so the
14989 default is static libraries only, and the OpenSSL programs
14990 are always statically linked for now, but there are
14991 preparations for dynamic linking in place.
14992 This has been tested on Linux and Tru64.
14996 * Randomness polling function for Win9x, as described in:
14997 Peter Gutmann, Software Generation of Practically Strong
15002 * Fix so PRNG is seeded in req if using an already existing
15007 * New options to smime application. -inform and -outform
15008 allow alternative formats for the S/MIME message including
15009 PEM and DER. The -content option allows the content to be
15010 specified separately. This should allow things like Netscape
15011 form signing output easier to verify.
15015 * Fix the ASN1 encoding of tags using the 'long form'.
15019 * New ASN1 functions, `i2c_*` and `c2i_*` for INTEGER and BIT
15020 STRING types. These convert content octets to and from the
15021 underlying type. The actual tag and length octets are
15022 already assumed to have been read in and checked. These
15023 are needed because all other string types have virtually
15024 identical handling apart from the tag. By having versions
15025 of the ASN1 functions that just operate on content octets
15026 IMPLICIT tagging can be handled properly. It also allows
15027 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
15028 and ASN1_INTEGER are identical apart from the tag.
15032 * Change the handling of OID objects as follows:
15034 - New object identifiers are inserted in objects.txt, following
15035 the syntax given in [crypto/objects/README.md](crypto/objects/README.md).
15036 - objects.pl is used to process obj_mac.num and create a new
15038 - obj_dat.pl is used to create a new obj_dat.h, using the data in
15041 This is currently kind of a hack, and the perl code in objects.pl
15042 isn't very elegant, but it works as I intended. The simplest way
15043 to check that it worked correctly is to look in obj_dat.h and
15044 check the array nid_objs and make sure the objects haven't moved
15045 around (this is important!). Additions are OK, as well as
15046 consistent name changes.
15050 * Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
15054 * Addition of the command line parameter '-rand file' to 'openssl req'.
15055 The given file adds to whatever has already been seeded into the
15056 random pool through the RANDFILE configuration file option or
15057 environment variable, or the default random state file.
15061 * mkstack.pl now sorts each macro group into lexical order.
15062 Previously the output order depended on the order the files
15063 appeared in the directory, resulting in needless rewriting
15068 * Patches to make OpenSSL compile under Win32 again. Mostly
15069 work arounds for the VC++ problem that it treats func() as
15070 func(void). Also stripped out the parts of mkdef.pl that
15071 added extra typesafe functions: these no longer exist.
15075 * Reorganisation of the stack code. The macros are now all
15076 collected in safestack.h . Each macro is defined in terms of
15077 a "stack macro" of the form `SKM_<name>(type, a, b)`. The
15078 DEBUG_SAFESTACK is now handled in terms of function casts,
15079 this has the advantage of retaining type safety without the
15080 use of additional functions. If DEBUG_SAFESTACK is not defined
15081 then the non typesafe macros are used instead. Also modified the
15082 mkstack.pl script to handle the new form. Needs testing to see
15083 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
15084 the default if no major problems. Similar behaviour for ASN1_SET_OF
15085 and PKCS12_STACK_OF.
15089 * When some versions of IIS use the 'NET' form of private key the
15090 key derivation algorithm is different. Normally MD5(password) is
15091 used as a 128 bit RC4 key. In the modified case
15092 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
15093 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
15094 as the old Netscape_RSA functions except they have an additional
15095 'sgckey' parameter which uses the modified algorithm. Also added
15096 an -sgckey command line option to the rsa utility. Thanks to
15097 Adrian Peck <bertie@ncipher.com> for posting details of the modified
15098 algorithm to openssl-dev.
15102 * The evp_local.h macros were using 'c.##kname' which resulted in
15103 invalid expansion on some systems (SCO 5.0.5 for example).
15104 Corrected to 'c.kname'.
15106 *Phillip Porch <root@theporch.com>*
15108 * New X509_get1_email() and X509_REQ_get1_email() functions that return
15109 a STACK of email addresses from a certificate or request, these look
15110 in the subject name and the subject alternative name extensions and
15111 omit any duplicate addresses.
15115 * Re-implement BN_mod_exp2_mont using independent (and larger) windows.
15116 This makes DSA verification about 2 % faster.
15120 * Increase maximum window size in `BN_mod_exp_...` to 6 bits instead of 5
15121 (meaning that now 2^5 values will be precomputed, which is only 4 KB
15122 plus overhead for 1024 bit moduli).
15123 This makes exponentiations about 0.5 % faster for 1024 bit
15124 exponents (as measured by "openssl speed rsa2048").
15128 * Rename memory handling macros to avoid conflicts with other
15130 Malloc => OPENSSL_malloc
15131 Malloc_locked => OPENSSL_malloc_locked
15132 Realloc => OPENSSL_realloc
15133 Free => OPENSSL_free
15137 * New function BN_mod_exp_mont_word for small bases (roughly 15%
15138 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
15142 * CygWin32 support.
15144 *John Jarvie <jjarvie@newsguy.com>*
15146 * The type-safe stack code has been rejigged. It is now only compiled
15147 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
15148 by default all type-specific stack functions are "#define"d back to
15149 standard stack functions. This results in more streamlined output
15150 but retains the type-safety checking possibilities of the original
15155 * The STACK code has been cleaned up, and certain type declarations
15156 that didn't make a lot of sense have been brought in line. This has
15157 also involved a cleanup of sorts in safestack.h to more correctly
15158 map type-safe stack functions onto their plain stack counterparts.
15159 This work has also resulted in a variety of "const"ifications of
15160 lots of the code, especially `_cmp` operations which should normally
15161 be prototyped with "const" parameters anyway.
15165 * When generating bytes for the first time in md_rand.c, 'stir the pool'
15166 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
15167 (The PRNG state consists of two parts, the large pool 'state' and 'md',
15168 where all of 'md' is used each time the PRNG is used, but 'state'
15169 is used only indexed by a cyclic counter. As entropy may not be
15170 well distributed from the beginning, 'md' is important as a
15171 chaining variable. However, the output function chains only half
15172 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
15173 all of 'md', and seeding with STATE_SIZE dummy bytes will result
15174 in all of 'state' being rewritten, with the new values depending
15175 on virtually all of 'md'. This overcomes the 80 bit limitation.)
15179 * In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
15180 the handshake is continued after ssl_verify_cert_chain();
15181 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
15182 can lead to 'unexplainable' connection aborts later.
15184 *Bodo Moeller; problem tracked down by Lutz Jaenicke*
15186 * Major EVP API cipher revision.
15187 Add hooks for extra EVP features. This allows various cipher
15188 parameters to be set in the EVP interface. Support added for variable
15189 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
15190 setting of RC2 and RC5 parameters.
15192 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
15195 Remove lots of duplicated code from the EVP library. For example *every*
15196 cipher init() function handles the 'iv' in the same way according to the
15197 cipher mode. They also all do nothing if the 'key' parameter is NULL and
15198 for CFB and OFB modes they zero ctx->num.
15200 New functionality allows removal of S/MIME code RC2 hack.
15202 Most of the routines have the same form and so can be declared in terms
15205 By shifting this to the top level EVP_CipherInit() it can be removed from
15206 all individual ciphers. If the cipher wants to handle IVs or keys
15207 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
15210 Change lots of functions like EVP_EncryptUpdate() to now return a
15211 value: although software versions of the algorithms cannot fail
15212 any installed hardware versions can.
15216 * Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
15217 this option is set, tolerate broken clients that send the negotiated
15218 protocol version number instead of the requested protocol version
15223 * Call dh_tmp_cb (set by `..._TMP_DH_CB`) with correct 'is_export' flag;
15224 i.e. non-zero for export ciphersuites, zero otherwise.
15225 Previous versions had this flag inverted, inconsistent with
15226 rsa_tmp_cb (..._TMP_RSA_CB).
15228 *Bodo Moeller; problem reported by Amit Chopra*
15230 * Add missing DSA library text string. Work around for some IIS
15231 key files with invalid SEQUENCE encoding.
15235 * Add a document (doc/standards.txt) that list all kinds of standards
15236 and so on that are implemented in OpenSSL.
15240 * Enhance c_rehash script. Old version would mishandle certificates
15241 with the same subject name hash and wouldn't handle CRLs at all.
15242 Added -fingerprint option to crl utility, to support new c_rehash
15247 * Eliminate non-ANSI declarations in crypto.h and stack.h.
15251 * Fix for SSL server purpose checking. Server checking was
15252 rejecting certificates which had extended key usage present
15253 but no ssl client purpose.
15255 *Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>*
15257 * Make PKCS#12 code work with no password. The PKCS#12 spec
15258 is a little unclear about how a blank password is handled.
15259 Since the password in encoded as a BMPString with terminating
15260 double NULL a zero length password would end up as just the
15261 double NULL. However no password at all is different and is
15262 handled differently in the PKCS#12 key generation code. NS
15263 treats a blank password as zero length. MSIE treats it as no
15264 password on export: but it will try both on import. We now do
15265 the same: PKCS12_parse() tries zero length and no password if
15266 the password is set to "" or NULL (NULL is now a valid password:
15267 it wasn't before) as does the pkcs12 application.
15271 * Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
15272 perror when PEM_read_bio_X509_REQ fails, the error message must
15273 be obtained from the error queue.
15277 * Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
15278 it in ERR_remove_state if appropriate, and change ERR_get_state
15279 accordingly to avoid race conditions (this is necessary because
15280 thread_hash is no longer constant once set).
15284 * Bugfix for linux-elf makefile.one.
15288 * RSA_get_default_method() will now cause a default
15289 RSA_METHOD to be chosen if one doesn't exist already.
15290 Previously this was only set during a call to RSA_new()
15291 or RSA_new_method(NULL) meaning it was possible for
15292 RSA_get_default_method() to return NULL.
15296 * Added native name translation to the existing DSO code
15297 that will convert (if the flag to do so is set) filenames
15298 that are sufficiently small and have no path information
15299 into a canonical native form. Eg. "blah" converted to
15300 "libblah.so" or "blah.dll" etc.
15304 * New function ERR_error_string_n(e, buf, len) which is like
15305 ERR_error_string(e, buf), but writes at most 'len' bytes
15306 including the 0 terminator. For ERR_error_string_n, 'buf'
15309 *Damien Miller <djm@mindrot.org>, Bodo Moeller*
15311 * CONF library reworked to become more general. A new CONF
15312 configuration file reader "class" is implemented as well as a
15313 new functions (`NCONF_*`, for "New CONF") to handle it. The now
15314 old `CONF_*` functions are still there, but are reimplemented to
15315 work in terms of the new functions. Also, a set of functions
15316 to handle the internal storage of the configuration data is
15317 provided to make it easier to write new configuration file
15318 reader "classes" (I can definitely see something reading a
15319 configuration file in XML format, for example), called `_CONF_*`,
15320 or "the configuration storage API"...
15322 The new configuration file reading functions are:
15324 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
15325 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
15327 NCONF_default, NCONF_WIN32
15329 NCONF_dump_fp, NCONF_dump_bio
15331 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
15332 NCONF_new creates a new CONF object. This works in the same way
15333 as other interfaces in OpenSSL, like the BIO interface.
15334 `NCONF_dump_*` dump the internal storage of the configuration file,
15335 which is useful for debugging. All other functions take the same
15336 arguments as the old `CONF_*` functions with the exception of the
15337 first that must be a `CONF *` instead of a `LHASH *`.
15339 To make it easier to use the new classes with the old `CONF_*` functions,
15340 the function CONF_set_default_method is provided.
15344 * Add '-tls1' option to 'openssl ciphers', which was already
15345 mentioned in the documentation but had not been implemented.
15346 (This option is not yet really useful because even the additional
15347 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
15351 * Initial DSO code added into libcrypto for letting OpenSSL (and
15352 OpenSSL-based applications) load shared libraries and bind to
15353 them in a portable way.
15355 *Geoff Thorpe, with contributions from Richard Levitte*
15357 ### Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
15359 * Make sure _lrotl and _lrotr are only used with MSVC.
15361 * Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
15362 (the default implementation of RAND_status).
15364 * Rename openssl x509 option '-crlext', which was added in 0.9.5,
15365 to '-clrext' (= clear extensions), as intended and documented.
15366 *Bodo Moeller; inconsistency pointed out by Michael Attili
15367 <attili@amaxo.com>*
15369 * Fix for HMAC. It wasn't zeroing the rest of the block if the key length
15370 was larger than the MD block size.
15372 *Steve Henson, pointed out by Yost William <YostW@tce.com>*
15374 * Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
15375 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
15376 using the passed key: if the passed key was a private key the result
15377 of X509_print(), for example, would be to print out all the private key
15382 * des_quad_cksum() byte order bug fix.
15383 *Ulf Möller, using the problem description in krb4-0.9.7, where
15384 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>*
15386 * Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
15389 *Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>*
15391 * For easily testing in shell scripts whether some command
15392 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
15393 returns with exit code 0 iff no command of the given name is available.
15394 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
15395 the output goes to stdout and nothing is printed to stderr.
15396 Additional arguments are always ignored.
15398 Since for each cipher there is a command of the same name,
15399 the 'no-cipher' compilation switches can be tested this way.
15401 ('openssl no-XXX' is not able to detect pseudo-commands such
15402 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
15406 * Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
15410 * For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
15411 is set; it will be thrown away anyway because each handshake creates
15413 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
15414 to parameters -- in previous versions (since OpenSSL 0.9.3) the
15415 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
15416 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
15420 * New s_client option -ign_eof: EOF at stdin is ignored, and
15421 'Q' and 'R' lose their special meanings (quit/renegotiate).
15422 This is part of what -quiet does; unlike -quiet, -ign_eof
15423 does not suppress any output.
15427 * Add compatibility options to the purpose and trust code. The
15428 purpose X509_PURPOSE_ANY is "any purpose" which automatically
15429 accepts a certificate or CA, this was the previous behaviour,
15430 with all the associated security issues.
15432 X509_TRUST_COMPAT is the old trust behaviour: only and
15433 automatically trust self signed roots in certificate store. A
15434 new trust setting X509_TRUST_DEFAULT is used to specify that
15435 a purpose has no associated trust setting and it should instead
15436 use the value in the default purpose.
15440 * Fix the PKCS#8 DSA private key code so it decodes keys again
15441 and fix a memory leak.
15445 * In util/mkerr.pl (which implements 'make errors'), preserve
15446 reason strings from the previous version of the .c file, as
15447 the default to have only downcase letters (and digits) in
15448 automatically generated reasons codes is not always appropriate.
15452 * In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
15453 using strerror. Previously, ERR_reason_error_string() returned
15454 library names as reason strings for SYSerr; but SYSerr is a special
15455 case where small numbers are errno values, not library numbers.
15459 * Add '-dsaparam' option to 'openssl dhparam' application. This
15460 converts DSA parameters into DH parameters. (When creating parameters,
15461 DSA_generate_parameters is used.)
15465 * Include 'length' (recommended exponent length) in C code generated
15466 by 'openssl dhparam -C'.
15470 * The second argument to set_label in perlasm was already being used
15471 so couldn't be used as a "file scope" flag. Moved to third argument
15476 * In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
15477 instead of RAND_bytes for encryption IVs and salts.
15481 * Include RAND_status() into RAND_METHOD instead of implementing
15482 it only for md_rand.c Otherwise replacing the PRNG by calling
15483 RAND_set_rand_method would be impossible.
15487 * Don't let DSA_generate_key() enter an infinite loop if the random
15488 number generation fails.
15492 * New 'rand' application for creating pseudo-random output.
15496 * Added configuration support for Linux/IA64
15498 *Rolf Haberrecker <rolf@suse.de>*
15500 * Assembler module support for Mingw32.
15504 * Shared library support for HPUX (in shlib/).
15506 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous*
15508 * Shared library support for Solaris gcc.
15510 *Lutz Behnke <behnke@trustcenter.de>*
15512 ### Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
15514 * PKCS7_encrypt() was adding text MIME headers twice because they
15515 were added manually and by SMIME_crlf_copy().
15519 * In bntest.c don't call BN_rand with zero bits argument.
15521 *Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>*
15523 * BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
15524 case was implemented. This caused BN_div_recp() to fail occasionally.
15528 * Add an optional second argument to the set_label() in the perl
15529 assembly language builder. If this argument exists and is set
15530 to 1 it signals that the assembler should use a symbol whose
15531 scope is the entire file, not just the current function. This
15532 is needed with MASM which uses the format label:: for this scope.
15534 *Steve Henson, pointed out by Peter Runestig <peter@runestig.com>*
15536 * Change the ASN1 types so they are typedefs by default. Before
15537 almost all types were #define'd to ASN1_STRING which was causing
15538 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
15543 * Change names of new functions to the new get1/get0 naming
15544 convention: After 'get1', the caller owns a reference count
15545 and has to call `..._free`; 'get0' returns a pointer to some
15546 data structure without incrementing reference counters.
15547 (Some of the existing 'get' functions increment a reference
15548 counter, some don't.)
15549 Similarly, 'set1' and 'add1' functions increase reference
15550 counters or duplicate objects.
15554 * Allow for the possibility of temp RSA key generation failure:
15555 the code used to assume it always worked and crashed on failure.
15559 * Fix potential buffer overrun problem in BIO_printf().
15560 *Ulf Möller, using public domain code by Patrick Powell; problem
15561 pointed out by David Sacerdote <das33@cornell.edu>*
15563 * Support EGD <http://www.lothar.com/tech/crypto/>. New functions
15564 RAND_egd() and RAND_status(). In the command line application,
15565 the EGD socket can be specified like a seed file using RANDFILE
15570 * Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
15571 Some CAs (e.g. Verisign) distribute certificates in this form.
15575 * Remove the SSL_ALLOW_ADH compile option and set the default cipher
15576 list to exclude them. This means that no special compilation option
15577 is needed to use anonymous DH: it just needs to be included in the
15582 * Change the EVP_MD_CTX_type macro so its meaning consistent with
15583 EVP_MD_type. The old functionality is available in a new macro called
15584 EVP_MD_md(). Change code that uses it and update docs.
15588 * `..._ctrl` functions now have corresponding `..._callback_ctrl` functions
15589 where the `void *` argument is replaced by a function pointer argument.
15590 Previously `void *` was abused to point to functions, which works on
15591 many platforms, but is not correct. As these functions are usually
15592 called by macros defined in OpenSSL header files, most source code
15593 should work without changes.
15597 * `<openssl/opensslconf.h>` (which is created by Configure) now contains
15598 sections with information on -D... compiler switches used for
15599 compiling the library so that applications can see them. To enable
15600 one of these sections, a pre-processor symbol `OPENSSL_..._DEFINES`
15601 must be defined. E.g.,
15602 #define OPENSSL_ALGORITHM_DEFINES
15603 #include <openssl/opensslconf.h>
15604 defines all pertinent `NO_<algo>` symbols, such as NO_IDEA, NO_RSA, etc.
15606 *Richard Levitte, Ulf and Bodo Möller*
15608 * Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
15613 * Change the 'other' type in certificate aux info to a STACK_OF
15614 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
15615 the required ASN1 format: arbitrary types determined by an OID.
15619 * Add some PEM_write_X509_REQ_NEW() functions and a command line
15620 argument to 'req'. This is not because the function is newer or
15621 better than others it just uses the work 'NEW' in the certificate
15622 request header lines. Some software needs this.
15626 * Reorganise password command line arguments: now passwords can be
15627 obtained from various sources. Delete the PEM_cb function and make
15628 it the default behaviour: i.e. if the callback is NULL and the
15629 usrdata argument is not NULL interpret it as a null terminated pass
15630 phrase. If usrdata and the callback are NULL then the pass phrase
15631 is prompted for as usual.
15635 * Add support for the Compaq Atalla crypto accelerator. If it is installed,
15636 the support is automatically enabled. The resulting binaries will
15637 autodetect the card and use it if present.
15639 *Ben Laurie and Compaq Inc.*
15641 * Work around for Netscape hang bug. This sends certificate request
15642 and server done in one record. Since this is perfectly legal in the
15643 SSL/TLS protocol it isn't a "bug" option and is on by default. See
15644 the bugs/SSLv3 entry for more info.
15648 * HP-UX tune-up: new unified configs, HP C compiler bug workaround.
15652 * Add -rand argument to smime and pkcs12 applications and read/write
15657 * New 'passwd' tool for crypt(3) and apr1 password hashes.
15661 * Add command line password options to the remaining applications.
15665 * Bug fix for BN_div_recp() for numerators with an even number of
15670 * More tests in bntest.c, and changed test_bn output.
15674 * ./config recognizes MacOS X now.
15678 * Bug fix for BN_div() when the first words of num and divisor are
15679 equal (it gave wrong results if `(rem=(n1-q*d0)&BN_MASK2) < d0)`.
15683 * Add support for various broken PKCS#8 formats, and command line
15684 options to produce them.
15688 * New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
15689 get temporary BIGNUMs from a BN_CTX.
15693 * Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
15698 * Change the `SSLeay_add_all_*()` functions to `OpenSSL_add_all_*()` and
15699 include a #define from the old name to the new. The original intent
15700 was that statically linked binaries could for example just call
15701 SSLeay_add_all_ciphers() to just add ciphers to the table and not
15702 link with digests. This never worked because SSLeay_add_all_digests()
15703 and SSLeay_add_all_ciphers() were in the same source file so calling
15704 one would link with the other. They are now in separate source files.
15708 * Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
15712 * Use a less unusual form of the Miller-Rabin primality test (it used
15713 a binary algorithm for exponentiation integrated into the Miller-Rabin
15714 loop, our standard modexp algorithms are faster).
15718 * Support for the EBCDIC character set completed.
15720 *Martin Kraemer <Martin.Kraemer@Mch.SNI.De>*
15722 * Source code cleanups: use const where appropriate, eliminate casts,
15723 use `void *` instead of `char *` in lhash.
15727 * Bugfix: ssl3_send_server_key_exchange was not restartable
15728 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
15729 this the server could overwrite ephemeral keys that the client
15734 * Turn DSA_is_prime into a macro that calls BN_is_prime,
15735 using 50 iterations of the Rabin-Miller test.
15737 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
15738 iterations of the Rabin-Miller test as required by the appendix
15739 to FIPS PUB 186[-1]) instead of DSA_is_prime.
15740 As BN_is_prime_fasttest includes trial division, DSA parameter
15741 generation becomes much faster.
15743 This implies a change for the callback functions in DSA_is_prime
15744 and DSA_generate_parameters: The callback function is called once
15745 for each positive witness in the Rabin-Miller test, not just
15746 occasionally in the inner loop; and the parameters to the
15747 callback function now provide an iteration count for the outer
15748 loop rather than for the current invocation of the inner loop.
15749 DSA_generate_parameters additionally can call the callback
15750 function with an 'iteration count' of -1, meaning that a
15751 candidate has passed the trial division test (when q is generated
15752 from an application-provided seed, trial division is skipped).
15756 * New function BN_is_prime_fasttest that optionally does trial
15757 division before starting the Rabin-Miller test and has
15758 an additional BN_CTX * argument (whereas BN_is_prime always
15759 has to allocate at least one BN_CTX).
15760 'callback(1, -1, cb_arg)' is called when a number has passed the
15761 trial division stage.
15765 * Fix for bug in CRL encoding. The validity dates weren't being handled
15770 * New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
15774 * New function BN_pseudo_rand().
15778 * Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
15779 bignum version of BN_from_montgomery() with the working code from
15780 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
15785 * Avoid a race condition in s2_clnt.c (function get_server_hello) that
15786 made it impossible to use the same SSL_SESSION data structure in
15787 SSL2 clients in multiple threads.
15791 * The return value of RAND_load_file() no longer counts bytes obtained
15792 by stat(). RAND_load_file(..., -1) is new and uses the complete file
15793 to seed the PRNG (previously an explicit byte count was required).
15795 *Ulf Möller, Bodo Möller*
15797 * Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
15798 used `char *` instead of `void *` and had casts all over the place.
15802 * Make BN_generate_prime() return NULL on error if ret!=NULL.
15806 * Retain source code compatibility for BN_prime_checks macro:
15807 BN_is_prime(..., BN_prime_checks, ...) now uses
15808 BN_prime_checks_for_size to determine the appropriate number of
15809 Rabin-Miller iterations.
15813 * Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
15814 DH_CHECK_P_NOT_SAFE_PRIME.
15815 (Check if this is true? OpenPGP calls them "strong".)
15819 * Merge the functionality of "dh" and "gendh" programs into a new program
15820 "dhparam". The old programs are retained for now but will handle DH keys
15821 (instead of parameters) in future.
15825 * Make the ciphers, s_server and s_client programs check the return values
15826 when a new cipher list is set.
15830 * Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
15831 ciphers. Before when the 56bit ciphers were enabled the sorting was
15834 The syntax for the cipher sorting has been extended to support sorting by
15835 cipher-strength (using the strength_bits hard coded in the tables).
15836 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
15838 Fix a bug in the cipher-command parser: when supplying a cipher command
15839 string with an "undefined" symbol (neither command nor alphanumeric
15840 *A-Za-z0-9*, ssl_set_cipher_list used to hang in an endless loop. Now
15841 an error is flagged.
15843 Due to the strength-sorting extension, the code of the
15844 ssl_create_cipher_list() function was completely rearranged. I hope that
15845 the readability was also increased :-)
15847 *Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>*
15849 * Minor change to 'x509' utility. The -CAcreateserial option now uses 1
15850 for the first serial number and places 2 in the serial number file. This
15851 avoids problems when the root CA is created with serial number zero and
15852 the first user certificate has the same issuer name and serial number
15857 * Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
15858 the new code. Add documentation for this stuff.
15862 * Changes to X509_ATTRIBUTE utilities. These have been renamed from
15863 `X509_*()` to `X509at_*()` on the grounds that they don't handle X509
15864 structures and behave in an analogous way to the X509v3 functions:
15865 they shouldn't be called directly but wrapper functions should be used
15868 So we also now have some wrapper functions that call the X509at functions
15869 when passed certificate requests. (TO DO: similar things can be done with
15870 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
15871 things. Some of these need some d2i or i2d and print functionality
15872 because they handle more complex structures.)
15876 * Add missing #ifndefs that caused missing symbols when building libssl
15877 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
15878 NO_RSA in `ssl/s2*.c`.
15880 *Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller*
15882 * Precautions against using the PRNG uninitialized: RAND_bytes() now
15883 has a return value which indicates the quality of the random data
15884 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
15885 error queue. New function RAND_pseudo_bytes() generates output that is
15886 guaranteed to be unique but not unpredictable. RAND_add is like
15887 RAND_seed, but takes an extra argument for an entropy estimate
15888 (RAND_seed always assumes full entropy).
15892 * Do more iterations of Rabin-Miller probable prime test (specifically,
15893 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
15894 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
15895 in crypto/bn/bn_prime.c for the complete table). This guarantees a
15896 false-positive rate of at most 2^-80 for random input.
15900 * Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
15904 * New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
15905 in the 0.9.5 release), this returns the chain
15906 from an X509_CTX structure with a dup of the stack and all
15907 the X509 reference counts upped: so the stack will exist
15908 after X509_CTX_cleanup() has been called. Modify pkcs12.c
15911 Also make SSL_SESSION_print() print out the verify return
15916 * Add manpage for the pkcs12 command. Also change the default
15917 behaviour so MAC iteration counts are used unless the new
15918 -nomaciter option is used. This improves file security and
15919 only older versions of MSIE (4.0 for example) need it.
15923 * Honor the no-xxx Configure options when creating .DEF files.
15927 * Add PKCS#10 attributes to field table: challengePassword,
15928 unstructuredName and unstructuredAddress. These are taken from
15929 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
15930 international characters are used.
15932 More changes to X509_ATTRIBUTE code: allow the setting of types
15933 based on strings. Remove the 'loc' parameter when adding
15934 attributes because these will be a SET OF encoding which is sorted
15939 * Initial changes to the 'req' utility to allow request generation
15940 automation. This will allow an application to just generate a template
15941 file containing all the field values and have req construct the
15944 Initial support for X509_ATTRIBUTE handling. Stacks of these are
15945 used all over the place including certificate requests and PKCS#7
15946 structures. They are currently handled manually where necessary with
15947 some primitive wrappers for PKCS#7. The new functions behave in a
15948 manner analogous to the X509 extension functions: they allow
15949 attributes to be looked up by NID and added.
15951 Later something similar to the X509V3 code would be desirable to
15952 automatically handle the encoding, decoding and printing of the
15953 more complex types. The string types like challengePassword can
15954 be handled by the string table functions.
15956 Also modified the multi byte string table handling. Now there is
15957 a 'global mask' which masks out certain types. The table itself
15958 can use the flag STABLE_NO_MASK to ignore the mask setting: this
15959 is useful when for example there is only one permissible type
15960 (as in countryName) and using the mask might result in no valid
15965 * Clean up 'Finished' handling, and add functions SSL_get_finished and
15966 SSL_get_peer_finished to allow applications to obtain the latest
15967 Finished messages sent to the peer or expected from the peer,
15968 respectively. (SSL_get_peer_finished is usually the Finished message
15969 actually received from the peer, otherwise the protocol will be aborted.)
15971 As the Finished message are message digests of the complete handshake
15972 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
15973 be used for external authentication procedures when the authentication
15974 provided by SSL/TLS is not desired or is not enough.
15978 * Enhanced support for Alpha Linux is added. Now ./config checks if
15979 the host supports BWX extension and if Compaq C is present on the
15980 $PATH. Just exploiting of the BWX extension results in 20-30%
15981 performance kick for some algorithms, e.g. DES and RC4 to mention
15982 a couple. Compaq C in turn generates ~20% faster code for MD5 and
15987 * Add support for MS "fast SGC". This is arguably a violation of the
15988 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
15989 weak crypto and after checking the certificate is SGC a second one
15990 with strong crypto. MS SGC stops the first handshake after receiving
15991 the server certificate message and sends a second client hello. Since
15992 a server will typically do all the time consuming operations before
15993 expecting any further messages from the client (server key exchange
15994 is the most expensive) there is little difference between the two.
15996 To get OpenSSL to support MS SGC we have to permit a second client
15997 hello message after we have sent server done. In addition we have to
15998 reset the MAC if we do get this second client hello.
16002 * Add a function 'd2i_AutoPrivateKey()' this will automatically decide
16003 if a DER encoded private key is RSA or DSA traditional format. Changed
16004 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
16005 format DER encoded private key. Newer code should use PKCS#8 format which
16006 has the key type encoded in the ASN1 structure. Added DER private key
16007 support to pkcs8 application.
16011 * SSL 3/TLS 1 servers now don't request certificates when an anonymous
16012 ciphersuites has been selected (as required by the SSL 3/TLS 1
16013 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
16014 is set, we interpret this as a request to violate the specification
16015 (the worst that can happen is a handshake failure, and 'correct'
16016 behaviour would result in a handshake failure anyway).
16020 * In SSL_CTX_add_session, take into account that there might be multiple
16021 SSL_SESSION structures with the same session ID (e.g. when two threads
16022 concurrently obtain them from an external cache).
16023 The internal cache can handle only one SSL_SESSION with a given ID,
16024 so if there's a conflict, we now throw out the old one to achieve
16029 * Add OIDs for idea and blowfish in CBC mode. This will allow both
16030 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
16031 some routines that use cipher OIDs: some ciphers do not have OIDs
16032 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
16037 * Simplify the trust setting structure and code. Now we just have
16038 two sequences of OIDs for trusted and rejected settings. These will
16039 typically have values the same as the extended key usage extension
16040 and any application specific purposes.
16042 The trust checking code now has a default behaviour: it will just
16043 check for an object with the same NID as the passed id. Functions can
16044 be provided to override either the default behaviour or the behaviour
16045 for a given id. SSL client, server and email already have functions
16046 in place for compatibility: they check the NID and also return "trusted"
16047 if the certificate is self signed.
16051 * Add d2i,i2d bio/fp functions for PrivateKey: these convert the
16052 traditional format into an EVP_PKEY structure.
16056 * Add a password callback function PEM_cb() which either prompts for
16057 a password if usr_data is NULL or otherwise assumes it is a null
16058 terminated password. Allow passwords to be passed on command line
16059 environment or config files in a few more utilities.
16063 * Add a bunch of DER and PEM functions to handle PKCS#8 format private
16064 keys. Add some short names for PKCS#8 PBE algorithms and allow them
16065 to be specified on the command line for the pkcs8 and pkcs12 utilities.
16066 Update documentation.
16070 * Support for ASN1 "NULL" type. This could be handled before by using
16071 ASN1_TYPE but there wasn't any function that would try to read a NULL
16072 and produce an error if it couldn't. For compatibility we also have
16073 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
16074 don't allocate anything because they don't need to.
16078 * Initial support for MacOS is now provided. Examine INSTALL.MacOS
16081 *Andy Polyakov, Roy Woods <roy@centicsystems.ca>*
16083 * Rebuild of the memory allocation routines used by OpenSSL code and
16084 possibly others as well. The purpose is to make an interface that
16085 provide hooks so anyone can build a separate set of allocation and
16086 deallocation routines to be used by OpenSSL, for example memory
16087 pool implementations, or something else, which was previously hard
16088 since Malloc(), Realloc() and Free() were defined as macros having
16089 the values malloc, realloc and free, respectively (except for Win32
16090 compilations). The same is provided for memory debugging code.
16091 OpenSSL already comes with functionality to find memory leaks, but
16092 this gives people a chance to debug other memory problems.
16094 With these changes, a new set of functions and macros have appeared:
16096 CRYPTO_set_mem_debug_functions() [F]
16097 CRYPTO_get_mem_debug_functions() [F]
16098 CRYPTO_dbg_set_options() [F]
16099 CRYPTO_dbg_get_options() [F]
16100 CRYPTO_malloc_debug_init() [M]
16102 The memory debug functions are NULL by default, unless the library
16103 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
16104 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
16105 gives the standard debugging functions that come with OpenSSL) or
16106 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
16107 provided by the library user) must be used. When the standard
16108 debugging functions are used, CRYPTO_dbg_set_options can be used to
16109 request additional information:
16110 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
16111 the CRYPTO_MDEBUG_xxx macro when compiling the library.
16113 Also, things like CRYPTO_set_mem_functions will always give the
16114 expected result (the new set of functions is used for allocation
16115 and deallocation) at all times, regardless of platform and compiler
16118 To finish it up, some functions that were never use in any other
16119 way than through macros have a new API and new semantic:
16121 CRYPTO_dbg_malloc()
16122 CRYPTO_dbg_realloc()
16125 All macros of value have retained their old syntax.
16127 *Richard Levitte and Bodo Moeller*
16129 * Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
16130 ordering of SMIMECapabilities wasn't in "strength order" and there
16131 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
16136 * Some ASN1 types with illegal zero length encoding (INTEGER,
16137 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
16139 *Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson*
16141 * Merge in my S/MIME library for OpenSSL. This provides a simple
16142 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
16143 functionality to handle multipart/signed properly) and a utility
16144 called 'smime' to call all this stuff. This is based on code I
16145 originally wrote for Celo who have kindly allowed it to be
16146 included in OpenSSL.
16150 * Add variants des_set_key_checked and des_set_key_unchecked of
16151 des_set_key (aka des_key_sched). Global variable des_check_key
16152 decides which of these is called by des_set_key; this way
16153 des_check_key behaves as it always did, but applications and
16154 the library itself, which was buggy for des_check_key == 1,
16155 have a cleaner way to pick the version they need.
16159 * New function PKCS12_newpass() which changes the password of a
16164 * Modify X509_TRUST and X509_PURPOSE so it also uses a static and
16165 dynamic mix. In both cases the ids can be used as an index into the
16166 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
16167 functions so they accept a list of the field values and the
16168 application doesn't need to directly manipulate the X509_TRUST
16173 * Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
16178 * Modify the way the V3 extension code looks up extensions. This now
16179 works in a similar way to the object code: we have some "standard"
16180 extensions in a static table which is searched with OBJ_bsearch()
16181 and the application can add dynamic ones if needed. The file
16182 crypto/x509v3/ext_dat.h now has the info: this file needs to be
16183 updated whenever a new extension is added to the core code and kept
16184 in ext_nid order. There is a simple program 'tabtest.c' which checks
16185 this. New extensions are not added too often so this file can readily
16186 be maintained manually.
16188 There are two big advantages in doing things this way. The extensions
16189 can be looked up immediately and no longer need to be "added" using
16190 X509V3_add_standard_extensions(): this function now does nothing.
16191 Side note: I get *lots* of email saying the extension code doesn't
16192 work because people forget to call this function.
16193 Also no dynamic allocation is done unless new extensions are added:
16194 so if we don't add custom extensions there is no need to call
16195 X509V3_EXT_cleanup().
16199 * Modify enc utility's salting as follows: make salting the default. Add a
16200 magic header, so unsalted files fail gracefully instead of just decrypting
16201 to garbage. This is because not salting is a big security hole, so people
16202 should be discouraged from doing it.
16206 * Fixes and enhancements to the 'x509' utility. It allowed a message
16207 digest to be passed on the command line but it only used this
16208 parameter when signing a certificate. Modified so all relevant
16209 operations are affected by the digest parameter including the
16210 -fingerprint and -x509toreq options. Also -x509toreq choked if a
16211 DSA key was used because it didn't fix the digest.
16215 * Initial certificate chain verify code. Currently tests the untrusted
16216 certificates for consistency with the verify purpose (which is set
16217 when the X509_STORE_CTX structure is set up) and checks the pathlength.
16219 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
16220 this is because it will reject chains with invalid extensions whereas
16221 every previous version of OpenSSL and SSLeay made no checks at all.
16223 Trust code: checks the root CA for the relevant trust settings. Trust
16224 settings have an initial value consistent with the verify purpose: e.g.
16225 if the verify purpose is for SSL client use it expects the CA to be
16226 trusted for SSL client use. However the default value can be changed to
16227 permit custom trust settings: one example of this would be to only trust
16228 certificates from a specific "secure" set of CAs.
16230 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
16231 which should be used for version portability: especially since the
16232 verify structure is likely to change more often now.
16234 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
16235 to set them. If not set then assume SSL clients will verify SSL servers
16238 Two new options to the verify program: -untrusted allows a set of
16239 untrusted certificates to be passed in and -purpose which sets the
16240 intended purpose of the certificate. If a purpose is set then the
16241 new chain verify code is used to check extension consistency.
16245 * Support for the authority information access extension.
16249 * Modify RSA and DSA PEM read routines to transparently handle
16250 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
16251 public keys in a format compatible with certificate
16252 SubjectPublicKeyInfo structures. Unfortunately there were already
16253 functions called *_PublicKey_* which used various odd formats so
16254 these are retained for compatibility: however the DSA variants were
16255 never in a public release so they have been deleted. Changed dsa/rsa
16256 utilities to handle the new format: note no releases ever handled public
16257 keys so we should be OK.
16259 The primary motivation for this change is to avoid the same fiasco
16260 that dogs private keys: there are several incompatible private key
16261 formats some of which are standard and some OpenSSL specific and
16262 require various evil hacks to allow partial transparent handling and
16263 even then it doesn't work with DER formats. Given the option anything
16264 other than PKCS#8 should be dumped: but the other formats have to
16265 stay in the name of compatibility.
16267 With public keys and the benefit of hindsight one standard format
16268 is used which works with EVP_PKEY, RSA or DSA structures: though
16269 it clearly returns an error if you try to read the wrong kind of key.
16271 Added a -pubkey option to the 'x509' utility to output the public key.
16272 Also rename the `EVP_PKEY_get_*()` to `EVP_PKEY_rget_*()`
16273 (renamed to `EVP_PKEY_get1_*()` in the OpenSSL 0.9.5 release) and add
16274 `EVP_PKEY_rset_*()` functions (renamed to `EVP_PKEY_set1_*()`)
16275 that do the same as the `EVP_PKEY_assign_*()` except they up the
16276 reference count of the added key (they don't "swallow" the
16281 * Fixes to crypto/x509/by_file.c the code to read in certificates and
16282 CRLs would fail if the file contained no certificates or no CRLs:
16283 added a new function to read in both types and return the number
16284 read: this means that if none are read it will be an error. The
16285 DER versions of the certificate and CRL reader would always fail
16286 because it isn't possible to mix certificates and CRLs in DER format
16287 without choking one or the other routine. Changed this to just read
16288 a certificate: this is the best we can do. Also modified the code
16289 in apps/verify.c to take notice of return codes: it was previously
16290 attempting to read in certificates from NULL pointers and ignoring
16291 any errors: this is one reason why the cert and CRL reader seemed
16292 to work. It doesn't check return codes from the default certificate
16293 routines: these may well fail if the certificates aren't installed.
16297 * Code to support otherName option in GeneralName.
16301 * First update to verify code. Change the verify utility
16302 so it warns if it is passed a self signed certificate:
16303 for consistency with the normal behaviour. X509_verify
16304 has been modified to it will now verify a self signed
16305 certificate if *exactly* the same certificate appears
16306 in the store: it was previously impossible to trust a
16307 single self signed certificate. This means that:
16308 openssl verify ss.pem
16309 now gives a warning about a self signed certificate but
16310 openssl verify -CAfile ss.pem ss.pem
16315 * For servers, store verify_result in SSL_SESSION data structure
16316 (and add it to external session representation).
16317 This is needed when client certificate verifications fails,
16318 but an application-provided verification callback (set by
16319 SSL_CTX_set_cert_verify_callback) allows accepting the session
16320 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
16321 but returns 1): When the session is reused, we have to set
16322 ssl->verify_result to the appropriate error code to avoid
16325 *Bodo Moeller, problem pointed out by Lutz Jaenicke*
16327 * Fix a bug in the new PKCS#7 code: it didn't consider the
16328 case in PKCS7_dataInit() where the signed PKCS7 structure
16329 didn't contain any existing data because it was being created.
16331 *Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson*
16333 * Add a salt to the key derivation routines in enc.c. This
16334 forms the first 8 bytes of the encrypted file. Also add a
16335 -S option to allow a salt to be input on the command line.
16339 * New function X509_cmp(). Oddly enough there wasn't a function
16340 to compare two certificates. We do this by working out the SHA1
16341 hash and comparing that. X509_cmp() will be needed by the trust
16346 * SSL_get1_session() is like SSL_get_session(), but increments
16347 the reference count in the SSL_SESSION returned.
16349 *Geoff Thorpe <geoff@eu.c2.net>*
16351 * Fix for 'req': it was adding a null to request attributes.
16352 Also change the X509_LOOKUP and X509_INFO code to handle
16353 certificate auxiliary information.
16357 * Add support for 40 and 64 bit RC2 and RC4 algorithms: document
16362 * Add the possibility to add extra information to the memory leak
16363 detecting output, to form tracebacks, showing from where each
16364 allocation was originated: CRYPTO_push_info("constant string") adds
16365 the string plus current file name and line number to a per-thread
16366 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
16367 is like calling CYRPTO_pop_info() until the stack is empty.
16368 Also updated memory leak detection code to be multi-thread-safe.
16372 * Add options -text and -noout to pkcs7 utility and delete the
16373 encryption options which never did anything. Update docs.
16377 * Add options to some of the utilities to allow the pass phrase
16378 to be included on either the command line (not recommended on
16379 OSes like Unix) or read from the environment. Update the
16380 manpages and fix a few bugs.
16384 * Add a few manpages for some of the openssl commands.
16388 * Fix the -revoke option in ca. It was freeing up memory twice,
16389 leaking and not finding already revoked certificates.
16393 * Extensive changes to support certificate auxiliary information.
16394 This involves the use of X509_CERT_AUX structure and X509_AUX
16395 functions. An X509_AUX function such as PEM_read_X509_AUX()
16396 can still read in a certificate file in the usual way but it
16397 will also read in any additional "auxiliary information". By
16398 doing things this way a fair degree of compatibility can be
16399 retained: existing certificates can have this information added
16400 using the new 'x509' options.
16402 Current auxiliary information includes an "alias" and some trust
16403 settings. The trust settings will ultimately be used in enhanced
16404 certificate chain verification routines: currently a certificate
16405 can only be trusted if it is self signed and then it is trusted
16410 * Fix assembler for Alpha (tested only on DEC OSF not Linux or `*BSD`).
16411 The problem was that one of the replacement routines had not been working
16412 since SSLeay releases. For now the offending routine has been replaced
16413 with non-optimised assembler. Even so, this now gives around 95%
16414 performance improvement for 1024 bit RSA signs.
16418 * Hack to fix PKCS#7 decryption when used with some unorthodox RC2
16419 handling. Most clients have the effective key size in bits equal to
16420 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
16421 A few however don't do this and instead use the size of the decrypted key
16422 to determine the RC2 key length and the AlgorithmIdentifier to determine
16423 the effective key length. In this case the effective key length can still
16424 be 40 bits but the key length can be 168 bits for example. This is fixed
16425 by manually forcing an RC2 key into the EVP_PKEY structure because the
16426 EVP code can't currently handle unusual RC2 key sizes: it always assumes
16427 the key length and effective key length are equal.
16431 * Add a bunch of functions that should simplify the creation of
16432 X509_NAME structures. Now you should be able to do:
16433 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
16434 and have it automatically work out the correct field type and fill in
16435 the structures. The more adventurous can try:
16436 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
16437 and it will (hopefully) work out the correct multibyte encoding.
16441 * Change the 'req' utility to use the new field handling and multibyte
16442 copy routines. Before the DN field creation was handled in an ad hoc
16443 way in req, ca, and x509 which was rather broken and didn't support
16444 BMPStrings or UTF8Strings. Since some software doesn't implement
16445 BMPStrings or UTF8Strings yet, they can be enabled using the config file
16446 using the dirstring_type option. See the new comment in the default
16447 openssl.cnf for more info.
16451 * Make crypto/rand/md_rand.c more robust:
16452 - Assure unique random numbers after fork().
16453 - Make sure that concurrent threads access the global counter and
16454 md serializably so that we never lose entropy in them
16455 or use exactly the same state in multiple threads.
16456 Access to the large state is not always serializable because
16457 the additional locking could be a performance killer, and
16458 md should be large enough anyway.
16462 * New file apps/app_rand.c with commonly needed functionality
16463 for handling the random seed file.
16465 Use the random seed file in some applications that previously did not:
16467 dsaparam -genkey (which also ignored its '-rand' option),
16470 x509 (when signing).
16471 Except on systems with /dev/urandom, it is crucial to have a random
16472 seed file at least for key creation, DSA signing, and for DH exchanges;
16473 for RSA signatures we could do without one.
16475 gendh and gendsa (unlike genrsa) used to read only the first byte
16476 of each file listed in the '-rand' option. The function as previously
16477 found in genrsa is now in app_rand.c and is used by all programs
16478 that support '-rand'.
16482 * In RAND_write_file, use mode 0600 for creating files;
16483 don't just chmod when it may be too late.
16487 * Report an error from X509_STORE_load_locations
16488 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
16492 * New function ASN1_mbstring_copy() this copies a string in either
16493 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
16494 into an ASN1_STRING type. A mask of permissible types is passed
16495 and it chooses the "minimal" type to use or an error if not type
16500 * Add function equivalents to the various macros in asn1.h. The old
16501 macros are retained with an `M_` prefix. Code inside the library can
16502 use the `M_` macros. External code (including the openssl utility)
16503 should *NOT* in order to be "shared library friendly".
16507 * Add various functions that can check a certificate's extensions
16508 to see if it usable for various purposes such as SSL client,
16509 server or S/MIME and CAs of these types. This is currently
16510 VERY EXPERIMENTAL but will ultimately be used for certificate chain
16511 verification. Also added a -purpose flag to x509 utility to
16512 print out all the purposes.
16516 * Add a CRYPTO_EX_DATA to X509 certificate structure and associated
16521 * New `X509V3_{X509,CRL,REVOKED}_get_d2i()` functions. These will search
16522 for, obtain and decode and extension and obtain its critical flag.
16523 This allows all the necessary extension code to be handled in a
16524 single function call.
16528 * RC4 tune-up featuring 30-40% performance improvement on most RISC
16529 platforms. See crypto/rc4/rc4_enc.c for further details.
16533 * New -noout option to asn1parse. This causes no output to be produced
16534 its main use is when combined with -strparse and -out to extract data
16535 from a file (which may not be in ASN.1 format).
16539 * Fix for pkcs12 program. It was hashing an invalid certificate pointer
16540 when producing the local key id.
16542 *Richard Levitte <levitte@stacken.kth.se>*
16544 * New option -dhparam in s_server. This allows a DH parameter file to be
16545 stated explicitly. If it is not stated then it tries the first server
16546 certificate file. The previous behaviour hard coded the filename
16551 * Add -pubin and -pubout options to the rsa and dsa commands. These allow
16552 a public key to be input or output. For example:
16553 openssl rsa -in key.pem -pubout -out pubkey.pem
16554 Also added necessary DSA public key functions to handle this.
16558 * Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
16559 in the message. This was handled by allowing
16560 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
16562 *Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>*
16564 * Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
16565 to the end of the strings whereas this didn't. This would cause problems
16566 if strings read with d2i_ASN1_bytes() were later modified.
16568 *Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>*
16570 * Fix for base64 decode bug. When a base64 bio reads only one line of
16571 data and it contains EOF it will end up returning an error. This is
16572 caused by input 46 bytes long. The cause is due to the way base64
16573 BIOs find the start of base64 encoded data. They do this by trying a
16574 trial decode on each line until they find one that works. When they
16575 do a flag is set and it starts again knowing it can pass all the
16576 data directly through the decoder. Unfortunately it doesn't reset
16577 the context it uses. This means that if EOF is reached an attempt
16578 is made to pass two EOFs through the context and this causes the
16579 resulting error. This can also cause other problems as well. As is
16580 usual with these problems it takes *ages* to find and the fix is
16581 trivial: move one line.
16583 *Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer)*
16585 * Ugly workaround to get s_client and s_server working under Windows. The
16586 old code wouldn't work because it needed to select() on sockets and the
16587 tty (for keypresses and to see if data could be written). Win32 only
16588 supports select() on sockets so we select() with a 1s timeout on the
16589 sockets and then see if any characters are waiting to be read, if none
16590 are present then we retry, we also assume we can always write data to
16591 the tty. This isn't nice because the code then blocks until we've
16592 received a complete line of data and it is effectively polling the
16593 keyboard at 1s intervals: however it's quite a bit better than not
16594 working at all :-) A dedicated Windows application might handle this
16595 with an event loop for example.
16599 * Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
16600 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
16601 will be called when RSA_sign() and RSA_verify() are used. This is useful
16602 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
16603 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
16604 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
16605 This necessitated the support of an extra signature type NID_md5_sha1
16606 for SSL signatures and modifications to the SSL library to use it instead
16607 of calling RSA_public_decrypt() and RSA_private_encrypt().
16611 * Add new -verify -CAfile and -CApath options to the crl program, these
16612 will lookup a CRL issuers certificate and verify the signature in a
16613 similar way to the verify program. Tidy up the crl program so it
16614 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
16615 less strict. It will now permit CRL extensions even if it is not
16616 a V2 CRL: this will allow it to tolerate some broken CRLs.
16620 * Initialize all non-automatic variables each time one of the openssl
16621 sub-programs is started (this is necessary as they may be started
16622 multiple times from the "OpenSSL>" prompt).
16624 *Lennart Bang, Bodo Moeller*
16626 * Preliminary compilation option RSA_NULL which disables RSA crypto without
16627 removing all other RSA functionality (this is what NO_RSA does). This
16628 is so (for example) those in the US can disable those operations covered
16629 by the RSA patent while allowing storage and parsing of RSA keys and RSA
16634 * Non-copying interface to BIO pairs.
16635 (still largely untested)
16639 * New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
16640 ASCII string. This was handled independently in various places before.
16644 * New functions UTF8_getc() and UTF8_putc() that parse and generate
16645 UTF8 strings a character at a time.
16649 * Use client_version from client hello to select the protocol
16650 (s23_srvr.c) and for RSA client key exchange verification
16651 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
16655 * Add various utility functions to handle SPKACs, these were previously
16656 handled by poking round in the structure internals. Added new function
16657 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
16658 print, verify and generate SPKACs. Based on an original idea from
16659 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
16663 * RIPEMD160 is operational on all platforms and is back in 'make test'.
16667 * Allow the config file extension section to be overwritten on the
16668 command line. Based on an original idea from Massimiliano Pala
16669 <madwolf@comune.modena.it>. The new option is called -extensions
16670 and can be applied to ca, req and x509. Also -reqexts to override
16671 the request extensions in req and -crlexts to override the crl extensions
16676 * Add new feature to the SPKAC handling in ca. Now you can include
16677 the same field multiple times by preceding it by "XXXX." for example:
16680 this is the same syntax as used in the req config file.
16684 * Allow certificate extensions to be added to certificate requests. These
16685 are specified in a 'req_extensions' option of the req section of the
16686 config file. They can be printed out with the -text option to req but
16687 are otherwise ignored at present.
16691 * Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
16692 data read consists of only the final block it would not decrypted because
16693 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
16694 A misplaced 'break' also meant the decrypted final block might not be
16695 copied until the next read.
16699 * Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
16700 a few extra parameters to the DH structure: these will be useful if
16701 for example we want the value of 'q' or implement X9.42 DH.
16705 * Initial support for DSA_METHOD. This is based on the RSA_METHOD and
16706 provides hooks that allow the default DSA functions or functions on a
16707 "per key" basis to be replaced. This allows hardware acceleration and
16708 hardware key storage to be handled without major modification to the
16709 library. Also added low level modexp hooks and CRYPTO_EX structure and
16710 associated functions.
16714 * Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
16715 as "read only": it can't be written to and the buffer it points to will
16716 not be freed. Reading from a read only BIO is much more efficient than
16717 a normal memory BIO. This was added because there are several times when
16718 an area of memory needs to be read from a BIO. The previous method was
16719 to create a memory BIO and write the data to it, this results in two
16720 copies of the data and an O(n^2) reading algorithm. There is a new
16721 function BIO_new_mem_buf() which creates a read only memory BIO from
16722 an area of memory. Also modified the PKCS#7 routines to use read only
16727 * Bugfix: ssl23_get_client_hello did not work properly when called in
16728 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
16729 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
16730 but a retry condition occurred while trying to read the rest.
16734 * The PKCS7_ENC_CONTENT_new() function was setting the content type as
16735 NID_pkcs7_encrypted by default: this was wrong since this should almost
16736 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
16737 the encrypted data type: this is a more sensible place to put it and it
16738 allows the PKCS#12 code to be tidied up that duplicated this
16743 * Changed obj_dat.pl script so it takes its input and output files on
16744 the command line. This should avoid shell escape redirection problems
16749 * Initial support for certificate extension requests, these are included
16750 in things like Xenroll certificate requests. Included functions to allow
16751 extensions to be obtained and added.
16755 * -crlf option to s_client and s_server for sending newlines as
16756 CRLF (as required by many protocols).
16760 ### Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
16762 * Install libRSAglue.a when OpenSSL is built with RSAref.
16764 *Ralf S. Engelschall*
16766 * A few more `#ifndef NO_FP_API / #endif` pairs for consistency.
16768 *Andrija Antonijevic <TheAntony2@bigfoot.com>*
16770 * Fix -startdate and -enddate (which was missing) arguments to 'ca'
16775 * New function DSA_dup_DH, which duplicates DSA parameters/keys as
16776 DH parameters/keys (q is lost during that conversion, but the resulting
16777 DH parameters contain its length).
16779 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
16780 much faster than DH_generate_parameters (which creates parameters
16781 where `p = 2*q + 1`), and also the smaller q makes DH computations
16782 much more efficient (160-bit exponentiation instead of 1024-bit
16783 exponentiation); so this provides a convenient way to support DHE
16784 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
16785 utter importance to use
16786 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16788 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
16789 when such DH parameters are used, because otherwise small subgroup
16790 attacks may become possible!
16794 * Avoid memory leak in i2d_DHparams.
16798 * Allow the -k option to be used more than once in the enc program:
16799 this allows the same encrypted message to be read by multiple recipients.
16803 * New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
16804 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
16805 it will always use the numerical form of the OID, even if it has a short
16810 * Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
16811 method only got called if p,q,dmp1,dmq1,iqmp components were present,
16812 otherwise bn_mod_exp was called. In the case of hardware keys for example
16813 no private key components need be present and it might store extra data
16814 in the RSA structure, which cannot be accessed from bn_mod_exp.
16815 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
16816 private key operations.
16820 * Added support for SPARC Linux.
16824 * pem_password_cb function type incompatibly changed from
16825 typedef int pem_password_cb(char *buf, int size, int rwflag);
16827 ....(char *buf, int size, int rwflag, void *userdata);
16828 so that applications can pass data to their callbacks:
16829 The `PEM[_ASN1]_{read,write}...` functions and macros now take an
16830 additional void * argument, which is just handed through whenever
16831 the password callback is called.
16833 *Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller*
16835 New function SSL_CTX_set_default_passwd_cb_userdata.
16837 Compatibility note: As many C implementations push function arguments
16838 onto the stack in reverse order, the new library version is likely to
16839 interoperate with programs that have been compiled with the old
16840 pem_password_cb definition (PEM_whatever takes some data that
16841 happens to be on the stack as its last argument, and the callback
16842 just ignores this garbage); but there is no guarantee whatsoever that
16845 * The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
16846 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
16847 problems not only on Windows, but also on some Unix platforms.
16848 To avoid problematic command lines, these definitions are now in an
16849 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
16850 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
16854 * MIPS III/IV assembler module is reimplemented.
16858 * More DES library cleanups: remove references to srand/rand and
16859 delete an unused file.
16863 * Add support for the free Netwide assembler (NASM) under Win32,
16864 since not many people have MASM (ml) and it can be hard to obtain.
16865 This is currently experimental but it seems to work OK and pass all
16866 the tests. Check out INSTALL.W32 for info.
16870 * Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
16871 without temporary keys kept an extra copy of the server key,
16872 and connections with temporary keys did not free everything in case
16877 * New function RSA_check_key and new openssl rsa option -check
16878 for verifying the consistency of RSA keys.
16880 *Ulf Moeller, Bodo Moeller*
16882 * Various changes to make Win32 compile work:
16883 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
16884 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
16885 comparison" warnings.
16886 3. Add `sk_<TYPE>_sort` to DEF file generator and do make update.
16890 * Add a debugging option to PKCS#5 v2 key generation function: when
16891 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
16892 derived keys are printed to stderr.
16896 * Copy the flags in ASN1_STRING_dup().
16898 *Roman E. Pavlov <pre@mo.msk.ru>*
16900 * The x509 application mishandled signing requests containing DSA
16901 keys when the signing key was also DSA and the parameters didn't match.
16903 It was supposed to omit the parameters when they matched the signing key:
16904 the verifying software was then supposed to automatically use the CA's
16905 parameters if they were absent from the end user certificate.
16907 Omitting parameters is no longer recommended. The test was also
16908 the wrong way round! This was probably due to unusual behaviour in
16909 EVP_cmp_parameters() which returns 1 if the parameters match.
16910 This meant that parameters were omitted when they *didn't* match and
16911 the certificate was useless. Certificates signed with 'ca' didn't have
16914 *Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>*
16916 * Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
16917 The interface is as follows:
16918 Applications can use
16919 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
16920 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
16921 "off" is now the default.
16922 The library internally uses
16923 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
16924 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
16925 to disable memory-checking temporarily.
16927 Some inconsistent states that previously were possible (and were
16928 even the default) are now avoided.
16930 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
16931 with each memory chunk allocated; this is occasionally more helpful
16932 than just having a counter.
16934 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
16936 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
16941 * Introduce "mode" for SSL structures (with defaults in SSL_CTX),
16942 which largely parallels "options", but is for changing API behaviour,
16943 whereas "options" are about protocol behaviour.
16944 Initial "mode" flags are:
16946 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
16947 a single record has been written.
16948 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
16949 retries use the same buffer location.
16950 (But all of the contents must be
16955 * Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
16958 * Fix problems with no-hmac etc.
16960 *Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>*
16962 * New functions RSA_get_default_method(), RSA_set_method() and
16963 RSA_get_method(). These allows replacement of RSA_METHODs without having
16964 to mess around with the internals of an RSA structure.
16968 * Fix memory leaks in DSA_do_sign and DSA_is_prime.
16969 Also really enable memory leak checks in openssl.c and in some
16972 *Chad C. Mulligan, Bodo Moeller*
16974 * Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
16975 up the length of negative integers. This has now been simplified to just
16976 store the length when it is first determined and use it later, rather
16977 than trying to keep track of where data is copied and updating it to
16979 *Steve Henson, reported by Brien Wheeler <bwheeler@authentica-security.com>*
16981 * Add a new function PKCS7_signatureVerify. This allows the verification
16982 of a PKCS#7 signature but with the signing certificate passed to the
16983 function itself. This contrasts with PKCS7_dataVerify which assumes the
16984 certificate is present in the PKCS#7 structure. This isn't always the
16985 case: certificates can be omitted from a PKCS#7 structure and be
16986 distributed by "out of band" means (such as a certificate database).
16990 * Complete the `PEM_*` macros with DECLARE_PEM versions to replace the
16991 function prototypes in pem.h, also change util/mkdef.pl to add the
16992 necessary function names.
16996 * mk1mf.pl (used by Windows builds) did not properly read the
16997 options set by Configure in the top level Makefile, and Configure
16998 was not even able to write more than one option correctly.
16999 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
17003 * New functions CONF_load_bio() and CONF_load_fp() to allow a config
17004 file to be loaded from a BIO or FILE pointer. The BIO version will
17005 for example allow memory BIOs to contain config info.
17009 * New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
17010 Whoever hopes to achieve shared-library compatibility across versions
17011 must use this, not the compile-time macro.
17012 (Exercise 0.9.4: Which is the minimum library version required by
17014 Note: All this applies only to multi-threaded programs, others don't
17019 * Add missing case to s3_clnt.c state machine -- one of the new SSL tests
17020 through a BIO pair triggered the default case, i.e.
17021 SSLerr(...,SSL_R_UNKNOWN_STATE).
17025 * New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
17026 can use the SSL library even if none of the specific BIOs is
17031 * Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
17032 for the encoded length.
17034 *Jeon KyoungHo <khjeon@sds.samsung.co.kr>*
17036 * Add initial documentation of the X509V3 functions.
17040 * Add a new pair of functions PEM_write_PKCS8PrivateKey() and
17041 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
17042 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
17043 secure PKCS#8 private key format with a high iteration count.
17047 * Fix determination of Perl interpreter: A perl or perl5
17048 *directory* in $PATH was also accepted as the interpreter.
17050 *Ralf S. Engelschall*
17052 * Fix demos/sign/sign.c: well there wasn't anything strictly speaking
17053 wrong with it but it was very old and did things like calling
17054 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
17055 unusual formatting.
17059 * Fix demos/selfsign.c: it used obsolete and deleted functions, changed
17060 to use the new extension code.
17064 * Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
17065 with macros. This should make it easier to change their form, add extra
17066 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
17071 * Add to configuration table a new entry that can specify an alternative
17072 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
17073 according to Mark Crispin <MRC@Panda.COM>.
17077 * DES CBC did not update the IV. Weird.
17081 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
17082 Changing the behaviour of the former might break existing programs --
17083 where IV updating is needed, des_ncbc_encrypt can be used.
17086 * When bntest is run from "make test" it drives bc to check its
17087 calculations, as well as internally checking them. If an internal check
17088 fails, it needs to cause bc to give a non-zero result or make test carries
17089 on without noticing the failure. Fixed.
17093 * DES library cleanups.
17097 * Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
17098 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
17099 ciphers. NOTE: although the key derivation function has been verified
17100 against some published test vectors it has not been extensively tested
17101 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
17106 * Instead of "mkdir -p", which is not fully portable, use new
17107 Perl script "util/mkdir-p.pl".
17111 * Rewrite the way password based encryption (PBE) is handled. It used to
17112 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
17113 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
17114 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
17115 the 'parameter' field of the AlgorithmIdentifier is passed to the
17116 underlying key generation function so it must do its own ASN1 parsing.
17117 This has also changed the EVP_PBE_CipherInit() function which now has a
17118 'parameter' argument instead of literal salt and iteration count values
17119 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
17123 * Support for PKCS#5 v1.5 compatible password based encryption algorithms
17124 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
17125 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
17126 KEY" because this clashed with PKCS#8 unencrypted string. Since this
17127 value was just used as a "magic string" and not used directly its
17128 value doesn't matter.
17132 * Introduce some semblance of const correctness to BN. Shame C doesn't
17137 * "linux-sparc64" configuration (ultrapenguin).
17139 *Ray Miller <ray.miller@oucs.ox.ac.uk>*
17140 "linux-sparc" configuration.
17142 *Christian Forster <fo@hawo.stw.uni-erlangen.de>*
17144 * config now generates no-xxx options for missing ciphers.
17148 * Support the EBCDIC character set (work in progress).
17149 File ebcdic.c not yet included because it has a different license.
17151 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17153 * Support BS2000/OSD-POSIX.
17155 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>*
17157 * Make callbacks for key generation use `void *` instead of `char *`.
17161 * Make S/MIME samples compile (not yet tested).
17165 * Additional typesafe stacks.
17169 * New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
17173 ### Changes between 0.9.3 and 0.9.3a [29 May 1999]
17175 * New configuration variant "sco5-gcc".
17177 * Updated some demos.
17179 *Sean O Riordain, Wade Scholine*
17181 * Add missing BIO_free at exit of pkcs12 application.
17185 * Fix memory leak in conf.c.
17189 * Updates for Win32 to assembler version of MD5.
17193 * Set #! path to perl in apps/der_chop to where we found it
17194 instead of using a fixed path.
17198 * SHA library changes for irix64-mips4-cc.
17202 * Improvements for VMS support.
17206 ### Changes between 0.9.2b and 0.9.3 [24 May 1999]
17208 * Bignum library bug fix. IRIX 6 passes "make test" now!
17209 This also avoids the problems with SC4.2 and unpatched SC5.
17211 *Andy Polyakov <appro@fy.chalmers.se>*
17213 * New functions sk_num, sk_value and sk_set to replace the previous macros.
17214 These are required because of the typesafe stack would otherwise break
17215 existing code. If old code used a structure member which used to be STACK
17216 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
17217 sk_num or sk_value it would produce an error because the num, data members
17218 are not present in STACK_OF. Now it just produces a warning. sk_set
17219 replaces the old method of assigning a value to sk_value
17220 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
17221 that does this will no longer work (and should use sk_set instead) but
17222 this could be regarded as a "questionable" behaviour anyway.
17226 * Fix most of the other PKCS#7 bugs. The "experimental" code can now
17227 correctly handle encrypted S/MIME data.
17231 * Change type of various DES function arguments from des_cblock
17232 (which means, in function argument declarations, pointer to char)
17233 to des_cblock * (meaning pointer to array with 8 char elements),
17234 which allows the compiler to do more typechecking; it was like
17235 that back in SSLeay, but with lots of ugly casts.
17237 Introduce new type const_des_cblock.
17241 * Reorganise the PKCS#7 library and get rid of some of the more obvious
17242 problems: find RecipientInfo structure that matches recipient certificate
17243 and initialise the ASN1 structures properly based on passed cipher.
17247 * Belatedly make the BN tests actually check the results.
17251 * Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
17252 to and from BNs: it was completely broken. New compilation option
17253 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
17254 key elements as negative integers.
17258 * Reorganize and speed up MD5.
17260 *Andy Polyakov <appro@fy.chalmers.se>*
17264 *Richard Levitte <richard@levitte.org>*
17266 * New option -out to asn1parse to allow the parsed structure to be
17267 output to a file. This is most useful when combined with the -strparse
17268 option to examine the output of things like OCTET STRINGS.
17272 * Make SSL library a little more fool-proof by not requiring any longer
17273 that `SSL_set_{accept,connect}_state` be called before
17274 `SSL_{accept,connect}` may be used (`SSL_set_..._state` is omitted
17275 in many applications because usually everything *appeared* to work as
17276 intended anyway -- now it really works as intended).
17280 * Move openssl.cnf out of lib/.
17284 * Fix various things to let OpenSSL even pass "egcc -pipe -O2 -Wall
17285 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
17286 -Wmissing-declarations -Wnested-externs -Winline" with EGCS 1.1.2+
17288 *Ralf S. Engelschall*
17290 * Various fixes to the EVP and PKCS#7 code. It may now be able to
17291 handle PKCS#7 enveloped data properly.
17293 *Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve*
17295 * Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
17296 copying pointers. The cert_st handling is changed by this in
17297 various ways (and thus what used to be known as ctx->default_cert
17298 is now called ctx->cert, since we don't resort to `s->ctx->[default_]cert`
17299 any longer when s->cert does not give us what we need).
17300 ssl_cert_instantiate becomes obsolete by this change.
17301 As soon as we've got the new code right (possibly it already is?),
17302 we have solved a couple of bugs of the earlier code where s->cert
17303 was used as if it could not have been shared with other SSL structures.
17305 Note that using the SSL API in certain dirty ways now will result
17306 in different behaviour than observed with earlier library versions:
17307 Changing settings for an `SSL_CTX *ctx` after having done s = SSL_new(ctx)
17308 does not influence s as it used to.
17310 In order to clean up things more thoroughly, inside SSL_SESSION
17311 we don't use CERT any longer, but a new structure SESS_CERT
17312 that holds per-session data (if available); currently, this is
17313 the peer's certificate chain and, for clients, the server's certificate
17314 and temporary key. CERT holds only those values that can have
17315 meaningful defaults in an SSL_CTX.
17319 * New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
17320 from the internal representation. Various PKCS#7 fixes: remove some
17321 evil casts and set the enc_dig_alg field properly based on the signing
17326 * Allow PKCS#12 password to be set from the command line or the
17327 environment. Let 'ca' get its config file name from the environment
17328 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
17333 * Allow certificate policies extension to use an IA5STRING for the
17334 organization field. This is contrary to the PKIX definition but
17335 VeriSign uses it and IE5 only recognises this form. Document 'x509'
17340 * Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
17341 without disallowing inline assembler and the like for non-pedantic builds.
17345 * Support Borland C++ builder.
17347 *Janez Jere <jj@void.si>, modified by Ulf Möller*
17353 * SHA-1 cleanups and performance enhancements.
17355 *Andy Polyakov <appro@fy.chalmers.se>*
17357 * Sparc v8plus assembler for the bignum library.
17359 *Andy Polyakov <appro@fy.chalmers.se>*
17361 * Accept any -xxx and +xxx compiler options in Configure.
17365 * Update HPUX configuration.
17369 * Add missing `sk_<type>_unshift()` function to safestack.h
17371 *Ralf S. Engelschall*
17373 * New function SSL_CTX_use_certificate_chain_file that sets the
17374 "extra_cert"s in addition to the certificate. (This makes sense
17375 only for "PEM" format files, as chains as a whole are not
17380 * Support verify_depth from the SSL API.
17381 x509_vfy.c had what can be considered an off-by-one-error:
17382 Its depth (which was not part of the external interface)
17383 was actually counting the number of certificates in a chain;
17384 now it really counts the depth.
17388 * Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
17389 instead of X509err, which often resulted in confusing error
17390 messages since the error codes are not globally unique
17391 (e.g. an alleged error in ssl3_accept when a certificate
17392 didn't match the private key).
17394 * New function SSL_CTX_set_session_id_context that allows to set a default
17395 value (so that you don't need SSL_set_session_id_context for each
17396 connection using the SSL_CTX).
17400 * OAEP decoding bug fix.
17404 * Support INSTALL_PREFIX for package builders, as proposed by
17409 * New Configure options "threads" and "no-threads". For systems
17410 where the proper compiler options are known (currently Solaris
17411 and Linux), "threads" is the default.
17415 * New script util/mklink.pl as a faster substitute for util/mklink.sh.
17419 * Install various scripts to $(OPENSSLDIR)/misc, not to
17420 $(INSTALLTOP)/bin -- they shouldn't clutter directories
17421 such as /usr/local/bin.
17425 * "make linux-shared" to build shared libraries.
17427 *Niels Poppe <niels@netbox.org>*
17429 * New Configure option `no-<cipher>` (rsa, idea, rc5, ...).
17433 * Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
17434 extension adding in x509 utility.
17438 * Remove NOPROTO sections and error code comments.
17442 * Partial rewrite of the DEF file generator to now parse the ANSI
17447 * New Configure options --prefix=DIR and --openssldir=DIR.
17451 * Complete rewrite of the error code script(s). It is all now handled
17452 by one script at the top level which handles error code gathering,
17453 header rewriting and C source file generation. It should be much better
17454 than the old method: it now uses a modified version of Ulf's parser to
17455 read the ANSI prototypes in all header files (thus the old K&R definitions
17456 aren't needed for error creation any more) and do a better job of
17457 translating function codes into names. The old 'ASN1 error code embedded
17458 in a comment' is no longer necessary and it doesn't use .err files which
17459 have now been deleted. Also the error code call doesn't have to appear all
17460 on one line (which resulted in some large lines...).
17464 * Change #include filenames from `<foo.h>` to `<openssl/foo.h>`.
17468 * Change behaviour of ssl2_read when facing length-0 packets: Don't return
17469 0 (which usually indicates a closed connection), but continue reading.
17473 * Fix some race conditions.
17477 * Add support for CRL distribution points extension. Add Certificate
17478 Policies and CRL distribution points documentation.
17482 * Move the autogenerated header file parts to crypto/opensslconf.h.
17486 * Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
17487 8 of keying material. Merlin has also confirmed interop with this fix
17488 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
17490 *Merlin Hughes <merlin@baltimore.ie>*
17492 * Fix lots of warnings.
17494 *Richard Levitte <levitte@stacken.kth.se>*
17496 * In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
17497 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
17499 *Richard Levitte <levitte@stacken.kth.se>*
17501 * Fix problems with sizeof(long) == 8.
17503 *Andy Polyakov <appro@fy.chalmers.se>*
17505 * Change functions to ANSI C.
17509 * Fix typos in error codes.
17511 *Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller*
17513 * Remove defunct assembler files from Configure.
17517 * SPARC v8 assembler BIGNUM implementation.
17519 *Andy Polyakov <appro@fy.chalmers.se>*
17521 * Support for Certificate Policies extension: both print and set.
17522 Various additions to support the r2i method this uses.
17526 * A lot of constification, and fix a bug in X509_NAME_oneline() that could
17527 return a const string when you are expecting an allocated buffer.
17531 * Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
17532 types DirectoryString and DisplayText.
17536 * Add code to allow r2i extensions to access the configuration database,
17537 add an LHASH database driver and add several ctx helper functions.
17541 * Fix an evil bug in bn_expand2() which caused various BN functions to
17542 fail when they extended the size of a BIGNUM.
17546 * Various utility functions to handle SXNet extension. Modify mkdef.pl to
17547 support typesafe stack.
17551 * Fix typo in SSL_[gs]et_options().
17553 *Nils Frostberg <nils@medcom.se>*
17555 * Delete various functions and files that belonged to the (now obsolete)
17556 old X509V3 handling code.
17560 * New Configure option "rsaref".
17564 * Don't auto-generate pem.h.
17568 * Introduce type-safe ASN.1 SETs.
17572 * Convert various additional casted stacks to type-safe STACK_OF() variants.
17574 *Ben Laurie, Ralf S. Engelschall, Steve Henson*
17576 * Introduce type-safe STACKs. This will almost certainly break lots of code
17577 that links with OpenSSL (well at least cause lots of warnings), but fear
17578 not: the conversion is trivial, and it eliminates loads of evil casts. A
17579 few STACKed things have been converted already. Feel free to convert more.
17580 In the fullness of time, I'll do away with the STACK type altogether.
17584 * Add `openssl ca -revoke <certfile>` facility which revokes a certificate
17585 specified in `<certfile>` by updating the entry in the index.txt file.
17586 This way one no longer has to edit the index.txt file manually for
17587 revoking a certificate. The -revoke option does the gory details now.
17589 *Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall*
17591 * Fix `openssl crl -noout -text` combination where `-noout` killed the
17592 `-text` option at all and this way the `-noout -text` combination was
17593 inconsistent in `openssl crl` with the friends in `openssl x509|rsa|dsa`.
17595 *Ralf S. Engelschall*
17597 * Make sure a corresponding plain text error message exists for the
17598 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
17599 verify callback function determined that a certificate was revoked.
17601 *Ralf S. Engelschall*
17603 * Bugfix: In test/testenc, don't test `openssl <cipher>` for
17604 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
17605 all available ciphers including rc5, which was forgotten until now.
17606 In order to let the testing shell script know which algorithms
17607 are available, a new (up to now undocumented) command
17608 `openssl list-cipher-commands` is used.
17612 * Bugfix: s_client occasionally would sleep in select() when
17613 it should have checked SSL_pending() first.
17617 * New functions DSA_do_sign and DSA_do_verify to provide access to
17618 the raw DSA values prior to ASN.1 encoding.
17622 * Tweaks to Configure
17624 *Niels Poppe <niels@netbox.org>*
17626 * Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
17631 * New variables $(RANLIB) and $(PERL) in the Makefiles.
17635 * New config option to avoid instructions that are illegal on the 80386.
17636 The default code is faster, but requires at least a 486.
17640 * Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
17641 SSL2_SERVER_VERSION (not used at all) macros, which are now the
17642 same as SSL2_VERSION anyway.
17646 * New "-showcerts" option for s_client.
17650 * Still more PKCS#12 integration. Add pkcs12 application to openssl
17651 application. Various cleanups and fixes.
17655 * More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
17656 modify error routines to work internally. Add error codes and PBE init
17657 to library startup routines.
17661 * Further PKCS#12 integration. Added password based encryption, PKCS#8 and
17662 packing functions to asn1 and evp. Changed function names and error
17663 codes along the way.
17667 * PKCS12 integration: and so it begins... First of several patches to
17668 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
17669 objects to objects.h
17673 * Add a new 'indent' option to some X509V3 extension code. Initial ASN1
17674 and display support for Thawte strong extranet extension.
17678 * Add LinuxPPC support.
17680 *Jeff Dubrule <igor@pobox.org>*
17682 * Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
17683 bn_div_words in alpha.s.
17685 *Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie*
17687 * Make sure the RSA OAEP test is skipped under -DRSAref because
17688 OAEP isn't supported when OpenSSL is built with RSAref.
17690 *Ulf Moeller <ulf@fitug.de>*
17692 * Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
17693 so they no longer are missing under -DNOPROTO.
17695 *Soren S. Jorvang <soren@t.dk>*
17697 ### Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
17699 * Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
17700 doesn't work when the session is reused. Coming soon!
17704 * Fix a security hole, that allows sessions to be reused in the wrong
17705 context thus bypassing client cert protection! All software that uses
17706 client certs and session caches in multiple contexts NEEDS PATCHING to
17707 allow session reuse! A fuller solution is in the works.
17709 *Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)*
17711 * Some more source tree cleanups (removed obsolete files
17712 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
17713 permission on "config" script to be executable) and a fix for the INSTALL
17716 *Ulf Moeller <ulf@fitug.de>*
17718 * Remove some legacy and erroneous uses of malloc, free instead of
17721 *Lennart Bang <lob@netstream.se>, with minor changes by Steve*
17723 * Make rsa_oaep_test return non-zero on error.
17725 *Ulf Moeller <ulf@fitug.de>*
17727 * Add support for native Solaris shared libraries. Configure
17728 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
17729 if someone would make that last step automatic.
17731 *Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>*
17733 * ctx_size was not built with the right compiler during "make links". Fixed.
17737 * Change the meaning of 'ALL' in the cipher list. It now means "everything
17738 except NULL ciphers". This means the default cipher list will no longer
17739 enable NULL ciphers. They need to be specifically enabled e.g. with
17740 the string "DEFAULT:eNULL".
17744 * Fix to RSA private encryption routines: if p < q then it would
17745 occasionally produce an invalid result. This will only happen with
17746 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
17750 * Be less restrictive and allow also `perl util/perlpath.pl
17751 /path/to/bin/perl` in addition to `perl util/perlpath.pl /path/to/bin`,
17752 because this way one can also use an interpreter named `perl5` (which is
17753 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
17754 installed as `perl`).
17756 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17758 * Let util/clean-depend.pl work also with older Perl 5.00x versions.
17760 *Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
17762 * Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
17763 advapi32.lib to Win32 build and change the pem test comparison
17764 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
17765 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
17766 and crypto/des/ede_cbcm_enc.c.
17770 * DES quad checksum was broken on big-endian architectures. Fixed.
17774 * Comment out two functions in bio.h that aren't implemented. Fix up the
17775 Win32 test batch file so it (might) work again. The Win32 test batch file
17776 is horrible: I feel ill....
17780 * Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
17781 in e_os.h. Audit of header files to check ANSI and non ANSI
17782 sections: 10 functions were absent from non ANSI section and not exported
17783 from Windows DLLs. Fixed up libeay.num for new functions.
17787 * Make `openssl version` output lines consistent.
17789 *Ralf S. Engelschall*
17791 * Fix Win32 symbol export lists for BIO functions: Added
17792 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
17793 to ms/libeay{16,32}.def.
17795 *Ralf S. Engelschall*
17797 * Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
17798 fine under Unix and passes some trivial tests I've now added. But the
17799 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
17800 added to make sure no one expects that this stuff really works in the
17801 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
17802 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
17805 *Ralf S. Engelschall*
17807 * Fix the generation of two part addresses in perl.
17809 *Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie*
17811 * Add config entry for Linux on MIPS.
17813 *John Tobey <jtobey@channel1.com>*
17815 * Make links whenever Configure is run, unless we are on Windoze.
17819 * Permit extensions to be added to CRLs using crl_section in openssl.cnf.
17820 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
17825 * Add a useful kludge to allow package maintainers to specify compiler and
17826 other platforms details on the command line without having to patch the
17827 Configure script every time: One now can use
17828 `perl Configure <id>:<details>`,
17829 i.e. platform ids are allowed to have details appended
17830 to them (separated by colons). This is treated as there would be a static
17831 pre-configured entry in Configure's %table under key `<id>` with value
17832 `<details>` and `perl Configure <id>` is called. So, when you want to
17833 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
17834 assembler stuff you can use `perl Configure "FreeBSD-elf:pgcc:-O6:::"`
17835 now, which overrides the FreeBSD-elf entry on-the-fly.
17837 *Ralf S. Engelschall*
17839 * Disable new TLS1 ciphersuites by default: they aren't official yet.
17843 * Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
17844 on the `perl Configure ...` command line. This way one can compile
17845 OpenSSL libraries with Position Independent Code (PIC) which is needed
17846 for linking it into DSOs.
17848 *Ralf S. Engelschall*
17850 * Remarkably, export ciphers were totally broken and no-one had noticed!
17855 * Cleaned up the LICENSE document: The official contact for any license
17856 questions now is the OpenSSL core team under openssl-core@openssl.org.
17857 And add a paragraph about the dual-license situation to make sure people
17858 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
17859 to the OpenSSL toolkit.
17861 *Ralf S. Engelschall*
17863 * General source tree makefile cleanups: Made `making xxx in yyy...`
17864 display consistent in the source tree and replaced `/bin/rm` by `rm`.
17865 Additionally cleaned up the `make links` target: Remove unnecessary
17866 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
17867 to speed processing and no longer clutter the display with confusing
17868 stuff. Instead only the actually done links are displayed.
17870 *Ralf S. Engelschall*
17872 * Permit null encryption ciphersuites, used for authentication only. It used
17873 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
17874 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
17879 * Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
17880 signed attributes when verifying signatures (this would break them),
17881 the detached data encoding was wrong and public keys obtained using
17882 X509_get_pubkey() weren't freed.
17886 * Add text documentation for the BUFFER functions. Also added a work around
17887 to a Win95 console bug. This was triggered by the password read stuff: the
17888 last character typed gets carried over to the next fread(). If you were
17889 generating a new cert request using 'req' for example then the last
17890 character of the passphrase would be CR which would then enter the first
17895 * Added the new 'Includes OpenSSL Cryptography Software' button as
17896 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
17897 button and can be used by applications based on OpenSSL to show the
17898 relationship to the OpenSSL project.
17900 *Ralf S. Engelschall*
17902 * Remove confusing variables in function signatures in files
17903 ssl/ssl_lib.c and ssl/ssl.h.
17905 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17907 * Don't install bss_file.c under PREFIX/include/
17909 *Lennart Bong <lob@kulthea.stacken.kth.se>*
17911 * Get the Win32 compile working again. Modify mkdef.pl so it can handle
17912 functions that return function pointers and has support for NT specific
17913 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
17914 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
17915 unsigned to signed types: this was killing the Win32 compile.
17919 * Add new certificate file to stack functions,
17920 SSL_add_dir_cert_subjects_to_stack() and
17921 SSL_add_file_cert_subjects_to_stack(). These largely supplant
17922 SSL_load_client_CA_file(), and can be used to add multiple certs easily
17923 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
17924 This means that Apache-SSL and similar packages don't have to mess around
17925 to add as many CAs as they want to the preferred list.
17929 * Experiment with doxygen documentation. Currently only partially applied to
17931 See <http://www.stack.nl/~dimitri/doxygen/index.html>, and run doxygen with
17932 openssl.doxy as the configuration file.
17936 * Get rid of remaining C++-style comments which strict C compilers hate.
17938 *Ralf S. Engelschall, pointed out by Carlos Amengual*
17940 * Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
17941 compiled in by default: it has problems with large keys.
17945 * Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
17946 DH private keys and/or callback functions which directly correspond to
17947 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
17948 is needed for applications which have to configure certificates on a
17949 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
17951 For the RSA certificate situation is makes no difference, but
17952 for the DSA certificate situation this fixes the "no shared cipher"
17953 problem where the OpenSSL cipher selection procedure failed because the
17954 temporary keys were not overtaken from the context and the API provided
17955 no way to reconfigure them.
17956 The new functions now let applications reconfigure the stuff and they
17957 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
17958 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
17959 non-public-API function ssl_cert_instantiate() is used as a helper
17960 function and also to reduce code redundancy inside ssl_rsa.c.
17962 *Ralf S. Engelschall*
17964 * Move s_server -dcert and -dkey options out of the undocumented feature
17965 area because they are useful for the DSA situation and should be
17966 recognized by the users.
17968 *Ralf S. Engelschall*
17970 * Fix the cipher decision scheme for export ciphers: the export bits are
17971 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
17972 SSL_EXP_MASK. So, the original variable has to be used instead of the
17973 already masked variable.
17975 *Richard Levitte <levitte@stacken.kth.se>*
17977 * Fix `port` variable from `int` to `unsigned int` in crypto/bio/b_sock.c
17979 *Richard Levitte <levitte@stacken.kth.se>*
17981 * Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
17982 from `int` to `unsigned int` because it is a length and initialized by
17983 EVP_DigestFinal() which expects an `unsigned int *`.
17985 *Richard Levitte <levitte@stacken.kth.se>*
17987 * Don't hard-code path to Perl interpreter on shebang line of Configure
17988 script. Instead use the usual Shell->Perl transition trick.
17990 *Ralf S. Engelschall*
17992 * Make `openssl x509 -noout -modulus`' functional also for DSA certificates
17993 (in addition to RSA certificates) to match the behaviour of `openssl dsa
17994 -noout -modulus` as it's already the case for `openssl rsa -noout
17995 -modulus`. For RSA the -modulus is the real "modulus" while for DSA
17996 currently the public key is printed (a decision which was already done by
17997 `openssl dsa -modulus` in the past) which serves a similar purpose.
17998 Additionally the NO_RSA no longer completely removes the whole -modulus
17999 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
18002 *Ralf S. Engelschall*
18004 * Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
18005 BIO. See the source (crypto/evp/bio_ok.c) for more info.
18007 *Arne Ansper <arne@ats.cyber.ee>*
18009 * Dump the old yucky req code that tried (and failed) to allow raw OIDs
18010 to be added. Now both 'req' and 'ca' can use new objects defined in the
18015 * Add cool BIO that does syslog (or event log on NT).
18017 *Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie*
18019 * Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
18020 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
18021 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
18022 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
18026 * Add preliminary config info for new extension code.
18030 * Make RSA_NO_PADDING really use no padding.
18032 *Ulf Moeller <ulf@fitug.de>*
18034 * Generate errors when private/public key check is done.
18038 * Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
18039 for some CRL extensions and new objects added.
18043 * Really fix the ASN1 IMPLICIT bug this time... Partial support for private
18044 key usage extension and fuller support for authority key id.
18048 * Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
18049 padding method for RSA, which is recommended for new applications in PKCS
18050 #1 v2.0 (RFC 2437, October 1998).
18051 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
18052 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
18053 against Bleichbacher's attack on RSA.
18054 *Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
18057 * Updates to the new SSL compression code
18059 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18061 * Fix so that the version number in the master secret, when passed
18062 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
18063 (because the server will not accept higher), that the version number
18064 is 0x03,0x01, not 0x03,0x00
18066 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18068 * Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
18069 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
18070 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
18074 * Support for RAW extensions where an arbitrary extension can be
18075 created by including its DER encoding. See apps/openssl.cnf for
18080 * Make sure latest Perl versions don't interpret some generated C array
18081 code as Perl array code in the crypto/err/err_genc.pl script.
18083 *Lars Weber <3weber@informatik.uni-hamburg.de>*
18085 * Modify ms/do_ms.bat to not generate assembly language makefiles since
18086 not many people have the assembler. Various Win32 compilation fixes and
18087 update to the INSTALL.W32 file with (hopefully) more accurate Win32
18088 build instructions.
18092 * Modify configure script 'Configure' to automatically create crypto/date.h
18093 file under Win32 and also build pem.h from pem.org. New script
18094 util/mkfiles.pl to create the MINFO file on environments that can't do a
18095 'make files': perl util/mkfiles.pl >MINFO should work.
18099 * Major rework of DES function declarations, in the pursuit of correctness
18100 and purity. As a result, many evil casts evaporated, and some weirdness,
18101 too. You may find this causes warnings in your code. Zapping your evil
18102 casts will probably fix them. Mostly.
18106 * Fix for a typo in asn1.h. Bug fix to object creation script
18107 obj_dat.pl. It considered a zero in an object definition to mean
18108 "end of object": none of the objects in objects.h have any zeros
18109 so it wasn't spotted.
18111 *Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>*
18113 * Add support for Triple DES Cipher Block Chaining with Output Feedback
18114 Masking (CBCM). In the absence of test vectors, the best I have been able
18115 to do is check that the decrypt undoes the encrypt, so far. Send me test
18116 vectors if you have them.
18120 * Correct calculation of key length for export ciphers (too much space was
18121 allocated for null ciphers). This has not been tested!
18125 * Modifications to the mkdef.pl for Win32 DEF file creation. The usage
18126 message is now correct (it understands "crypto" and "ssl" on its
18127 command line). There is also now an "update" option. This will update
18128 the util/ssleay.num and util/libeay.num files with any new functions.
18130 perl util/mkdef.pl crypto ssl update
18131 it will update them.
18135 * Overhauled the Perl interface:
18136 - ported BN stuff to OpenSSL's different BN library
18137 - made the perl/ source tree CVS-aware
18138 - renamed the package from SSLeay to OpenSSL (the files still contain
18139 their history because I've copied them in the repository)
18140 - removed obsolete files (the test scripts will be replaced
18141 by better Test::Harness variants in the future)
18143 *Ralf S. Engelschall*
18145 * First cut for a very conservative source tree cleanup:
18146 1. merge various obsolete readme texts into doc/ssleay.txt
18147 where we collect the old documents and readme texts.
18148 2. remove the first part of files where I'm already sure that we no
18149 longer need them because of three reasons: either they are just temporary
18150 files which were left by Eric or they are preserved original files where
18151 I've verified that the diff is also available in the CVS via "cvs diff
18152 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
18153 the crypto/md/ stuff).
18155 *Ralf S. Engelschall*
18157 * More extension code. Incomplete support for subject and issuer alt
18158 name, issuer and authority key id. Change the i2v function parameters
18159 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
18160 what that's for :-) Fix to ASN1 macro which messed up
18161 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
18165 * Preliminary support for ENUMERATED type. This is largely copied from the
18170 * Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
18172 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18174 * Make sure `make rehash` target really finds the `openssl` program.
18176 *Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>*
18178 * Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
18179 like to hear about it if this slows down other processors.
18183 * Add CygWin32 platform information to Configure script.
18185 *Alan Batie <batie@aahz.jf.intel.com>*
18187 * Fixed ms/32all.bat script: `no_asm` -> `no-asm`
18189 *Rainer W. Gerling <gerling@mpg-gv.mpg.de>*
18191 * New program nseq to manipulate netscape certificate sequences
18195 * Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
18200 * Fixes to BN code. Previously the default was to define BN_RECURSION
18201 but the BN code had some problems that would cause failures when
18202 doing certificate verification and some other functions.
18204 *Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)*
18206 * Add ASN1 and PEM code to support netscape certificate sequences.
18210 * Add ASN1 and PEM code to support netscape certificate sequences.
18214 * Add several PKIX and private extended key usage OIDs.
18218 * Modify the 'ca' program to handle the new extension code. Modify
18219 openssl.cnf for new extension format, add comments.
18223 * More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
18224 and add a sample to openssl.cnf so req -x509 now adds appropriate
18229 * Continued X509 V3 changes. Add to other makefiles, integrate with the
18230 error code, add initial support to X509_print() and x509 application.
18234 * Takes a deep breath and start adding X509 V3 extension support code. Add
18235 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
18236 stuff is currently isolated and isn't even compiled yet.
18240 * Continuing patches for GeneralizedTime. Fix up certificate and CRL
18241 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
18242 Removed the versions check from X509 routines when loading extensions:
18243 this allows certain broken certificates that don't set the version
18244 properly to be processed.
18248 * Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
18249 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
18250 can still be regenerated with "make depend".
18254 * Spelling mistake in C version of CAST-128.
18256 *Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>*
18258 * Changes to the error generation code. The perl script err-code.pl
18259 now reads in the old error codes and retains the old numbers, only
18260 adding new ones if necessary. It also only changes the .err files if new
18261 codes are added. The makefiles have been modified to only insert errors
18262 when needed (to avoid needlessly modifying header files). This is done
18263 by only inserting errors if the .err file is newer than the auto generated
18264 C file. To rebuild all the error codes from scratch (the old behaviour)
18265 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
18266 or delete all the .err files.
18270 * CAST-128 was incorrectly implemented for short keys. The C version has
18271 been fixed, but is untested. The assembler versions are also fixed, but
18272 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
18273 to regenerate it if needed.
18274 *Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
18275 Hagino <itojun@kame.net>*
18277 * File was opened incorrectly in randfile.c.
18279 *Ulf Möller <ulf@fitug.de>*
18281 * Beginning of support for GeneralizedTime. d2i, i2d, check and print
18282 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
18283 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
18284 al: it's just almost always a UTCTime. Note this patch adds new error
18285 codes so do a "make errors" if there are problems.
18289 * Correct Linux 1 recognition in config.
18291 *Ulf Möller <ulf@fitug.de>*
18293 * Remove pointless MD5 hash when using DSA keys in ca.
18295 *Anonymous <nobody@replay.com>*
18297 * Generate an error if given an empty string as a cert directory. Also
18298 generate an error if handed NULL (previously returned 0 to indicate an
18299 error, but didn't set one).
18301 *Ben Laurie, reported by Anonymous <nobody@replay.com>*
18303 * Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
18307 * Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
18308 parameters. This was causing a warning which killed off the Win32 compile.
18312 * Remove C++ style comments from crypto/bn/bn_local.h.
18314 *Neil Costigan <neil.costigan@celocom.com>*
18316 * The function OBJ_txt2nid was broken. It was supposed to return a nid
18317 based on a text string, looking up short and long names and finally
18318 "dot" format. The "dot" format stuff didn't work. Added new function
18319 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
18320 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
18321 OID is not part of the table.
18325 * Add prototypes to X509 lookup/verify methods, fixing a bug in
18326 X509_LOOKUP_by_alias().
18330 * Sort openssl functions by name.
18334 * Get the gendsa program working (hopefully) and add it to app list. Remove
18335 encryption from sample DSA keys (in case anyone is interested the password
18340 * Make *all* `*_free` functions accept a NULL pointer.
18342 *Frans Heymans <fheymans@isaserver.be>*
18344 * If a DH key is generated in s3_srvr.c, don't blow it by trying to use
18347 *Anonymous <nobody@replay.com>*
18349 * s_server should send the CAfile as acceptable CAs, not its own cert.
18351 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18353 * Don't blow it for numeric -newkey arguments to apps/req.
18355 *Bodo Moeller <3moeller@informatik.uni-hamburg.de>*
18357 * Temp key "for export" tests were wrong in s3_srvr.c.
18359 *Anonymous <nobody@replay.com>*
18361 * Add prototype for temp key callback functions
18362 SSL_CTX_set_tmp_{rsa,dh}_callback().
18366 * Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
18367 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
18371 * X509_name_add_entry() freed the wrong thing after an error.
18373 *Arne Ansper <arne@ats.cyber.ee>*
18375 * rsa_eay.c would attempt to free a NULL context.
18377 *Arne Ansper <arne@ats.cyber.ee>*
18379 * BIO_s_socket() had a broken should_retry() on Windoze.
18381 *Arne Ansper <arne@ats.cyber.ee>*
18383 * BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
18385 *Arne Ansper <arne@ats.cyber.ee>*
18387 * Make sure the already existing X509_STORE->depth variable is initialized
18388 in X509_STORE_new(), but document the fact that this variable is still
18389 unused in the certificate verification process.
18391 *Ralf S. Engelschall*
18393 * Fix the various library and apps files to free up pkeys obtained from
18394 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
18398 * Fix reference counting in X509_PUBKEY_get(). This makes
18399 demos/maurice/example2.c work, amongst others, probably.
18401 *Steve Henson and Ben Laurie*
18403 * First cut of a cleanup for apps/. First the `ssleay` program is now named
18404 `openssl` and second, the shortcut symlinks for the `openssl <command>`
18405 are no longer created. This way we have a single and consistent command
18406 line interface `openssl <command>`, similar to `cvs <command>`.
18408 *Ralf S. Engelschall, Paul Sutton and Ben Laurie*
18410 * ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
18411 BIT STRING wrapper always have zero unused bits.
18415 * Add CA.pl, perl version of CA.sh, add extended key usage OID.
18419 * Make the top-level INSTALL documentation easier to understand.
18423 * Makefiles updated to exit if an error occurs in a sub-directory
18424 make (including if user presses ^C) [Paul Sutton]
18426 * Make Montgomery context stuff explicit in RSA data structure.
18430 * Fix build order of pem and err to allow for generated pem.h.
18434 * Fix renumbering bug in X509_NAME_delete_entry().
18438 * Enhanced the err-ins.pl script so it makes the error library number
18439 global and can add a library name. This is needed for external ASN1 and
18440 other error libraries.
18444 * Fixed sk_insert which never worked properly.
18448 * Fix ASN1 macros so they can handle indefinite length constructed
18449 EXPLICIT tags. Some non standard certificates use these: they can now
18454 * Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
18455 into a single doc/ssleay.txt bundle. This way the information is still
18456 preserved but no longer messes up this directory. Now it's new room for
18457 the new set of documentation files.
18459 *Ralf S. Engelschall*
18461 * SETs were incorrectly DER encoded. This was a major pain, because they
18462 shared code with SEQUENCEs, which aren't coded the same. This means that
18463 almost everything to do with SETs or SEQUENCEs has either changed name or
18464 number of arguments.
18466 *Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>*
18468 * Fix test data to work with the above.
18472 * Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
18473 was already fixed by Eric for 0.9.1 it seems.
18475 *Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>*
18477 * Autodetect FreeBSD3.
18481 * Fix various bugs in Configure. This affects the following platforms:
18485 unixware-2.0-pentium
18490 * Eliminate generated files from CVS. Reorder tests to regenerate files
18491 before they are needed.
18495 * Generate Makefile.ssl from Makefile.org (to keep CVS happy).
18499 ### Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
18501 * Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
18502 changed SSLeay to OpenSSL in version strings.
18504 *Ralf S. Engelschall*
18506 * Some fixups to the top-level documents.
18510 * Fixed the nasty bug where rsaref.h was not found under compile-time
18511 because the symlink to include/ was missing.
18513 *Ralf S. Engelschall*
18515 * Incorporated the popular no-RSA/DSA-only patches
18516 which allow to compile a RSA-free SSLeay.
18518 *Andrew Cooke / Interrader Ldt., Ralf S. Engelschall*
18520 * Fixed nasty rehash problem under `make -f Makefile.ssl links`
18521 when "ssleay" is still not found.
18523 *Ralf S. Engelschall*
18525 * Added more platforms to Configure: Cray T3E, HPUX 11,
18527 *Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>*
18529 * Updated the README file.
18531 *Ralf S. Engelschall*
18533 * Added various .cvsignore files in the CVS repository subdirs
18534 to make a "cvs update" really silent.
18536 *Ralf S. Engelschall*
18538 * Recompiled the error-definition header files and added
18539 missing symbols to the Win32 linker tables.
18541 *Ralf S. Engelschall*
18543 * Cleaned up the top-level documents;
18544 o new files: CHANGES and LICENSE
18545 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
18546 o merged COPYRIGHT into LICENSE
18547 o removed obsolete TODO file
18548 o renamed MICROSOFT to INSTALL.W32
18550 *Ralf S. Engelschall*
18552 * Removed dummy files from the 0.9.1b source tree:
18553 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
18554 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
18555 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
18556 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
18557 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
18559 *Ralf S. Engelschall*
18561 * Added various platform portability fixes.
18565 * The Genesis of the OpenSSL rpject:
18566 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
18567 Young and Tim J. Hudson created while they were working for C2Net until
18570 *The OpenSSL Project*
18572 ### Changes between 0.9.0b and 0.9.1b [not released]
18574 * Updated a few CA certificates under certs/
18578 * Changed some BIGNUM api stuff.
18582 * Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
18583 DGUX x86, Linux Alpha, etc.
18587 * New COMP library [crypto/comp/] for SSL Record Layer Compression:
18588 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
18593 * Add -strparse option to asn1pars program which parses nested
18596 *Dr Stephen Henson <shenson@bigfoot.com>*
18598 * Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
18602 * DSA fix for "ca" program.
18606 * Added "-genkey" option to "dsaparam" program.
18610 * Added RIPE MD160 (rmd160) message digest.
18614 * Added -a (all) option to "ssleay version" command.
18618 * Added PLATFORM define which is the id given to Configure.
18622 * Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
18626 * Extended the ASN.1 parser routines.
18630 * Extended BIO routines to support REUSEADDR, seek, tell, etc.
18634 * Added a BN_CTX to the BN library.
18638 * Fixed the weak key values in DES library
18642 * Changed API in EVP library for cipher aliases.
18646 * Added support for RC2/64bit cipher.
18650 * Converted the lhash library to the crypto/mem.c functions.
18654 * Added more recognized ASN.1 object ids.
18658 * Added more RSA padding checks for SSL/TLS.
18662 * Added BIO proxy/filter functionality.
18666 * Added extra_certs to SSL_CTX which can be used
18667 send extra CA certificates to the client in the CA cert chain sending
18668 process. It can be configured with SSL_CTX_add_extra_chain_cert().
18672 * Now Fortezza is denied in the authentication phase because
18673 this is key exchange mechanism is not supported by SSLeay at all.
18677 * Additional PKCS1 checks.
18681 * Support the string "TLSv1" for all TLS v1 ciphers.
18685 * Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
18686 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
18690 * Fixed a few memory leaks.
18694 * Fixed various code and comment typos.
18698 * A minor bug in ssl/s3_clnt.c where there would always be 4 0
18699 bytes sent in the client random.
18701 *Edward Bishop <ebishop@spyglass.com>*
18705 [CVE-2020-1971]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1971
18706 [CVE-2020-1967]: https://www.openssl.org/news/vulnerabilities.html#CVE-2020-1967
18707 [CVE-2019-1563]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1563
18708 [CVE-2019-1559]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1559
18709 [CVE-2019-1552]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1552
18710 [CVE-2019-1551]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1551
18711 [CVE-2019-1549]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1549
18712 [CVE-2019-1547]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1547
18713 [CVE-2019-1543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2019-1543
18714 [CVE-2018-5407]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-5407
18715 [CVE-2018-0739]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0739
18716 [CVE-2018-0737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0737
18717 [CVE-2018-0735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0735
18718 [CVE-2018-0734]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0734
18719 [CVE-2018-0733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0733
18720 [CVE-2018-0732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2018-0732
18721 [CVE-2017-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3738
18722 [CVE-2017-3737]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3737
18723 [CVE-2017-3736]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3736
18724 [CVE-2017-3735]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3735
18725 [CVE-2017-3733]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3733
18726 [CVE-2017-3732]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3732
18727 [CVE-2017-3731]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3731
18728 [CVE-2017-3730]: https://www.openssl.org/news/vulnerabilities.html#CVE-2017-3730
18729 [CVE-2016-7055]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7055
18730 [CVE-2016-7054]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7054
18731 [CVE-2016-7053]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7053
18732 [CVE-2016-7052]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-7052
18733 [CVE-2016-6309]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6309
18734 [CVE-2016-6308]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6308
18735 [CVE-2016-6307]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6307
18736 [CVE-2016-6306]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6306
18737 [CVE-2016-6305]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6305
18738 [CVE-2016-6304]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6304
18739 [CVE-2016-6303]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6303
18740 [CVE-2016-6302]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-6302
18741 [CVE-2016-2183]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2183
18742 [CVE-2016-2182]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2182
18743 [CVE-2016-2181]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2181
18744 [CVE-2016-2180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2180
18745 [CVE-2016-2179]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2179
18746 [CVE-2016-2178]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2178
18747 [CVE-2016-2177]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2177
18748 [CVE-2016-2176]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2176
18749 [CVE-2016-2109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2109
18750 [CVE-2016-2107]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2107
18751 [CVE-2016-2106]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2106
18752 [CVE-2016-2105]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-2105
18753 [CVE-2016-0800]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0800
18754 [CVE-2016-0799]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0799
18755 [CVE-2016-0798]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0798
18756 [CVE-2016-0797]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0797
18757 [CVE-2016-0705]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0705
18758 [CVE-2016-0702]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0702
18759 [CVE-2016-0701]: https://www.openssl.org/news/vulnerabilities.html#CVE-2016-0701
18760 [CVE-2015-3197]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3197
18761 [CVE-2015-3196]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3196
18762 [CVE-2015-3195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3195
18763 [CVE-2015-3194]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3194
18764 [CVE-2015-3193]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-3193
18765 [CVE-2015-1793]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1793
18766 [CVE-2015-1792]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1792
18767 [CVE-2015-1791]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1791
18768 [CVE-2015-1790]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1790
18769 [CVE-2015-1789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1789
18770 [CVE-2015-1788]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1788
18771 [CVE-2015-1787]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-1787
18772 [CVE-2015-0293]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0293
18773 [CVE-2015-0291]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0291
18774 [CVE-2015-0290]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0290
18775 [CVE-2015-0289]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0289
18776 [CVE-2015-0288]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0288
18777 [CVE-2015-0287]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0287
18778 [CVE-2015-0286]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0286
18779 [CVE-2015-0285]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0285
18780 [CVE-2015-0209]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0209
18781 [CVE-2015-0208]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0208
18782 [CVE-2015-0207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0207
18783 [CVE-2015-0206]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0206
18784 [CVE-2015-0205]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0205
18785 [CVE-2015-0204]: https://www.openssl.org/news/vulnerabilities.html#CVE-2015-0204
18786 [CVE-2014-8275]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-8275
18787 [CVE-2014-5139]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-5139
18788 [CVE-2014-3572]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3572
18789 [CVE-2014-3571]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3571
18790 [CVE-2014-3570]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3570
18791 [CVE-2014-3569]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3569
18792 [CVE-2014-3568]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3568
18793 [CVE-2014-3567]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3567
18794 [CVE-2014-3566]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3566
18795 [CVE-2014-3513]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3513
18796 [CVE-2014-3512]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3512
18797 [CVE-2014-3511]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3511
18798 [CVE-2014-3510]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3510
18799 [CVE-2014-3509]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3509
18800 [CVE-2014-3508]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3508
18801 [CVE-2014-3507]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3507
18802 [CVE-2014-3506]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3506
18803 [CVE-2014-3505]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3505
18804 [CVE-2014-3470]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-3470
18805 [CVE-2014-0224]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0224
18806 [CVE-2014-0221]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0221
18807 [CVE-2014-0195]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0195
18808 [CVE-2014-0160]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0160
18809 [CVE-2014-0076]: https://www.openssl.org/news/vulnerabilities.html#CVE-2014-0076
18810 [CVE-2013-6450]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-6450
18811 [CVE-2013-4353]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-4353
18812 [CVE-2013-0169]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0169
18813 [CVE-2013-0166]: https://www.openssl.org/news/vulnerabilities.html#CVE-2013-0166
18814 [CVE-2012-2686]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2686
18815 [CVE-2012-2333]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2333
18816 [CVE-2012-2110]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-2110
18817 [CVE-2012-0884]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0884
18818 [CVE-2012-0050]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0050
18819 [CVE-2012-0027]: https://www.openssl.org/news/vulnerabilities.html#CVE-2012-0027
18820 [CVE-2011-4619]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4619
18821 [CVE-2011-4577]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4577
18822 [CVE-2011-4576]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4576
18823 [CVE-2011-4109]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4109
18824 [CVE-2011-4108]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-4108
18825 [CVE-2011-3210]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3210
18826 [CVE-2011-3207]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-3207
18827 [CVE-2011-0014]: https://www.openssl.org/news/vulnerabilities.html#CVE-2011-0014
18828 [CVE-2010-4252]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4252
18829 [CVE-2010-4180]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-4180
18830 [CVE-2010-3864]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-3864
18831 [CVE-2010-1633]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-1633
18832 [CVE-2010-0740]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0740
18833 [CVE-2010-0433]: https://www.openssl.org/news/vulnerabilities.html#CVE-2010-0433
18834 [CVE-2009-4355]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-4355
18835 [CVE-2009-3555]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3555
18836 [CVE-2009-3245]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-3245
18837 [CVE-2009-1386]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1386
18838 [CVE-2009-1379]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1379
18839 [CVE-2009-1378]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1378
18840 [CVE-2009-1377]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-1377
18841 [CVE-2009-0789]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0789
18842 [CVE-2009-0591]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0591
18843 [CVE-2009-0590]: https://www.openssl.org/news/vulnerabilities.html#CVE-2009-0590
18844 [CVE-2008-5077]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-5077
18845 [CVE-2008-1678]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1678
18846 [CVE-2008-1672]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-1672
18847 [CVE-2008-0891]: https://www.openssl.org/news/vulnerabilities.html#CVE-2008-0891
18848 [CVE-2007-5135]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-5135
18849 [CVE-2007-4995]: https://www.openssl.org/news/vulnerabilities.html#CVE-2007-4995
18850 [CVE-2006-4343]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4343
18851 [CVE-2006-4339]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-4339
18852 [CVE-2006-3738]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-3738
18853 [CVE-2006-2940]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2940
18854 [CVE-2006-2937]: https://www.openssl.org/news/vulnerabilities.html#CVE-2006-2937
18855 [CVE-2005-2969]: https://www.openssl.org/news/vulnerabilities.html#CVE-2005-2969
18856 [CVE-2004-0112]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0112
18857 [CVE-2004-0079]: https://www.openssl.org/news/vulnerabilities.html#CVE-2004-0079
18858 [CVE-2003-0851]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0851
18859 [CVE-2003-0545]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0545
18860 [CVE-2003-0544]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0544
18861 [CVE-2003-0543]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0543
18862 [CVE-2003-0078]: https://www.openssl.org/news/vulnerabilities.html#CVE-2003-0078
18863 [CVE-2002-0659]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0659
18864 [CVE-2002-0657]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0657
18865 [CVE-2002-0656]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0656
18866 [CVE-2002-0655]: https://www.openssl.org/news/vulnerabilities.html#CVE-2002-0655